5. International Experience
Wind energy integration is a subject receiving a significant amount of attention around the world and much has been published on the topic. The wind turbine industry is also growing at a rapid rate and continually developing and marketing new technology.
The global installed wind capacity increased from 2,500 MW in 1992 to over 40,000 MW in 2003, according to the European Wind Energy Association (EWEA).24 This represents annual growth of almost 30%.
In the early 1980's, WTGs were typically around 50 kW (0.05 MW) with rotor diameters of 20 m, but today WTGs of up to 5 MW and 100 m rotor diameters are being tested for operation in off-shore wind farms. The rapid growth in the sector has provided manufacturers with opportunities for research and development leading to lower cost per MW of installed WTG capacity, significant gains in the efficiency of conversion of wind energy to electricity and the addition of safety and grid support features which are considered normal in conventional generators.
5.1 Wind Energy Penetration
Using the measure of wind energy penetration adopted in section 3, the Western half of Denmark tops the figures with over 60%.25 However, this system also has intertie capacity equal to almost 70% of total peak generation, providing additional flexibility in managing the inevitable "unders and overs" that arise from high levels of wind energy penetration.
Germany, Spain and New Mexico all have penetration of around 15% with intertie capacity of approximately 75%, 13% and 100%, respectively. While it would be incorrect to conclude that high levels of wind energy penetration and comparable intertie capacity must go together, it nevertheless appears that the presence of interties has assisted the development of wind energy in these countries. There are, of course, other factors that go with interties, the most notable being the policies in a number of European countries that have favoured renewable energy - and noting that Europe is relatively highly interconnected anyway.
5.2 Relevance to New Zealand
Much of the international experience and studies are relevant to New Zealand as there are many common issues around connection and operation, but this country has some characteristics that require careful assessment. The most important differences between New Zealand and other countries are not unique on an individual basis, but taken together they represent a supply system with particular challenges:
- Geographical isolation
There are no links to other electricity grids, e.g. New Zealand does not have a link to the Australian grid, while grids in Europe, for example, are highly interlinked in many cases. - Small size
The total size of our supply system, measured in GWh per annum, is not large especially given the length of our transmission grid. - Two islands
New Zealand's supply system is split into two islands, connected by the HVDC link which currently isolates the islands for the purposes of providing regulation and instantaneous reserves.
A study by Transend,26 the owner and operator of the grid in Tasmania, is relevant to the New Zealand situation, mainly because of the small size of the Tasmanian electricity system. The study concludes that the major operational issue for wind energy integration in Tasmania is maintaining control of system frequency. However, the Transend study does not state the type of WTG its analysis is based on and appears to include only asynchronous WTGs which are connected through power electronic devices. It does not appear to consider the double-fed induction generator (DFIG) and synchronous WTGs which we assume in sections 8 and 9 for an initial assessment of the limits on wind energy integration in New Zealand.
Though management of system frequency is also a prime concern in New Zealand, the assumptions made in this study about trends in WTG technology differ from those made in the Transend study and indicate that the issue is potentially manageable even with high levels of wind energy integration.
5.3 Trends in WTG Technology
Wind energy is a classic case of technology driving demand driving technology. The growth of wind energy in Europe and the United States has taken WTGs from being an alternative energy source to being mainstream, requiring manufacturers and the supply industry to tackle integration issues head on. The latest technology trends are driven by economics and also by the need for WTGs to directly support grid security (refer sections 6 and 7.2).
There is currently a trend27 towards WTGs that have variable speed rotors which is achieved using pitch control of the turbine blades, particularly for the increasingly common megawatt class WTGs. The pitch of the blades is constantly adjusted to achieve the required operating point in the most efficient manner, and to limit the output of the WTG to its safe operating range. EWEA28 states that the preferred generator technology for variable speed operation is the DFIG in which the stator windings are directly connected to the grid and supply about two thirds of the power output while the rotor windings are connected to the grid via an AC-DC-AC converter and supply the balance of the output power.
The DFIGWTGs are increasingly equipped29 with fault ride-through capabilities and have the ability, or the potential, to contribute passively and actively to managing under-frequency events on the grid. Both of these issues will be of crucial importance to the New Zealand grid if wind energy is to realise anywhere near its potential.
A WTG produced locally by Windflow Technology of Christchurch has its turbine connected mechanically to a synchronous generator via a "torque limiting gearbox" which ensures that the generator operates at constant speed and stays in synchronisation with the grid. This particular WTG is inherently similar to conventional synchronous generators and hence has the potential to provide fault ride-through and to maintain constant output during under frequency events.
The issue of fault ride-through has recently received much attention world-wide. According to GE Power Systems,30 "the bottom line is that the wind industry is moving towards low voltage ride-through being a standard requirement." GE also notes that "a related topic, but one which has received much less attention, is under-frequency ride-through" but goes on to say that this "criteria should be applied to wind systems as well".
While the fault ride-through capabilities of modern turbines are generally quite clear from the manufacturer's literature, the contribution of WTGs to system security during an under-frequency event is not always clear. In future, the authors anticipate that this will change as WTGs are made increasingly able to contribute to managing frequency. In the meantime, the performance of WTGs during under-frequency events is an area which requires further study in New Zealand given the nature of our grid and the way that frequency is managed.
