Balancing India's 2030 Electricity Grid Needs Management of Time Granularity and Uncertainty: Insights from a Parametric Model

Rahul Tongia

DOI: https://doi.org/10.48550/arXiv.2207.02151

2022-07-01

General Economics

Abstract:With some of the world's most ambitious renewable energy (RE) growth targets, especially when normalized for scale, India aims more than quadrupling wind and solar by 2030. Simultaneously, coal dominates the electricity grid, providing roughly three-quarters of electricity today. We present results from the first of a kind model to handle high uncertainty, which uses parametric analysis instead of stochastic analysis for grid balancing based on economic despatch through 2030, covering 30-minute resolution granularity at a national level. The model assumes a range of growing demand, supply options, prices, and other uncertain inputs. It calculates the lowest cost portfolio across a spectrum of parametric uncertainty. We apply simplifications to handle the intersection of capacity planning with optimized despatch. Our results indicate that very high RE scenarios are cost-effective, even if a measurable fraction would be surplus and thus discarded ("curtailed"). We find that high RE without storage as well as existing slack in coal- and gas-powered capacity are insufficient to meet rising demand on a real-time basis, especially adding time-of-day balancing. Storage technologies prove valuable but remain expensive compared to the 2019 portfolio mix, due to issues of duty cycling like seasonal variability, not merely inherent high capital costs. However, examining alternatives to batteries for future growth finds all solutions for peaking power are even more expensive. For balancing at peak times, a smarter grid that applies demand response may be cost-effective. We also find the need for more sophisticated modelling with higher stochasticity across annual timeframes (especially year on year changes in wind output, rainfall, and demand) along with uncertainty on supply and load profiles (shapes).