8th Solar & 17th Wind Integration Workshop // E-Mobility Integration Symposium 2018 (SIW / WIW / E-Mobility 2018 // Stockholm, Sweden)
Stockholm / Sweden, 15 to 19 October 2018

Wind Integration Workshop 2018

Modeling the arrangement of turbines for onshore wind power plants under varying wind conditions
Submission-ID 004
Maja Celeska 1, Krste Najdenkoski 1, Vladimir Dimchev 1, Vlatko Stoilkov 1, Lothar Fickert 2, Robert Schuerhuber 2
1 Ss. Cyril and Methodius University in Skopje, Faculty of Electrical Engineering and Information Technologies, Macedonia
2 Institute of Electrical Power Systems Technical University of Graz, Austria, Austria
Optimization of wind farm layout concerning various parameters is a major point in planning and will influence the revenue for the whole life of the installation. Besides the obvious impact of wind distribution also other parameters like connection costs and levelized costs of energy influence the optimum layout and have to be included in a realistic optimization algorithm. In this paper the sophisticated optimization of wind farm layout with of two fundamentally different heuristic algorithms is investigated. To do so, detailed real-world data from an existing wind farm in Bogdanci, Macedonia is utilized by employing real wind farm data we are able to calibrate model adequacy and ascertain a model that will serve as a referent guidance in the planning of future onshore wind farms. The major unique feature of the research is the simultaneous optimization taking into account all major technical influence and cost factors, including: (i) detailed and advanced models for power modeling due to bivariate distribution of wind speed and direction; (ii) accurate estimation of levelized cost of energy (LCOE); (iii) analysis of the shortest electrical interconnections among wind turbines and (iv) correction of hub height on each wind turbine in the wind farm with taking also the wake effect into consideration. Different layouts were designed using sophisticated algorithms for handling the resulting high-dimensional, highly non-linear optimization problem. In particular, a non-dominated sorting genetic algorithm (NSGA) and a mixed‑discrete particle swarm optimization algorithm (MD-PSO) were applied. Both optimization algorithms established bi-objective fitness functions, in particular- minimizing the levelized cost of energy and maximizing the capacity factor. By comparing the results obtained with the existing layout, it is established that both optimization algorithms are adequate in determination of wind power plant layouts. Results show also a remarkable improvement of 4.6 % and 23.4 % for levelized costs and capacity factor, respectively, compared to the as built wind farm layout. So, it is proven that the implementation of sophisticated optimization methods can results in essential savings during the whole lifetime of the wind farm.

Keywords—Onshore wind power plant layout optimization, Non-dominated Genetic Algorithm, Mixed-discrete Particle Swarm, Bogdanci wind farm

Impact of Inertia Emulation on Damping of Inter-Area Power Oscillations during Under-Frequency Events in the Nordic Power System
Submission-ID 005
Tuire Kujansuu 1, Antti Harjula 1, Richard Ogiewa 2, Yuwa Chompoobutrgool 3
1 Fingrid Oyj, Finland
2 ENERCON GmbH, Sweden
3 ENERCON GmbH, Germany
This paper reviews the impact of Inertia Emulation control provided by converter-connected wind power plants on the damping of inter-area power oscillations during under-frequency events in the Nordic power system. For this purpose the paper summarises, the Inertia Emulation function of Type 4 wind power plants and the current state of the Nordic power system. Based on a future scenario a model of the Nordic power system is set up, containing a certain amount of wind power plants with Inertia Emulation function. Then, the effect of Inertia Emulation on the course of the frequency is reviewed for this future scenario. The presented simulations examine the effect of the delay between the frequency measurement and Inertia Emulation control response on the damping of the power and frequency oscillations. This issue is especially important to consider when using Inertia Emulation function with a power increase dependent on the frequency measurement. Namely, with a power increase dependent on the measured frequency and with a reasonable delay, Inertia Emulation control either does not affect the oscillations or can even dampen the oscillations during the under-frequency event. This leads to the conclusion that the delay between the frequency measurement of the Inertia Emulation function and the actual active power response is crucial to the damping of the inter-area power oscillations but also to the Inertia Emulation performance to support the power system.

Japan's R&D Project of Ramp Forecasting Technology: Deterministic Forecast with Post-processing Using Real-time Monitoring Data
Submission-ID 006
Shinji Kadokura, Daisuke Nohara, Masamichi Ohba, Atsushi Hashimoto, Keisuke Nakao, Yasuo Hattori, Takeshi Watanabe, Hiromaru Hirakuchi
Central Research Institute of Electric Power Industry, Japan
Deployment of power generation by renewable energy such as photovoltaics (PV) and wind turbine (WT) is progressing worldwide. Since PV and WT power generation is markedly fluctuated due to the change of the weather, there is concern about the stability of the electric power system.

Therefore, we have been developing a prediction method of wind power generation output to contribute for the stabilization of electric power system. In this method, the numerical forecast value provided operationally by the Japan Meteorological Agency is downscaled using the Weather Research and Forecasting model (WRF) and CFD model to predict the wind speed at the wind turbine, and the wind power output is projected using the power curve. As the summation of the wind power output projected for each wind turbine, the total output of the area is estimated.

In this study, we have improved the prediction of wind speed and power generation output by using the monitoring data for each wind turbine, by correction based on multiple regression and by using empirical power curve, which are derived by the relationship between the wind speed and wind power output observed at each wind turbine. By this improvement, the prediction error of the wind power generation output area total value is reduced by about 30%.

Short term forecasting of wind turbine production whith Machine Learning methods : direct approach and integrated approach.
Submission-ID 011
Mamadou Dione 1, Eric Matzner-Lober 2, Philippe Alexandre 3
1 Engie Green France- CREST (ENSAE ParisTech), France
2 CREST (ENSAE ParisTech), France
3 Engie Green France, France
In this paper, we compared the bi-variate distributions of the component pairs (U, V) of the ECMWF and GFS model wind forecasts taking as a reference the measurements of a pylon on a wind farm. Then, we applied some machine learning algorithms for the short-term forecast (24h to 48h) of wind generation. We do two approaches: a direct approach that predicts wind generation directly from weather data, and an integrated approach that estimates wind from weather data and converts it into wind power by power curves. The models tested are random forests, CART + Bagging, CART + Boosting, SVM (Support Vector Machine). The application is made on a wind farm of 22MW (11 wind turbines) of the La Compagnie du Vent (that became Engie Green France). Our results are very conclusive compared to the literature.

Passive houses as power and heat storage in the smart grid
Submission-ID 013
Axel Bretzke, Richard Höfer
Univ. of appl. Sc. Biberach VDI, Germany
Due to the expansion of wind and solar power, there are major fluctuations in power generation. In order to make better use of this electricity, the consumption must be adapted to the generation of electricity. For this purpose, it is possible to use the thermal storage mass of massive, well-insulated buildings. For this purpose, the building is, in excess of electricity before a foreseeable winter low, e.g. heated by a heat pump to over 20 degrees, in the context of individual comfort. This energy is stored due to the insulation over several days in the building and lost slowly. Therefore there is in the following period of power shortage and despite lack of solar radiation no need for heating. For this purpose, the heating demand was simulated on the basis of real weather data and the reheating demand was assessed.


The most critical weather with the highest heating demand is the winter low, since the outside temperatures are only moderately low but there are hardly any solar and wind gains. If the heating days are considered, it should be noted that, due to the solar radiation, even in winter often occur days without heating demand. For days without solar radiation but with wind surplus (wind D> 23 GW and electricity export), the building can be heated with regenerative electricity to 2-3 Kelvin over 20 degrees, if subsequently a winter low with covered and wind poor weather is announced. Then, the building only needs heating after 4-5 days again by the heat pump. The building therefore acts as heat and electricity storage for the network for a period of 4-5 days. Overall, the heating effort with this driving with 407 kWh / a power consumption is slightly higher than at even 20 degrees. Depending on the need for comfort, the building can so be heated to almost 100% with regenerative electricity from surplus wind (wind D> 23 GW and electricity export) and passively with the sun. The demand for non-renewable electricity in the example is very low at 73 kWh / a or 0.34 kWh / a m². This also makes it possible to relieve the local networks by passive house buildings and their storage options, which could be used differently within the scope of smart meters and corresponding electricity price offers, depending on the users' need for comfort. In-depth studies on the possibilities based on nationwide and local generation and grid situations are underway or completed by various parties (including ibp, network-reactive buildings). The inclusion of a building's own PV would also make sense. Conclusion: It is possible to use the heat storage mass of buildings with PH standard as variable power consumers to permanently relieve the power grid.

A Fault Handling Current Control Strategy for Offshore Wind Power Plants with Diode Rectifier HVDC Transmission
Submission-ID 016
Christian Neumann 1, Hans-Günter Eckel 1, Sven Achenbach 2
1 University of Rostock Institute of Electrical Power Engineering, Germany
2 Siemens AG Energy Management Division, Germany
This paper presents an extension of the Fixed Reference Frame (FixReF) current control strategy for wind turbine converters in diode rectifier based HVDC links. By adding a negative sequence FixReF current control loop, the line-side converters can cope with symmetrical as well as asymmetrical faults within the offshore AC grid. The implemented fault ride through (FRT) control strategy, based on this decoupled double FixReF current control scheme, is capable of low voltage and high voltage FRT which is validated at different operation points using PSCADTM/EMTDCTM simulations. Furthermore, the paper illustrates a possible solution for the auxiliary power supply during calm wind situations utilizing the existing DC cable together with a paralleled medium-voltage modular multilevel converter (MMC). The mentioned FRT strategy is used for the transition between HVDC power export and MVDC auxiliary power supply mode.

Keywords-HVDC; rectifier; wind power plant; FixReF; grid-forming control; FRT; asymmetrical fault; auxiliary power

An Impedance-Based Active Filter for Harmonic Damping by Type-IV Wind Turbines
Submission-ID 020
Emerson Guest 1, 2, Tonny Rasmussen 2, Kim Jensen 1
1 Siemens Gamesa Renewable Energy, Denmark
2 Technical University of Denmark, Denmark
In this paper, an impedance-based active filter is developed for the three-phase grid-connected converter used in a Type-IV wind turbine. The active filter utilizes the voltage and current measurement to provide programmable narrowband shaping of the converter impedance. This has application in mitigating voltage or current amplification in wind power plants by increasing the harmonic damping applied by the wind turbines. Moreover, the active filter inherently compensates converter generated voltage harmonics, ensuring the voltage quality of the wind turbine. The performance of the active filter is verified through experiments on a 0.6MW grid-connected converter.

Analysis of Harmonic Resonance Stability in Power System with Renewable Generations
Submission-ID 026
Lijun Cai 1, Ulas Karaagac 2, Jean Mahseredjian 3, Hans-Günter Eckel 1, Harald Weber 1
1 University of Rostock, Germany
2 Hong Kong Polytechnic University, Hong Kong
3 Polytechnique Montréal, Canada
Key words: renewable generation; harmonic impedance; subsynchronous resonance; supersynchronous resonance; harmonic resonance stability.


With the development of renewable energy in the power systems, converters that interface with renewable generations are becoming new potential sources of harmonics. Harmonics may lead to problems such as damage of capacitors due to overheat, increased mechanical vibration of inductors, wrong trig of power system controllers and unintended shutdown of renewable generations.

Moreover, the renewable generations are increasingly connected to the distribution weak systems and voltage source converter (VSC) based HVDC systems, where the whole system harmonic resonance behaviors should be considered in detail. Therefore, this paper concerns harmonic resonance stability analysis in power system with renewable generations.

First, an overview of harmonic impedance modeling and harmonic resonance analysis will be given.

Then harmonic impedance of both doubly fed induction generator (DFIG) and full converter (FC) based wind turbines are analyzed in detail: the impedance scan of DFIG and FC wind turbines are performed, where the detailed generator model, rotor side converter (RSC) control, grid side converter (GSC) control and filters are considered.

Concerning the harmonic impedances of wind turbines and power systems, traditional resonance stability (Nyquist) criteria will be improved by concerning the harmonic resistance (damping) and harmonic reactance separately.

Both subsynchronous and supersynchronous resonance scenarios between wind farm and compensated power system are simulated to verify the proposed modified harmonic stability criteria.

Finally, measures for damping the harmonic resonances are given.

Comparison of Connection Policies for the Planning of Generation Integration into a Distribution Network
Submission-ID 027
Anthony Chabrol, John Wallace, Conor Molloy
ESB Networks, Ireland
Comparison of Connection Policies for the Planning of Generation Integration into a Distribution Network

Over the past two decades the Irish distribution system has experienced large-scale integration of renewable generators, such as Wind Farms and small scale synchronous generators. This has posed a challenge for distribution system operator (DSO) planners, with the distribution network originally designed generally as a one-way feed to meet load downstream, this is no longer the case.

The role of the planner when assessing proposed generator connections, such as wind or solar, is to ensure that the connections can be made at least cost whilst keeping the system reliable and capable of operating within the SO limits. In order to connect these generators in a timely fashion, a process must be used. There are two different approaches that are used. This paper will compare and contrast these two different methods, allowing for others to determine which method would work best on their system if required.

Modelling of Large Size Electrolyser for Electrical Grid Stability Studies - A Hierarchical Control Approach
Submission-ID 037
P.K.S Ayivor 1, J.L. Rueda Torres 1, M.A.M.M. van der Meijden 1, 2
1 Delft University of Technology, Netherlands
2 TenneT TSO B.V, Netherlands
This paper proposes a control approach which extends the capabilities of large scale commercial electrolysers (>1MW) to support power system stability. The search for new sources of ancillary services and the projected demand for hydrogen as a medium of energy storage has aroused considerable interest in the use of large scale electrolysers for power system ancillary services. In order to fully leverage the potential of large scale electrolysers to support power system stability, effective control schemes are needed. To illustrate the feasibility of the proposed approach, case studies are built on a model of a 1 MW plant augmented with a high level controller. Simulations are performed using the Real-Time Digital Simulator (RTDS) platform, to investigate the response of the high level control scheme to external command signals and power system disturbances. The results show that the proposed control scheme can enable an optimal operation of the electrolyser load on the basis of market and power system conditions.

Global Geospatial Optimization of the Locations of Wind Farms and the Configuration of Transmission Networks
Submission-ID 040
Kazuaki Iwamura, Ryo Kobayashi, Kohei Nishiyama, Yosuke Nakanishi
Graduate School of Environment and Energy Engineering, Waseda University, Japan
1. Introduction

Herein, we present a global geospatial optimization for the location and transmission network (TN) configuration of wind farms (WFs). We aim to minimize construction costs by considering geographical conditions using a cost surface. To select the best WF locations and TNs in large areas, penetration Voronoi division (PVD) is applied to group the WFs. Then, a genetic algorithm (GA) is applied with a relaxation process on the minimum spanning tree (MST) to minimize the construction cost and avoid overloads. Moreover, we simulated a large area in Japan to test the proposed technique.

2. 4D-GIS and Cost Surface

A geographic information platform, 4D-GIS, is used to process 2D, 3D, and time-series data and manage the cost surface. The cost surface data is a digital elevation model (DEM) that included topographical factors (e.g., height constraints), environmental factors (e.g., construction-prohibited areas), and economic factors (e.g., power transmission losses).

3. Extraction of feasible areas for WFs

To identify the feasible areas for building WFs, we proposed a method to merge WF areas. Originally, small candidate areas in the DEM are selected based on the wind speed and cost surface constraints. These areas are merged as WF-feasible areas by merging small neighboring areas with high wind speeds.

