3.4 Changes required to accommodate high penetration IRG

3.4.1 Forecasting

To manage increased levels of IRG, changes to the power grid are predicted to be necessary. Changes to operations, infrastructure, and planning and power quality management are suggested in the literature.

It is widely agreed that accurate forecasting is an essential element for the successful integration of large amounts of intermittent generation. It was mentioned in [6] that accurate forecasting is necessary for solar power to be economically viable. Forecasting at various timescales is required. More accurate day-ahead prediction of renewable resources is required for more accurate unit commitment [13]. Satellite images can be used to track cloud movement for forecasting at short timescales while numerical weather models can be used to predict insolation out to a number of days [1].

Figure 26 shows the impact of incorporating an hour-ahead wind forecast on the duty imposed on a proxy Automatic Generation Control (AGC) unit [10]. The power output of this proxy unit approximated the amount of regulation required of all units on AGC between 5-minute re-dispatches of the system. The blue line represents the original economic dispatch model with a persistence-based wind forecast and the yellow line represents the improved economic dispatch model. The green lines represent the minimum up and down regulation procured during the study period. Originally, the proxy AGC unit output was offset by an average of about -180 MW, as indicated by the heavy blue line. With a hour-ahead forecast included, the offset is reduced to about -60 MW, as indicated by the heavy yellow line. This shows that the Economic Dispatch (ED) units are better able to follow load with the improved wind forecast incorporated into the dispatch function.

Figure 26 Graph showing impact of incorporating hour-ahead wind forecast on AGC unit [10]

3.4.2 System flexibility

Increasing system flexibility or decreasing the flexibility requirements of the system is another important determinant for increasing levels of intermittent generation. A study discussed in [10] claims a required increase in system flexibility of 40% for the 2020 scenario (33% renewables) for normal load conditions and 50% for light load due to greater impact of renewables during such conditions. Due to increased variability, the number of start/stops increases and sustained load ramps (up and down) steepen. The load following capability will also need to increase above what is required for variation in load alone due to renewable [10].

It is claimed in [13] that system flexibility can be increased through

  • balancing the generation portfolio
  • introduction of more flexible conventional generation
  • redesign of power system to enable it to handle reverse power flow from distributed PV.

Reducing net load variability reduces the required flexibility of the system. Measures suggested in [13] are:

  • energy storage
  • load control
  • increased control and communication
  • ability to curtail Intermittent Renewable Generation (IRG) would reduce required flexibility
  • spatial diversity of the resource.

Experience with systems containing large amounts of variable renewable generation shows that flexibility of the portfolio balance is crucial for economic and stable operation [6]. For a CSP system, introduction of thermal storage would introduce greater flexibility to the production profile. It is predicted that more flexible generators, (ones which are able to not only vary their output but cycle on and off quickly), will find more situations where they can respond to network requirements and hence are likely to be dispatched more often. This is to manage what will be variable net load due to the large penetration of solar power.

A measure of the flexibility of a system is its minimum load capability. In [10], it is suggested that Californian grid operators should plan for a combination of flexible generation and import-export agreements to allow for a smaller minimum net load (load minus wind and solar) and greater net load variability. A few different methods of system flexibility which would allow a larger penetration of intermittent renewable generation could consist of:

Minimum turndown: Generation able to operate at a lower minimum power enables greater flexibility in the system and reduces the amount of renewable generation curtailment required. New generation will be required to operate at lower power levels and existing generation may be required to upgrade plant in order to enable operation at a lower power level.

Diurnal start/stop: More generation able to operate economically on a diurnal cycle (run at the same time for the same period each day) could be scheduled. These generators could run at the morning and evening peaks only, and be off during what would otherwise be uneconomical times.

Load participation: Large loads could take advantage of inexpensive power periods, further reducing the minimum operational load. Load shifting, for example in cold stores, and energy storage will also contribute.

According to [10], the evaluation of the generation flexibility of a system should be done at the ‘load following’ time scale in relation to the variability of the net load. It is also suggested that the selection of sites for renewable generation should also take into account the costs of required transmission to accommodate the new generation. It may be more feasible to place the renewable generation at a site where yield is less but connection costs are lower. It is also expected that existing contributors of frequency stability at the ‘regulation’ timescale will step up to manage the increased variability greater IRG penetration will present.

3.4.3 Curtailment

Curtailment requirements for ERCOT for varying levels of minimum load are analysed in [19] with penetration of up to 80% renewables. Curtailment of generation from variable renewable sources would be required if the generation portfolio is not sufficiently flexible to manage the increased fluctuations in net load introduced by the intermittent generation. Figure 27 shows the simulation results of two scenarios at ERCOT, one where the minimum load is 21 GW (65% below annual peak load of about 60 GW) and another with minimum load of 13 GW (78% below annual peak load). In the first case, 21% of the intermittent generation must be curtailed due to the minimum generation constraints of inflexible generation with wind and solar only contributing 20% of the energy demand. Less than about 3% curtailment is required in the second case by increasing flexible generation with renewables contributing 25% of the system’s annual energy. Insufficient transmission capacity also impacts on flexibility. A real world example of this occurred in 2009 where insufficient transmission from West Texas to loads in the Eastern USA resulted in 17% curtailment of wind generation [80].

Figure 27 Impact of system flexibility on curtailed energy for ERCOT [19]

Further simulations on the ERCOT system shows the relationship between wind penetration level and required curtailment for different levels of ‘System Flexibility’ (minimum load as a percentage of annual peak load). Figure 28 shows that the required amount of wind curtailment is significantly reduced with the presence of a larger amount of flexible generation. It can be seen that when the flexibility of the system is increased by 10% from 80% to 90% for a wind penetration level of 50%, the amount of wind generation curtailment required drops from about 45% to 20%. This shows how the level of system flexibility affects the amount of intermittent renewable generation that can be accommodated and utilised in a power system. It has to be noted that increased system flexibility level can incur increased operational costs which have to be taken into consideration during planning stages.

