7. Factors Affecting Wind Energy
In this section we provide an overview of the range of factors which impact on the potential for wind energy to integrate into the New Zealand electricity supply system. Each factor is covered in more detail in subsequent sections.
In 1982, Elgerd40 wrote that "essentially one hundred percent of the world's bulk electric power is being produced in three-phase synchronous generators". But for a variety of reasons, most WTGs are not synchronous generators. As the penetration of wind energy has increased, this has required the wind industry to deal with a number of issues where WTGs and synchronous generators differ in their capability for supporting the grid.
7.1 The Wind Resource
A fundamental limit on wind integration is the availability of the wind resource, or simply the total amount of energy that can physically be generated from wind in this country. EECA's review41 of the wind potential in New Zealand is widely quoted on this topic and states that the physical potential for wind energy is of the order of 100,000 GWh per annum.
To put this potential in context, the nation's total annual electricity consumption is currently running at about 36,00042GWh, to which distribution losses of about 5% in local networks and transmission losses on the grid of a further 3% must be added, bringing the total annual expected generation to about 39,000 GWh for 2005. Thus the total potential wind energy resource does not appear to place a limit on wind energy integration in New Zealand in the near future.
7.2 Connection
Given the relatively low density of wind farms (approx 10 MW/km²), the nature of the New Zealand topology and the probable staged nature of wind farm developments, it is considered unlikely that single wind farms in excess of 150 MW will be developed. Proposed farms are likely to be in the size range of 40 - 70 MW.
It is therefore probable that the majority of farms will be connected either at the distribution system level or onto the main transmission network at voltages lower than 220 kV. Large numbers of 220 V and higher connections are not considered likely owing to the high cost of developing the associated connection substations. The Transpower 110 kV network is also generally in relatively close proximity to the most likely wind farm locations.
7.2.1 Electricity Governance Rules Requirements
The regulations for connection of generators to the New Zealand transmission and distribution networks are published by the EC, which has responsibility for the regulation of the electricity market. The rules that affect the technical connection aspects for generators are included in the EGRs' common quality obligations. The key requirements for generators are:
- A requirement for contribution to system frequency;
- A requirement to be able to control the rate of change in output;
- All generating units to have a speed governor;
- All generating units to have a voltage control system;
- An ability to export net reactive power at full load over a range of grid voltages;
- An ability to import net reactive power at full load over a range of grid voltages;
- An ability to operate to support system voltage.
The rules allow generating stations less than 30 MW to be excluded from complying with the rules, but the EC has the right to impose these conditions on any sized generator connected to the electricity transmission or distribution system. No specific rules currently exist for the connection of IG, but the EC is currently consulting on this matter.
There are some minor areas of technical conflict between the EGRs and Transpower's documentation. This is hardly surprising given the short period of time that has passed since formation of the EC. While it is encouraging to see the EC consult on wind power and other forms of intermittent generation, we believe it desirable for any rules or policies for the connection of this form of generation to provide requirements that support, or at least are consistent with, higher levels of penetration of wind energy. Otherwise a situation could well arise where the technical connection requirements need to be changed in the future. This could disadvantage wind generators who connect at this later time. More importantly, failure to adopt appropriate standards while penetration is low, could result in lower limits to penetration in the long run than could otherwise be attained.
7.2.2 Frequency Control Requirements
The EGRs includes two rules in Section III or Part C that relate to system frequency contribution:
Rule 2.1 states that each generator shall make the maximum possible injection contribution to maintain frequency to the normal band. This requirement effectively requires all generators to include a governor system.
Rule 2.3.1 requires generators to stay connected to the grid, or "ride-through" under frequency events as follows:
North Island
- At all times when the frequency is above 47.5 Hz
- For at least 120 seconds when the frequency is 47.5 Hz
- For at least 20 seconds when the frequency is 47.3 Hz
- For at least 5 seconds when the frequency is 47.1 Hz
- For at least 0.1 seconds when the frequency is 47.0 Hz
South Island
- At all times when the frequency is above 47.0 Hz
- For at least 30 seconds when the frequency is below 47.0 Hz but not below 45 Hz
These requirements have been defined to ensure that the frequency of the grid is stable under any generation scenario.
It is important to note that the EGR requirements have been determined based on the dynamic performance of the grid. Should the nature of the inertia of the rotating components of the grid change considerably through the addition of large quantities of, for example, VSD connected wind turbines, then the dynamic stability of the grid may change requiring different frequency criteria to be adopted. This issue is examined in more detail in section 7.7 but full system stability studies would be required to assess the extent of this impact.
A typical stall regulated, induction generator wind turbine cannot contribute to system frequency stability, but will be able to stay connected during under frequency events. However it is likely that these events will be coupled with an under-voltage condition, which will result in the generator being disconnected.
A DFIG generator can contribute to system frequency and at least one manufacturer can cope with sustained under frequency events down to 47 Hz before disconnecting. Therefore the North Island requirements can be met by this particular unit, but not those for the South Island. It should be noted that under frequency events where frequency drops below 47 Hz are very rare. A DFIGWTG will also contribute to system inertia.
A VSD connected generator can, in theory, increase output as system frequency falls, although it would appear that this is not a commonly implemented feature. A VSD connected generator will not contribute to system inertia.
Note that the EGRs do not define a maximum rate of change of frequency that must also be tolerated by connected generators. Wind generators that use power converters are likely to have a rate of change limit that could affect their ability to ride through under frequency events.
7.2.3 Output Control Requirements
The EGRs require all generators to be able to control the rate of change in output. Wind farms, and the associated generators, are usually designed to maximise the utilisation of the wind resource and therefore do not incorporate rate of change control.
The requirement for this is being argued between the wind farm developers, manufacturers and SOs throughout Europe. The wind farm proponents argue that output control imposes significant and unnecessary production losses, thereby affecting the economics of these developments. The SOs appear to be looking further into the future when wind turbines could make up a large component of the generating resource.
In the case of New Zealand, it is potentially important that, with a large wind power penetration, wind farm output control is available so that the SO is able to manage the grid effectively, especially during post fault events. Whether individual generators need to have output control provisions is an issue that requires further research.
Should output control at a "per generator" level be required, DFIG, VSD and synchronous machines all have the capability of this facility.
7.2.4 Speed Governing Requirements
As with output control, speed governing is usually not provided on wind farm developments. This is because the governing function can only be provided if the turbine operates at a lower power output than is possible at any given time so as to create a power "reserve" that can support the governor functions. This is usually called "delta control".
