论文部分内容阅读
在对ART-2和并行BP人工神经网络模型进行分析的基础上,提出了一种针对大型复杂设备进行故障诊断的复合神经网络诊断策略:先运用ART-2识別单一类故障,再使用并行BP网络实现并发故障和新故障的分离,解决了对复杂设备并发故障诊断和新故障都能有效识別的问题。该模型利用柴油机缸盖振动数据进行诊断验证,能够有效识別并发故障和新故障等复杂故障类型。
Based on the analysis of ART-2 and parallel BP artificial neural network model, a composite neural network diagnosis strategy for fault diagnosis of large and complex equipment is proposed. Firstly, ART-2 is used to identify a single type of fault and then parallel BP network to achieve the separation of concurrent failures and new failures, to solve the complicated equipment concurrent failure diagnosis and new problems can effectively identify the problem. The model uses diesel cylinder head vibration data for diagnostic verification, which can effectively identify complicated fault types such as concurrent faults and new faults.