4.Configuration Design of Transmission Networks

Challenges associated with designing the configuration of a TN in a large area exist:

[1] Selecting WFs that exceed their target capacities.

[2] Configuring TNs among WFs and access points (APs).

PVD and a GA are applied to address these challenges.

4.1 PVD

In large areas, WFs should be classified into small groups to avoid long-distance connections. PVD is used to classify WFs based on the cost surface. The areas on the cost surface that are reachable from an AP are merged by an area-range expansion process called penetration.

4.2 GA with a MST Relaxation Process

All transmission lines among the WFs and an AP in each group are found using a shortest-route search via PVD. The WFs whose total capacities exceed their targets are selected and a TN among these WFs and APs is configured. With 100 WFs (practically), 2100 combinations of these WFs exist; thus, a GA is used. In the GA, WFs are treated as genes in a chromosome. The objective function is the sum of the construction costs of WFs, transmission lines, and transformers. The GA is executed iteratively to minimize the objective function. At the optimal point, the TNs are configured with the MST, which is a tree graph structure without loops. Then, MST relaxation is applied to partially release and reorganize the MST topology to avoid overloads.

5. Simulation Results

A simulation of a 9,600 km2 area in Japan is performed. In this area, 110 feasible WF areas are identified. The target capacity is 2.0 GW; thus, the introduction capacity is 2.09 GW. This result demonstrates that the proposed method yields an optimal result.

Enabling 80+ percent share of wind and solar in Lithuanian power sector: role of storage technologies
Submission-ID 048
Justinas Jasiūnas, Lennart Söder
KTH Royal Institute of Technology, Stockholm, Sweden, Sweden
For wind and solar to play a major role in deep decarbonization of power sector costs of systems with very high shares of variable supply need to become lower than that of existing systems. Balancing challenges for systems with high shares of variable supply is expected to differ significantly from systems with intermediate levels of supply variability. To get a better understanding on the former study is performed on insulated, 100% renewable Lithuanian power sector with pumped hydro and hydrogen energy storage systems. The study is performed using deterministic optimization model with linear formulation, which simultaneously varies installed capacity and its dispatch. Model is run at hourly resolution for a single year with existing technologies.

Among the aspects considered in the study none were found to be prohibiting feasibility of the system analysed. Performed analysis shows seasonal energy shortages to account for more than 5% of annual consumption in non-oversized system. System balancing in the model was achieved with oversizing supply, use of hydrogen and pumped hydro storage systems at 3-5 times of electricity costs in today's wholesale market. The same scenarios show wind and solar curtailment levels of 11% to 46%. Given limited number of technologies included in optimization, large cost reduction potential remains. Among included technologies wind power capacity cost reductions have the largest impact to overall system costs. However, wind cost reductions alone are highly unlikely to achieve overall system competitiveness.

Submission-ID 061
Markus Lehmann 1, Dr. Martin Pieschel 1, Dr. Łukasz Kocewiak 2, Miguel Juamperez 2, Sridhar Sahukari 2, Karsten Kabel 2
1 Siemens AG Freyeslebenstr. 1-3 91058 Erlangen, Germany
2 Ørsted A/S Kraftværksvej 53 – Skærbæk 7000 Fredericia, Denmark
The application of active filtering (AF) in a large-scale STATCOM for offshore wind power integration is presented in this paper. The project topology and its main parameters are introduced to give an understanding of the basic layout of an offshore wind power plant (WPP) integration using an AC connection. The impact of the AF functionality on the measurement hardware of the STATCOM system is described. The methodology and operating principle of the AF within the STATCOM is shown and explained. Harmonic propagation studies illustrate that low-frequency resonances in WPP components , such as HVAC export power cables, reactive power compensating devices and park transformers may introduce amplification of harmonic distortion (HD) at the point of connection (POC). The impact of the AF functionality during operation is shown in real-time simulation cases.

Wind generation in adequacy calculations and capacity markets in different power system control zones
Submission-ID 063
Lennart Söder 1, Ana Estanqueiro 2, Damian Flynn 3, Bri-Mathias Hodge 4, Juha Kiviluoma 5, Magnus Korpås 6, Emmanuel Neau 7, Kaushik Das 8, Danny Pudjianto 9, Goran Strbac 9, Daniel Burke 10, Tomás Gómez 11, Antonio Couto 2
1 Department of Electric Power and Energy Systems, KTH, Sweden
2 Laboratório Nacional de Energia e Geologia, Portugal
3 Electrical and Electronic Engineering University College Dublin, Ireland
4 National Renewable Energy Laboratory, United States
5 VTT Technical Research Centre of Finland, Finland
6 NTNU, Norway
7 EDF, France
8 Wind Energy, DTU, Denmark
9 Imperial College London, United Kingdom
10 National Grid, United Kingdom
11 Universidad Pontificia Comillas, Spain
Generation capacity adequacy is a major issue in most power systems, but there are many approaches which can be assessed. Power system planners often define target values for the capacity adequacy, which may be achieved through capacity markets/auctions, capacity reserves, or capacity purchases. Wind power contributes to the generation capacity adequacy of the power system since there is a possibility that wind power will generate in high load situations and thereby decreases the risk of generation capacity deficit compared to the system without this source. The contribution is probabilistic – as it is with any other source, since nothing is 100% reliable - but the capacity value of wind power is significantly smaller compared to the capacity value of conventional fossil-fueled plants.

In this article, an overview of the fundamental challenges in the regulation of capacity adequacy as well as how wind power is treated in some selected existing jurisdictions is presented. The jurisdictions that are included are Sweden, Great Britain , France, Ireland, United States (PJM), Finland, Portugal, Spain Norway and Denmark.

System Services by Wind Power Plants Supporting 75% Wind Penetration in Ireland
Submission-ID 067
Mark Gilsenan 1, David McMullin 1, Sönke Engelken 2
1 ENERCON GmbH - Irish Branch, Ireland
2 WRD GmbH, Germany
This paper briefly reviews a number of markets where system services by wind are active or in the early stages of development, before turning its focus to Ireland's DS3 System Services market. The paper includes a review of the technical requirements of relevant system services in Ireland, while providing context as to why the System Operators, EirGrid and SONI, have developed such requirements. Finally, this paper presents and assesses the outcomes of system services trials conducted in 2017 on Wind Power Plants connected to the Irish transmission system. Since the trials' conclusion, the system operator has commenced issuing contracts to Wind Power Plants of operating reserve services.

Requirements for control strategies of grid connected converters in the future power system
Submission-ID 068
Hanna Emanuel, Roberto Rosso, Konstantinos Pierros, Johannes Brombach
ENERCON GmbH, Germany
The ongoing increase in penetration of converter-based generation in many power systems around the world has sparked a discussion about how to operate these power systems with the usual levels of efficiency, reliability and cost-effectiveness. Current grid-feeding converter-based generators have proven to run stably in parallel to one another, even if there are thousands of them connected in a power system, and even in very small isolated power systems with extremely low system inertia. Discussions around the necessity of additional converter performance, usually under the “grid-forming” and “Virtual Synchronous Machines” concepts, have recently been transferred from the academic sphere to national and international industry fora. Formal discussions have started in Great Britain, in Germany and at ENTSO-E level. However, there is still a lot of uncertainty about the real and not simulated performance of grid-forming converters and the needs case for requiring this radically different control method has not been adequately justified. With the present paper we raise key questions that will serve towards an objective discussion about power system needs, grid infeed technologies, and their interaction.

Operation of Hydrothermal System with Increased Wind Generation
Submission-ID 074
Luke Schwartfeger 1, Alan Wood 1, Gari Bickers 2
1 University of Canterbury, Christchurch, New Zealand, New Zealand
2 Transpower New Zealand, New Zealand
New Zealand, like many countries, aspires to replace fossil fuel based generation with renewable resource based generation. Wind generation is a likely candidate for this in New Zealand. This paper describes a case study of New Zealand’s hydrothermal electricity system in response to additional wind generation. A modelling tool using implicit stochastic optimization with dynamic programming was developed along with a model of New Zealand hydrological and electricity systems to conduct this analysis.

Frequency Response of Energy Storage Systems in Grids with High Level of Wind Power Penetration- Gotland Case Study
Submission-ID 075
Firas Daraiseh, Vincent Gliniewicz, Erica Lidström
Vattenfall Research and Development, Sweden
Gotland is a Swedish island that is connected to the synchronous grid only by two HVDC cables. While being known for its high level of wind power penetration, the growth of installed wind power on Gotland has been temporarily stopped due to concerns regarding the reliability of the electrical grid operations. This paper investigate the capability of a centralized energy storage system along with or without wind curtailment to support the growth of installed wind power capacity. The energy storage system is tested for maintaining the frequency stability during an unintentional islanding through dynamic studies using PSS/E. The results assess the ability of an energy storage to prevent frequency instabilities and provide primary frequency response albeit of the absence of any form of rotating inertia in highly wind power penetrated electrical grid. The analysis determines the requirements for the power and energy capacity of the energy storage system in relation to the exported power from the HVDC cables at the instant of fault, which eventually relates to the installed wind power capacity. Moreover, the study examines wind power curtailment as primary and/or secondary frequency response technology and the impact on the energy capacity of the energy storage system. Keywords- HVDC; Energy Storage System; Frequency Response; Gotland; Wind Curtailment

Frequency Support Provision to Power Systems from HVDC-Connected Offshore Wind Power Plants
Submission-ID 081
Ali Bidadfar 1, Oscar Saborío-Romano 1, Eduardo Prieto-Araujo 2, Oriol Gomis-Bellmunt 2, Müfit Altin 1, Nicolaos A. Cutululis 1, Poul E. Sørensen 1
1 Wind Energy Department, DTU, Denmark
2 CITCEA-UPC, Barcelona, Spain
High penetration of power electronic devices in generation and transmission networks results in lower inertia in power systems, which can increase the risk of frequency instability. Thereby, the fast frequency support is anticipated to be the most important ancillary service provided by renewable generation systems in the near future. Frequency support from offshore wind power plants (WPPs) connected to high-voltage dc (HVDC) links is a challenging issue because of the complexities in their control and operation. In this paper, some technical challenges regarding fast frequency support from HVDC-grid-connected offshore WPPs are addressed. Two different methods–communication-based and droop based–of implementing such support are explained first, then the advantages and disadvantages of each are detailed by means of mathematical and numerical (simulations) analyses. A four terminal HVDC grid, including two offshore WPPs and two onshore ac systems, is used for analyses and simulations. Moreover, a point-to-point HVDC connection is also studied to understand how in an HVDC grid converters interactions can impact the frequency control.

Grid Code Certification in Germany – A recipe for Europe?
Submission-ID 087
Bernhard Schowe, Christian Scheefer, Alina Hoppmann, Christoph Lütke-Lengerich
FGH Zertifizierungsgesellschaft mbH (FGH Certification Body), Germany
The liberalization of the energy markets demanded for guidelines regulating the access to the market as well as defining technical requirements to guarantee a safe and reliable operation of the network grids. Following the harmonization inside the European Union, the Network Code on Requirements for Grid Connection Applicable to all Generators (RfG) defines the technical requirements for the connection of power generating units (PGU), plants (PGP) and their components. In Germany, certification processes for power generating units and plants are in place for about 14 years. Europe, why don’t you take lesson’s learned besides starting from scratch?

Considering curtailments in wind power forecasting
Submission-ID 093
Jonas Koch, Dominik Jost, Axel Braun, Jan Dobschinski
Fraunhofer Institute for Energy Economics and Energy System Technology IEE, Germany
Curtailments (often called EinsMan or Eisman) are not only a hot topic for grid operators but also for providers of wind power forecasts. Regarding different use cases curtailments and other reductions have to be considered within the forecast model or not. The difference between the possible power output and the actual feed-in can be of significant magnitude and makes a differentiation between available and actual feed-in power in forecasts necessary. In categorizing reductions of wind power generation output first, we suggest a clearer definition of the differentiation between possible, available and actual power forecasts (for further discussions). With these definitions and with focus on curtailments an algorithm will be presented allowing the detection of curtailment events from power data and providing the ability to improve two hour short-term power forecasts by nearly 30 %.

Wind Farm Grid Code Compliance Testing: Different approaches, biggest hurdles and the foreseeable future
Submission-ID 096
Tobias Rösner, Jonas Fleischhauer, Manuel Fernandez
Nordex Energy GmbH, Germany
Many markets require grid code compliance testing for wind farms to prove their electrical capabilities and control behaviour. For a global acting wind turbine manufacturer it can be difficult to deal with the different approaches in different markets for performing a Grid Code Compliance Test (GCCT). In dependence of the grid code and network operator such testing can be done within a few hours or needs a few weeks. This paper gives an idea of what keeps a test simple and what makes it extensive. Topics like scripted tests compared to manual testing and compliance testing performed due to a third party consultant are being discussed.

A further objective of this paper is to give an idea of common hurdles on the way to a grid code compliance certificate for a wind farm project. It explains what the usual mistakes are when it comes to evaluate the performance of different control functionalities like reactive power, voltage or frequency control. The main purpose is a discussion regarding accuracy requirements related to set point change or target values which often leads to technically not achievable tolerances. It reflects simple written requirements within grid codes and gives an example of how the same sentence used in two different grid codes could lead to two complete different requirements.

GCCT will grow and it will grow big. It is the best way to prove grid code compliance and to demonstrate the contribution to a sustainable network. Some of the reasons of greater importance of GCCT in the future are the rising continuous monitoring of wind farms by network operators and the approach of repetitive testing every few years. Additionally costumers ordering their own GCCT for the project acceptance even if not required within the grid code.

Control solutions for Blackstart capability and Islanding operation of Offshore wind power plants
Submission-ID 097
Anubhav Jain, Kaushik Das, Ömer Göksu, Nicolaos Antonio Cutululis
Department of Wind Energy, Denmark Technical University (DTU), Denmark
Environmental sustainability concerns make renewable energy systems (RES) integrated into the grid crucial for future power systems. Amongst RES, wind energy especially offshore wind power plants (OWPP) show huge promise. Increasing penetration of RES requires re-thinking of critical operation states that could lead to an increased risk of generation tripping that ultimately triggers blackouts. Thus maintaining reliability, robustness and stability of grid operation has become more complex and so blackstart (BS) and islanded (Is) operation requirements are being considered as options for WPPs in the grid codes. Additionally, advanced control functionalities provided by modern wind turbines (WT) owing to their power electronics converter (PEC) interface, enables them to provide fast, high power environment-friendly BS capability that facilitates grid recovery & reduces the impact of a blackout. In this paper, the motivation for BS capabilities in OWPPs has been presented, and the different stages of restoration using OWPPs identified. Finally the existing control solutions and potential challenges for BS&Is using OWPPs have also been investigated.

How large wind parks contribute to harmonic waveform distortion
Submission-ID 101
Daphne Schwanz, Math Bollen, Sarah Rönnberg, Anders Larsson
Electric Power Engineering, Luleå University of Technology, Sweden
This paper summarizes the important issues related to waveform distortion and large wind parks. The emission at the classical frequencies (low-order odd harmonics) is low. Instead emission is higher than for most other installations at even harmonics, especially at high order even harmonics, at interharmonics and at supraharmonics. Theoretical studies have shown that secondary emission (emission driven by sources outside of the turbine or outside of the park) can be a substantial part of the harmonic voltages and currents with a large wind park. Other issues to be considered are aggregation and resonances..