For very high levels of wind penetration, a governing function may be required. Current wind generator technologies can support this, although it may be difficult to retrofit this technology on to wind facilities not designed specifically for speed governing. The difficulty this presents is that the additional expense of requiring wind farms to include governor functions is unlikely to be justified at present as the facilities are not yet required to support grid operations. If a large wind penetration is developed over time, governing will almost certainly be required. The question that needs to be addressed is whether only future wind farms should pay the cost of this or whether it should be a requirement for all farms irrespective of the current need. Clearly this issue requires further assessment by the industry, policy makers and regulators.
7.2.5 Voltage Control Requirements
The EGR requirement for generators is that "each generator with a point of connection to the grid will at all times ensure that its assets are capable of being operated, and do operate, when the grid is operated within the range of voltages set out in rule 3.1.1:-" (shown in Table 3.)
Table 3: EGR Voltage Operating Limits
| Nominal Grid Voltage | Voltage Limits |
| 220kV | 198kVmin | 242kVmax |
| 110kV | 99kVmin | 121kVmax |
| 66kV | 62.7kVmin | 69.3kVmax |
| 50kV | 47.5kVmin | 52.5kVmax |
Direct connected induction generators as are used in stall regulated machines are not able to control their voltage and therefore will not comply with this requirement. The other three types of generators considered can all control voltage to some extent.
This requirement should be considered on the basis of the wind farm as a single generator, rather than on an individual machine basis, and should not provide any insurmountable difficulties. It is probable that the grid connection transformer will require on-load tap changers with voltage regulating provisions.
7.2.6 Reactive Power Requirements
Consumption and generation of reactive power must be matched in order to maintain a stable system voltage. As with frequency control, a reserve capacity is required to maintain system voltage stability during fault events, such as a generator tripping or other system disturbance.
Although there are other power system components that can be used to provide variable reactive power production or absorption, these are for the most part controllable only in steps of a fixed magnitude. Continuously variable reactive compensation is usually provided by synchronous generators using their automatic voltage regulator equipment.
The EGRs require that:-
- Generators must be able to export net reactive power equal to 50% of the maximum continuous MW output power when operating at full load;
- Generators must be able to import net reactive power equal to 33% of the maximum continuous MW output power when operating at full load.
This effectively requires that generators are able to have a power factor of 0.9 for export and 0.95 for import.
Figure 3: Reactive Capability Chart
→ Figure 3: Reactive Capability Chart [9 KB GIF]
Most modern wind generators can provide reactive power compensation of comparable capability to similar hydro or thermal generating sets. Direct connected induction machines, such as are used in stall regulated turbines, cannot provide reactive power and are in fact a consumer of reactive power.
Figure 3 provides a typical capability curve for a small (2MW at 0.9 power factor) synchronous generator, and for a double fed induction wind generator.
It can be seen that the DFIG generator provides similar reactive power capability to a standard synchronous machine. Note that at the DFIG maximum rated output, the supported power factors are 0.96 (export) and 0.98 (import). Therefore, to meet the full requirements of the EGR, the example wind turbine would need to be de-rated so that its maximum continuous output is about 90% of its nameplate rating.
The current EGR's also require the connected generators to be able to import and export reactive power over a full range of system voltages. In the case of wind farms, this requirement should be applied on the farm as a whole, rather than on a per turbine basis.43 For wind farms, it is likely that the voltage range requirement will need to be met by using transformers fitted with voltage regulating on-line tap changers at the grid connection interface.
Provided the EGR requirements for reactive support are enforced for all new wind farm developments future difficulties should not arise.
7.2.7 Additional Transpower Requirements
Transpower also has two documents that define the technical requirements for generators connected to their grid:
- Connection Policy 2001
- Connection and Dispatch Guide 2004
The Connection Policy includes a number of specific requirements for generating sets. These requirements have been determined to ensure the overall security and stability of the national electricity transmission network. The key requirements are:
- The generator must be capable of isolated operation. This requires the generator to be capable of controlling the system frequency when the generator is supplying a load isolated from the network and other generators. To control frequency the generator must be able to regulate generation to match the load at any particular instance.
- The output voltage quality must be a true sinusoid within the deviations permitted by IEC34-1.
- Be capable of operating under voltage imbalance conditions
Other general clauses in the Connection Policy also require all users to contribute to, and not adversely affect, the performance of the grid. Users must also "meet any other special requirements identified by planning studies undertaken by Transpower".
Note that the Connection Policy was published in 2001, before the EC was established by the Government.
The Connection and Dispatch Guide expands on the connection requirements, and also includes some of the requirements of the EGRs' common quality obligations. In particular the frequency and voltage control requirements are reiterated. The guide also discusses fault ride-through capability which, although not explicitly stated in the EGRs, is a function of the requirement to support system voltage. This requirement has clearly been added in response to the connection of large wind farms onto the grid.
7.2.8 Isolated Operation
The Transpower isolated operation requirement is in line with the EGR requirement for speed governing.
It is likely that a large proportion of wind farms will, at least initially, be embedded within the local distribution network. This is because most farms currently under planning appear to be adopting a staged approach so that the wind resource can be proven before full capital outlay is committed. It is possible that individual farms will transfer across to grid connection at a later stage as they outgrow the capacity of the local distribution network. Connection to the distribution system introduces some additional complexity to the connection arrangement owing to the radial nature of most distribution feeders.
Transpower defines a generator as "A person who owns and/or manages one or more generating sets which are physically connected to the grid assets or to a network or to other assets connected to the grid assets". In other words, generators not directly connected to the Transpower system but connected to a distribution network are expected to comply with the Transpower requirements.
This requirement causes an immediate issue as it is usually a requirement for generators connected to a distribution network to trip off during system faults and not allow the local network to operate as an island.
For a large embedded wind farm44 where the requirements of the EGRs are deemed to apply, the situation is more complex. For a fault on the transmission grid, the wind farm should stay connected and contribute to system restoration. However for a fault on a distribution system, where a circuit breaker connecting the distribution feeder to the transmission grid opens, it is important that the wind farm trips offline to prevent creating an "islanded" network. To cover both instances the wind farm needs to provide voltage ride through capability, but also requires anti-islanding protection. It can be difficult to meet both of these objectives satisfactorily.
7.2.9 Fault Ride-Through
The Transpower requirement is for the generator to stay connected to the grid for the duration of a fault on the system based on 3 phase-earth faults being cleared in primary time and 2 phase-earth and 1 phase-earth faults being cleared in backup time. Although these values vary from site to site, worst case conditions are summarised in the connection guide.
The requirement for generators to include fault ride-through provisions is initiating change in the design of wind turbines as machines with direct connected induction generators (e.g. stall regulated turbines) cannot meet this requirement. This issue is the subject of much discussion in the industry and wind turbine suppliers are able to provide solutions to this issue through the use of blade pitch control and static VAr compensation devices.
