Back to the programme printer.gif Print



Thursday, 13 March 2014
09:00 - 10:30 Electrical aspects and grid integration
Science & Research  


Room: Llevant
Session description

The grid integration of wind power has always been a challenge. In this session a range of speakers from academia and industry will address the impact of wind on network operation and the development of a large scale offshore HVDC grid, offshore systems, the state-of-the-art in HVDC technologies, as well as more specific topics including the design of offshore networks from a reliability perspective and the transient response of HVDC links.

Learning objectives

  • Understand the effect of wind power on grid operation and development via case studies
  • Get an overview of the state of the art in HVDC offshore grid research
  • Get acquainted with specific topics of offshore grids, including reliability and transient response aspects
Lead Session Chair:
Stavros Papathanassiou, National Technical University of Athens, Greece

Co-chair(s):
John Olav Tande, SINTEF, Norway
Callum MacIver University of Strathclyde, United Kingdom
Co-authors:
Callum MacIver (1) F P Keith Bell (1) Dusko Nedic (2)
(1) University of Strathclyde, Glasgow, United Kingdom (2) Siemens PTI, Manchester, United Kingdom

Printer friendly version: printer.gif Print

Presenter's biography

Biographies are supplied directly by presenters at EWEA 2014 and are published here unedited

Mr MacIver is currently studying for a PhD in the Wind Energy Systems Centre for Doctoral Training at the University of Strathclyde. He graduated in 2010 with a Masters (with Distinction) in Electrical & Mechanical Engineering also from Strathclyde. His research is focused on the reliability of offshore network designs to facilitate offshore wind generation and the interconnection of regions. He has further interests in Power Systems Analysis and the facilitation of increased wind power on grid systems.

Abstract

A comparison of design options for offshore HVDC networks through a sequential Monte-Carlo reliability analysis

Introduction

With large scale expansion of offshore wind power planned in the coming decades there is a need to review possible options around the grid connections to supply this to the end user. The use of VSC HVDC technology is likely to be used and there are few technical barriers to the creation of multi-terminal HVDC grids that could not only transmit power to shore but also act as interconnection between regions. Such designs are broadly untested so an evaluation of potential reliability is required. This study compares the reliability of different options for future offshore transmission networks for wind power.

Approach

The study makes use of a Sequential Monte Carlo Simulation technique to model the reliability of different network designs under a lifetime of component fault conditions. This allows faults to be stochastically applied to the network in chronological order, meaning the influence of weather conditions on repair time can be explored through the use of concurrent wind and wave data. One hundred years worth of simulated wind and wave time series were derived to match the characteristics and correlations observed in the publicly available FINO offshore research station [1]. The technique outlined in [2] was used to create the simulated data. The use of a sequential simulation technique allows for estimates of repair rates for offshore components to be based, not only, on published data but also to include weather and logistical constraints, as faced in reality. For example repair times can include, if relevant, the need for an appropriate weather access window or expected lead time in obtaining a specialist vessel or repair component. This study compares this technique with a standard stochastic repair time analysis to investigate the influence of seasonal variations in the weather on system reliability.
In conjunction with the reliability analysis, cost and loss modelling can be applied to provide a full cost-benefit analysis and would be investigated further in a full paper. Modelling has taken place in two stages, the first being to model individual wind farm networks down to wind turbine level such that every element of the electrical transmission system is represented. This allows studies of the internal wind farm reliability. This method has also been used to compare the reliability of simple radial solutions to connect two offshore wind farms to shore with a more coordinated design that connects the two wind separate wind farms together as well. To analyse multi-terminal grids incorporating multiple offshore wind farms a second stage of modelling is required that utilises results from the first to condense the wind farms into reduced lumped models. This is required to avoid computational constraints. A full paper will investigate results from the second stage of analysis.


Main body of abstract

Using the simulated wave data it was observed that access windows required to get to and perform repair operations were likely to be significantly higher in autumn and winter seasons than they were at other times of year. Using the industry standard that mean hourly significant wave height should be below 1.5m for the duration of the required weather window, the average length of time to wait for different sized weather windows could be calculated for each month of the year. Figure 1 shows that on average faults occurring in November could take upwards of three times as long to repair as faults in June. As such, this paper will first look to examine how seasonal restrictions on repair rates can influence the calculated megawatt hour losses for particular network designs.