Importance of voltage-dip knowledge for improving fault-ride-through of wind-power installations
Submission-ID 102
Azam Bagheri 1, Math Bollen 1, massimo Bongiorno 2, Chen Cheng 3, Sarah Rönnberg 1
1 Electric Power Engineering, Luleå University of Technology, 931 87 Skellefteå, Sweden
2 Dept of Electrical Engineering, Chalmers University of Technology, 412 96 Gothenburg, Sweden
3 KTH Royal Institute of Technology Stockholm, Sweden
This paper verifies the dynamic behavior of DFIG-based wind turbines in presence of voltage dips or transients due to various origins: electrical faults, transformer energizing, capacitor energizing, and cable energizing. These different origins result in different characteristics of voltage dips in term of, residual voltage, duration, dip type, harmonic distortion, and so on. Each of these characteristics has different impacts on electrical parameters of the wind turbines. The voltage dip ride-through capability of the wind turbine installations is provided by grid codes which are derived based on electrical fault origin voltage dips, considering only two dip characteristics: residual voltage and duration.

The double-fed induction generator (DFIG)-based wind turbine is simulated in MATLAB/Simulink environment to investigate the dynamic behavior of the critical electrical parameters including DC-link voltage and rotor current. The simulation results show that some shallow dips which are aligned with grid-code requirements may lead to unwanted trips because of heating problem and crowbar system activation.

Experiences in the NEM: Practical considerations for the successful integration of utility-scale renewable storage solutions
Submission-ID 105
Harley Mackenzie, Jonathon Dyson
Dispatch Solutions, Australia
Many existing and proposed wind and solar plant generators in Australia and internationally are presently considering the integration of storage solutions with their generation assets to provide a means of arbitraging energy spot prices and potentially participating in market ancillary services, or capturing energy that might be lost due to curtailment. This paper will document and discuss in practical detail some of the recent experiences in the competitive Australian National Electricity Market (NEM). Whilst the regulator and system operator have all been supportive of the recent developments, a number of challenges have been encountered that test many of the existing specifications and requirements that were formulated during market design discussions 20 to 25 years ago. During recent project developments and implementation, a number of practical constraints and limitations have been encountered such as:

* The NEM Dispatch Engine (NEMDE) required the battery to be registered in the market, bid and operated as a separate load and generator, as a storage model had not been implemented,

* The ancillary service performance has exceeded expectations but effective payment for those services appears less than desired, and

* The half hour trading interval settlements (that average the 5 minute dispatch periods) potentially lead to economic incentives not matching the system requirements, which in turn lead to very challenging system implementation challenges. Similarly, given storage devices as low as 5MW are registered in the market as dispatched generators, many of the inherent control and trading controls required by large scale fossil fuel or hydro generators are now required by renewable generators. The considerable costs of running 24/7 operations and trading centres are not well suited to the prevailing economic thinking of small scale operations that have been a significant contributor to developing and constructing renewable generation in Australia.

This paper will also discuss a number of the information and dispatch system challenges that occur with utility-scale storage implementation in 5-minute dispatch competitive markets.

Steady-state characteristics of substation-free wind power plant composed of series-connected wind turbine generators and current-source thyristor inverter
Submission-ID 106
Shoji Nishikata, Fujio Tatsuta
Tokyo Denki University, Japan
For offshore wind power plants it is desirable not to equip offshore substations because of initial costs and maintenance costs. From this point of view, we have already proposed a substation-free wind power plant consisting of series-connected wind turbine generators, HVdc transmission line, current-source thyristor inverter and synchronous compensator with duplex reactor. The basic studies on the series-connected wind power plant have been reported in the previous papers.

In this paper the steady-state characteristics of the series-connected wind power plant are investigated in some detail.

In order to clarify the performances of the whole system as accurately as possible, the performances of the system should be investigated based on the equations established on the system, and a set of such equations is first derived. Based on these equations the steady-state characteristics of the series-connected wind power plant are discussed in this paper.

In this system since the commutation of the inverter thyristors is accomplished by the electro-motive-forces induced in the armature windings of the synchronous compensator, the margin angle of commutation for the thyristors (which is the difference of leading angle of commutation of the inverter and overlapping angle of inverter output currents) must be always positive to secure safe commutation of the inverter thyristors. Hence, the steady-state characteristics of these angles have to be discussed to clarify the operation limit of the system. In addition, the characteristics of the field current and the armature current of synchronous compensator, as well as the system output current, have to be discussed.

To confirm not only the usefulness of the proposed system but also the validity of the system equations derived in this paper we have developed an experimental set-up (system output: three-phase 200 V, 50 Hz, 4 kVA) composed of two simulated wind power generators (three-phase 200 V, 100 Hz, 2 kVA, 8 poles x 2 sets) connected in series. The experimental results of the effective value and THD of the output current, the DC link voltage, the DC link current, and the margin angle of the inverter will be included when the power factor of the system output changes. The value of the power factor was chosen as specified by the Danish grid code.

Finally, steady-state characteristics of the system when dc input voltage is changed due to the changes in the wind speed are discussed using the equations introduced in this paper, and the operation limit of the system is revealed for various power factors of the system output.

This paper is based on results obtained from a project commissioned by the New Energy and Industrial Technology Development Organization (NEDO).

Japan's R&D Project of Ramp Forecasting Technology: Meteorological Pattern Analysis Method
Submission-ID 107
Maki OKADA 1, Tomohiro ICHIZAWA 1, Yuko NAKAMURA 1, Koji YAMAGUCHI 1, Ryo KODAMA 1, Hisashi KATO 2, Yoshinori NAGANO 2, Ryosaku IKEDA 3, Van Q. DOAN 3, Hiroyuki KUSAKA 3, Takamitsu ARAKI 3, Noriko N. ISHIZAKI 3, Norimitsu OGASAWARA 1
1 Japan Weather Association, Japan
2 Nihon University, Japan
3 University of Tsukuba, Japan
Abrupt wind power output fluctuation, which is known as wind power ramp, tends to negatively influence electric grid operations. To understand and counteract these influences, a national research and development (R&D) project commissioned by the New Energy and Industrial Technology Development Organization (NEDO) is underway to develop forecasting technology for wind power ramps. In this project, authors have been developing forecasting technology and systems for wind power ramps from the viewpoint of meteorological pattern analysis. The developed forecast system used the Weather Research and Forecasting Model (WRF) as a numerical prediction model and selected the different setup WRF model depending on the weather patterns. Firstly, the forecast system evaluated the analogy between sea level pressure (SLP) patterns in the latest global-scale NWP model and past SLP patterns associated with abrupt wind speed changes. According to the latest SLP patterns, secondly, the forecast system selected the optimal WRF setup. Wind forecast data from the WRF model were statistically corrected and converted into wind power output using an empirical power curve. This power curve was constructed using past wind forecast and power output monitoring data that were collected during the NEDO R&D project. The developed forecast system also outputted the probability of ramp occurrence. The probability was calculated from the means and standard deviations of wind power output on and around two transit locations (the start and after six hours) in the principal component space related with SLP forecast data. Ramp alert, which informed the occurrence of wind power ramp in advance, was defined to occur when the ramp probability exceeded 20%. We will explain our forecasting system and its performance for the three target areas (Hokkaido, Tohoku, and Tokyo) of Japan over a set timeframe.

Active Power Control for Mitigation of Very-Short-term and Short-term Fluctuation of Wind Power
Submission-ID 108
Chiyori T. Urabe, Tetsuo Saitou, Kazuhiko Ogimoto
The University of Tokyo, Japan
At the 21st session of the Conference of the Parties (COP21), participating countries agreed to achieve a balance between sources and sinks of greenhouse gases in the second half of this century. The countries are required to meet strict targets to reduce carbon emissions. In electric sector, for reduction of carbon emission, deployment of renewable energies (RE) can contribute. While variable renewable energy (VRE) sources, such as wind and solar power, occupy a large part of deployed RE, output of them can change rapidly depending on wind speed and insolation. Therefore, challenges of VRE penetration have emerged depending on the level of VRE share [1].

One of the challenges is large fluctuations of VRE. To compensate fluctuation, flexibility which conventional thermal and hydroelectric power supply is increasingly required depending on integration level of VRE in the electric system. From operational perspective on electrical system, fluctuations of demand and VRE can be classified according time scale: very-short-term, short-term, and long-term fluctuations, or primary, secondary, and tertiary control. For short-term fluctuation, power plants changes their generation by control signal from a central supply center or an aggregator. To provide such flexibility, some operations sacrifices economic efficiency for providing flexibility.

We focus on very-short-term and short-term fluctuations wind power output. Using high-resolution data of wind power output in FY2012 in Hokkaido, the northern Japanese island where large wind capacity is expected to be deployed, first we separated the very-short-term fluctuations and the short-term fluctuations from time series of the original wind power output. Fluctuations in Hokkaido is relatively large because the capacity of wind power is unevenly distributed and is a half of Tohoku which is the nearest balancing area.

To mitigate fluctuations in Hokkaido, we proposed a fluctuation mitigation control method that active power ramp rate limitation and setpoint control which the most modern wind turbines are equipped are incorporated to mitigate the very-short-term and short-term fluctuations respectively. There are two parameters related the active power ramp rate limitation and the setpoint control. In this study, we limited energy loss with the control, because there is tradeoff between the mitigating fluctuations and energy loss. Under energy loss is less than a few percent, we seek effective combination of the parameters for two types of fluctuation mitigation.

[1] International Energy Agency (IEA): “System Integration of Renewables”, IEA (2018)

Unconventional High-voltage Ride-Through Technical Retrofitting Scheme for Certain Imported Old Generating Units of the Northeast China Power Grid
Submission-ID 109
Qi Lv 1, Hong Peng Zhang 2, Guang Hui Shao 2, Jia Qing Liu 2, Yang Liu 2, Shi Bo Du 2
1 Columbia University, United States
2 Northeast Power Dispatching Center of SGCC, China
Since 2017, the ±800kV extra-high-voltage DC transmission project extending from Tongliao in Inner Mongolia to Qingzhou in Shandong Province, with a transmission capacity up to 10000MW, has been put into operation. Over the past year, the high-voltage ride-through retrofit of the wind power generating units has been carried out in the adjacent region of the EHV convertor station in the Grid. Generally, as for most of wind power generating units manufactured by China and put into operation in recent years, the conventional retrofitting approaches are applied. The control strategy of the converter of the wind power generating units needs to be retrofitted; and then, the retrofit of control logic of the wind power generating units is also required, so as to realize the grid-connected controlled operation of the wind power generating units during the overvoltage period of the Grid through coordination between the main control system and the wind turbine converter. As for a few of old generating units imported from Vestas in Denmark and Suzlon in India in the last decade, and those wind turbines manufactured by China in the early period using the imported frequency converters supplied by Converteam in France, ABB in Switzerland and SEG in Germany, if the high-voltage ride-through retrofit is to be conducted by the original equipment suppliers, the hardware upgrading and software optimization of the main control system, variable pitch system and converter are also required, which will not only result in a high cost and a long period, and in particular, when these three systems are not supplied by the same supplier, there will be some technical communication problems between different manufacturers. To solve this problem economically and efficiently, some wind power enterprises of the Grid have adopted the high-voltage ride-through retrofitting technical route with parallel connection of SCR controlled shunt reactor at the wind turbine terminal, which can realize the high-voltage ride-through capability without necessarily changing the equipment of the wind power generating units. Through the field test, it is proven that after using the retrofit scheme, the wind power generating units are able to meet the standard requirements of the high-voltage ride-through. This technical scheme is a kind of technically and economically feasible solution for the high-voltage ride-through technology retrofit for a great number of complete wind power generating units imported in the early period of the wind power development, and those wind power generating units manufactured by China using the imported frequency converters in the early period. This paper will present in detail the unconventional high-voltage ride-through technology retrofit scheme for the old generating units as well as the results of field tests.

Japan's R&D Project of Ramp Forecasting Technology: Probabilistic Forecast Based on Dynamical and Statistical Ensemble Methods
Submission-ID 112
Daisuke Nohara, Shiji Kadokura, Masamichi Ohba, Takeshi Watanabe
Central Research Institute of Electric Power Industry, Japan
Abrupt change in wind power generation, known as ramp events, due to fluctuating wind speeds present challenges to the stability of the electric power supply. In Japan, the ramp events are generally induced by extratropical cyclones along their track. Since prediction for the behavior of the cyclones entails uncertainty caused by nonlinear dynamics, probabilistic prediction is more effective. For this study, we developed a regional ensemble prediction method using the Weather Research and Forecasting model (WRF). To obtain dynamically consistent perturbations with a synoptic weather pattern, both initial and lateral boundary perturbations were determined by differences between the control and an ensemble member of the Japan Meteorological Agency (JMA)'s operational one-week ensemble forecast. This method provides 11 ensemble members with a horizontal resolution of 15 km for 75 hours at 30 minutes interval outputs by downscaling JMA's operational global forecast along with the perturbations. Wind power is projected using power curve which is estimated by relationship between area averaged wind speed and wind power generation. In addition to the dynamical ensemble, we expanded the ensemble member using Monte Carlo method to take account of the impact on uncertainty in the power curve. The expanded ensemble has 100 ensemble members. Forecast for the ramp events using the expanded ensemble exhibits high statistical consistency and reliability of ability to capture the ramp events compared with the dynamical ensemble prediction.

Case Study: Reliability of the Summation Method to Assess the Harmonic Current due to a Wind Power Plant
Submission-ID 116
Koldo Redondo 1, Izaskun Azcarate 1, José Julio Gutierrez 1, Purificación Saiz 1, Luis Alberto Leturiondo 1, Stefano Lodetti 2
1 Department of Communications Engineering University of the Basque Country (UPV/EHU), Spain
2 CIRCE Universidad de Zaragoza, Spain
The measurement of Wind Power Plant (WPP) Harmonic Current (HC) is an important issue in terms of power quality. The IEC 61400-21 standard specifies the procedure to measure the HC produced by a grid connected Wind Turbine (WT). The standard also proposes an aggregation method to estimate HC from all the WTs connected to the Point of Common Coupling of the WPP. The aim of this work is to assess the reliability of the summation method proposed by the standard, based on current signals recorded at a WPP in Spain. The results show that the summation method does not accurately estimate the HC of the WPP. The obtained deviations between the measured and estimated HC values are significant.

Comparison of different estimation methods for the grid frequency using the example of a system split in the interconnected electrical power system
Submission-ID 119
Holger Becker 1, Hendrik Sölter 1, Lutz Hofmann 1, 2
1 Fraunhofer Institute for Energy Economics and Energy System Technology (IEE), Germany
2 Institut for Electrical Energy Systems, Leibniz University Hannover, Germany
According to experience, system splits in the European electrical system happen every few years (as examples see Italy in 2003, Europe in 2006 or Turkey in 2015). Compared to normal operation, high frequency gradients occur due to spontaneous occurance of disturbances causing a system split and the power imbalances within the different network islands. The chance for surviving of these islands rises with a fast and reliable detection of the situation and an appropriate reaction of the generating units. Due to direct coupling of generator’s rotor and its mechanical inertia, synchronous machines provide instantaneous frequency support to the frequency, while inverter based generators do not do so automatically. Additionally, the grid frequency can be easily determined from generator’s speed of rotation, while inverters need to determine the frequency by analysing the voltage progress. In the event of a system split with different separated island grids, it is important to stabilize them by balancing active power prompt. Quick frequency estimation is important for fast adjustment of the feed-in power in relation to the frequency.

