摘要
针对基于粒子群优化算法的粒子滤波(PSO-PF)算法精度不高,容易陷入局部最优,难以满足电厂温控系统故障诊断的需求,提出一种适用于故障诊断的新型粒子群优化粒子滤波(NPSO-PF)算法。该算法引入社会个体对群体的认知规律优化了粒子更新的方式,并且完善了粒子速度的更新策略,对优势速度赋有较小概率的变异,提高了粒子的寻优能力,同时随机初始化劣势速度,保证了样本的多样性。实验结果表明,与PSO-PF相比,NPSO-PF提高了故障检测的精度和鲁棒性,可以有效地应用于温控系统故障的诊断。
Particle Fiher based on Particle Swarm Optimization （PSO-PF） algorithm is not precise and easily trapped in local optimum, which can hardly satisfy the requirement of fault diagnosis of temperature control system in power plant. To solve these problems, a new particle swarm optimization particle filter named NPSO-PF suitable for fault diagnosis was proposed. This algorithm introduced the cognition rule of individuals to groups to optimize the method for updating particles and improved the speed update strategy. As a result, the superior particle velocity can mutate with a small probability and improve the search ability. Meanwhile, due to the random, initialization of on inferior particle, the diversity of samples is ensured. The simulation results show that NPSO-PF improves the precision and robustness compared with PSO-PF, and it is suitable for fault diagnosis of temperature control system.
出处
《计算机应用》
CSCD
北大核心
2012年第2期432-435,439,共5页
journal of Computer Applications
基金
国防重点预研项目(40405020201)
高等学校博士学科点专项科研基金资助项目(200802881017)
南京理工大学自主科研专项计划自主项目(2010ZYTS051)
南京理工大学紫金之星基金资助项目(AB41381)
关键词
粒子群优化
粒子滤波
温控系统
变异
故障诊断
Particle Swarm Optimization （PSO）
Particle Fiher （PF）
temperature control system
mutation
faultdiagnosis
