论文标题
平衡印度的2030电网需要时间粒度和不确定性的管理：参数模型的见解
Balancing India's 2030 Electricity Grid Needs Management of Time Granularity and Uncertainty: Insights from a Parametric Model
论文作者
论文摘要
到2030年，印度的目标是世界上一些最雄心勃勃的可再生能源（RE）增长目标，尤其是在规模归一化时，印度的目标超过了四倍的风和太阳能。同时，煤炭占主导地位，在今天提供了大约四分之三的电力。我们提出了处理高不确定性的第一个模型的结果，该模型使用参数分析而不是基于2030年经济发行的随机分析，涵盖了全国一级的30分钟分辨率粒度。该模型假设需求，供应期权，价格和其他不确定的投入。它计算了一系列参数不确定性的最低成本投资组合。我们应用简化来处理能力计划与优化发货的交集。我们的结果表明，即使将可测量的分数剩余并因此被丢弃（“缩减”），即使可测量的分数非常高（“缩减”）。我们发现，没有存储空间以及煤炭和汽油驱动能力的现有懈怠不足以实时满足需求不足，尤其是增加日平衡。由于季节性变异性（不仅仅是固有的高资本成本），因此存储技术证明是有价值的，但与2019年的投资组合相比，与2019年的投资组合混合相比仍然昂贵。但是，检查电池的未来增长替代方案发现所有有关峰值功率的解决方案都更加昂贵。对于在高峰时段平衡，应用需求响应的更聪明的网格可能具有成本效益。我们还发现需要在年度时间表（尤其是风产量，降雨量和需求年度变化）以及供应和负载概况（形状）的不确定性的情况下进行更复杂的建模。
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).