According to today’s grid codes in continental Europe, wind energy converter (WEC) currently do not provide fast frequency response functionalities. With an increasing proportion of WEC, it is important for grid stabilization during the system split that these generation plants also contribute frequency support rapidly. For being able to do so, WEC control unit need to detect high frequency gradients while being robust against local disturbances like phase angle jumps. Different frequency estimation methods like Zero Crossing, Clarke-Transformation (αβ-transformation) and discrete Fourier-Transformation based methods are described in detail and analysed by investigation of synthetic test signals showing voltage progress during different realistic and extreme situations. The created synthetic test signals represent voltage progress during a system split as well as during other disturbance phenomena. The goal of the frequency estimation algorithm is to detect a system split reliable by being robust against other disturbances like phase angle jumps caused by line tripping or short circuits. In comparison to the Zero Crossing method and the Clarke-Transformation it has been shown, that the discrete Fourier-Transformation is a robust and reliable frequency estimation method under realistic conditions. In contrast to the other methods, the discrete Fourier-Transformation is robust against harmonics and noise. With this method it is possible to clearly distinguish a frequency gradient from a phase jump. Under ideal conditions and without any measuring filters, the Clarke-Transformation could be an appropriate estimation method in consideration of speed and measuring error.

Reducing operational costs of offshore HVDC energy export systems through optimized maintenance
Submission-ID 121
Jan Frederick Unnewehr 1, Hans-Peter (Igor) Waldl 2, Thomas Pahlke 2, Iván Herráez 3
1 Department of Sustainable Systems Engineering, University of Freiburg, Freiburg im Breisgau, Germany
2 Overspeed GmbH & CO. KG, Oldenburg, Germany
3 University of Applied Science Emden/Leer, Emden, Germany
For the grid connection of offshore wind farms, today, in most cases an HVDC connection to the shore is implemented. The scheduled maintenance of the offshore and onshore HVDC stations makes up a significant part of the operational costs of the connected wind farms. The main factor for the maintenance cost is the lost income from the missing energy export. In this study, we show an in-depth analysis of the used components, maintenance cycles, maintenance work and the risks assigned to prolong the maintenance cycle of the Modular Multilevel Converter (MMC). In addition, we investigate the potential, to shift the start date of the maintenance work based on a forecast of the energy production. The results presented in this contribution indicate that an optimized maintenance of an HVDC-export-systems can decrease the maintenance-related losses for an offshore wind farm to almost the half.

Japan's R&D Project of Ramp Forecasting Technology: Correction method with additive model for NWP-based wind speed forecast
Submission-ID 122
Takamitsu Araki, Ryosaku Ikeda, Van Q. Doan, Noriko N. Ishizaki, Hiroyuki Kusaka
Center for Computational Sciences, University of Tsukuba, Japan
To further spread the wind generation, an accurate wind forecast is an essential process. To improve the accuracy of a short term wind forecasts given by a numerical weather prediction (NWP) model, nonlinear regression methods using observed wind speed data and the NWP outputs have been applied. In this study, we introduce an additive model based on B-spline basis expansion into the NWP wind forecast correction, which consists of nonlinear and flexible coordinate regression functions of input variables. We also estimate parameters of the functions using the maximum penalized likelihood method to obtain smooth and flexible regression functions. We compared the proposed method with two nonlinear regression methods that have been widely used, an artificial neural network (ANN) and analog ensemble (AnEn) method through the correction of the wind forecasts of the Weather Research and Forecasting (WRF) model. As the results, our method corrected the WRF wind speed forecasts more accurately than the ANN and AnEn method and revealed the relationships between the WRF wind speed forecasts and the observed wind speed. The relationship was not able to be obtained by the ANN and AnEn method.

Method for Harmonic and TOV Connection Impact Assessment of Offshore Wind Power Plants - Part I: Harmonic distortion
Submission-ID 124
Richard de Groot 1, Frans van Erp 1, Kees Jansen 1, Jeroen van Waes 1, Manon Hap 2, Luc Thielman 2
1 TenneT TSO BV, Netherlands
2 Tractebel Engineering, Belgium
This paper presents the method used by TenneT, the Dutch transmission system operator, to assess the risk of high harmonic distortion due to amplification of background distortion (Part I) and high temporary overvoltages (Part II) in transmission networks that connect (large) offshore wind power plants to the onshore transmission system using long high-voltage AC cables. Studies have shown that the impact of the wind power plant, in particular the capacitance of the long high-voltage AC cable, on the resonance frequencies of the post‑connection network is a potential source for amplification of the pre‑connection harmonic distortion observed at the point of common coupling. This could cause harmonic distortion to exceed compatibility levels, forming a potential threat against the system’s reliability and normal operation. The policy, assessment method and a case study are presented, including the design of a harmonic passive filter for mitigating the risk of too high harmonic distortion.

Keywords-harmonic analysis, harmonic passive filter, connection assessment, offshore wind power, power grid

Analysis and mitigation of storm and ramping risks from offshore wind power in Belgium
Submission-ID 125
Ruben Baetens 1, Simon De Clerq 1, Achim Woyte 1, Carlos Guerrero 1
1 3E, Belgium
2 Elia, Belgium
Once the current concessions will be operational, the total installed capacity of wind power in the Belgian exclusive economic zone will increase from 0.7 GW in 2017 to 2.3 GW in 2020 (and up to 4 GW before 2030), this in an area of only 3447 km² and a control area with a peak load of 13.4 GW and a low of 5.8 GW. By then, unexpected variations in wind power generation may caused by high wind speeds or sudden changes in wind speed or direction could trigger a substantial imbalance in the control area.

To understand the possible impact of such events, ELIA and 3E assessed the observed events of 2015 and 2016, and extrapolated them towards the expected installed capacity in 2020.

It is concluded that by 2020 and in the most realistic scenarios, the power loss caused by a storm event (i.e. cut-out) often goes beyond 1000 MW, while a major event with wind speeds above 30 m/s will always cause a power deviation of more than 2000 MW. In addition, when looking at the maximal ramps observed in both cut-out and cut-in phases, it is observed that deviations around 1000 MW can happen in both directions (up and down) within a period of 30 minutes. Also ramping becomes more significant: power variation of 150 MW within 15 min will happen during approximately 3% of all 15-min time steps.

Method for Harmonic and TOV Connection Impact Assessment of Offshore Wind Power Plants – Part II: TOV Impact Assessment
Submission-ID 129
Kees Jansen 1, Richard de Groot 1, Bart van Hulst 1, Konstantinos Velitsikakis 2, Christiaan Engelbrecht 2
1 TenneT TSO BV, Netherlands
2 DNV GL, Netherlands
Various methods exist to assess the risk of high harmonic distortion due to amplification of background distortion (Part I) and high temporary overvoltages (Part II) in transmission networks that connect (large) offshore wind power plants to the onshore transmission system using long HVAC power cables. Studies have shown that resonance frequencies between the 2nd and 3rd harmonic might occur at the Point-of-Common-Coupling (PCC). Transient conditions around the PCC, either scheduled switching events (e.g. energizing a power transformer) or fault conditions, could be a potential source for exciting these resonances. Once excited, the resonances might result in Temporary Overvoltages (TOVs), which could form a potential threat against the system’s reliability and normal operation. This paper presents the assessment method for TOVs used by TenneT TSO BV, the Dutch Transmission System Operator (hereafter TenneT). The method is based on the evaluation of simulation results, as obtained via detailed Electromagnetic Transient (EMT) studies and its outcome indicates whether remedial measures against excessive TOVs need to be further considered. Last but not least, the results of a case study are presented, including the application and effectiveness of harmonic passive filters for mitigating TOVs.

Keywords / Indexing terms

Temporary overvoltages, TOV impact assessment, harmonic passive filter, C-type filter, connection assessment, offshore wind power, power grid

Submission-ID 133
Corinna Möhrlen
WEPROG, Denmark
In this paper, the progress of the first part of a recommended practices guideline “Forecast Solution Selection Process” is presented that deals with the selection and background information necessary to collect and evaluate when developing or renewing a wind power forecasting solution for the power market. This is part of the IEA Wind Task 36 project 2016-2018. The work aims to provide a series of three recommended practices that deal with the development and operation of forecasting solutions of wind (and solar power) in the power market.

The effectiveness of forecasts in reducing the variability management costs of power generation from wind and solar resources is dependent upon both the accuracy of the forecasts and the ability to effectively use the forecast information in the operational decision-making processes. With increasing amounts of forecasting methods and vendors, it has become more difficult to obtain forecasts of high quality with a fit-for-purpose that can effectively be used as input to operational processes in system operation, trading, market management, unit commitment etc. The recommended practices guideline is intended to provide guidance to forecast users who are seeking a forecasting solution that fits their purpose and enables them to work efficient and economically responsible. In this paper we summarize some of the important aspects in this respect from the document under review and explain, how the decision support tool to establish procedures for the selection process, can be applied.

Lyapunov-Based Control for Grid Side Inverters of Wind Turbine Systems
Submission-ID 134
Alexander Schöley, Magdalena Gierschner, Wolfgang Drewelow, Torsten Jeinsch
University of Rostock, Germany
The control of the grid side inverter of a wind turbine system is addressed in this paper. The main objective of the grid side converter is to feed an appropriate amount of electrical power into the grid in order to keep the DC-link voltage constant. At first, a nonlinear model of the inverter is developed in the ABC-frame consisting of the DC-link circuit, the LCL output filter and the phase-locked loop (PLL). Then a control strategy is proposed that is based on the Lyapunov stability theory. Contrary to the well-known VOC control scheme, the new strategy does not rely on Park's Transformation. Two superordinated PI-controllers are added to the Lyapunov control law to track the reference values for active and reactive power as well as to cope for uncertainties. The designed controller is tested in simulations with varying references and measurement noise.

Meteorological Categorization of Wind Power Ramp Events - Case Study of Three Areas of Japan -
Submission-ID 136
Maki OKADA 1, Koji YAMAGUCHI 1, Ryo KODAMA 1, Norimitsu OGASAWARA 1, Kazuhiko OGIMOTO 2
1 Japan Weather Association, Japan
2 University of Tokyo, Japan
Although renewable energies such as wind power continue to spread globally, electrical output from wind power fluctuates constantly with wind changes. Thus, the installation of large wind power capacities tends to result in negative impacts on grid stability. In particular, sudden wind power changes (wind power ramps) have significant effects, and making the forecasting of wind power ramps is one of the major issues in wind power usage. Since a wind power ramp is most commonly associated with abrupt wind changes, knowledge about the connections between power output changes and weather conditions has the potential to contribute to the development of wind power ramp forecasting. Accordingly, this study meteorologically categorized wind power ramp events occurring in Japan. Occurrence tendency of ramp events and weather patterns that cause ramp events will be introduced. We used total power output, which interval was 10-s, in three electrical power areas (Hokkaido, Tohoku, and Tokyo) collected by a project of the New Energy and Industrial Technology Development Organization (NEDO). The total power output data were normalized by the total output capacity (PU) because the total output for each area were different. A wind power ramp was defined as a 30% or more fluctuation in the total output capacity between the start and end times, and an average fluctuation rate of 5% / h or more during that time period [Kataoka et al., 2013, Proc. of 2013 Annual Conference of Power & Energy Society, IEE Japan]. Weather patterns for each power areas in 2016 was categorized by weather forecasters, who focused on the track, location, and strength of low- and high-pressure systems in the surface weather chart. For example, a weather pattern similar to “winter monsoon pattern” occurred frequently in Hokkaido from October to April. Here, “winter monsoon pattern” means a distribution of sea level pressure in which the high-pressure system is to the west and the low-pressure system is to the east of Japan. In this categorization type, as the low-pressure system developed and was moving to the east of the Hokkaido, the interval between constant-pressure bars narrowed. Then, the north wind gradually strengthened, and resulted in a sudden increase of wind power (ramp-up). Other categorized results will be presented.

Japan's R&D Project of Ramp Forecasting Technology: Metrics for Evaluating Ramp Forecast
Submission-ID 140
Kenji Yoshida 1, Nobuyuki Hayasaki 1, Naoto Ushigami 1, Norimitsu Ogasawara 2, Maki Okada 2, Yuko Nakamura 2
1 ITOCHU Techno-Solutions Corporation, Japan
2 Japan Weather Association, Japan
In the R&D project of the New Energy and Industrial Technology Development Organization in Japan, the authors participate in the group focused on forecasting of wind generation ramps in eastern Japan. In the group, an enhanced technology to forecast wind power output ramps is being developed. To evaluate the ramp forecasting performance of developed technologies, we select various metrics suitable for each forecast output (wind power time series, prediction intervals, ramp forecast and ramp probability forecast). Critical Success Index (CSI) for ramp forecast is defined as the primary metric in the project. We conduct ramp forecasting by the existing and developed technologies over the past period and compare their ramp forecasting performance. The developed technologies show better ramp forecasting performance than the existing technologies.

North Sea offshore Grid development: Combined optimization of grid and generation investments towards 2050
Submission-ID 142
Matti Koivisto 1, Juan Gea-Bermudez 2, Poul Sørensen 1
1 DTU Wind Energy, Denmark
2 DTU Management Engineering, Denmark
The North Sea area has already significant offshore wind generation, and it offers large amounts of investable generation with high capacity factors (CFs) in the future. In addition, several transmission lines are located in the area. This paper models optimal transmission and generation investments in scenarios towards 2050, and investigates the viability of connecting future transmission and offshore wind generation investments to create an offshore grid.

A baseline scenario is created using the traditional solution of connecting offshore wind power plants (OWPPs) directly to shore and using country-to-country lines for transmission (radial case). Results from the integrated approach of combining offshore wind generation and transmission (meshed case) are then compared to the radial case. The modelling is carried out using the Balmorel energy system model. The countries with investment optimisation are Denmark, Norway, Germany, UK, Netherlands and Belgium. Surrounding countries participate in the electricity dispatch optimisation.

The benefits of a meshed North Sea offshore grid have been shown in previous research. In many studies, the generation investments were fixed parameters. This study contributes to the literature by allowing both generation and transmission investments to be optimized by Balmorel. In addition, scenario years 2030 and 2050 are optimized concurrently, which allows optimal investments in 2030 considering planned investments in the future.

To compare offshore wind to other renewables, onshore wind and solar photovoltaic (PV) investments are optimized simultaneously in the model. Other generation types are also part of the optimization, but only some gas turbine investments are seen necessary as back-up capacity.

DTU Wind Energy’s CorRES tool is used to simulate the wind and solar PV generation time series used in Balmorel modelling. CorRES models the varying CFs depending on installation locations, and the spatiotemporal dependencies in VRE generation. Especially offshore wind is modelled in detail, starting from the planned locations of individual OWPPs.

All analysed scenarios show significant transmission investments, with increased connection of Norwegian hydro generation to the other countries. The results indicate that going to a meshed solution can increase the total offshore wind investments by several GW. In addition to offshore wind, the scenarios include large amounts of onshore wind and solar PV. A large share of country-to-country transmission is provided by radial transmission lines also in the meshed case. However, the presented analyses show that integrating the offshore hubs as part of the transmission system can be beneficial.

Modeling the dynamics and control of power systems with high share of renewable energies
Submission-ID 143
Sabine Auer, Tim Kittel
Potsdam Institute for Climate Impact Research (PIK), Germany
A challenge for renewable and hybrid power systems is the dynamically stable integration of Renewable Energy Sources (RES). This paper specifically investigates the influence of intermittent RES and measurement delays from power electronic resources on frequency stability. In addition it presents an Open-Source framework to undertake dynamic RES modeling.

First, for local intermittent fluctuations in lossy distribution grids we find a remarkable and subtle but robust interplay of dynamical and topological properties, which is largely absent for lossless grids.

