TRANSMISSION & DISTRIBUTION
Living with wind power in Germany1 March 2007
Increasing wind energy in the national mix threatens the stability of the supply and transmission system. To gain a thorough knowledge of the demands likely to be made on it, the German Energy Agency commissioned a study that, in the light of the causes of the 4 November 2006 blackouts, proved uncannily percipient.
By the end of 2006, over 20 900 MW of wind turbine capacity had been installed in Germany. The power derived from wind can cover, at least temporarily, the total grid load in some regions of the country. But if grid congestion is to be avoided the further development of wind power will require concomitant network reinforcement. Further, the sporadic availability of wind power will require spinning and secondary reserve generation to be used in new ways, while an increased reserve will be necessary if imports are to be kept low.
In response to these concerns the German Energy Agency (DENA) a couple of years ago decided to commission an assessment of the impact of the integration of wind power into the EHV grid for the period to 2020, and from it derive a long-term management strategy. A consortium of Energiewirtschaftliches Institut Köln (EWI), Deutsches Windinstitut (DEWI), the TSO E.ON Netz, RWE Transportnetz Strom and Vattenfall Europe Transmission carried out the study, the results of which were reported* at Cigré Session 2006 in August, and are described here.
The proportion of electricity consumption supplied from renewable energy sources is expected to rise to at least 12.5% by 2010 and to at least 20% by 2020, with wind energy likely to make the greatest contribution. This will result in a pronounced concentration of in-feed from wind energy in northern Germany, with its marked daily and seasonal fluctuations and unpredictability. This gives rise to new challenges, in particular for transmission system operators: first, more electrical energy must be transported over greater distances; second, the balance between electricity consumed and that fed into the grid must be maintained. New and adapted operating methods in power stations and transmission systems will be called for.
An ageing power plant population and the agreed phase-out of nuclear energy mean that an estimated 40 GW of new capacity must be in place by 2020. This renewal coincides with the planned expansion of wind energy, allowing the possibility of adapting the structure of new power plants to meet changing conditions without compromising the present level of reliability.
The German Wind Energy Institute (DEWI) has determined the areas most suitable for wind energy development. In Germany, wind energy projects (WEPs) are established almost exclusively in designated areas classified as 'priority', 'conditional’ or 'suitable'. It was also possible to determine the potential capacity remaining in the ‘suitable’ areas. Factors that have the effect of reducing the suitability of assigned areas, such as lack of grid connections, were taken into account by reducing the calculated growth potential by a flat 20%.
In respect of offshore wind energy development, DEWI carried out an independent feasibility study of individual projects. It estimated that the installation of approximately 20 GW of capacity ought to be possible by 2020, and concluded that a planned 16 500 MW could not be installed until after 2020. In respect of the repowering of wind turbines, it was assumed that of those installed after 1998, 50% would be replaced after 15 years, with the remainder replaced after 20 years with capacity increased by a factor of 1.2 (Table 1).
Strengthening the grid
The technical investigation was based on two assumptions – that the reliability of Germany's electrical supply system would remain unchanged, and that safe and reliable interconnected operation with its European partners would be maintained. The technical criteria, analyses and evaluations assumed agreed framework conditions which also covered determination of the grid extensions necessary for N-1 secure transmission.
The wind related transmission functions of the interconnected system are determined by the characteristics of the regional network into which the wind capacity is installed, and which conventional generation units are freed by its uptake. As the economic costs of the residual generation have been considered minimised in all scenarios, power plants are assumed to be used in merit order. A temporary use of neighbouring foreign transmission systems in heavy-load periods was regarded as satisfactory for the purposes of the study. The individual extensions were defined on the basis of an optimised final extension stage in 2015 (Figure 1).
The Saßnick et al report* indicates that approximately 850 km of new 380 kV transmission lines will be needed before 2015 to transport wind power to the load centres (Table 2). In addition, numerous 380 kV installations will need to be fitted with new components for active power flow control and reactive power generation. With thus improved regional distribution, the integration of a total of 36 GW of WEP capacity should be possible.
The total costs for the transmission system extensions up to 2015 are estimated at approximately r1.1bn, and those to 2020, r3bn. The specific installation costs for installed WEP power of 20–40 GW are approximately r50/kW, but this figure does not take into account the land and marine cable connections to the offshore plant. The cost of connecting some 10 GW in the North Sea and the Baltic Sea before 2015 is estimated at R5bn, or R500/kW.
The impact on supply reliability
The study shows that even today, particularly during strong winds, German TSOs are forced to operate their grids close to the permissible operating limits. Grid-related problems arise when wind energy is not available in the right location or at the right time. Even grids in neighbouring countries are significantly affected by the energy generated by German WEPs.