Figure 4 below outlines the Transpower minimum design target requirements, the maximum actual range and the voltage tolerance values for a typical modern dual fed induction generator that includes the necessary hardware for fault ride-through. It is noted that a standard DFIG does tolerate voltage fluctuations and will disconnect from the grid in order to protect the power converter electronics.
Points to note are:
- The generator voltage tolerance equals or exceeds the Transpower minimum requirements;
- The generator will not be able to tolerate the actual worst case (maximum) requirements, especially during the time period 0.2s - 1.0s after fault initiation.
Figure 4: DFIG Voltage Tolerances
→ Figure 4: DFIG Voltage Tolerances [7 KB GIF]
This second point is a potential concern for very high levels of wind power penetration. WTGs that provide better voltage tolerance between 0.2s and 1.0s after fault initiation are likely to become available. But if not, it may become necessary to improve the fault clearance time on large sections of the transmission network. Technical issues associated with main and backup protection coordination, plus the potential cost of upgrades to the transmission system protection and protection signalling systems, may make extensive improvements impractical.
From a brief review of fault ride-through provisions for other countries, it would appear that the Transpower requirements are slightly more onerous than those for other networks, but there are grids that have much more demanding requirements - for example, Scotland's. We believe that the voltage tolerance of wind generators will improve over time in response to the demands of the grid operators.
7.2.10 Harmonics
The EGR's require all connected generators to have a true sinusoidal output with deviations within the requirements of IEC34-1 "Rotating electrical machines". This standard outlines the method of assessment to be measurement of the harmonic currents from the generator when supplied by a system of balanced and sinusoidal voltages. A generator is deemed suitable if it has a harmonic current factor (HCF) of less than 5%. The HCF is a measure of the harmonic currents (up to 13th harmonic) relative to the generator rated current.
For a direct connected induction or synchronous machine, this requirement is not difficult to meet. However in order to meet the fault ride-through, frequency control and reactive support requirements, many wind turbine manufacturers are now using power electronic devices in the grid connection. The EGRs are written on the assumption that connected generators will be of conventional rotating synchronous or asynchronous type without power converters.
The generation of harmonics in the electricity system is covered by electrical code of practice NZECP36:1993. The code has clearly been developed from a consumer perspective and defines actual harmonic current injection limits (in volts and amps) rather than a percentage of installation rating. The NZECP requirements are more arduous than IEC34-1, for example a 30 MW wind farm operating at the current injection limits allowed in the NZECP would have an HCF of about 2.7%, well within the requirements of the EGRs.
Electronic converter devices can produce undesirable levels of harmonics, but the systems in use today generally include advanced converter and harmonic filter arrangements to limit these to acceptable levels. Suppliers' information does not relate these details in terms of compliance with IEC34-1 or NZECP36 and it would be necessary to evaluate these on a case by case basis.
7.2.11 Connection Summary
From the limited perspective of the rules and policies relating to connection alone, there would not appear to be any unsurmountable technical issues that would prevent arbitrary amounts of wind energy being connected to the New Zealand grid. Indeed, market pressures from European countries are forcing wind turbine manufacturers to develop solutions for most of the issues that are relevant in the New Zealand context.
However, wind farms must do more than just connect to the grid, either directly or via a local network. They must also be able to operate in a manner consistent with the safe and secure supply of electricity and the orderly operation of the electricity market. Operational issues are the subject of the following subsections in this section 7.
7.3 Grid Dynamic Stability
The dynamic stability of the grid is a complex issue that potentially covers issues such as fault ride-through and under frequency events. Nevertheless there are some aspects of stability which typically require further analysis using modelling tools to analyse particular scenarios. These are outside the scope of this study as relevant wind data is currently not available.
7.4 Interconnection
In Europe and the United States many countries or states, respectively, have one or more interties to other countries or regions. The interties provide system operators with more flexibility when it comes to making up for the "unders and overs" potentially caused by wind generation.
For example, if the output of all wind farms in a region falls, power may be imported over the intertie to make up the difference, if there is not enough plant available within the region and the interconnected region has spare plant. Likewise, a region with too much wind power may be able to export some or all of the excess to the connected region.
In essence, the presence of an intertie provides the SO with additional flexibility which, all other things being equal, could allow the region to support more wind generation.
In New Zealand, we have no interties to other countries so this additional degree of flexibility does not exist. The North and South Islands are connected by the HVDC link so that energy can be transferred one way or the other, but the link currently has limitations when it comes to transferring or sharing ancillary services including regulation, IR and voltage support.
7.5 HVDC Link
Limitations which currently apply to the HVDC link in respect of regulation and IR, in particular, were mentioned in previous sections. However, there are other properties of the HVDC link which could impose additional limitations on its ability to service increased amounts of wind generation in either island.
AC transmission lines are just wires that connect to each other passively, but the HVDC link is a direct current line which connects to AC lines at either end. AC and DC are not directly compatible so between the AC lines and the HVDC link, at each of its ends, there is sophisticated power electronic equipment which acts as either converter or inverter depending on which way power is transferring.
Transpower45 has already identified that increased switching of the direction of power flow on the HVDC link, occurring with increased activity in the wholesale electricity market, is likely to reduce the useful life of some equipment at either end of Pole 1, this being older than the equipment installed on Pole 2. In the last year, in fact, the HVDC link has operated for long periods at reduced capacity due to problems with control equipment on Pole 1, though these problems are not necessarily the result of having a market.
The useful life of certain items of key equipment could be further shortened if the presence of large scale wind generation increases the amount of direction switching occurring on the HVDC link, reducing reliability, increasing maintenance costs and potentially even requiring constraints on switching, for example. Increased switching is most likely to occur when the HVDC link is lightly loaded. Table 4 shows the proportion of the year in which the HVDC link transfers power in both north and south directions, indicating that switching occurred during these particular periods.
Table 4 : Proportion of the Year When HVDC Link Switches Direction
| Year | Periods with Switching |
| 2004 | 2.7% |
| 2003 | 3.5% |
| 2002 | 2.6% |
| 2001 | 6.8% |
| 2000 | 1.4% |
Because of the limitations on switching of the HVDC link, the SO currently changes the direction on the HVDC link only if the flow will go above a threshold value of a few tens of MW.
Assessment of the potential for high levels of penetration of wind energy to increase HVDC switching depends on a number of factors and requires detailed modelling of the operation of the grid and the wind farms that might be connected in future.
7.6 Regulation
In this country, regulation is often referred to as frequency keeping, but in this study we refer to it as regulation in line with overseas convention.