Second, we show how delays from measurement and reaction times of power electronic devices may induce resonance catastrophes in power networks.

Third, the above research topics illustrate the necessity to transition to a new open-source software framework for dynamic power grid modeling. For this we want to present PowerDynamics.jl, which is in the process of being published in the programming language Julia. It will cover the rich novel dynamics caused by the integration of RES.

Altogether, this paper investigates the stability of future power grids moving towards integrating more aspects of renewable energy dynamics and presents an adequate modeling framework for RES integration studies.

Onshore Grid Frequency Control using DC Capacitor in Full-Scale Converter for Offshore Wind Generator and Adjustable Speed Motor for Offshore Plant Connected by Multi-Terminal HVDC
Submission-ID 144
Hiroshi Matsuda, Yutaka Ota, Tatsuhito Nakajima
Tokyo City University, Japan
Renewable energy resources (RERs) are being widely introduced recently. A number of offshore wind farms are being commissioned in the world. R&Ds on multi-terminal HVDC (MT-HVDC) links for connecting between onshore power grids and offshore wind farms are being conducted in research projects. The offshore MT-HVDC will be applied not only for power transfer from the wind farms, but also for power supply to offshore plants such as oil or natural gas platforms. With increase in the RERs, on the other hand, conventional thermal power stations using synchronous generators will be out of operation in the future. Power system inertia will be decreased, and that will make the grid frequency more fluctuate consequently. Effective countermeasures must be established for keeping power grid stability.

This paper proposes a new control method for supplementary supporting the onshore grid frequency, using both the dc capacitor in a full-scale converter (FSC) for the offshore wind generator, and an adjustable-speed induction motor drive (ASMD) for the offshore plant connected to the HVDC link. In the proposed method, when a disturbance occurs in the onshore grid, in accordance to the frequency signal transmitted from the onshore grid, the generated power of wind generator system is controlled by making the FSC dc capacitor charge or discharge. The consumed power of the ASMD is also controlled by changing the motor speed using the motor-driving inverter.

In the FSC for the wind generator, a frequency droop for the grid frequency deviation is added to the dc voltage control block. When the grid frequency is decreased, the dc voltage is changed slightly lower, and the dc capacitor is discharged for the frequency support. In the ASMD system, the grid frequency deviation is converted to a change in the motor consuming power reference, based on the grid power frequency constant. It is assumed in the paper that the motor has a pump load, of which torque is proportional to the square of the motor speed, and thus the motor consuming power is approximately proportional to the third power of the motor speed. The required motor speed and the required driving-inverter frequency are obtained. When the grid frequency is decreased, the motor speed is regulated slightly lower, and the motor inertia energy is released for the frequency support.

The proposed control method was validated by PSCAD simulation case studies. The simulation model consists of a four-machine grid, a three-terminal HVDC link, one-machine wind generator with one FSC, and a ASMD with a inverter. The grid disturbance is made by connecting a load to the grid. The simulation results show that, during the grid disturbance, power is released or absorbed by both the dc capacitor and induction motor, without any control interference problem, through HVDC. The results also prove that the grid frequency nadir is improved to be higher, and the oscillated grid frequency is damped effectively.

A dispatch methodology to secure power system inertia in future power systems
Submission-ID 146
Henning Thiesen, Clemens Jauch
Flensburg University of Applied Sciences - Wind Energy Technology Institute, Germany
Power system inertia is a fundamental part of grid frequency control. Present power markets do not take account of a sufficiently high amount of power system inertia. In this paper an extension for the conventional economical dispatch is presented, which can be applied once the power system falls of a certain level of power system inertia. Non-inertia providing generators of the merit order are replaced with synchronous generators. A new market equilibrium is achieved under the condition of sufficient inertia. Additionally, this novel approach allows for a quantification of system costs for additional inertia due to the application of the market extension. Certain impacts on the instrument like external costs, inertia from loads and synthetic inertia are discussed.

Analysis of Power System Oscillation Stability with Large Integration of Renewable Generations
Submission-ID 150
Lijun Cai 1, Harald Weber 1, Hans-Günter Eckel 1, Ulas Karaagac 2, Jean Mahseredjian 3
1 Institute of Electrical Power Engineering, University of Rostock, Germany
2 Department of Electrical Engineering, Hong Kong Polytechnic University, Hong Kong
3 Department of Electrical Engineering, Polytechnique Montréal, Canada
Key words: renewable generation; electro-mechanical oscillation; angle stability.


Electromechanical oscillations in power systems are classified by the system components that they affect and they are mainly the result of the interaction of generation units. Different modes of electromechanical oscillations could occur in power systems: i.e. intra-plant mode, local mode, inter-area mode, control model and torsional mode. The oscillation modes are strongly dependent on the characteristics of the generation units, the network topology and the power flow scenario. To ensure stabile operation of power systems, oscillation modes must present an acceptable damping, so that those oscillations are well damped after small perturbations. The classical method for analyzing oscillation stability is the modal analysis and it is well approved in the large power systems with synchronous generators.

Recently, with the rapid development of renewable energies in the power systems, converters that interface with renewable generations are becoming more popular. Due to the new characteristics of these renewable generations, power systems will experience changes in oscillatory dynamics due to the following reasons:

- Changes in the architecture of transmission system to connect wind farms;

- Replacement part of the synchronous generators by large renewable generations;

- Alteration in the dispatch of synchronous generators in order to meet the strongly varying renewable generations and

- The reduction of the power system inertia due to the large amounts of renewable generations.

All these changes will result in the change of the synchronizing and damping behavior of the power system. Furthermore, the modal analysis could not be directly applicable to the renewable generations, since these generations partially (i.e. doubly-fed induction generator wind turbines) or fully (i.e. full converter wind turbines) decouple the mechanical side of generators to the grid by the converters. Therefore, the dynamic interaction between renewable generations and synchronous generators should be considered in detail.

This paper analyzes electro-mechanical oscillations in a power system with large integration of renewable generations. After system disturbances, oscillatory modes are analyzed. Also, the main factors that influence the damping of the oscillation modes are given.

Furthermore, different scenarios are considered (i.e. different active and reactive power exchange between wind farms and synchronous generator) to analyze their effects on oscillatory behavior.

Finally, measures for damping of the electro-mechanical oscillations will be given.

Understanding Uncertainty: the difficult move from a deterministic to a probabilistic world
Submission-ID 155
Corinna Möhrlen 1, Ricardo Bessa 2
1 WEPROG, Germany
2 INESC TEC, Germany
Forecast methodologies have advanced over the past 20 years along side the needs of system operation and trading of energy at the power exchange markets. Like every discipline in development, also forecasting of renewable generation has evolved disruptive and chaotic at times when new ways to handle these variable sources were sought.

The forecasts of these sources inherit an uncertainty in their operation due to the uncertainty of the underlying weather forecast. Once these uncertainties are understood the future outcome at the time scale required to operate our electric grids and trade the energy on our power exchanges can be forecasted much more efficient than with deterministic methods. Uncertainty forecasts are filling a gap of information missing in deterministic approaches and are gradually moving into the control rooms and trading floors.

Nevertheless, there are a number of barriers in the industrial adaptation of uncertainty forecasts that have their root in a lack of understanding of the methodologies and their respective applicability. There is a complication level that needs to be overcome in order to move forward. The IEA Wind Task 36 has been carrying out a number of expert round discussions picking up a number of the loose ends of integration and application issues. The applications presently used in industry, suggestions how to apply and integrate uncertainty forecasts into operation and an outlook from this discussion are presented and discussed in this paper.

Japan's R&D Project of Ramp Forecasting Technology: A Machine Learning Scheme for Ramp Forecast
Submission-ID 156
Yu Fujimoto, Kazutoshi Higashiyama, Yasuhiro Hayashi
Waseda University, Japan
In Japan, the R&D project “Grid Integration of Variable Renewable Energy: Mitigation Technologies on Output Fluctuations of Renewable Energy Generations in Power Grid” promoted by the New Energy and Industrial Technology Development Organization (NEDO) has been started since 2014. Forecast of ramp events in wind generation is one of important topics of technical development in this project; this topic aims to break through difficulties in forecast of largescale changes within a short time in wind generation. The authors study an applicability of machine learning methodology to the ramp forecast of wind generation and implemented a hierarchical forecast scheme composed of the wind power predictor and the ramp event predictor in this project. This paper shows supervised setups for learning predictors and discuss the difficulties in these learning schemes.

Submission-ID 160
Corinna Möhrlen 1, Jeffrey Lerner 2, Jethro Browell 4, Craig Collier 7, Gregor Giebel 8, John Zack 6, Jakob Messner 3, Aidan Tuohy 5
1 WEPROG, Denmark
2 Vaisala, United States
3 Anemos Analytics, Denmark
4 University of Strathclyde, United Kingdom
5 EPRI, United States
6 AWS Truepower, United States
7 DNVGL, United States
8 DTU Wind Energy, Denmark
In this paper, we summarize the second and third part of a series of three recommended practice documents for the power industry that deal with how to setup and run a trial or benchmark as well as verifying the goodness of forecast solutions.

The Recommended Practice is intended to serve as a set of standards that provide guidance for private industry, academics and government for the process of obtaining an optimal forecast solution for specific applications as well as the ongoing evaluation of the performance of the solution to increase the probability that it continues to be an optimal solution as forecast technology evolves. The work is part of the IEA Wind Task 36 on Wind Power Forecasting.

The guideline provides an overview of the factors that should be addressed when conducting a benchmark or trial and present the key issues that should be considered in the design as well as describe the characteristics of a successful trial/benchmark. We also discuss how to execute an effective benchmark or trial and specify common pitfalls that a Forecast User should try to avoid.

Part 3 of the recommended practices guideline deals with the effective evaluation and verification of forecasting solutions, benchmarks and trials. The core of any effective evaluation and verification is ``fairness'', ``repeatability'' and ``representativeness.'' The evaluation paradigm is another aspect that needs consideration. Accuracy metrics need to be weighed versus the value of a solution, benefits of blended forecasts versus strategic forecasts, and how to verify complex solutions that feed into various processes inside an organisation. Recommendations on the design and execution of incentive schemes, their pros and cons for the development and improvement of forecast solutions is also part of the guideline and will be presented and discussed briefly.

On-line Markov Chain Based Thermal Risk Estimation for Offshore Wind Farm Cables
Submission-ID 163
Maria Angelica Hernandez Colin, James Pilgrim
Electronics and Computer Science (ECS), University of Southampton, United Kingdom
This paper proposes the use of a Markov Chain and Monte Carlo Analysis to estimate likely thermal risk of cable overheating in submarine export cables 6 hours ahead. The methodology can account for the uncertainty in the load current transferred by offshore cables which is the product of wind variations. The estimated conductor temperature and thermal risks calculated by the methodology shown a close resemblance to the real conductor temperature and thus a high percentage of successful thermal risk estimations 6 hours ahead of up 98% over one year of testing was estimated.

Analysis of the Flicker Estimation at PCC of a Wind Power Plant
Submission-ID 166
Koldo Redondo, Izaskun Azcarate, José Julio Gutierrez, Luis Alberto Leturiondo, Purificación Saiz
Department of Communications Engineering University of the Basque Country (UPV/EHU), Spain
In terms of power quality, the measurement of the flicker emission of a wind power plant is an important issue. The procedure to measure the flicker produce exclusively by a grid-connected wind turbine is specified in the IEC 61400-21 standard. The standard also proposes a summation method to estimate the flicker produced by a wind power plant at the point of common coupling. This work studies the reliability of the summation method, based on actual current and voltage waveforms recorded at a wind power plant in Spain. The results show that the summation method significantly overestimates the flicker emission of the wind power plant.

Japan’s R&D Project of Ramp Forecasting Technology: Project Overview
Submission-ID 168
Takahiro Suga 1, Nobuyuki Hayasaki 1, Kazuhiko Ogimoto 2
1 ITOCHU Techno-Solutions Corporation, Japan
2 The University of Tokyo, Japan
Renewable generation has been increasingly deployed under the global trend of reduction of CO2 emission. It is necessary to use more enhanced forecast and control technologies to integrate them in power systems operation planning and real-time operation in order to increase the use of renewable energy. This is the objective of the R&D project “Grid Integration of Variable Renewable Energy: Mitigation Technologies on Output Fluctuations of Renewable Energy Generations in Power Grid”, which has been going on since 2014. We aim to break through the important issue regarding the forecast error of wind power output: that it is too large to properly estimate up/down ramps within short time. This project focused on wind power ramp forecasting have not tackled yet even though it is important issue to forecast wind power ramp phenomena. This paper shows an overview of the Project.

Grid Forming Control for Stable Power Systems with up to 100 % Inverter Based Generation: A Paradigm Scenario Using the IEEE 118-Bus System
Submission-ID 172
Mario Ndreko, Sven Rüberg, Wilhelm Winter
TenneT TSO GmbH, Germany
This paper presents a grid forming control for grid connected inverter based renewable-generation and grid connected storage units that could facilitate stable and robust operation of power systems with even up to 100 % inertia-less inverter based renewable generation. The proposed grid forming control directly regulates the voltage magnitude and the angle of the respective grid side inverter in the synchronous dq-reference frame. During severe three-phase grid faults, the control scheme provides current limitation capability and injection of reactive fault current using a modified current control block. EMT time domain simulations are performed on two different test systems. The first test case is on a simple inverter-based power plant connected to a small and weak grid (with very low short-circuit power ratio). The second test case is on the modified 118-bus system in which the penetration of inverter based generation connected to the 33-kV level may reach up to 100 %. For the latter case, grid faults are presented in order to demonstrate the stable response of the system. The paper concludes that bulk-transmission systems with high penetration of inverter based generation up to 100 % are stable when a portion of the inverter based active-power infeed units (renewable generation or grid connected storage) apply grid forming control scheme.

Optimal Allocation of Wind Power Considering its Contribution to Security of Supply
Submission-ID 178
Jakob Peter, Johannes Wagner
Institute of Energy Economics, Department of Economics, University of Cologne, Germany
Reliability of supply has due to its high economic value always been a major concern in electricity systems while ongoing climate change creates the need to switch to low-carbon electricity generation such as wind power with its fluctuating stochastically dependent electricity supply. Thus the following questions arise: What is the optimal allocation of wind capacity throughout Europe in order to benefit from balancing effects both in wind generation and equivalent firm capacity provision from wind, considering interconnection constraints between countries and resulting system effects?

Against this background, this paper builds on a new methodological approach which endogenously determines the contribution of wind power to security of supply in an optimization model for electricity markets. We deploy the capacity credit as a well-established measure to depict the contribution of fluctuating power generation to reliability of supply. In contrast to many existing modeling approaches, we use a capacity credit formulation which accounts for its dependence on the amount of installed wind capacity, its spatial distribution and the available interconnection capacity. Thereby, a more precise assessment of the contribution of wind power to security of supply is reached, allowing for an optimal system configuration to reach a required level of security of supply. In a next step, the proposed methodology builds on an iterative approach, which captures the non-linear dependency of the capacity credit of wind power on installed capacity and interconnection while keeping computational tractability in a large-scale application. We apply our methodology to the European electricity system to determine an optimal decarbonization pathway until 2050. We base the analysis on a new dataset which is based on meteorological reanalysis data and has a high spatial and temporal resolution, capturing the stochastic properties of wind power generation.