Owing to the speed of adoption of wind energy, information from existing studies and operational experience has not contributed significantly to the technical development of WEPs. As a result UCTE rules are more likely to be broken, usually because WEPs connected to the grid are immediately disconnected in the event of a fault, to prevent their being damaged, while conventional generators are obliged to maintain their supply and to support system stability in line with the system connection conditions.
Saßnick* et al indicates that with advanced technology and better grid integration this problem can be solved for new and repowered plants, and should result in a continuous reduction up to 2010 in WEP outages caused by grid problems. However, by 2015 this positive effect will be partially negated by the shutdown and the wind-related output reduction of conventional capacity, actually exacerbating the problem owing to the large number of remaining older plants which cannot contribute to voltage support. High enough levels of generation outage can put interconnected European operation at risk. It remains to be seen to what extent the upgrading of obsolete plants and additional installations to support the transmission system can improve the situation.
Fault current and voltage support
With increasing wind energy generation and thus increasing exclusion of conventional power station output in the HV and EHV networks, support such as back up voltage operation plays a very important role not only in the event of faults, but also during normal system operation. Wind turbines therefore should be capable of taking over the necessary voltage support characteristics of conventional power stations. Accomplishing this will not be easy, because of the greater electrical distance between the location of the fault in an HV network and the WEP bus (ie between the distribution network and the 110 kV network, or offshore connection).
Two phases can be identified as a fault progresses, the voltage drop during the time between the occurrence of the fault and the fault clearance, and voltage recovery directly after the fault clearance. To ensure a safe reserve margin at the network protection devices, the magnitude of the fault current is important, not only during the 10 – 30 ms following an occurrence, but also up to 150 ms afterwards or, in the case of reserve protection, up to times measured in seconds.
The behaviour of conventional synchronous generators differs from that of WEPs. Wind power plants using asynchronous machines or doubly fed asynchronous generators provide a fault-current contribution in the first 10 – 30 ms, which is comparable to the contribution of synchronous generators on a similar scale and with similar subtransient reactance values. If the relevant data are available, this fault contribution must be taken into account in the planning phase of a wind park connection when checking the short circuit capacity of network elements.
After the first 10-30 ms the short circuit waveforms differ from one other (Figure 2). In the case of generators with asynchronous machine behaviour, the fault current contribution decreases, since the magnetic storage of the machine is exhausted. But that of the synchronous generator is maintained through the transient reactance, so that its excitation state is resumed. In the case of longer fault times the voltage regulator drives the excitation of the synchronous machine to the maximum possible value (its ceiling voltage), which ensures a sufficient fault current contribution.
The behaviour of doubly fed induction generators is more complex to assess. There are various contributory factors, for example, whether the Crowbar fires for the protection of the plant, or whether an automatic voltage support (voltage regulation) device is included in the machine controller.
The maximum fault current contribution that can be provided by design concepts implying synchronous generators is approximately equal to the magnitude of the plant rated current.
The fault current portion injected by a wind park at the connection point in the high or extra-high voltage network depends not only on the generator type but also on the medium voltage network and the network transformers at the wind park
In the phase directly after the fault clearance it is important that the recovery of the network voltage is not hindered by the WEP (Figure 3). Those with asynchronous-machine behaviour have unfavourable effects, since they absorb inductive reactive power from the power system after fault clearance, thus extending the voltage recovery phase and sometimes compromising conventional power stations in the vicinity. For doubly fed induction generators the behaviour is also complex and depends on the voltage support (voltage regulation), Crowbar firing, and some manufacturer-dependent effects.
The possibility of applying FACTS technology at the connections to large WEPs should also be investigated to determine whether they can achieve the proven voltage support characteristics of conventional power plants.
Effect on grid stability
Initial investigations showed that further integration of WEPs leads to the continued reduction of network support. This occurs because the WEPs, despite their technological improvements for dealing with short circuits, do not contribute sufficiently to 380 kV support. Furthermore, the proportion of the old plants can probably not be sufficiently reduced despite the steady increase in repowering projects.
There are two key instances in which generation loss, resulting from network faults, is to be expected. First, in strong-wind areas in which the voltage support of conventional power stations is reduced or completely missing in the extra-high voltage level, larger voltage drops are likely in the case of disturbances. Second, with increasing integration of wind energy and simultaneous reduction of voltage support in the EHV network, a weaker voltage recovery performance is to be expected after fault clearance owing to the deeper voltage drop during the fault.
Effects on HV and EHV networks
With increased WEP integration, the 380 kV voltage level decreases by approximately 15% more during a network fault than the 2003 reference case (Figure 4). The result is an expansion of the voltage funnel and an increase in generation trips by old plants. After fault clearance the network voltage recovers less well compared to the reference case.
