基于优化BP神经网络的冷水机组故障监测研究  

Research on Fault Monitoring of Chillers Unit Based on Optimized BP Neural Network

作　　者：郑力达 张玉成 

机构地区：[1]上海宝信软件股份有限公司

出　　处：《信息技术与标准化》2024年第7期32-35,52,共5页Information Technology & Standardization

摘　　要：为提升冷水机组故障监测的准确性和实时性,结合粒子群优化算法并引入动量因子,提出一种基于优化BP神经网络的实时在线监测模型。主要介绍了BP神经网络的优化方式、冷水机组故障分析及参数选取、在线状态管理,通过Java编程语言和Encog神经网络库构建BP神经网络模型,并部署于某城市智慧运维平台,该模型有效提升了平台对冷水机组故障的监测能力。In order to improve the accuracy and real-time performance of chiller failure monitoring,a real-time online monitoring model based on optimized BP neural network is proposed by combining the particle swarm optimization algorithm and the introduction of momentum factor.The optimization method of BP neural network,chiller fault analysis and parameter selection,and online state management are mainly introduced,and the BP neural network model is constructed by Java programming language and Encog neural network library and deployed in a city's intelligent operation and maintenance platform,which effectively improves the platform's ability to monitor chiller faults.

关 键 词：BP神经网络 冷水机组 粒子群算法 动量因子 在线监测 

分 类 号：TU831.4[建筑科学—供热、供燃气、通风及空调工程] TP183[自动化与计算机技术—控制理论与控制工程]