基于ENN-AOSMA算法的集中式温差发电系统最大功率跟踪控制  

MAXIMUM POWER POINT TRACKING CONTROL OF CENTRALIZED THERMOELECTRIC POWER GENERATION SYSTEMS BASED ON ENN-AOSMA ALGORITHM

作　　者：冯璐璐 陈燕[1] 何鑫颖 张艺博 程源 Feng Lulu;Chen Yan;He Xinying;Zhang Yibo;Cheng Yuan(School of Electrical and Power Engineering,Taiyuan University of Technology,Key Laboratory of Clean Intelligent Control of Coal Power,Ministry of Education,Taiyuan 030024,China)

机构地区：[1]太原理工大学电气与动力工程学院煤电清洁智能控制教育部重点实验室,太原030024

出　　处：《太阳能学报》2025年第4期125-132,共8页Acta Energiae Solaris Sinica

基　　金：山西省基础研究计划(202203021211175)。

摘　　要：针对非均匀温度下集中式温差发电阵列呈现出的非线性多峰值P-V曲线,提出一种改进的最大功率跟踪(MPPT)算法。该算法基于改进黏菌算法结合Elman神经网络,以控制器IGBT占空比为输入、系统功率为输出,建立前馈神经网络,得到输入-输出拟合曲线。基于拟合曲线,更新边界条件,并通过算术优化算法(AOA)更新黏菌算法(SMA)局部搜索公式,实现快速逼近全局最大功率点(GMPP)。通过启动测试、阶跃温度测试、时变温度测试、准确度分析4个动态算例验证所提算法的有效性和优越性。仿真结果表明,所提MPPT方法不仅追踪速度快,且追踪精度达到99.94%。A novel maximum power point tracking(MPPT)algorithm based on an improved slime mould algorithm(SMA)combined with Elman neural network is proposed to address the nonlinear multi-peak P-V curves exhibited by centralized thermoelectric generation systems under nonuniform temperature distribution(NTD)condition.The feedforward neural network is established with duty cycle as input and system power as output,and the input-output fitting curve is obtained.Based on the fitting curve,the boundary conditions are updated,and the local search formula of the slime mold algorithm(SMA)is updated using arithmetic optimization algorithm(AOA),achieving a fast approximation to the global maximum power point(GMPP).The effectiveness and superiority of the proposed algorithm are verified through examples including start-up tests,temperature step changes,time-varying temperature tests,and accuracy analysis.Simulation results demonstrate that the proposed MPPT method not only achieves fast tracking speed,but also attains a tracking accuracy of 99.94%.

关 键 词：最大功率跟踪 温差发电 神经网络 非均匀温度分布 

分 类 号：TM913[电气工程—电力电子与电力传动]