Figure 28 Wind penetration level vs. required curtailment [19]

The required level of curtailment when solar power is added to the generation mix is shown in Figure 29 [19]. It is seen that the required curtailment of renewable generation is reduced with the integration of solar into the renewable mix, with minimum curtailment seen when the solar/wind ratio is at 30/70. The required curtailment is seen to increase beyond that of the wind-only scenario when the proportion of solar exceeds that of wind in the renewable mix. According to the author, this is due to the limited spatial diversity of the solar resources compared to the wind used in the model. With greater spatial diversity higher penetration of solar with less curtailment should be possible. Still, 80% penetration of intermittent renewable generation requires 42% curtailment in the wind-only scenario while the 30/70 mix only requires 33% curtailment. This shows the advantage of mixed renewable generation.

Figure 29 Required curtailment with various levels of solar power added to the renewable mix [19]

The effect of storage on the required curtailment of the above system was also simulated and the result is shown in Figure 30 [19]. For 24 hours of storage with 80% intermittent renewable penetration, the required curtailment reduces from 33% to 10%. Nonetheless, there is only so much storage can do, as shown in Table 6. Diminishing returns can be seen for the same increase in storage. Twelve hours of storage in the ERCOT region is equivalent to 34 GW of power capacity and 414 GWh of energy capacity. Current storage in all the US is about 21 GW, nearly all of which is pumped hydro. This analysis shows the scale of storage required to accommodate high penetration levels of intermittent renewable resources.

Figure 30 Required curtailment with storage [19]

Table 6 Reduction in curtailment required for various storage capacity

Storage (hours) Drop in required curtailment (%)
4 45
8 61
12 65
24 70

3.4.4 Interconnection

According to [18], interconnection is a crucial aspect of integrating large amounts of intermittent generation. Pöyry, the authors, show the increase in load balancing area which comes from increased interconnection reduces overall variability without an increase in price in all countries with the exception of the Nordic countries. A price rise occurs there because the usually cheap hydro generation from this part of Europe is able to fetch higher prices elsewhere in other countries through interconnection.

Pöyry also found the amount of backup plant required to account for variability is not offset greatly by increased interconnection. This is because weather systems at times stretch across thousands of kilometres, thereby impacting renewable generation across the area. The effect of the increase in load balancing area is nullified and operators are still required to maintain sufficient back-up plant to manage demand during these periods of very low renewable generation. Although the amount of backup plant required might not change, the frequency of deployment of such plants is likely to increase to account for the variability of intermittent renewable generation.

3.4.5 Increased control and communication

It is suggested in [13] that as penetration levels of PV increase, to manage power quality issues arising due to its intermittent nature, communication between central control and distributed PV sources will be necessary. It has been generally recommended that an increase in control and communication will allow for more:

  • effective management at the distribution level
    • reducing losses
    • improving power quality
  • flexible system configuration
  • increased capacity for system restoration
  • more selective protection.

New communications and control infrastructure will be required for the installation of new renewables. Minimising the cost associated with this new infrastructure will make renewables more attractive economically. According to [10], transmission installation costs can also be reduced via reduced required capacity through local control of renewables, including real-time power flow monitoring and local short-term forecasting. Further benefits and savings can be made through renewables providing ancillary services such as local frequency control and Var support.

3.4.6 Planning

Planning and assessment of adequate resources to meet expected demand needs to take into account the requirement for flexibility [10]. Generation planning is shifting from planning for peak load towards planning for system energy [13]. System energy is centred on using net load as a basis for capacity planning which requires accurate renewable resource data.

Traditionally, planning involved predicting the future demand and extrapolating the peak demand from this number. Required generation and transmission were then calculated with intermittent renewable generation (IRG) effectively ignored. Conventional generation would need to reduce output due to IRG leading to lower efficiency and greater operating costs. The emerging planning approach looks at net load; which incorporates the addition of IRG into planning. Demand is

forecast and offset by predicted installations of IRG. The contribution of this IRG and its expected impact on the variability of net load is estimated based on historical renewable resource data. Generation and transmission is then planned to meet forecast net load and associated variability. These traditional and emerging planning approaches are summarised in Figure 31.

Figure 31 Traditional and emerging practice in capacity planning [10]

3.4.7 Voltage regulation

It is indicated in [20] that voltage fluctuations caused by cloud transients, amongst other sources, are of concern. Methods for mitigating this include the use of fast acting energy storage to smooth out voltage fluctuations. It is suggested the type of storage should be chosen to manage the power profile shown in Figure 32, an example where the PV power output drops to about 20% of its value under clear sky conditions because of cloud cover. The size of the required storage depends on the duration of the cloud transient. The irradiance is assumed to ramp at a rate of 200 W/m2/s in the study (making t1 = t3 = 4 seconds in Figure 32) and it can be seen from Figure 33 that to level a cloud transient of a 20 kW system whose t2 aspect (flat part of Figure 32) is eight seconds long, would require about 53 Wh of energy from a fast acting energy storage source. It is seen that the size of energy storage required to level a 16-second cloud transient instead for the 20 kW system is 90 Wh, almost double that for the 8-second transient. The transient duration an energy storage can manage reduces hyperbolically with system size. A 20 Wh stored energy can manage five second transients for a 10 kW system but less than one second (80% reduction) for only a doubling in system size to 20 kW.

Figure 32 Power profile required of an energy storage unit to level a cloud transient in a PV system [20]

Figure 33 Energy storage required to provide the power profile above, as a function of PV system rating and duration of the cloud transient [20]