The analysis shows that wind power can substantially contribute to security of supply in a decarbonized European electricity system, with regional capacity credits ranging from 1 - 40%. Additionally the results show that the capacity credit of wind power depends on the specific wind properties in a country or region as well as the installed capacity of wind power and available interconnections to neighbouring countries. Consequently the capacity credit is heterogeneous across different regions and years. Existing modeling approaches, which typically assign constant values for the capacity credit of wind power therefore over- or underestimate back-up capacities, which are required to guarantee security of supply in an electricity systems with high shares of wind power. Especially if the potential of wind power to provide secure capacity is neglected in simulations, an inefficient allocation of wind capacity arises, leading to an overestimation of required back-up capacities and total system costs.

Energy Control of Modular Multilevel Converters in MTDC Grids for Wind Power Integration
Submission-ID 179
Kosei Shinoda 1, Ragavendran Ramachandran 1, Abdelkrim Benchaib 1, Jing Dai 1, 2, Bruno François 3, Seddik Bacha 1, 4, Xavier Guillaud 3
1 SuperGrid Institute SAS, France
2 Group of Electrical Engineering - Paris (GeePs), UMR CNRS 8507, CentraleSupélec, Univ. Paris-Sud, Université Paris-Saclay, Sorbonne Universites, UPMC Univ. Paris 06, France
3 Universite Lille, Centrale Lille, Arts et Metiers, HEI EA 2697 - L2EP, France
4 Universite Grenoble Alpes, CNRS, Grenoble INP, G2Elab, France

The role of Modular Multilevel Converters (MMCs) in HVDC grid greatly differs depending on whether it is an offshore or an onshore station. From the common point in their control schemes, an unexploited ability of the MMC—the controllability of the internally stored energy—is identified in both offshore and onshore applications. The virtual capacitor control, previously proposed by the authors, makes use of this degree of freedom to provide energy contribution to the DC grid. The impact of this control is demonstrated by time-domain simulations of a five-terminal HVDC grid.

Development of Wind Ramp Forecasting Technology in the National R&D Project (in Japan): Evaluation of Developed Forecasts by Power System Operation Simulation
Submission-ID 180
Yuki Nishitsuji 1, 2, Yusuke Udagawa 1, 2, Kazuhiko Ogimoto 1, Katsuyuki Ukegawa 2, Suguru Fukutome 3
1 the University of Tokyo, Japan
3 JP Business Service Corporation, Japan
Renewable Energy shows large grow in the global stream of reduction on CO2 emission. In Japan, a lot of deployment of renewable energy facilities has continued, boosted by the disastrous earthquake in 2011 damaging nuclear power plants, and FIT begun in 2012. Although penetration is dominated by photovoltaics at present, it is expected that wind power will raise its impact with the completion of environmental assessment and construction.

To ensure the security of efficient power system operation with high penetration of variable renewable energy, it is vital to enhance power generation forecast and control technology.

To tackle with this task, Japanese R&D project “Grid Integration of Variable Renewable Energy: Mitigation Technologies on Output Fluctuations of Renewable Energy Generations in Power Grid” was started in 2014. This project has four working groups, consisting of the group for development of forecast on wind power generation and on occurrence of steep ramp up/down of its output, mitigation technologies for power output fluctuation by energy storage, power supply-demand balancing simulation, and field experiment of developed technologies. This paper is written for reporting forecast evaluation using power system operation simulation in the working group of forecast development.

Developed forecasts are required to be beneficial to power system operation. To evaluate those forecasts from this viewpoint, we developed a simulation model of advanced power system operation which can utilize developed “ramp alert”, which is a binary signal, with developed wind power output forecast. Ramp alert can inform power system operators of the possibility of large total wind power output fluctuation exceeding a threshold in a balancing area. In addition to the conventional constraints such as supply-demand balance and reserve requirement for load frequency control (secondary reserve), our simulation has some constraints that require reserves to address steep ramp up/down of wind power output, when ramp alert is active.

Using this simulation, the investigation into the impact of developed ramp alerts and wind power output forecasts on power system operations is conducted, not by statistical criteria such as Root Mean Square Error, but by the economical indices such as operational costs and indices for supply security such as energy generated by the last resort generator. The details of simulation results, which describe how each generator deals with wind power forecast error, are also analyzed. Furthermore, we discuss the characteristics and qualities of developed ramp alerts and wind power output forecast associated with power system operations.

Experimental Results of a Wind Power Plant Scheduling Method Considering State-of-Charge Transition for an Electricity Market with the Compressed Air Energy Storage System
Submission-ID 181
Masakazu Ito 1, Aki Kikuchi 2, Yu Fujimoto 1, Masataka Mitsuoka 1, Hideo Ishii 1, Yasuhiro Hayashi 2
1 Advanced Collaborative Research Organization for Smart Society (ACROSS), Waseda University, Japan
2 Dept. of Electrical Engineering and Bioscience, Waseda University, Japan
Wind power plant (WPP) operators may have to consider selling electricity at an electricity market when the feed-in-tariff (FIT) term finish or the law of the FIT finish. In this situation, WPP operators may need to introduce energy storage systems (ESSs) to control amount of electricity to sell at the electricity market. The authors proposed the state-of-charge (SOC) control considering SOC transition (SCST) previously which enhance controllability of the SOC, and effectiveness of the ESS increase. To verify the effectiveness of the SCST, it has been mounted on an actual ESS. It is the Kawazu compressed air energy storage (CAES) system test site installed in Higashi-Izu beside of an actual WPP which is in the south of Tokyo. Then the experiment assuming an electricity market had been executed by using actual wind power output and the CAES system. As a result, imbalance between generation schedule and total output which is wind power output and CAES power output was less than 0.5% of the generation schedule while SOC kept between 0.0 to 1.0. Furthermore, post-experiment simulation had been executed to compare effectiveness of operation with and without the proposed method by using measured wind power output and wind generation forecast (WGF). In case of a variable wind speed day, imbalance without the proposed method was 6.9%, however the imbalance with the proposed SCST reduced to 6.0%. The result shows that the SCST can reduce imbalance and WPP operators can reduce imbalance fee, in consequence, the SCST increase WPP operator’s profit.

Wind Farm Fault Ride Through – An Irish Context
Submission-ID 187
Aaron McDonnell, John Kelleher, Stephen Hunt, John Whelan
ESB International, Ireland
In 2017 Ireland installed 426 MW of new wind turbines and had an average power consumption of 3 GW. This made Ireland the European state with the highest level of installed wind capacity relative to its power consumption. Ireland had an annual penetration rate of 26 % from wind in 2017 and has a target of 40 % by 2020. Furthermore Ireland is trialling a System Non-Synchronous Penetration (SNSP) limit of 65 %. This has resulted in an increased focus on the ability of wind farms to contribute to system stability during system disturbances such as faults or other step changes in voltage. In this paper the key Fault Ride Through (FRT) requirements specified in the Irish Grid Code are outlined. The modelled performance of four wind farms connected to the Irish distribution system are compared against the clauses relevant to the FRT requirements. The wind farms in question comprise of turbines with different technologies and from different manufacturers, allowing for a comprehensive overview of the common clauses where compliance may not be fully achieved.

Benchmark of Spatio-temporal Shortest-Term Wind Power Forecast Models
Submission-ID 190
Stephan Vogt, Axel Braun, Jan Dobschinski, Dominik Jost, Jonas Koch
Fraunhofer Institute for Energy Economics and Energy System Technology IEE, Germany
Many European energy supply systems are increasingly penetrated by wind energy. In order to be able to act optimally on the market or in the operation of electricity grids, it is necessary to have high-quality intraday forecasts of the expected wind power production. For this purpose, numerical weather forecasts and recent power measurements transmitted in real time are used. This provides a lot of information to the forecaster. It is, on the one hand, necessary to be able to decide which information are beneficial, and on the other hand, to be able to handle proper forecasting models. Suitable models to calculate wind power forecasts are power curve-based models and models from the field of statistics as well as machine learning models.

In this work we benchmark (first) different models ranging from power curve based to machine learning like random forests, artificial neural networks and extreme learning machines, and (second) the value of spatio-temporal information from surrounding wind parks .

A Machine Learning Approach to Low System Strength Grid Identification for Large Scale Power Systems
Submission-ID 191
Angel Clark 1, 2, Yang Zhang 3, Shun Hsien Huang 3, Le Xie 2
1 KTH, Sweden
2 Texas A&M University, United States
3 ERCOT, United States
This research proposes a screening method to identify low system strength, i.e. weak grid with low short circuit current level, portions of power systems where potential voltage instabilities could occur in a system with high penetration of inverter-based resources (IBRs), like wind and solar. The proposed method uses a random forest algorithm by clustering the transmission buses with interconnected IBRs based on features extracted from the short circuit current and electrical distance. This proposed screening method can be used as a screening tool to identify the system strength based on short circuit current and connected IBRs’ capacity without computationally intensive dynamic simulations. Therefore, the impact of system outages and different system scenarios can quickly be analyzed, with the resulting areas identified and quantified in each scenario. The developed method was applied to an ERCOT case under a developed long-term system condition and a Synthetic Texas Network, demonstrating the robustness of the screening tool and ability to identify the areas with low system strength challenges.

How to combine state-of-the-art multi-scale numerical wind power forecasts and benefits of a human meteorological expertise?
Submission-ID 197
Olivier VANNIER, Aurélien BEN DAOUD, Alexandre FALGON, Guillaume BONTRON
Compagnie Nationale du Rhône, France
As France’s leading producer of energy exclusively generated from renewable sources, Compagnie Nationale du Rhône (CNR) has the need to forecast intermittent energy production for its own assets as well as for those belonging to external producers. CNR has benefited from human expertise on hydrometeorological forecasts for a long time, and has now to face the challenge of keeping using its meteorological knowledge for wind power forecasting, in a context of constantly growing production capacity. This study presents the results of a live trial performed over three months, aiming at evaluating a potential new method of expertised wind power forecasts, based on daily confidence indices given by CNR meteorologists to different Numerical Weather Prediction (NWP) models for the meteorological situation at hand. During these three months, the method was not found to give better performances than automatic forecasts, but such results need to be confirmed on a longer evaluation period, with normal wind conditions. These results may possibly also highlight a potential expertise bias that would need to be further post-treated.

Type IV Wind Turbine System Impedance Modelling for Harmonic Analysis: On the Use of a Double Synchronous Reference Frame and Notch Filter
Submission-ID 198
Lucia Beloqui Larumbe, Zian Qin, Pavol Bauer
Delft University of Technology, Netherlands
Several efforts are being done nowadays to improve the modelling of Wind Turbine Systems (WTSs) for harmonic analysis in Offshore Wind Power Plants (OWPPs). Due to the high influence of the different control structures in the Power Electronic Converters (PECs) on the dynamic response of a WTS, each structure needs to be modelled specifically.

Following these lines, this paper addresses the impedance part of the harmonic model in the case of a double Synchronous Reference Frame control structure. The main focus lies on the correct modelling of one of the main elements of this structure: the notch filter tuned at twice the fundamental frequency.

The inclusion or disregard of this notch filter is very important because, as shown in works by other authors, this filter can have a big influence in the shaping of the WTS output impedance and such of the OWPP. However, the modelling procedure for the notch filter followed previously ignores the cross-couplings that this element creates in the αβ frame, which leads to a wrong calculation of the WTS impedance in the lower frequency range.

The proper modelling of this notch filter and its implications are detailed in the paper first theoretically and then by numerical simulations.

Model-based Control of grid-side Converter: An LMI approach
Submission-ID 202
Horst Schulte, Nico Goldschmidt
University of Applied Sciences Berlin (HTW Berlin), Germany
This paper presents the basic approach of a generic method for the robust control of grid-side converters. For this purpose, the concept of an LMI-based representation of stability regions with pole specifications is used. A special feature of this approach is that the desired closed loop behavior is determined by the selection of a pole region. The pole regions are formulated via constraints as linear matrix inequality. The solving of this LMI guaranteed the stable solution of the control design.

As a model basis, a grid-side converter is represented in d-q coordinates. The resulting non-linear state space model is converted into a linear model structure with Taylor expansion. The linear model is the design basis for the pole region LMI approach. In order to decouple the influence of the disturbances on the control loop dynamics, suitable control laws are presented.

The design is supported and validated by simulation studies based on the various types of model uncertainties. Finally, the controller performance is compared with state-of-the-art d-q Inverter control using PI controller. The improved robustness and performance properties compared to the standard approach are presented.

Analysis of HVDC and Wind Turbine Converters Response during Offshore Asymmetrical Faults
Submission-ID 229
Ömer Göksu, Nicolaos Cutululis, Poul Sørensen
DTU Wind Energy, Denmark
One of the main short-circuit analysis for the HVDC connected offshore wind power plants is the asymmetrical faults at the offshore network between the offshore HVDC converter and the wind turbines. There have been few requirements specified for these offshore asymmetrical faults, e.g. in the recent ENTSO-E HVDC network code, and few studies in the literature, focusing on the control algorithms for the HVDC converter and the wind turbines during asymmetrical faults. However, a fundamental theoretical analysis focusing on the underlying physics behind the offshore asymmetrical faults has been missing. In this paper, physical phenomenon occurring in the offshore network with the HVDC converter and the wind turbine converters during asymmetrical faults has been explored. It has been showed that the negative sequence fault current flow via the HVDC converter is strictly necessary and there is an interaction between wind turbine positive sequence fault current and the HVDC negative sequence current. The classical fault analysis approach has been utilized and complemented with time-domain simulations. The output provides understanding for the offshore asymmetrical faults and insights for related future grid code requirements and converter control design.

Mixed grid forming and grid following wind power plants for black start operation
Submission-ID 232
Jaime Martinez-Turegano 1, Salvador Añó-Villalba 1, Soledad Bernal-Perez 1, Ruben Peña 2, Ramon Blasco-Gimenez 1
1 Universitat Politecnica de Valencia, Spain
2 Universidad de Concepcion, Chile
Grid forming wind turbines allow for additional services to be provided by wind power plants. These additional services include self-start operation, islanding operation of off- shore ac-grids and contribution to system restoration after a black-out.

As the complete electric system will consist of both grid-following and grid-forming wind power plants, the aim of this paper is to study possible interactions between grid forming and grid following wind turbines during black-start operation.

The presented case study includes a 400 MW off-shore wind power plant connected by means of a 75km HVac cable to the on-shore transmission grid. The results show adequate islanded and black-start operation with less than 25% grid forming power.

It has been found that whereas stable steady state operation can be achieved for relatively low amount of grid-forming capability, large differences in speed of response between grid-forming and grid-following wind turbines imply the need of more grid-forming power to operate as a "slack bus" during transients, particularly during load rejection.

FRT Test System Compact for 27 MVA with less Grid Burdens is now in Operation
Submission-ID 240
Rainer Klosse
WindGuard Certification, Germany
The completely newly developed FRT test system of WindGuard Certification is now completed and in use since March 2018. The test system is designed for networks up to 30 kV and rated currents up to 630 A. This results in possible DUT up to 27 MVA rated apparent power. Due to a very small number of components, a single ISO container is sufficient to store everything together.

Technically, this test system switches from an series impedance to an autotransformer at the moment of fault simulation. By using the transformer effect, the network load is considerably lower for most numbers of test cases compared to conventional test equipment. For example, in the event of a 50% drop, the ½ short-circuit current at the same longitudinal impedance must be expected compared to a conventional voltage divider. In addition to UVRT, OVRT can also be simulated. Due to the not used capacitors the usual problem with resonance points is not given. The sin form of the voltage is much more consistent. This test system is also suitable for simulating not only pure amplitude changes but also vector jumps of the voltages.

