摘要
将改进的粒子群优化(PSO)算法与误差反向传播(BP)算法相结合构成混合算法训练人工神经网络。改进的PSO算法中,惯性权重从最大到最小线性减小,以平衡局部和全局搜索能力,并将类似“选择”的概念引入PSO算法,使该算法更好地协调全局和局部搜索能力,有利于更快寻找到全局最优点。该算法有效地解决了常规BP算法学习网络权值和阈值收敛速度慢、易陷入局部极小和GA算法独立训练神经网络速度缓慢等缺点。将该算法应用于变压器故障诊断,仿真结果表明了该算法具有较快的收敛速度和较高的计算精度,满足电力变压器故障诊断的要求。
The hybrid algorithm combining improved PSO（Particle Swarm Optimization） algorithm with BP（error Back Propagation） algorithm is used to train the artificial neural network. To balance and reconcile the global and local searching capability,the inertia weight of improved PSO is reduced linearly from maximum to minimum and the cencept of ＂selection＂ is introduced into the PSO to find the global optimal point more quickly. Defects of conventional BP algorithm,i.e, the slow convergence of weight and threshold learning,premature result,and the slow training speed of GA,are settled by it. Its application in power transformer fault diagnosis is simulated. Results show that it meets the requirements of power transformer fault diagnosis for both convergence speed and calculation accuracy.
出处
《电力自动化设备》
EI
CSCD
北大核心
2006年第5期35-38,共4页
Electric Power Automation Equipment
关键词
改进PSO算法
人工神经网络
故障诊断
电力变压器
improved particle swarm optimization
artificial neural network
fault diagnosis
power transformer