Maintaining a constant system frequency is vital to grid security because:46
- thermal generators will reduce output when frequency falls and some thermal turbines can be damaged by vibration at low frequencies;
- synchronous generators are controlled by governors which use frequency as the control variable which determines their output or, in other words, output is varied to maintain frequency;
- industrial processes and equipment can be damaged by excessive frequency excursions;
- equipment often senses system frequency in order to keep time.
Figure 5: Regulating Station
When synchronous generators connect to the grid they synchronise with the grid frequency and, with relatively few exceptions, their governors act to adjust power output to maintain constant output frequency. Thus all synchronous generators contribute to maintaining a constant frequency.
The first issue for wind in respect of regulation is that currently WTGs do not typically act to control frequency. Rather, they act to synchronise with whatever frequency the grid is operating at. WTGs will therefore simply follow the system frequency down during a plant outage.
The North and South Islands are connected only by the HVDC link which has AC-DC and DC-AC conversion equipment at either end. Since the HVDC link is DC, the frequency in the two islands is maintained independently and therefore there must be at least one regulating station in each island, acting independently of each other. These stations have special governors which react faster than others to frequency changes within their control bands, typically ±50 MW, centred around a value at least as large as their minimum output for regulation duty, plus 50 MW - refer Figure 5. Minimum control output is therefore typically 50 - 100 MW.
The SO monitors the frequency in each island and dispatches the appropriate regulating station up if frequency falls or vice versa. The SO also monitors time error in each island, which is the accumulated frequency error, and dispatches the regulating stations to eliminate this at least once per day.
Hence regulation requires at least one station in each island to be operating at 50 - 100 MW, placing a fundamental limit on how much wind generation there can be at any given time, i.e. load plus losses minus 100 MW. Wind farms do not provide regulation, as it is beyond the current limits of the technology and would, in any case, require a guaranteed response from a wind farm.
In the presence of large-scale wind generation, it is possible, though not indicated by overseas studies, that the variability of wind would take the regulating station outside of its ±50 MW control range. If this were to occur, then it is possible to have more than one regulating station.47
7.7 Instantaneous Reserves
The two regulating stations, one in each island, provide the fine control of system frequency during normal operations. IR, on the other hand, provides additional generation required to meet demand and bring frequency back to 50 Hz after a contingency, e.g. a generator outage, or the loss of a line, or one or two poles of the HVDC link. Similarly, ILR disconnects when the frequency drops below pre-set limits, thus reducing the rate at which the system frequency decays.
Table 5 : 2004 IR Summary
| 2004 IR Summary | North Island | South Island |
| Offered MW | Cleared MW | % Cleared | Offered MW | Cleared MW | % Cleared |
| FIR | Average | 603 | 150 | 26% | 306 | 66 | 23% |
| Maximum | 861 | 443 | 92% | 456 | 381 | 86 |
| Standard Deviation | 101 | 70 | 13% | 70 | 18 | 10% |
| SIR | Average | 754 | 319 | 42% | 421 | 121 | 29 |
| Maximum | 1156 | 513 | 100% | 540 | 325 | 64% |
| Standard Deviation | 113 | 65 | 11% | 50 | 10 | 4% |
Table 5 shows the offered and cleared FIR and SIR summary for 2004. The "% Cleared" column shows, in percentage terms, the average, maximum and standard deviation of reserves actually dispatched ("cleared") relative to the amount offered when taken over all trading periods in 2004, so it is not just the ratio of the "Cleared MW" and the "Offered MW" shown above.
The table shows on average there is a significant surplus of both SIR and FIR, though there are also periods when reserve supply is tight. For example, there are two trading periods when the SIR "% Cleared" is 100% - this means that all the SIR offered for that trading period was actually dispatched.
At dispatch time, SR and ILR are dispatched according to a relatively simple principle: there should be enough IR dispatched to make up for the largest contingency on the grid at the time.
Generators are often referred to as either stations or units. A generating unit is simply a single generator, but a station may contain a number of units. For example, the 1,000 MW Huntly power station is actually made up of four units of 250 MW each, housed in a single large building.
The risks associated with generator contingencies are based on units, not stations. For example, if the largest unit operating is 300 MW, the contingent risk that has to be covered by IR is 300 MW. In this example the SO would dispatch 300 MW of IR to cover the risk.
But, it's not quite that simple in reality. An N-1 contingent event results from the loss of one generating unit or one line. After an N-1 event which causes a sudden drop in frequency, it takes of the order of 60 seconds to work through from initial event to the recovery of the system frequency to 50 Hz. But some IR is better for providing immediate response and some is not. Hence there are two classes of IR in the New Zealand market - fast IR (FIR) and sustained IR (SIR). FIR can be thought of as primarily helping to arrest the fall in system frequency, while SIR can be thought of as primarily restoring frequency to 50 Hz.
The HVDC link is also counted as a potential risk, e.g. one or both poles might suddenly disconnect, so its ability to share energy between islands immediately after a contingent event is limited. As a result, the reserve risk and IR dispatch are calculated separately for each island in RMT.
Could a wind farm become the largest risk and set the reserve risk? Quite possibly, as long as it is possible that the whole farm could be lost in one go and that the farm is also sufficiently large. At present, the risk in the North Island is typically set by the large gas-fired stations such as Otahuhu B (up to about 365 MW), the Taranaki combined cycle plant (up to about 355 MW), a unit at Huntly (up to 250 MW) or the risk of a pole tripping on the HVDC link (up to 500 MW.) It is not beyond the realms of possibility that a very large wind farm could set the risk as it would only have to be operating at more than the output of the largest thermal unit.
But wind speed is variable and so is the output of wind farms. So this begs the question - how would the SO know that the wind farm would have the highest output, say 5 minutes in advance of when dispatch is done? The answer is that persistence forecasting provides a reasonable estimate of wind farm output up to at least 10 minutes ahead.
A more difficult question arises in respect of the risk of a large wind farm shutting down suddenly when the wind speed significantly exceeds its wind cut-out speed. This would be equivalent to a large unit disconnecting. More importantly, if a number of wind farms are clustered in a relatively small geographical region a large storm with high winds could shut down a much larger amount of wind generation. This type of event would not only create significant requirements for IR but might also be difficult to predict in terms of timing and magnitude.
This type of event has occurred overseas48 and been tolerated by the system. It has happened in Germany - which features a number of interties to other countries. In New Zealand, severe weather events occur with surprising regularity, so an approach to IR management must be adopted which will safely accommodate such events. In particular, the assessment of risk for the purposes of dispatching IR may need to include the risk of wind farms being shut down simultaneously by severe weather events, or conversely rapidly increasing their combined output.
