改进二进制-实数编码混合蛙跳算法在水电机组短期发电调度中的应用

Application of Short-Term Hydropower Generation Scheduling in a Hydropower Station Based on Improved Binary—Real-Coded Shuffled Frog Leaping Algorithm

本文将改进实数编码混合蛙跳算法(IR-SFLA)和二进制编码的(IB-SFLA)方法分别应用到水电站经济负荷分配(ELD)和机组组合(UC)问题,提出解决STHGS问题的IBR-SFLA方法.实数编码版本IR-SFLA利用混沌学遍历性、随机性特征生成初始种群,采用更新的局部搜索和位置更新策略实现青蛙更新换代,并在迭代后期通过自适应青蛙激活机制重新激发青蛙搜索能力;在二进制编码IB-SFLA中引入改良青蛙子种群分组方式,将青蛙种群分为领导蛙、追随蛙和变异蛙3类蛙群,各类蛙群分别基于正态云模型的精英进化策略、改进的局部搜索机制和混沌理论的蛙群变异操作进行更新迭代.运行结果显示IBR-SFLA相较对比算法,在低、中、高水头下最高缩减耗水量1.14×10^7、1.22×10^7、7.52×10^6m^3,有效提升水能资源利用效率;在保证运算精度、稳定性的同时,平均运行时间最高缩减178、173和172 s,进一步,改进策略性能分析显示,各改进策略可有效增强搜索性能,提升精度,且耗时增幅较小,在较小种群规模下便可获取较高质量的解,为解决大规模机组短期电力调度优化课题提供有效了新思路.

Short-term hydropower generation scheduling(STHGS)in a hydropower station is a complex multidimensional,non-linear optimization problem in non-continuous space with multiple constraints. The STHGS problem can be decomposed into unit commitment(UC)and economic load distribution(ELD)sub-problems. The traditional shuffled frog leaping algorithm(SFLA)has low efficiency on solution precision and reliability. This research focuses on incorporating the improved binary and real-coded SFLA into solving STHGS. The improved real-coded SFLA(IR-SFLA)and binary-coded SFLA(IB-SFLA)are applied to ELD and UC sub-problems,respectively. The IR-SFLA initializes the population with ergodicity and randomness in chaos theory and adopts renewed local search and location update strategies to realize frog evolution. In addition,the adaptive frog activation mechanism is introduced into reactive frog vitality during later iteration stage. In terms of IB-SFLA,the new frog grouping strategy tends to divide the frog population into leader,follower,and mutation frog sub-populations. Furthermore,each subpopulation adopts the elite evolution strategy based on the normal cloud model as well as the improved local search pattern and frog mutation operation on basis of chaos theory to complete the iteration process. Overall,the final IBRSFLA is established and applied to solve the STHGS problem. Comparison of the simulation results indicates that the IBR-SFLA can dramatically reduce the water consumption during power generation,with values of 1.14× 10^7,1.22× 10^7,and 7.52×10^6 m^3,corresponding to 75,88,and 107 m water heads,and improve the utilization efficiency on hydropower resources. Moreover,with computational reliability and precision guaranteed,the maximum mean computation time reduction values are 178,173,and 172 s. Furthermore,performance analysis indicates that the strategies can effectively enhance the search performance and improve the accuracy by increasing the computation time with a small amount. The high-quality solutions can be obtained at a small population size. Thus,the IBR-SFLA is verified to provide new basis to solve the STHGS problem in large hydropower stations.