Depending on the mains conditions, voltages of up to 150% of the input voltage can be achieved by the transformer. Initial measurements have already been carried out on the medium-voltage grid. These show that the basic assumptions are correct. However, the test system also uncovered properties that make correction factors necessary in addition to rough calculation. Due to the mixture of a simple coil and a transformer, there are no standardized models available. In recognition, the new system is compared with a conventional voltage divider. Due to the cost-effective design with the enormously increased application possibilities, a quick replacement of conventional voltage dividers is expected.

Kriegers Flak Combined Grid Solution – Principles of Voltage and Reactive Power Control for HVAC/HVDC Meshed Offshore Grids
Submission-ID 245
Vladislav Akhmatov 1, Anne-Katrin Marten 2, Reinhard Stornowski 2, Thomas Bentzon Soerensen 1
1 Energinet, Transmission System Operator of Denmark, Denmark
2 50Hertz-Transmission, German Transmission System Operator, Germany
Kriegers Flak Combined Grid Solution (KF CGS) will be an interconnector between East Denmark and Germany and utilize the already existing and under commissioning grid-connections of the large offshore wind farms in the Baltic Sea.

The already established 150 kV AC submarine and underground cable connections of the Baltic 1 (48.3 MW) and Baltic 2 (288 MW) wind farms with a total connection length of approx. 135 km to Bentwisch in Germany will be extended by the two 24 km long 150 kV AC cables to the Kriegers Flak B extension platform (KFE). On the Danish side, the two platforms Kriegers Flak A (KFA) and Kriegers Flak B (KFB) will collect 200 MW and 400 MW offshore wind power respectively. The KFA and KFB platforms will be linked through an approx. 9 km long 220 kV AC submarine cable and connected to the onshore, double-busbar compensation substation Bjæverskov via two approx. 80 km long 220 kV AC submarine and land cables. In Bjæverskov there will be the first 400/220 kV step-up transformation with connection to the Danish 400 kV transmission grid. From Bjæverskov to Ishøj there will be another 220 kV land cable with the second 400/220kV step-up transformation to the 400 kV Danish transmission grid. On the KFE platform there will be the 220/150 kV transformation.

Since East Denmark (the Nordic system) and Germany (the Continental European system) are not synchronised, there will be an HVDC Back-to-Back converter (BtB) in the substation Bentwisch connecting the 150 kV AC offshore system with the 380 kV German onshore system. The KF CGS will connect wind power infeed and permit as much as possible energy market trade between the countries utilizing the already existing equipment for the wind farm grid-connections and the additional equipment of the interconnector. The KF CGS holds the status of a “Project of Common Interest” (PCI), given by the European Commission, and is granted financial support from the European Energy Programme for Recovery (EEPR).

The KF CGS network resembles an HVAC/HVDC meshed offshore transmission system which requires careful, well-tuned and advanced voltage and reactive-power control. Because the KF CGS is a meshed offshore grid, the overall control shall be robust and working in different operational regimes such as switching of the power transport equipment, i.e. intended and unintended disconnection and reconnection of cables and transformers as well as connected and separated regimes of the interconnector.

This paper will present the major principles of the voltage and reactive-power control to be applied within the KF CGS, which have been designed and verified by simulations. Each control area includes several equipment and control functions, which will be presented and discussed.

Challenges with the design of cost effective series DC collection network for sea-based wind-farm
Submission-ID 246
Mohammad Kharezy 1, Torbjörn Thiringer 2
1 Rise Research Institutes of Sweden, Sweden
2 Chalmers University of Technology, Sweden
Today, when a sea-based wind park reaches a size of 100 MW or more, a platform, acting as the hub in the collection grid and a point of connection for the cable to land is needed. Among other equipment, the platform carries a transformer. Since ordinary AC cables becomes inefficient over nearly 100 km, it has become necessary to move over to DC transmission for some of the new wind parks. However, within the wind park, still a collection grid running at 50 Hz AC is used and 2 platforms of gigantic sizes are needed for 50 Hz AC/AC and AC/DC conversion increasing the cost substantially for the wind energy installation.

A possibility here is to use the DC technology also for the energy collection grid within the wind park. To do that, a key component is missing, the high-power DC/DC converter. Such a device can transform the voltage from the wind turbine to a high DC voltage using a much smaller converter unit compared to a 50 Hz transformer. An idea is to fit the converter system into a container on the outside of a wind turbine, thus utilizing the existing foundation out in the sea.

A highly interesting solution is then to connect the output of the wind turbines in series, and in this way making the voltage level to reach 100, 150 or even 200 kV. The idea would then be to continue the connection directly to shore without the need of a large transformer platform. A cable can transport up to 2 kA, and using a bipolar set-up, 800 MW can be reached without a platform, for the 200 kV case. This is a huge investment saving. However, here comes a highly important factor: The wind turbine that is located closest to the DC-transmission cables going away to shore must take up the full insulation on its high-voltage side. Today, the DC/DC converter technology is far away from such capabilities. This is where the proposed project comes in.

The proposed design and optimization approach was introduced earlier in a PhD work where it was applied on two down-scaled 50 kW, 1/3 kV, 5 kHz prototype medium-frequency transformers. These optimized designs have later been manufactured, and successfully measured, fulfilling the efficiency, power density and leakage inductance requirements that the prototypes were designed for. To move further up in insulation level, more insulation material must be added reducing the power density of the unit, increasing the length of the winding, thus decreasing the efficiency.

In this research project, financed by Swedish Energy Agency, the aim is to study the size of such a DC/DC converter accounting for the increased voltage strength and thus insulation level. Both obtainable voltage withstand levels as well as life-time of the insulation are issues that must be investigated carefully. The success of this project puts academia, research institutes and industry in a very good position to take a lead into the development of this key enabler for the cost-effective harvesting of wind energy from offshore wind turbine installations.

Impact of technical parameters and data quality on wind energy modeling in Germany
Submission-ID 255
Daniel Beulertz, Armin Schnettler
Institute for High Voltage Technologies at RWTH Aachen Universtity, Germany
This paper presents a model framework that is capable of determining the generation from wind power plants on a large scale while taking technical parameters into account. Turbine characteristics and hub heights as well as data transformation methods are included in the approach. Most relevant input data, especially wind speed data and power plant data are analyzed regarding their comprehensiveness and level of detail. To account for the impact of different wind speed data sources the model is applied to a set of single wind power plants and the results are compared to historical daily generation data. Afterwards the model is extended to german scope, focusing on the impact of power plant data and model parameters on the modelled wind generation compared to historical data for Germany.

A correction method to improve the quality of the wind forecast-A case study for Wangjiangping station
Submission-ID 256
Yan Shen
National Meteorological Information Center, China Meteorological Administration, China

As a substitute to burning fossil fuels, hydropower and wind power belong to clean, renewable, abundant power and produce no hothouse gas radiations during operation compared with the non-renewable power sources. So more and more countries and organizations pay attention to the development of hydropower and wind power with no exception of China. Hydropower (wind power) is mostly dependent upon precipitation and elevation changes (wind speed and wind direction); high precipitation (wind speed) levels and large elevation changes are necessary to generate significant quantities of electricity. In it, meteorological data including hi-resolution, long-term precipitation and wind observations and forecast can do more contribution on the power site selection, power monitoring, prediction and early warning.

Based on the WRF (Weather Research and Forecasting) model, the wind speed at the 80 meters above the ground surface is forecasted ahead of 3 days with the 15-min temporal interval at the Wangjiangping station. However, the forecasted wind speed contains large errors owing to the problems in the complex physical process, improper boundary and first-guess values and high stochastic effect of the wind speed at the 15-min temporal resolution. Therefore, an efficient way to improve the quality of the forecasted wind speed is to correct the errors based on the in-situ wind speed values. In this research, a Probability Density Function (PDF) method is employed to correct the errors in the forecasted wind speed for the Wangjiangping station located in the Sichuan Province of China. The key to conduct the PDF is to obtain the sufficient samples to co-pair the observed and forecasted wind speed. In this study, the different number of co-pair samples is selected and conducted the correction. We used the 3-month data from August to October, 2017 to collect the samples and 1-month independent forecasted data in March 2018 to validate the PDF correction results. The independent validation result is indicated that the bias of the forecasted wind speed of one-day ahead has been improved from 1.113 m/s to 0.084 m/s after the PDF correction.

Sector coupling: Renewable gas from offshore wind and offshore electrolysers to decarbonise heat and transport
Submission-ID 257
Malte Jansen, Iain Staffell, Samuel Hill
Imperial College London, Centre for Environmental Policy, United Kingdom
Heat and transport are much harder to decarbonise than the electricity sector with little progress made in the past decade. Hydrogen can offer the energy system a low-carbon fuel that has direct exchangeability with natural gas, a higher specific energy density and more flexibility than electrification. However, this ‘sustainable’ fuel of the future is currently produced using a high-carbon process. This study proposes a more sustainable alternative. A techno-economic model of an offshore wind farm, offshore water electrolysers (both alkaline and proton exchange membrane) and salt-cavern storage is built. The model output is the Levelized Cost of Gas (LCOG), or the lifetime costs portioned across the lifetime hydrogen production. The LCOG has been calculated using deterministic and stochastic approaches. The deterministic model is based on discounted cashflow analysis whereas the stochastic model uses Monte Carlo analysis to calculate the expected LCOG by varying input parameters. Three scenarios were modelled, and alkaline electrolysis cost the least at 8.38 EUR/kgH2. This is, at minimum, four times the cost of the most common conventional hydrogen generation method and twelve times that of natural gas. States are in a unique position to make renewable hydrogen more competitive by reducing the uncertainty around private investment with a supportive policy environment. Reducing the risk of investment alone could see the LCOG of alkaline electrolysis fall to 5.32 EUR/kgH2, near competitive with conventional generation methods.

Optimal Provision of Frequency Containment Reserve with Hybrid Power Plants
Submission-ID 265
George Alin Raducu, Jonas Funkquist, Claudiu Ionita, Nikolaos Styliaras
Vattenfall R&D, Denmark
This paper presents the development of a controller, used to steer renewable hybrid power plants (RHPP), consisting of wind power plants (WPP), solar power plants (SPP) and battery energy storage systems (BESS) that are connected in the same point of common connection (PCC). This controller, referred to as Renewable Park Controller (RPC), will communicate directly with controllers of the individual components of the hybrid plant via set-point commands.

The development of such a controller is expected to facilitate the integration of new units (generation and/or storage) to already existing sites. Some scientific publications regarding optimization-based control can be found in the literature, but basic questions regarding the integration of new units to existing sites have not been addressed often. The exact functionality of this controller can vary depending on the requirements for the specific site, e.g. primary frequency control, power ramp limitation, power limitation, increased renewable energy utilization, etc. Batteries and solar panels are being installed in existing wind farms at an increased rate. It is crucial that the new units will be synchronized with the operation of the wind farm and that the power quality at the PCC will satisfy the requirements imposed by the grid codes.

The design of the controller and a basic dispatch functionality have been presented in the article “Design and Implementation of a Hybrid Power Plant Controller” - Hybrid Power Systems Workshop, Tenerife, 2018. Three functionalities can be mentioned: following a setpoint, power limitation and curtailment strategies. The power limitation functionality ensures that the RHPP never produces more power than the allowed power in the PCC. By defining a curtailment strategy the RPC will select an appropriate priority for power reduction in order to follow the PCC reference set-point. As an example, the wind turbines can be set with low curtailment priority, in order to reduce the mechanical loads.

With the proposed RPC architecture, different market-related operations can easily be performed, such as primary frequency control (PFC), arbitrage and imbalance market. Moreover, a proper combination and coordination of these functionalities has a great potential to increase revenue. The aim of this work is to develop and integrate the already mentioned market related functionalities in the overall RHPP structure for an optimized operation. It is desired to validate the control and optimization algorithm through a Hardware-In-The-Loop (HIL) test and to deploy it in Vattenfall’s future hybrid power plants.

The RPC development adds significant value to the operation of hybrid power plants, proving that such a controller can solve integration issues for newly added components to existing installations. It will also enable further enhanced optimization functionalities based on energy market spot prices, weather forecast and grid demands.

Operational Planning Strategies of Wind-Powered Electric Vehicle Charging Stations for Charging Demand Dispersion
Submission-ID 269
Yerim Lee 1, Yunsung Cho 2, Jin Hur 1
1 Sangmyung University, Korea, Republic of (South)
2 Daegu Catholic University, Korea, Republic of (South)
Jeju island in South Korea plans to deploy wind generating resources into power grids with electrical vehicles (EVs) according to the Carbon-free island Jeju by 2030. Due to the dynamic characteristics of the electric vehicle charging demands, the charging demands can be concentrated in local areas. If electric vehicle charging time is converged, the charging demand connected to power grids will increase and system instability may be induced. Wind Generating Sources (WGRs) can provide as much of the charging energy as possible. In this paper, we propose the operational planning strategies of wind-powered electric vehicle charging stations in order to distribute the charging demand of the electric vehicle with the wind generating resources. To validate the proposed approach for electric vehicle charging dispersion, we use the empirical data from the Jeju Island’s wind farms in South Korea.

Wind Farm Fault Ride Through – An Irish Context
Submission-ID 275
Aaron McDonnell, John Kelleher, Stephen Hunt, John Whelan
Renewable Energy Institute, Japan
In 2017 Ireland installed 426 MW of new wind turbines and had an average power consumption of 3 GW. This made Ireland the European state with the highest level of installed wind capacity relative to its power consumption. Ireland had an annual penetration rate of 26 % from wind in 2017 and has a target of 40 % by 2020. Furthermore Ireland is trialling a System Non-Synchronous Penetration (SNSP) limit of 65 %. This has resulted in an increased focus on the ability of wind farms to contribute to system stability during system disturbances such as faults or other step changes in voltage. In this paper the key Fault Ride Through (FRT) requirements specified in the Irish Grid Code are outlined. The modelled performance of four wind farms connected to the Irish distribution system are compared against the clauses relevant to the FRT requirements. The wind farms in question comprise of turbines with different technologies and from different manufacturers, allowing for a comprehensive overview of the common clauses where compliance may not be fully achieved.

A Study on Future Power System Database Construction According to the Renewable Energy Expansion of Korean Electric Power System
Submission-ID 280
Seongjun Park 1, Junghoon Han 1, Hyukil Kwon 1, Yunsung Cho 1, Jin Hur 2, Hyunjin Kim 3, Heasu Sin 2
1 Daegu Catholic University, Korea, Republic of (South)
2 Sangmyung University, Korea, Republic of (South)
3 Korea Electric Power Corporation, Korea, Republic of (South)
In Korea, reduction of nuclear power and coal power generation equipment and expansion of renewable energy were planned according to the renewable energy 3020 plan. It is planned to expansion of the renewable generation that amount to 20% of the total power generation by 2030. So it is necessary to build the future power system Database(DB). This paper describes the construction of future power system DB according to expansion of renewable. DB construction, system stabilization and stability analysis are performed automatically by using python-based simulator. The simulator is linked to the PSS/E, the power system analysis tool, through the application programming interface. Algorithm modeling is carried out about system DB construction and stability analysis for generating simulator. And about the PSS/E initial DB construction method of the power system raw data through the generated simulator will be described. First, construction a topology composed of existing generators without renewable sources and stabilize the power flow. And then describe the addition of a renewable generator and a method for power system analysis and stabilization of the constructed initial power system. This paper describes the modeling of renewable energy source and the control of shunt and the addition of FACTS devices for stabilize the power flow. And Transformer Tap ratio adjustment method for stabilizing grid bus voltage are described. Fault analysis is performed to check whether the fault current exceeds the allowable range according to being added of the renewable energy source. In addition, we describe a steady-state and contingency accident state analysis simulator. and describe method to overcome line overload. In this paper, we describe the line linkage method by calculating WSCR and SCR index. A renewable center point is selected in the area with low WSCR index, and SCR index calculation is used to describe the HVDC linkage with the heavy load area. Once the powerflow convergence, line overload and fault analysis are completed, dynamic analysis stimulation is described. It will be checked the stability of the system by checking the initialization of the dynamic stimulation and the divergence of the result graph.