Studies undertaken overseas focus on an N-1 contingent event for the purposes of evaluating the impact of wind generation on IR, but in this country the analysis has a subtle difference. The "SPD" software used by the SO for dispatching generation calculates the dispatch of energy and IR that minimises the total cost of providing energy and IR, based on the offers for energy and IR submitted by generators and IR providers for each half hour. SPD calculates the reserve risk according to a formula which has the basic form
where RAF stands for "risk adjustment factor" and the risk offset allows for factors which may reduce or increase the amount of IR actually required. In simple terms, RAFs are calculated for FIR and SIRIR in each island and applied to the basic risks used in SPD. Prior to 2004 the RAFs potentially varied by half hour but from 2004 onward the adjustments to the basic N-1 risk made by the RAFs appear to have largely been replaced by the adjustments made by the risk offsets. Data is not readily available for the value of the risk offsets, so the following discussion uses RAFs as a surrogate. The net effect is the same for the purposes of this study.
Up until 2003, each RAF was the ratio of the reserve requirement calculated by another software model called "RMT" and the reserve initially dispatched by SPD.49RMT contains an accurate model50 of how the system frequency responds to a contingent event or extended contingent event. It takes a snapshot of the state of the grid and calculates the reserve required to maintain frequency above preset limits. Thus
The process is shown diagrammatically in Figure 6.
Figure 6: Calculation of Risk Adjustments
In New Zealand then, the impact of large scale wind generation on IR has three dimensions, each of which must be considered carefully:
- the ability of the system to provide IR;
- the impact of wind farms on the basic N-1 reserve risk;
- the impact of wind farms on the adjustment factors from RMT.51
Figure 7 shows that RAFs varied across a day, where the time axis is split into half hourly trading periods. On New Year's Day of 2003, the RAFs varied between 0.43 and 1.06.
Figure 7: RAFs Varying Across a Day
Figure 8 and Figure 9 show all RAFs for 2003 for North Island FIR and SIR, respectively, and illustrate the significant range over which the RAFs varied. Since December 2003, the North Island RAFs have all been equal to one and the adjustments to SPD's assessment of reserve risk have been taken over by the risk offset values. Despite the change in the role of the RAFs from 2004, they nevertheless illustrate the point that significant adjustments are made to SPD's basic reserve risk calculations.
Figure 8: 2003 North Island FIRRAFs
Figure 9: 2003 North Island SIRRAFs
Figure 10 shows the response of the system to a contingent event or, in fact, an extended contingent event. The system frequency initially decays at a rate which is influenced by the amount of generation and load. Generators that provide FIR start to increase their power output, and ILR may disconnect if the frequency falls sufficiently. Eventually, plant scheduled for both FIR and SIR, along with other conventional plant, increase their combined output to restore the frequency to 50 Hz.
Figure 10: System Frequency Response after a Contingency
The change in frequency f in a short time interval ∆t is given approximately by
where H is the average "constant of inertia" over all of the generation connected to the system and generating in an island.52 Immediately after an event involving loss of a generating unit, for example, Generation - Load is negative. The formula shows us that frequency falls faster as the size of the loss relative to the amount of load increases. For example, loss of 300 MW when the load is 1,500 MW will produce a frequency decay at twice the rate of the same 300 MW loss when the load is 3,000 MW.
The formula also tells us that frequency falls more slowly when the constant of inertia is higher. Inertia is a measure of how much energy is stored in rotating machinery connected to the grid. For any particular generator, H is a function of the speed of rotation, its mass and the square of the radius of the generator. For example, a uniform cylinder of mass m, rotating at w rpm and having radius r has kinetic energy 
which is then divided by the rated output of the generator to give the H value.
This raises the question - does the presence of wind generation significantly change the average inertia of the system? If it does, then wind generation could significantly change the risk offset values that come from RMT and are entered into SPD.
Transpower's RMT model includes calculations of the system inertia, of how load responds to changes in frequency and how generator governors respond to changes in frequency. For this study, Energy Link developed a very much simplified model, similar in principle to RMT, in order to estimate the magnitude of any effect on the risk adjustments coming from RMT with various amounts of wind generation.
As WTGs have got larger, so has their inertia, primarily due to the increase in blade length. The Vestas V90 3 MWWTG, for example, has a rotor diameter of 90 m. Based on estimates made during the course of this study and confirmed by discussions with WTG manufacturers and hydro electric generators, typical hydro generators and modern WTGs have inertia that averages just under 3.53
But two questions arise:
- Is the rotor inertia available to the grid in the same way as it is for a synchronous generator, and thus able to assist in reducing the rate of decay of frequency after a contingency?
- Does the WTG control system respond to the frequency drop by increasing WTG output, thus assisting in bringing frequency back up to 50 Hz?
Given current technology, the answer to the second question appears to be that WTGs do not increase output in response to frequency drops, though they may be able to in future with adaptations in their controls, and by running below the maximum output possible in any given wind conditions, i.e. by holding some output in reserve.
The contribution of WTGs to maintaining system frequency, either under normal conditions or after a contingency, depends on the technology in use. Synchronous WTGs such as the Windflow Technology WTG do contribute to inertia, although without modifications to the turbine controls the contribution falls as the frequency falls.
Any WTG configured such that output passes entirely through a power electronic converter will not allow inertia to support frequency.
For DFIGWTGs, only about one-third of the output passes through a converter and the other two thirds54 is connected directly to the grid. So in this case, about two thirds of the WTG inertia is available to the grid and reduces the rate of decay of frequency after a contingency. Currently DFIG turbines do not increase output in response to a drop in system frequency, they simply follow the frequency down. In future, with improvements in control systems, combined with operation of a DFIG turbine at less than its potential output, there is the potential for these WTGs to increase output during an under-frequency event.
The simple model of system frequency response developed for this study was used to asses the impact on risk adjustments for the purposes of making the initial assessments of penetration and market share. With the levels of penetration estimated in sections 8 and 9.2 the risk adjustments increased by between 20% and 40% relative to the risk adjustments with no wind energy at all. These are relatively modest increases given the significant surplus of reserves apparent at present, so overall this would not be expected to significantly impact on the operation of the grid or on the cost of providing reserves, at least on average.
However, this is a crude initial assessment and more work is definitely required in this area. Even now, there are periods when reserves appear to be in short supply, but it is not known whether this is because there are not enough reserves available or because there is just not much offered into the market.
7.8 Other Ancillary Services
In section 7.6 and 7.7 we dealt with regulation and IR, respectively, and in section 7.2 we covered voltage support, which leaves black start and over-frequency arming ancillary services.
The potential for large amounts of wind generation raises the prospect of there being too much generation, particularly likely during periods of low load, and having to disconnect or limit the output of individual WTGs or perhaps whole wind farms. The facility to disconnect or limit output could conceivably have use in the over-frequency arming service if it can react sufficiently quickly to increases in frequency. A related concern, however, is that the variability of wind generation could actually make over-frequency events more likely in future.
