论文部分内容阅读
提出了一种基于粒子群和最近邻的热力系统变工况动态过程故障诊断方法,该方法利用改进的粒子群优化(IPSO)算法获取典型故障原型,运用进化Elman神经网络对特征参数应达值进行实时预测,计算故障征兆,依据最近邻分类规则计算实时故障征兆与各典型故障原型的相似性,对故障进行识别.以某600MW超临界机组高压加热器给水系统为例,借助电站全范围仿真系统进行了详细的故障诊断仿真实验.结果表明:该方法应用于热力系统动态过程故障诊断可取得较满意的效果.
A dynamic fault diagnosis method based on particle swarm optimization and nearest neighbor is proposed in this paper. The method uses an improved Particle Swarm Optimization (IPSO) algorithm to obtain typical fault prototypes. The evolutionary Elman neural network Real-time prediction is used to calculate the fault symptom, and the similarity between the real-time fault symptom and each typical fault prototype is calculated according to the nearest neighbor classification rule to identify the fault.Taking a high-pressure heater water supply system of a 600MW supercritical unit as an example, The system has carried on the detailed fault diagnosis simulation experiment.The results show that this method can be applied to the dynamic system fault diagnosis of the thermodynamic system to obtain more satisfactory results.