加速遗传算法在电厂动态负荷优化分配中的应用

Application of AGA in dynamic load dispatch

针对目前电厂负荷分配非实时最优、能源浪费巨大的问题,提出了一种动态分配负荷的算法。在考虑电厂机组实际运行特点后,建立了包含辅机、启停等约束条件下的连续多时段动态负荷优化分配的数学模型;分析了标准遗传算法(SGA)在解决负荷分配问题时收敛速度慢的原因,据此对SGA进行改进,提出了基于实数编码、双倍初始种群、自适应遗传参数和压缩搜索空间法的加速遗传算法(AGA)。仿真结果表明,AGA能稳定地收敛到全局最优解,并且有效缩短了运算时间,基本达到动态分配的要求。

A dynamic load dispatch algorithm is proposed for the optimal load distribution of power plant in real time and the mathematical model of continuous multi-period dynamic load dispatch for power plant is built with the constraints of auxiliary machinery and unit start-stop. As the convergence speed of standard genetic algorithm for load dispatch is slow,AGA（Accelerating Genetic Algorithm） is proposed based on real -code,diploid initial population,adaptive genetic parameters and narrow search range. Simulative results show that AGA can effectively reduce the computation time and stably converge to the global optimal solution, meeting basically the requirements of dynamic load dispatch.