Black start is hardly ever required, so overall its impact on wind generation, and vice versa, is probably negligible. Although WTGs could conceivably provide this service for much of the time, as long as the wind is blowing, black start is more likely to be contracted to generators that can guarantee to start at any time.
7.9 Constraints on the Grid
Transmission lines making up the grid are there for one reason only - to transfer power from one geographical location to another. We say that a line is constrained, or reaches its limit, when the power flowing through the line reaches the maximum amount that is allowed to flow through the line. The SPD model is used to dispatch generation and IR at the minimum total cost which simultaneously ensures that there is sufficient generation to meet demand, that system frequency is maintained before and after contingent events, and that there are no lines which have more power flowing through them than their allowable limits.
Lines typically constrain because there is more generation upstream of the constraint that is "cheaper" than downstream generation, than can be supplied across the line. When lines constrain then spot prices on either side of the line can separate to a large degree. For example, a constraint in the half hour ending at 11 am on 21 August 2004 produced a "spot price" in excess of $11,000/MWh and less than -$1,300/MWh at either end of the constrained line.
These large price separations signal the value of generation at various points on the grid. If a line often constrains, the pricing signal can be strong enough to encourage someone to build new generation downstream of the constraining line, which is a way of ensuring that generation is built where it is needed. The alternative is for the line to be upgraded, which again may be signalled by a continued price separation.
But participants in the spot market sometimes find it difficult to interpret and manage their operations given the often extreme pricing signals from constrained lines. For instance, the example cited above was the result of a temporary constraint on a line that was under maintenance, rather than a signal that a line was nearing its useful limit.
Any new generation, depending on where it is connected to the grid and how and when it runs, could either increase or reduce the occurrence of constraints at various points on the grid. What is peculiar to wind, however, is the variability in the output of wind farms and the potential for this to trigger constraints more often, making it more difficult for the SO and spot market participants to manage their operations around constraints. Table 2 shows potential wind farm development scattered around the country. All other things being equal, if wind farms of modest size are developed more or less evenly around the country, then the potential for a significant increase in the occurrence of line constraints is small.
Figure 11: Lower North Island Grid
But the key factor is the degree of concentration of the installed wind base in New Zealand, a case in point being the rapid growth of wind generation around the Manawatu, in which region 97.5% of the country's wind energy is concentrated. This is not the only region with a high degree of potential for wind energy - the Wellington region is even more attractive - but it has three features which are encouraging rapid growth in the installed wind base:
- a good wind resource - a number of sites have average wind speeds around 10 m/s;
- a number of sites for wind farms that are otherwise attractive, e.g. not too close to populated areas, and having good access;
- there are already sites there which can be expanded or which are close to other potential sites - this means that existing access and infrastructure may be able to be used, thus reducing the incremental cost of the next wind farm or wind farm expansion.
The Tararua and Te Apiti wind farms inject into or near the BPE55 and WDV56 grid nodes as shown in Figure 11. This is a particularly important part of the North Island grid. Power from South Island hydro-electric systems can flow north past Wellington, through BPE and towards Auckland. Or power can flow south through BPE to Wellington and then south on the HVDC link, as it does during times when water stored in southern hydro lakes is being conserved for later use.
Because the wind farms that inject, or may inject, at or near BPE are geographically close, their outputs may also correlate to produce large swings in output - possibly over short time-frames under storm conditions, for example.
Southward flow is particularly important as it is required for long periods during dry years, especially if southern storage is low when southern hydro inflows are low and demand is high, as happened during the winters of 1992 and 2001. The lines between BPE and the southern end of the HVDC link at Benmore, can reach their limit and restrict the amount of power that can be sent south, so the ability to load these lines up to their maximum can be critical. If large amounts of wind generation inject at or near BPE, and these are also highly variable and unpredictable, then it could be very difficult for the SO to dispatch generation accurately enough to keep maximum amounts of power flowing south from BPE without risking overloading the same lines. Under such conditions, it may be desirable for the total output of Manawatu wind farms to be limited to a constant or slowly changing value below their maximum possible output, given the ambient wind conditions.
The only way to adequately estimate the impact of wind farms, either in one or two clumps or dispersed about the country, on constraints is to undertake detailed modelling studies which will require a database of regional wind speeds. This data is not available at present so the impact has been assessed for the Manawatu only, and then on a static basis for one or two important cases.
For the purposes of this study, a static analysis was conducted which indicated that large amounts of wind generation at BPE, for example, could increase the likelihood and frequency of constraints on the BPE-TKU lines, which runs north from BPE, as shown in Figure 11. This static analysis is consistent with the findings in Transpower's review of Manawatu wind generation which noted that there may be implications for the limits on power flowing in these lines given the large swings that have so far been observed in total wind farm output. The Transpower report also noted the potential implications for the limits on southward transfers on the HVDC link, again due to large swings in Manawatu wind farm output.
The Te Rere Hau wind farm project proposes to add 52 MW of WTGs in 2008, and Stage III of the Tararua wind farm could add a further 120 MW, both of which are likely to add to the magnitude of the swings cited in the Transpower report. It is conceivable that the addition of wind farms in the Manawatu could require upgrades of the grid in that region.
7.10 Market Issues
The potential for large-scale wind development in the Manawatu region to constrain key lines also has the potential to create extreme price volatility in the wholesale electricity market from time to time. This and a number of other market issues are described below. Some of these factors, either individually or in combination, could conceivably limit the potential for wind integration in New Zealand.
7.10.1 Regulation
In real time, overseas studies tend to indicate that the impact of wind on regulation is manageable, because:
- to a large extent the output of wind farms can be predicted relatively easily in real time and beyond to dispatch time;
- assuming that wind farms are developed at sites which are geographically dispersed, because wind speed correlates poorly over larger distances, the aggregate output of all wind farms exhibits significantly less variability than the output of any single farm.
The latter assumption could break down if it turns out that New Zealand's wind farms are built primarily in, or near to, one region in each island.
7.10.2 Dispatch Time
Dispatch time is the domain of the marginal generator or generators. Wind farms are offered into the market at $0.01/MWh so it is unlikely, though not impossible, that they will be on the margin. Assuming there is sufficient plant available to be on the margin - typically plant that offered more than $0.01/MWh - then there should be no problem in matching generation to demand.
However, despite the best pre-scheduling efforts, the SO could get caught with insufficient marginal plant, or with too much wind generation, and therefore need to turn the marginal generator off, or turn wind output down in order to keep the marginal generator running. Here is a potential role for hydro under extreme conditions as it fits the role of highly flexible marginal generator almost perfectly.
