Multi-objective Unit Commitment Optimization with Ultra-Low Emissions under Stochastic and Fuzzy Uncertainties
Li You,Li Huaxiong,Wang Bo,Zhou Min,Jin Mei
DOI: https://doi.org/10.1007/s13042-020-01103-9
2020-01-01
International Journal of Machine Learning and Cybernetics
Abstract:Low cost, high reliability and low pollution are prime targets when performing current unit commitment optimization. As an extension of previous works, this study establishes a multi-objective unit commitment model which takes into account all of the above targets. The main content includes: First, the pricing support for thermal units with ultra-low emissions is involved when analyzing the operation cost of generation systems, which accords with the current policy of power markets. Second, a conditional Value-at-Risk-based measurement is formed to estimate system reliability considering the stochastic and fuzzy uncertainties existed in future load, renewable generation and equipment failures, which is sensitive to tail risks and provides easy-to-adjust conservativeness against worst-case scenarios. Third, to deal with the proposed model, a practical approach is applied to develop a multi-objective particle swarm optimization algorithm, which improves the Pareto fronts obtained by existing methods. The effectiveness of this research is exemplified by two case studies, which demonstrate that the model finds appropriate pricing support for the reformed units, and the proposed reliability measurement is able to realize a number of trade-offs between cost effective and solution robustness, thus providing decision support for system operators. Finally, the comparisons on performance metrics such as spacing and hyper-volume also justify the superiority of the algorithm.