Strengthening the Resource Adequacy Framework for an RE-Rich Future

Executive Summary

With greater additions of renewable energy (RE) to the generation mix, ensuring reliability of the power system will be a challenge. Resource Adequacy (RA) is an assessment to determine whether a power system has sufficient resources to meet the demand for electricity at all times. While resource adequacy has always been of concern, the addition of non-dispatchable and non-controllable resources, such as RE, to the system is making it more challenging to ensure resource adequacy. Recognising this challenge, the Ministry of Power (MoP) recently issued guidelines that give a recommended framework for ensuring resource adequacy. The MoP framework relies on three reliability metrics to ensure resource adequacy: (1) loss of load probability (LOLP); (2) planning reserve margin (PRM); and (3) normalised energy not served (NENS). This paper argues that, while the framework is a good starting point, it requires significant modifications to effectively address the challenges posed by a rapidly transforming power sector increasingly reliant on RE. 

The MoP framework and similar ones used in other parts of the world were developed for power systems driven by fossil fuels. This paper suggests several modifications that will make it more suitable for the RE-rich system of the future. While reliability is paramount for power sector planners, minimising its cost is also crucial. This paper proposes modifications to the MoP framework with this dual goal in mind, and most of the recommended modifications to the framework fall into one of two categories: (1) those that enhance reliability; and (2) those that increase the cost-effectiveness of the resource adequacy measures. We also suggest measures to improve implementation by making a few changes to the roles and responsibilities of the institutions involved in the current framework. 

For enhancing reliability, we recommend that, because LOLP is not a very intuitive metric, planners use an alternative metric—the number of loss of load hours (LOLH) in a year. In addition, we point out that the depth (in MW) and duration of individual generation shortfalls, and the frequency of shortfalls, are important because they affect the level of distress consumers experience from outages. Therefore, we recommend that, in addition to LOLH, the following metrics, along with their probability distribution over the year, be assessed in any RA planning exercise: (1) duration of individual shortfalls; (2) depth of shortfalls in MW; and (3) frequency of shortfalls given by the metric, loss of load events (LOLEv).

For enhancing reliability, we recommend caution when using PRM and capacity credits because these are much more appropriate for fossil-fuel-driven power systems and not for RE-rich systems of the future. Another major drawback of using PRM and capacity credits is that it assumes that failures or generation shortfalls at individual plants are independent of each other. However, that is not always the case. Common-mode or correlated failures or shortfalls can occur, particularly during extreme weather events. Extreme weather events can lead to shortfalls across entire regions. One example is the extreme winter storm in Texas, USA in February 2021. The forecast for the winter had predicted there would be reserves of 28% of the expected peak load after accounting for planned and estimated unplanned outages. But the peak load exceeded the forecast, and 32% of the generation capacity failed to operate, leading to widespread blackouts and extreme distress for people.

RA planning must also account for extreme weather events. As such events become more frequent, we recommend that the proposed system be stress-tested through modelling a few potential high-impact, low-probability events. Changing weather patterns and the increasing frequency of extreme weather events will also affect future patterns of RE generation and electricity demand, and these are likely to be very different from the past. Therefore, we suggest that planning models not rely on historical data alone, particularly for RE generation patterns and electricity demand. We suggest that, instead, electricity planners collaborate with climatologists to develop better forecasts of weather patterns.

In order to enhance cost-effectiveness of measures to ensure resource adequacy, we recommend that there be a re-evaluation of the economic justification for the selected RA criteria, such as an LOLP of 0.2% and NENS of 0.05% over the year. This is because the relationship between reliability and cost is highly non-linear, and a small relaxation in the reliability metrics can lead to a significantly larger reduction in costs. In addition, because consumers are indifferent to the cause of any outage, we suggest that it will be good to compare, on an energy basis (say, GWh), the outages caused by bulk system outages versus those caused by distribution network faults. If the outages, in GWh terms, due to distribution network outages are much larger than bulk power system outages, then it would be an indication that we are overspending on grid resource adequacy and underspending on upgrading the distribution network.

For enhancing cost-effectiveness, we also recommend that instead of developing just a single plan and subjecting it to various uncertainties, planners should evaluate a few alternate plans so that the preferred plan is the one that best balances value and risk. Furthermore, because good resource planning can reduce system costs considerably, we recommend that there be sufficient time and training for effective long-term resource planning. More specifically, we recommend that long-term resource plans be required only once every two years, as usually done in the US, instead of requiring it to be completed in two months as mandated in the MoP framework.

The MoP framework puts the onus of resource adequacy planning on the discoms alone. We think that, given the extent of consumer migration to other suppliers, it may be fairer and also cost-effective that all load serving entities (LSEs) be required to do resource adequacy planning. Furthermore, to capture synergies between all LSEs, including discoms, in the approval process for resource plans, SERCs should review them in a holistic manner for the entire State to ensure that electricity is delivered in the most optimal manner for the entire State.

We recognise that some of these changes to the RA framework may increase the complexity of the RA process, and some could also be challenging for discoms to carry out. Discoms in India are just beginning to consider resource planning. Therefore, we have recommended a gradual transition to our recommended framework. However, it is important that these changes are not ignored because doing so could lead to decreased reliability and increased costs that consumers will have to pay for electricity service. Power procurement costs constitute 70–80% of the costs of electricity that consumers pay, and effective resource planning can help significantly reduce those costs. Neglecting these recommended changes could lead to the entrenchment of outdated practices, making future framework revisions more difficult.