5.4 Factors Affecting Wind Energy Penetration
The Kema-Xenergy report for the California Energy Commission31 claims that, because wind energy penetration in Europe is substantial already and will only increase over time, the impacts of large-scale wind energy integration "are viewed not as an obstacle to development, but rather as obstacles that must be overcome". It also identifies the main factors that should be studied in detail as wind energy penetration exceeds 20%.
These include:
- the impact of wind on the occurrence of line constraints on the grid;
- the impact of additional wind generation on the short circuit current ratings of existing connected equipment;
- the impact of more wind generation on grid stability;
- how islanding and reverse power flow could impact on existing grid protection schemes;
- how the variability of wind energy can be accommodated in the operational time scale;
- how wind energy could impact on power quality including voltage and regulation.
Kema-Xenergy also makes reference to problems with "unmitigated" development of large-scale wind energy where penetration is higher than 20%, requiring innovative solutions to be developed.32
GE Power Systems33 recommends that New York State adopts requirements that have been proven in other areas, particularly Europe, where wind energy has achieved high penetration. These include:
- voltage regulation at the point where the wind farm connects to the grid;
- fault ride-through;
- a minimum level of monitoring and recording;
- the ability to reduce wind farm output from a remote location.
GE also recommends additional features be required as the technology becomes widely available. These include:
- control of the rate at which the wind farm output increases or decreases - known as "ramp rate" control;
- control features similar to the functions incorporated in the governors of synchronous generators, for example the ability to increase power output as system frequency falls;
- reserve capability - the ability to control a wind farm's output so that it operates at less than its maximum output given the prevailing wind speed, leaving reserve capacity available in the event of a contingency such as an under-frequency event;
- voltage regulation at zero power output, a feature not normally available from conventional generators but potentially able to be incorporated into modern wind farms.
5.5 Variability of Wind Generation
A number of international papers review the data from existing wind farms, or from data synthesised to give a database of many potential wind farms in a country or state. These all show that the variability of the total output of many wind farms is lower than the variability of any individual wind farm. Based on actual wind farm data, hourly variations in wind farm output34 fall from about ±30% of installed capacity when the area is in the order of 40,000 km² (about the size of Denmark) to about ±20% for an area of 160,000 km² (e.g. Germany or the state of Iowa) and then to about ±10% in larger areas covering several countries (e.g. the Nordic states).
Summarised data provided to the authors by Meridian Energy during the course of the study show a similar pattern in New Zealand. The data relates to six potential wind farm sites around the country, and is updated every 10 minutes. For one wind farm, the standard deviation of the change in wind farm output over a 10 minute period is over 8% of the installed capacity. For three farms this drops to 6% and for six it drops further to 4%.
5.6 Wind Speed Forecasting
A number of international papers emphasise the importance of developing accurate wind speed forecasting techniques, especially as wind energy penetration increases. The science of wind speed forecasting has received much attention but is still in its infancy. The MISG35 found that the simple persistence method produces good results for short time periods but loses accuracy when applied further into the future. The persistence method simply assumes the current output of a wind farm will be maintained in future and this method is written into the new IG rules in the EGRs. The MISG tested other methods but large improvements were not found.36
The best overseas forecasting models developed to date exhibit an error37 of 15% - 25% and settle to somewhere in this vicinity after around T + 6 hours.
It is obvious that wind speed forecasting in the operational time frame, out to the end of pre-dispatch time, is heavily reliant on meteorological forecasts of wind speed. Of particular interest are those weather events that could cause large changes in output over short periods. Here the issue is not only how much wind farm output could change but also when. For example, an active southerly storm moving up the country toward the Manawatu could cause total Manawatu wind farm output to increase quickly, but it could also precipitate an even more rapid reduction in output if wind speeds are above wind farms cut-out speeds of 25 - 30 m/s (108 km/h). As wind integration increases, the existing network of weather stations may require augmentation with additional, strategically placed, wind recording stations.
In Europe, where wind energy penetration levels are currently highest and there has been heavy investment in wind forecasting systems, forecasts are undertaken centrally. If this approach were applied here it would probably mean that Transpower as SO would be forecasting the output of all wind farms in New Zealand.38 Historically, the electricity market has tended toward the view that market participants are best placed to forecast the output and capability of their own plant, a view that is reflected in the new IG rules in the EGRs. To continue this approach with high levels of penetration for wind energy would need market rules requiring wind generators to:
- make accurate forecasts for their wind farms' output out to the end of tomorrow - these forecasts would be encapsulated in offers to generate;
- meet certain minimum standards of accuracy for their wind farm output forecasts, or alternatively use specified forecasting techniques or models;
- update their forecasts on a regular basis;
- submit their forecasts to the market in a standard format. This would also be achieved by submitting offers to generate, but might need to be augmented with additional information, for example an estimate of the error and spread in the forecasts for each offer submitted. It might also require near real time updates of forecasts or offers in extreme situations.
Back to Top