7.10.3 Spot Price Volatility
Assuming that hydro plant and the flexible thermal stations can act as marginal stations, the load followers, then one impact on the market will remain with an increase in penetration of wind energy: spot prices will become more volatile in the short term. This will be due to two effects:
- an increase in the frequency of constraints, and
- an increase in range of output of the marginal stations which will cause a wider range of marginal offer prices to reflect through into spot prices.
The latter effect needs to be understood in the context of how generators offer into the spot market. Each offer can consist of 5 separate price-quantity pairs. A marginal generator might offer, for example, five bands of 50 MW each increasing evenly in price from $20/MWh to $100/MWh. Suppose that without wind energy the marginal generator is dispatched from its 2nd to 4th offer band, then spot prices will vary from $40 to $80/MWh. But with more wind the variability in its output will increase so that changes in spot price could become more common and larger, e.g. from the 1st to the 5th offer band, or from $20/MWh to $100/MWh.
In theory, this impact on the spot market should not limit the penetration of wind energy since spot prices are highly volatile anyway. But increased levels of price volatility have an impact on perceptions of risk around electricity which could conceivably result in risk premiums in supply and hedging contracts for wind and other generation.
In addition to greater volatility in spot prices on an operational time scale, there will also be greater quantities of generation constrained on and constrained off created with large scale wind generation, simply due to the increased difficulties in scheduling and dispatching plant with increasing overall volatility in wind farm output.
In the medium term, high levels of wind integration could conceivably actually reduce spot price volatility. For example, consider the impact of dry years on spot prices - they have proven in the past to create price "spikes" over the course of several months. But depending on how wind speeds vary by month and by season, wind generation could conceivably enhance dry year security of supply and reduce the spot price volatility currently associated with dry years.
7.10.4 Pre-Dispatch Time
At pre-dispatch time, however, the issues around large scale wind generation start to mount:
- How well can the output of wind farms be forecast up to 36 hours ahead, the time-frame over which the PDS is published? Forecasts this far out will rely heavily on weather data, in particular predictions of wind speed;
- How will the SO allow for cover if a storm is forecast which could cause a number of wind farms to sharply increase output or to turn off almost simultaneously?
- Is there enough marginal plant offered into the market to cover the potential variation in wind farm output?
- Do we need to pre-schedule more plant than is currently required just in case wind output is low?
- Can we provide generators with large thermal units that take a long time to start up from cold, with adequate warning that their units will be needed?
There is currently no requirement for generators to offer any plant whatsoever into the spot market. As wind energy penetration increases, one can foresee that the SO may have increasing difficulty obtaining enough offers to complete the PDS. In addition, the accuracy of forecast prices published by the spot market could reduce significantly due to the variability of wind farm output.
The EC may need to consider making it mandatory for generators to offer all plant when it is reasonably available so that the pre-dispatch schedule can be completed.
Taken to the limit, there may be more situations than there are now when there is simply not enough plant available to fill the potential gaps between wind farm output and demand.
Some indication of the difficulty in managing availability of plant in pre-dispatch time comes from the Huntly power station which currently draws its cooling water from the Waikato River, thus heating the river downstream of Huntly. The river temperature is not allowed to exceed approximately 25°C, which has the effect of reducing Huntly's output by hundreds of MW on some hot summer afternoons.
Figure 12: Impact of Huntly Cooling Constraint
→ Figure 12: Impact of Huntly Cooling Constraint [23 KB GIF]
Figure 12 shows the impact of the Huntly cooling constraint on 11 February 2005. This drop of nearly 600 MW in output was manageable, though spot prices increased significantly. The magnitude of the output drop was similar to what might be expected from very high levels of wind generation, but the drop could occur much more quickly with wind.
7.10.5 Block Dispatch
The factor that offers perhaps the most implicit support for wind energy in the New Zealand market is the large amount of hydro electric plant. It has long been recognised that the operation of a wind farm in tandem with a hydro system allows the wind energy to be used when the wind blows, and stored hydro energy at other times.
In the New Zealand market, block dispatch is a departure from the "ideal" model of each generator being offered and dispatched separately. In theory, the market will provide plant to make up for the variability in wind in an efficient manner, without the need for close physical proximity or even for common ownership. However, to provide generators with greater flexibility, the EC may need to consider extending the concept of block dispatch to include a wind farm and a hydro system located adjacent to each other. The wind-hydro block would offer as usual, but would have flexibility to meet its exact dispatched quantities with either wind or hydro or both. This would effectively eliminate the variability of the output from wind farms. Similar arrangements could be contemplated for a wind farm and a thermal station.
As noted in section 7.13.2 network companies could be significant investors in smaller, embedded wind farms. Network companies also control a significant amount of load in their networks, so it is conceivable that load control and wind farms could be controlled in tandem to reduce variability in output, or to achieve other aims. In cases where a wind farm is embedded in a local network this could be achieved without using an extended block dispatch, unless the wind farm is large enough that the SO requires it to be offered into the spot market.
7.10.6 Market Summary
Of all of the market issues discussed in this section 7.10, the only one that potentially limits wind integration, within the limits defined for this study, is the ability of the market to provide sufficient plant to cover for the cases when wind farm output is considerably less than forecast up to 36 hours ahead. Plant might not be available, for example, because it cannot be started in time. This is the case of the large thermal unit which may take several hours to start from cold. Plant might also not be available because it is not considered economic to offer it into the market on the off-chance that it might be needed.
7.11 Wind Speed Data
Many of the potential factors that could limit wind energy can only be investigated thoroughly with a reasonable wind speed database, in particular:
- how the variability in wind output changes given the time over which it is measured, e.g. by minute, by hour, by day, by month and by year;
- how wind generation correlates with demand;
- how wind generation correlates with other sources of variability including Huntly's cooling constraint and inflows into hydro electric systems;
- how the variability of the total wind generation output decreases as more wind farms are built, including the impact of how concentrated the wind farms are in one or two regions;
- the potential for wind generation to trigger line constraints;
- the degree to which wind generation can be accommodated in real time, dispatch time and pre-dispatch time.
As an initial assessment, this study has not had access to any wind data and so has relied primarily on the results from overseas studies and on anecdotal evidence. In order to study these factors with any degree of accuracy, a wind dataset is required. The absence of such a dataset will limit the industry's ability to assess the impact of individual and aggregated wind farms and also the ability to spot opportunities and anticipate and prepare for problems.
7.12 Geographical Dispersion
All the available evidence from New Zealand and overseas indicates that the variability of total aggregated wind farm output reduces as the geographical dispersion of wind farms increases. This applies both to variability on average and to the largest swings in total output.