The network support characteristics of new WEP are apparent at the 110 kV voltage level, compared to the reference case, because a higher proportion of the on-shore WEP is connected at this level and below. The missing network support of conventional power stations causes a further breakdown of the network voltage in the 110 kV net after approximately 60 ms. This occurs because the network support contribution of the plant continuously decreases, and old plants, which have supplied a contribution for short-circuit capacity, are disconnected from the network due to the low network voltage. The voltage level in the 110 kV net falls by approximately 15%. Replacing the fault role of conventional power stations cannot easily be achieved, because the energy needed for the reactive power supply of a synchronous machine is magnetic field energy stored independently of voltage and can be made available immediately; while if static VAr compensation is employed the energy must be obtained from, and is dependent on, the network voltage.
Impact on conventional plant
The study attempted to predict the likely demands on conventional plant, on generation as a whole and on the cost of generation imposed by the expected WEP capacity for 2007, 2010 and 2015. Underlying factors included information about the conventional power plants currently in operation, the available conventional plant technologies, the agreed shutdown of nuclear power plants, the maximum transmission capacities to neighbouring countries and the feed-in from other renewable energy sources. Some regulatory and organisational framework conditions were assumed which did not fully correspond to the contemporary situation in Germany. For example, there was no separate examination of the four German TSOs, and the separation of responsibility between TSO and power plant operators in compensating for power plant outages was not taken into account.
Demands on conventional plants
Owing to the dependence of the electricity supply on the variable wind availability, only a small proportion of the installed WEP capacity can contribute to the reliable capacity in a mix of conventional and renewables plant. Depending on the time of year, the gain in guaranteed capacity from WEPs (as a proportion of the total installed WEP) is between 6 and 8 % for installed WEP of around 14.5 GW (in 2003), and between 5 and 6 % for around 36 GW (in 2015).
Regulating and reserve power
The forecast errors for the WEP feed-in give rise to an additional requirement for regulating power and reserve power capacity provision so as to guarantee the balance between in-feeds and tapping. Despite an assumed improvement in the predictability for the WEP in-feeds, the required regulating and reserve power capacity (RRP) increases disproportionately as the installed WEP capacity increases. As the wind related RRP requirement is dependent on the level of the predicted wind in-feed, the following day requirement can be defined as a function of the forecast WEP in-feed level, having taken into account the effect of optimisation. This provides an average “day ahead“ RRP. However, the power stations must be collectively configured to provide the required maximum RRP at all times.
It is anticipated that by 2015 an additional maximum of 7 064 MW of positive regulating and reserve power capacity will be needed, of which on average 3 227 MW would be contracted on a day ahead basis. In addition, a maximum of 5 480 MW of negative RRP will be needed, of which on average 2 822 MW would have to be contracted a day ahead. In 2003, the corresponding values for positive RRP capacity were at most 2 077 MW, and on average, 1 178 MW. While for negative RRP, values were at most 1 871 MW and on average, 753 MW.
The installed capacity of WEPs will accelerate, particularly offshore, and will continue to do so after 2015. The total WEP capacity of the forecast 48 GW by 2020 will require a further extension of the transmission systems, mainly to carry power to distant load centres, with 380 kV three-phase lines with series compensation or HVDC links. Determining the best technical and economic options will require particular attention to be paid to the effect on stability.
The replacement of conventional power stations by high WEP in-feed under windy conditions will increase after 2015. The result of this will be that during periods of high wind and low electricity demand, the grid load will be lower than the fed-in WEP power. To maintain power balance as defined by the EC priority rule conventional generation units would have to be disconnected completely from the grid. The extent to which such a reduction is permissible, while maintaining supply reliability, requires further consideration.
It seems that before any significant expansion in wind energy is put into effect certain questions must be answered, related to, among other things, maintaining the present level of supply reliability, incorporating new transmission technology into the interconnected offshore system, the technical and legal requirements for WEP in-feed management (obtaining the fed-in power levels, a manageable control mechanism), the possibilities of future WEP technologies in supporting system stability and the future demands placed on the residual conventional generation.
In future studies it will also be necessary to consider the effects on the European ‘grid’ of increasing wind energy use in Germany and the associated developments in neighbouring European grids. The present efforts on a political level in the European Commission, such as the desire to promote the expansion of European electricity trading, must also be examined.
DENA Study II is already under way and will cover the years 2015 to 2020. It will be based on the DENA I results, and extended to cover new developments. It should take two years, and will involve close co-operation with the planned European Wind Integration Study (EWIS) by UCTE and ETSO.
Figure 1. Required extensions to 380 kV transmission routes up to 2010 and 2015 Figure 2. Fault current contribution by generating units (50 MVA) at the 110 kV side in the case of a nearby fault (simulation) Figure 3. Possible voltage support during and after fault clearance (simulation) Figure 4. Waveform of the network voltage in the extra-high voltage network in case of a network disturbance – comparison between the current situation and the situation in case of further integration of wind generation (simulation)