燃气-蒸汽联合循环机组的发展对我国电力系统实现“双碳”目标起着至关重要的作用。由于燃气-蒸汽联合循环机组负荷对象具有强耦合、非线性等特点，传统辨识方法与粒子群（PSO）算法在燃气-蒸汽联合循环机组负荷对象模型辨识方面存在寻优精度低和收敛速度慢的问题。本文提出了一种基于维度学习策略的双群体学习改进粒子群优化算法（TSL-IPSO）来优化PSO算法的全局搜索能力和局部改良能力。通过开环阶跃实验得到的燃气-蒸汽联合循环机组257.4MW、436.07MW负荷点处数据对TSL-IPSO算法与PSO等算法辨识得到的负荷对象模型进行验证对比。结果表明，TSL-IPSO算法与PSO等算法相比，其辨识模型的均方根误差、平均绝对百分误差均最小，适应度变化曲线收敛效果最好，具有更好的模型辨识精度与寻优性能。

The development of gas-steam combined cycle units plays an important role in realizing the goal of "double carbon" in China's power system. Because the load object of gas-steam combined cycle unit has the characteristics of strong coupling and nonlinear, the traditional identification method and particle swarm optimization (PSO) algorithm have the problems of low optimization accuracy and slow convergence speed in the identification of gas-steam combined cycle unit load object model. In this paper, a two-population learning improved particle swarm optimization algorithm (TSL-IPSO) based on dimensional learning strategy is proposed to optimize the global search ability and local improvement ability of PSO algorithm. The load object models identified by TSL-IPSO algorithm and PSO algorithm were verified and compared with the data of 257.4MW and 436.07MW load points of gas-steam combined cycle units obtained by open-loop step experiment. The results show that compared with PSO algorithm, TSL-IPSO algorithm has the smallest root-mean-square error and the average absolute percentage error of the identification model, the best convergence effect of the fitness curve, and has better identification accuracy and optimization performance.

上海发电过程智能管控工程技术研究中心资助项目