Given that the impact of the variability of wind farm output on the operation of the grid and electricity spot market is a major factor in determining if and how operation is possible or feasible, it follows that the degree of dispersion is potentially a key limiting factor, albeit indirect, for wind energy integration.
7.13 Other Factors
A range of other factors may act to limit the total penetration of wind energy in New Zealand, but these are outside the scope of this study under our definition of wind energy integration. A number of these factors are briefly described below.
7.13.1 Relative Economics of Wind Energy
Electricity prices have risen recently in New Zealand at the same time as the cost of WTGs, measured in $/MWh of output, has decreased, to the point where wind farms have become economic. Another contributing factor is the ability of wind farms to bid for tradeable carbon credits which further improve the economics of marginal emission-reducing projects.
On the other hand, and despite recent concerns about a lack of spare capacity in dry years, the electricity market has since 1996 provided new capacity more or less in line with growth in demand. The need to build new plant is signalled to a large extent by expectations of rising prices as the gap between supply and demand reduces, and vice versa.
The rate of growth of wind generation will inevitably be limited by the balance of supply and demand and the impact on prices. In the case of wind energy, it currently adds to base-load generation and adding large amounts of it could produce long periods when spot prices collapse, making it hard to justify as a good investment.
In the longer term, driven by demand around the world, especially now that the Kyoto Protocol has come into effect, and by advances in technology, we expect the relative $/MWh cost of WTGs to fall further, thus allowing wind farms to be developed at sites with lower average wind speeds than the 9 - 10 m/s currently required to make a wind farm economically viable.
7.13.2 Smaller Wind Farms
The New Zealand electricity industry is split into two groups: energy suppliers, dominated by the five or six largest generator-retailers, and network companies. Based on anecdotal evidence, a number of network companies are investigating wind farms or other smaller generation projects scattered around the country. Under the Electricity Industry Reform Act of July 1998 (EIRA) networks companies are allowed to directly own up to 5 MW of generation, or 2% of the maximum loading on their network, whichever is the greater.
The large generator-retailers might take up smaller wind farm projects in their own right,57 or if they prove to be uneconomic for network companies given the constraints of the EIRA. The trend to date however has been for wind farms to be built in the areas with the most favourable conditions, e.g. the cluster that has developed in the Manawatu. It is, however, desirable to have greater geographical dispersion in wind farm location, thus helping to minimise the variability of aggregate wind farm output.
In section 7.10 we also noted that network companies are in a position where they could combine wind farm output and their load control to reduce variability in the demand their network places on the grid.
Networks and other small potential wind farm owners also face difficulties in selling their production once their wind farm is built. While it is relatively easy to arrange to sell the output of a small generator into the spot market to receive spot prices, most small generators, and their financial backers and stakeholders, would prefer to have the benefit of some hedging contracts in order to make their revenue streams more predictable.
Hedging contracts can at times be difficult to obtain at a price on a par with other hedge contracts sold in the wider hedge market. This phenomenon is not universal, and depends on a number of factors including timing and how highly hedged the market is in general. Some smaller generators may have the time and resources to offer their own hedges direct to large retail consumers, but some will not.
The risk of having to accept hedge contracts at a lower than ideal price, combined with additional costs or lack of scale, is likely to deter some smaller generators and therefore potentially limit the number and geographical diversity of wind farms in this country.58
7.13.3 Wind Energy and Dry Year Security
What will be the impact of large scale wind generation on dry year security? Given the recurrence of dry years - 1992, 2001, 2003 - which have caused significant difficulties for the electricity market, this is a key question, though not on our operational time scale.
If wind speed correlates closely with hydrological inflows into the large storage lakes, then wind energy could aggravate future hydro crises. But at present we simply cannot answer questions such as - did the wind blow around the country all through the first half of 1992? If it did not, then adding large amounts of wind energy could require a considerable degree of caution.
Anecdotal evidence suggests that wind speed is correlated with hydrological inflows, with average wind speeds peaking in spring and early summer when it also usually rains around the Southern Alps and fills southern storage lakes. Both wind and inflows may be closely associated with the prevailing westerly airflows across New Zealand.
But the seasonal variation in wind speeds appears to be much less than the seasonal variation in inflows into the southern storage lakes, which could indicate that wind energy could be a much steadier supply of energy over the medium term, than generation from hydro inflows.
These questions can only be answered by producing a wind speed dataset and comparing this with the hydrological inflow database that the electricity industry has had for many years.
7.13.4 Reliability of WTGs
The load factors achieved by wind farms in this country are high by international standards,59 which raises the possibility that WTGs installed in New Zealand may be less reliable due to the high duties they must perform.
Anecdotal evidence suggests that the reliability and maintenance costs of WTGs here are consistent with the duty they achieve. The global installed base of WTGs is also ballooning as the technology matures. So reliability does not appear to be a limiting issue for wind energy in New Zealand.
7.13.5 Public Acceptance of Wind Energy
There has been strong and widespread public opposition to some forms of generation in this country - for example, further hydro-development on the Clutha and Waitaki rivers. The evidence so far is that the public largely supports wind energy, but also that not everyone wants a wind farm across their back fence.
With the notable exception of single WTGs in Wellington and Christchurch, wind turbines and wind farms built to date have tended to be distant from populated areas. The Wellington region has enormous potential for wind energy but there appears to be significant opposition to having wind farms close to the city. As more remote wind sites are used up, it may become increasingly difficult and expensive to gain resource consents60 to build or extend wind farms, thus limiting the potential integration of wind energy.
7.13.6 Competition with Other Base-load Generation
Wind energy is offered into the spot market at $0.01/MWh and therefore competes with other base-load generation including:
- hydro plant that must operate to achieve minimum flows or lake levels required under the terms of their respective resource consents;
- slow-start thermal units running at minimum output in expectation of running at higher output later on;
- thermal plant running to use up fuel that must be paid for whether or not it is used;
- plant that also produces energy for an industrial process.
When demand is low, as occurs at night especially in summer, the base-load sector of the market can be crowded, so crowded in fact that there can be more base-load plant than demand. In this case the spot price is likely to be very low and some base-load plant will be dispatched down or off. If the growth in wind generation was very rapid, exceeding the underlying growth in demand for example, then competition for base-load could potentially see large amounts of water spilled from hydro-electric storage lakes from time to time. While there are no laws against spilling from hydro systems, spill figures must be reported by the large hydro generators. It hardly makes sense for wind energy to displace hydro energy to the point of creating excessive hydro spill.
Assuming that overly rapid growth in wind generation would in turn reduce prices, through the natural balance of supply and demand in a market, then this would also reduce the incentives to invest in more wind generation. The impact of this factor, therefore, is to moderate the rate of growth of wind generation.
Back to Top