Is cheap electricity storage essential to build an energy system based on wind and solar power?
Submission-ID 282
Emil Nyholm, Lina Reichenberg
Chalmers University of Technology, Sweden
This paper provides a counter-argument to the claim that the cost for storage is crucial to the economic viability of a power system based on renewables. We test the effect of cost reductions for two different generic storage types on the system cost for a European power system based on renewables. The generic storage types are (i) those where we vay the energy investment cost €/kWh], represented by Li-ion batteries, and (ii) those where we vary the capacity investment cost [€/kW], represented by flow batteries. The term greenification cost, the difference between the system cost with X% renewable energy and a business as usual case with very little renewable energy, is introduced in order to compare effects of different cost scenarios. It is found that a cost for Li-Ion batteries of 150 €/kWh, i.e. approximately half of the current cost, will lead to a 10% lower greenification cost, compared to a baseline scenario with a cost of 200 €/kWh. For flow batteries, a 5% lower greenification cost is reached with a cost of 300 €/kW, which amounts about one fourth of the current investment cost. In the most extreme low-cost scenario, where Li-ion batteries are assumed to cost 75 €/kWh and flow batteries 100 €/kW, the reduction in greenification cost compared to baseline is 25%.

Reactive Power Management of a Large Scale Wind Power Cluster in Northern Sweden
Submission-ID 284
Ingmar Leisse
E.ON Energy Networks, Sweden
In the Nysäter cluster in northern Sweden it is planned to establish eight wind farms with a total capacity of 1167 MW. The cluster will be connected to the transmission network at the 400 kV level by a dedicated 400/130 kV substation. The Swedish transmission network operator, Svenska Kraftnät, did put the requirement to control the reactive power exchange at the connection point continuously. In normal operation the reactive power exchange should be controlled to 0 Mvar. However, the wind farm cluster should also be able to absorb or feed-in reactive power corresponding to 10 percent of the actual active power production.

For the Nysäter cluster new power lines need to be raised to transport the power from the wind farms to the 400/130 kV substation. As the power lines are dimensioned according to the maximum power transfer from the wind farms, they will be heavily loaded during periods with maximum power production and a lot of reactive power will be consumed. Calculations have shown that in total 500 Mvar of reactive power are consumed by the 130 kV overhead lines and the adjacent transformers. Moreover, considerable voltage variations will arise in the 130 kV grid as a large amount of reactive power is transferred.

Providing a total of approximately 620 Mvar reactive power at its maximum (500 Mvar network losses and 120 Mvar to the TSO) and maintaining the voltage at the same time is a quite challenging control task. The requirement of continuous control of reactive power increases the complexity further as the need for reactive power is changing with the variation of the active power production.

In previous projects reactive power compensation was mainly achieved by shunt capacitors. However, due to the voltage variation caused by their switching the size needs to be limited as well.

One possible solution would be to install a large number of small shunt capacitors which would be an expensive solution as a large number of bays would be needed, a lot of switching operations are expected and the control is not really continuously.

Another solution would be the installation of an SVC, Static Var Compensor. This type of equipment is based on power electronics which are able to control the reactive power continuously but their cost per Mvar is quite high.

However, modern wind turbines grid connected by electronic power converters normally have the possibility to provide reactive power with both inductive and capacitive power factor. As the wind turbines are located far out in the grid their possible reactive power contribution is limited due to voltage variations.

To fulfil the requirements for the voltage but also the requirements from the TSO and keep the connection costs for the wind farm developers low, in this project it has been chosen to combine traditional shunt capacitors with the ability of continuous reactive power control from modern wind turbines.

Sneak Preview: PowerDynamics.jl – An Open-Source library for analyzing dynamic stability in power grids with high shares of renewable energy
Submission-ID 290
Tim Kittel, Sabine Auer, Christina Horn
Potsdam Institute for Climate Impact Research, Germany
PowerDynamics.jl is an Open-Source library for dynamic power grid modeling built in the latest scientific programming language, Julia. It provides all the tools necessary to analyze the dynamical stability of power grids with high share of renewable energy. In contrast to conventional tools, it makes full use of the simplicity and generality that Julia combines with highly-optimized just-in-time compiled Code. Additionally, its ecosystem provides DifferentialEquations.jl, a high-performance library for solving differential equations with built-in solvers and interfaces to industrial grade solvers like Sundials. PowerDynamics.jl provides a multitude of dynamics for different node/bus-types, e.g. rotating masses, droop-control in inverters, and is able to explicitly model time delays of inverters. Furthermore, it includes realistic models of fluctuations from renewable energy sources. In this paper, we demonstrate how to use PowerDynamics.jl for the IEEE 14-bus distribution grid feeder.

Machine Learning Approach for Probabilistic Wind Power Forecasts with Discrete Probability Density Function
Submission-ID 292
Anton Kaifel, Martin Felder, Frank Sehnke, Kay Ohnmeiß, Leon Schröder
Zentrum für Sonnenenergie- und Wasserstoff-Forschung Baden-Württemberg (ZSW), Germany
Neural networks trained on historical feed-in time series of wind turbines or parks are able to deterministically predict power over the next hours to days, by minimizing a scalar cost function on a training data set. Yet similar to the analog ensemble method, the training algorithm can also be adapted to analyze the uncertainty of the power output from the spread of possible targets found in the historical data for a certain meteorological situation. For this new approach of probabilistic forecasts, the uncertainty estimate is achieved by discretizing the continuous time series of power targets into probability density functions (PDF). For each forecast horizon, a deep neural network then predicts the PDF of power output. The resulting empirical probability distribution can then be analyzed to determine percentiles or statistical moments. The advantage of the proposed method is that it avoids the use of costly numerical weather prediction (NWP) ensemble runs. Although a selection of several deterministic NWP forecasts as input can be used. We demonstrate the application of this new method in case studies for a German wind power feed-in as well as a wind farm in Germany and Chile using state-of-the-art deep learning technology.

Bounded Probabilistic Wind Power Forecasting using Mixture Density Recurrent Neural Network
Submission-ID 302
Sebastian Haglund El Gaidi, Mihai Chiru
Greenlytics, Sweden
Uncertainty of wind power forecasts is typically only assessed as the resulting forecast spread from using ensembles of numerical weather predictions as input to the forecast model. Machine learning methods have proven efficient in quantifying forecast uncertainty even when ensembles are not available. Mixture density recurrent neural networks learn time-dependent uncertainty patterns from historical input-output mappings and subsequently estimate the uncertainty of forecasts when confronted with unseen input data. However, mixtures of unconstrained distributions are not suited for expressing bounded variables such as wind power production. Furthermore, the on-off nature of wind power production suggests it is a stochastic variable that could appropriately be described using a mixed discrete-continuous type distribution. To address these issues, we propose a mixture of bounded truncated Normal distributions and point masses at the interval end-points to model the uncertainty of wind power production in a principled manner. We apply the developed model on the GEFCom2014 dataset consisting of wind power production data from 10 wind farms in Australia. The model is compared with baseline forecast models and gives high forecast skill, especially after the first few forecast hours.

Submission-ID 306
Tsuji Takao
Yokohama National University, Japan

Although the forecast systems are developed to estimate the renewable energy sources generation, the predicted error is high, causing many problems. Therefore, Probabilistic Power Flow techniques are applied to calculate the risk of introduced Renewables. In this paper probabilistic power flow is created by Monte Carlo Simulations and developed further by Quasi-Monte Carlo Simulations. Monte Carlo method has the most accurate results but has enormous computation burden, while this is exclusively important when dealing with large systems and real-time applications. Thus, in this paper improvement on computation time is shown by changing the pseudo-random numbers to low-discrepancy sequence. Here, computation time is decreased to the desired value, but the accuracy became reduced than Monte Carlo method within an acceptable range. In addition, in the traditional PPF calculations, renewable energy sources and their forecast errors are taken separately as random variables by their probabilistic density functions, but in this paper relationship between them are taken into consideration. Multivariate functions are used to combine them.

Keywords – Probabilistic Power Flow (PPF), Monte Carlo Simulations (MCS), Quasi-Monte Carlo Simulations (QMCS), probabilistic density function (PDF), Sobol sequence, low-discrepancy sequences.

Operational Experience with a Type 3 WPP in a Weak Power System Region with Type 2 WPPs and Synchronous Condensers
Submission-ID 316
Amr Abdellaoui, Mohamed Asmine
Hydro-Québec (TransÉnergie), Canada
With the increasing integration level of wind generation over time, different Wind power plants (WPPs) types were integrated in a weak transmission system region. This raises challenges regarding voltage control schemes that have to be implemented in order to maintain the reliability of the system. Thus, thorough dynamic studies have to be performed to optimize the voltage control schemes of the WPPs which should be followed by field testing to ensure a reliable operation. This paper presents the Hydro-Québec operational experience with the implementation of a voltage control strategy scheme to coordinate the voltage regulation between a type-III WPP and two synchronous condensers at a common point of coupling in a region with two type-II WPPs.

Keywords- Wind power plants, synchronous condensers, voltage control, field tests, stability studies.

MIGRATE project and Future Power Systems
Submission-ID 329
Jako Kilter
Elering/Tallinn University of Technology, Estonia
MIGRATE stands for Massive InteGRATion of power Electronic devices and is an EU-funded project under the framework of Horizon 2020. The aim of MIGRATE is to find solutions for the technological challenges the grid is currently and especially in future faced with. By 2020, several areas of the HVAC pan-European transmission system will be operated with extremely high penetrations of Power Electronics(PE)- interfaced generators, thus becoming the only generating units for some periods of the day or of the year – due to renewable (wind, solar) electricity. This will result in growing dynamic stability issues for the power system (possibly a new major barrier against future renewable penetration), the necessity to upgrade existing protection schemes, and measures to mitigate the resulting degradation of power quality due to harmonics propagation.

European TSOs from Estonia, Finland, France, Germany, Iceland, Ireland, Italy, Netherlands, Slovenia, Spain and UK have joined to address such challenges with manufacturers (GE, Schneider Electric) and universities/research centres. They propose innovative solutions to progressively adjust the HVAC system operations.

Firstly, a replicable methodology is developed for appraising the distance of any EU 28 control zone to instability due to PE proliferation and for monitoring it in real time, along with a portfolio of incremental improvements of existing technologies (the tuning of controllers, a pilot test of wide-area control techniques and the upgrading of protection devices with impacts on the present grid codes). Next, innovative power system control laws are designed to cope with the lack of synchronous machines. Numerical simulations and laboratory tests deliver promising control solutions together with recommendations for new PE grid connection rules and the development of a novel protection technology and mitigation of the foreseen power quality disturbances. Technology and economic impacts of such innovations are quantified together with barriers to be overcome in order to recommend future deployment scenarios. Dissemination activities support the deployment schemes of the project outputs based on knowledge sharing among targeted stakeholders at EC level. Project website for more information: https://www.h2020-migrate.eu/

Large Disturbance Rotor Angle Stability Analysis in Power Systems with High Penetration Levels of Wind Power
Submission-ID 330
Da Wang
TU Delft, Netherlands
Unlike synchronous generators used in conventional power plants, wind turbines cannot directly respond to large disturbances, due to their power electronic converter based interface and maximum power control strategy. Therefore, the decommissioning of synchronous generators and the integration of wind turbines will influence the large disturbances rotor angle stability of existing power systems. How to solve this challenge? Three questions should be answered: what is the influence of wind turbine? How to assess/monitor this influence? and finally how to mitigate this influence? This presentation introduces the work to answer the above 3 questions, carried out in the framework of MIGRATE.

Operation of Wide-Area-Controls in Iceland
Submission-ID 331
Birkir Heimisson
Landsnet, Iceland
The Icelandic power system experiences large and rapid frequency deviations because of its low inertia and large loads relative to the system size. Using fast synchrophasor based signals to trigger new flexible demand- and supply-side resources, fast response services can reduce frequency deviations and the probability and impact of islanding. This presentation reports on the experience of the first live operational fast response services addressing angle and frequency stability. The approach improves reliability and access to the grid in constrained locations.

Experimentation results : Grid Forming Control Interoperability Tests and Current Limitation
Submission-ID 332
Thibault Prevost
RTE, France
Different grid forming controls have been developed/improved within WP3 of MIGRATE project and simulations are being performed to validate their behavior on small networks. Interoperability issue will also be addressed between the different controls on such small network. Moreover, current limitation strategies, which is of critical importance for such inverters will presented and first tests on reduced scale hardware will be presented.Different grid forming controls have been developed/improved within WP3 of MIGRATE project and simulations are being performed to validate their behavior on small networks. Interoperability issue will also be addressed between the different controls on such small network. Moreover, current limitation strategies, which is of critical importance for such inverters will presented and first tests on reduced scale hardware will be presented.

Power Quality and Grid Codes in View of Massive Integration of Power Electronic Devices
Submission-ID 334
Marta Val Escudero
Eirgrid, Ireland
This presentation will present some of the results from work package dealing with power quality in MIGRATE. The main aspects of power quality will be discussed considering the changes taking place in current and future power systems. Results from various power quality related questionniares and studies made in MIGRATE will be presented considering technical and legislative framework. Results and their influence to grid connection rules and codes will be discussed.

Review of European Grid Codes for Wind Farms and Their Implications for Wind Power Curtailments
Submission-ID 335
Elis Nycander, Lennart Söder
KTH Royal Institute of Technology, Sweden
In order to accommodate the increasing wind power penetration in power systems grid codes for wind power plants are being continuously updated by TSOs. In this paper we review several European grid codes for wind power released from 2015-2018. Specifically, we focus on the comparison of the new ENTSO-E grid code Requirements for Generators, released in 2016, with national grid codes to see to what extent these are in agreement and how this contributes towards harmonization of grid codes within Europe. Also, we discuss the implications of the grid codes for performing curtailments of wind power plants.

Spine Toolbox and Spine Model for Open Source Energy System Analysis
Submission-ID 336
Juha Kiviluoma 1, Mikael Amelin 2, Erik Delarue 3, Jody Dillon 5, Maren Ihlemann 3, Steffen Kaminski 3, Toni Lastusilta 1, Manuel Marin 2, Jon Olauson 2, Fabiano Pallonetto 4, Kris Poncelet 3, Erkka Rinne 1, Pekka T Savolainen 1, Lennart Söder 2, Per Vennström 1
1 VTT Technical Research Centre of Finland, Finland
2 KTH - Royal Institute of Technology, Sweden
3 KU Leuven, Belgium
4 University College Dublin, Ireland
5 Energy Reform Ltd, Ireland
EU project Spine is developing an open source toolbox and model for energy system modelling. The purpose is to offer a platform for data acquisition tools, data processing tools, and models of different temporal and/or geographic scope. The toolbox will help users to manage data, scenarios and modelling projects. It will allow connecting models of different scope through the toolbox interface. The poster presents the Spine Toolbox user interface, data structure and the interface for connecting external tools and models. It will also introduce the Spine Model, which is an adaptable energy systems optimization model integrated with the Spine Toolbox.