This is investigated by comparing two different techniques within the reliability model for calculating the repair time of components. The first is to use a standard stochastic calculation where the repair time is randomised around a mean value using the following formula: TTR=-MTTR*ln(x) where TTR is time to repair, MTTR is mean time to repair and x is a randomly generated number between 0 and 1. The second technique is to base repair times on the required weather window to access and repair the fault calculated as discussed. Perfect forecasting is assumed such that the repair is completed as soon as a weather window of sufficient length has been reached after the point of repair. For comparison the mean time to fail values for the first case and the weather windows for the second case were matched such that the availability for each run should converge to be equal. Applying this test to a standard single 500MW wind farm design the results of the two methods can be seen in Table I for a 1600 year run of the model.



The results show the availability using the two methods are close to equal so the simulation length is adequate and it can be seen that the losses, calculated using the simulated wind time series, in the case using weather windows were around 7.5% higher than the purely random method. This shows the significance of the fact that faults occurring in times of high sea states which delay repair time are also likely to correspond to times of high wind speed and so the magnitude of any lost generation is also likely to be higher.
Given that employing weather windows proved to capture some of the environmental influences on repair rates, this method was utilised to exam some of the different network options being proposed for offshore grid designs. One of the key questions that this project would like to quantify is what value there is in adding redundancy into the offshore network in terms of increased availability and reduced losses. The simplest case to study in this respect is the joining of two separate radially connected wind farms with a redundant transmission path between the wind farms. This allows power to be transmitted to shore even in the case that one of the offshore radial connections experiences an outage. Two 350MW wind farms were examined for this scenario with and without a redundant transmission path. The results are shown in Table 2.



For this case with the given set of input assumptions it was found that availability could be significantly improved through added redundancy which translated to around 16% fewer losses. For large scale offshore wind farms the monetary savings through a reduction in losses of this order are likely to be substantial over the expected lifetime of the wind farm. A full paper would provide a full cost benefit analysis as to the trade-off between savings and increased capital expenditure for a number of offshore design options.

Conclusion

This abstract has shown a method whereby the significance of seasonal variations in the wind speed and significant wave height at offshore wind farm sites can be captured and translated into figures for the overall reliability for offshore transmission network designs. The impact of this on losses due to faults on offshore network components was found to be significant. Given the financial returns for delivered energy are substantial, even small percentage changes in projected undelivered energy could have a significant impact on financial margins over the lifespan of an offshore wind project. A more accurate representation of expected transmission system availability and its impact on system losses is of clear interest to developers and investors of offshore wind.
It has also been shown that there is a quantifiable benefit to designing offshore grids in a co-ordinated fashion such that there are multiple routes for power transmission. Initial analysis showed the potential gains for even the simplest form of co-ordinated offshore design could significantly reduce total system losses compared with the business as usual option of radial connecting each offshore wind farm to shore. A full paper would explore the impact of this upon more complex example offshore networks including proposed multi-terminal HVDC solutions. An examination of the trade-off between capital expenditure to invest in more complex but more robust and reliable offshore networks and the financial returns that may bring in terms of reduced losses will be carried out. Further to this a discussion of the technical, financial and regulatory barriers that would need to be addressed to allow entities to invest in co-ordinated offshore network designs will also be included.



Learning objectives
A novel method of estimating offshore transmission system reliability and has been presented that allows for comparison of offshore network design options to be compared. This will allow for greater certainty to given to potential investors in offshore networks.


References
[1] - Bundesamt für Seeschifffahrt und Hydrographie (BSH). FINO - Datenbank [Online]. Available:
http://fino.bsh.de/

[2] - V. Catterson, I. Dinwoodie, D. McMillan, 'Wave height forecasting to improve off-shore access and maintenance scheduling', 2013 IEEE Power & Energy Society General Meeting, Vancouver, BC, Canada, 21-25 Jul 2013.