India’s electricity sector is undergoing a rapid transformation. The country’s renewable energy (RE) capacity has doubled in the last five years, to 89.2 GW, as of September 2020 (CEEW-CEF, 2020). The share of RE in the electricity mix also increased from 9 per cent to 11.8 per cent between the first quarters of 2018–19 and 2020–21 (POSOCO, 2020). The share of RE in the Indian grid is expected to reach 55 per cent by 2030, with a total installed capacity of 450 GW. As the share of RE increases, Its impact on the grid will grow. Key challenges include (a) higher grid integration costs due to the variable and unpredictable nature of solar–wind resources; (b) the need for greater system flexibility and improved voltage control; (c) difficulty forecasting and planning for RE procurement; (d) limited demand-side management by discoms; (e) the focus on the levelised costs of energy from RE instead of the overall cost of supply to consumers; and (f) increased cost of power generation for thermal projects due to lower utilisation rates. In this CEF Analysis, we aim to examine how India’s electricity sector is evolving to tackle these challenges using innovative RE procurement models. We also dissect one of the models to identify critical areas for these models.
Increasing the share of RE in the Indian grid requires designing new approaches to ensure grid stability. These possible solutions include reducing generation intermittency by deploying energy storage and increasing the flexibility of thermal projects. Besides, wholesale market redesign, advanced forecasting, smart metering, and demand-side management may also play an important role in ensuring a smooth transition. However, due to the nascency of smart-metering infrastructure in India and the low adoption of time-of-use tariffs, the impact of demand-side management has been limited, unlike in most developed countries. Therefore, as a response, India’s chief renewable project tendering agency, Solar Energy Corporation of India (SECI), has started to roll out innovative procurement models (or tender designs) to procure firm or less intermittent RE on behalf of discoms. These include the following (see Table 1 for details):
In these tenders, generation from solar and wind projects are blended at energy evacuation point. A hybrid project can ensure relatively smooth diurnal generation as compared to a solo solar or wind project. The capacity utilisation factor (CUF) of hybrid projects is typically high (~40 to 50 per cent). Such a model reduces discoms’ flexibility requirements and improves transmission infrastructure utilisation.1
Such tenders aim to meet a discom’s peak demand by deploying energy storage with solar and/or wind projects. Such models ensure firm supply during peak hours (typically for six hours a day) and increase dispatchability by reducing intermittency.
These models aim to further increase the dispatchability of solar and wind projects by oversizing and combining them with energy storage, thermal, gas, and other technologies. The recent tender (Pan-India (SECI) solar–wind–storage, RTC-I, 400 MW) requires a minimum CUF of 80 per cent, but based on the discom’s requirements, the SECI can increase the CUF. Such tenders reduce flexibility requirements for discoms.
Every model is designed to meet different objectives, directed by the specific needs of the offtaker (discom). An objective could include reducing evacuation infrastructure costs by improving its utilisation or meeting the RPO and peak demand simultaneously. After gathering the requirements of the discom, the SECI typically designs a model (or tender) with conditions for capacity utilisation, technology, injection points, sale of surplus power, and energy dispatch. Developers—selected on a least cost per unit of energy basis—are expected to optimise key project components such as generation profile, geographical location, technology, sale of surplus energy, financing, etc. based on these conditions.
To understand some of these components, we dissected a recently concluded bid for 400 MW round-the-clock energy supply (see Table 1 for bid conditions). To figure out the best suited technology combination for this bid, we looked at the yearly and daily generation profiles of solar and wind power projects in India. For our analysis, we considered a wind project in Kutch (Gujarat) and a solar project in Bhadla (Rajasthan), as these locations feature high wind speeds and solar insolation, respectively.
The projects clocked an annual utilisation of 37 and 27 per cent, respectively. From Figure 1, it is clear that wind generation varies significantly across seasons, with utilisation as high as 49–50 per cent during May–June and as low as 17–18 per cent during September–October. Therefore, significant oversizing of the project capacity (around 3–4x) may be needed to meet the monthly minimum utilisation requirement (70 per cent), particularly if the project has only wind capacity. Alternatively, energy storage can be deployed to smoothen generation across seasons; this would reduce the need to oversize. Electricity can be stored during high wind and solar months (May–June) and dispatched during lean months (September–October).
Another way to address seasonal variations in wind generation is through the hybridisation of the project with solar capacity, as solar’s generation profile is relatively stable across the year (Figure 1). However, power generation from solar varies diurnally, with operations typically restricted to 6 am–6 pm. Energy storage can be deployed to meet daily energy shifting needs, especially from daytime solar generation hours to evening hours. Therefore, an optimal mix of solar, wind, and energy storage capacities may be needed to meet project conditions economically.
The winning bidder of this project is not required to disclose its project capacity configuration. Therefore, to determine the optimum capacity mix, we simulated daily generation profiles for a wind project located in Kutch, Gujarat, and a solar project located in Bhadla, Rajasthan. Our main objective was to identify the optimum mix of wind, solar, and energy storage capacities while minimising project capital costs. Key assumptions for the analysis were as follows:
As per our analysis, the optimal capacity would be a mix of 1,200 MW of wind and 300 MW of solar capacity. Figure 2 captures the average hourly generation profile and monthly utilisation based on our simulated analysis. Therefore, the project is expected to generate around 4,506 GWh annually, 28 per cent of which shall be surplus that may be sold in the open market. This surplus energy shall be generated primarily from April to July. The generation profile leaves little room for daily energy shifting, as on most days (except during September and October), the project shall generate equal to or above the hourly limit of 400 MWh. Besides, monthly energy shifting (say from July to September) would require more than 5,000 MWh of energy storage capacity, significantly raising project costs. However, some amount of energy storage may be deployed to store excess generation during low price periods over power exchanges; it can be sold during high price periods in case the developer decides to trade it in power exchanges to increase revenue.
The success of these innovative models depends on various factors and considerations that require immediate attention from policymakers and industry. The relevant factors are as follows:
1. Reducing intermittency and firming RE requires energy storage. Offtake of energy by discoms through these models depends a lot on the tariffs discovered, which in turn are majorly determined by storage costs and lifecycles. Storage bids in the US indicate that tenders for higher storage capacity will help achieve economies of scale, thereby reducing costs.
2. We can no longer consider an RE project’s levelised cost of energy (LCOE) as the only indicators of its viability. Instead, policymakers and discoms should look at overcall costs to consumers by quantifying the grid integration costs of procuring RE. Typically these include transmission infrastructure costs, flexibility costs, deviation settlement mechanism costs, etc.
3. The success of these innovative models also depends on the liquidity of wholesale markets (e.g., day-ahead, green term–ahead, and real-time markets), as they involve developers selling surplus RE generation and discoms buying it to offset shortfalls. A key reason for the current low volumes of power exchange is the lack of awareness and related-capacity building among discoms.
4. As the dispatchability of RE increases through the use of these models, the viability of thermal projects may be affected (resulting particularly in higher variable costs). This may force these projects to go into early retirement. Therefore, policymakers and electricity sector planners may need to design an appropriate strategy to facilitate a smooth transition while phasing out thermal projects.