基于正反向混合推理的故障仿真预测模型

根据故障预测的特点和要求,分析了传统预测方法用于故障预测的不足,提出了动态模糊综合评判方法,建立了基于动态模糊综合评判和多Agent并行诊断的故障预测模型。该模型将正向推理与反向推理、不同的知识表达方式和推理机制融合于同一推理框架中,再现了故障的发生和发展过程。由基于动态模糊综合评判的正向推理得到候选故障集,基于多Agent并行诊断的反向推理确认实际发生的故障,找出故障原因,实现故障定位。

Fault Forecast and Simulation Model of Hybrid Inference Integrated Forward and Backward Inference

Considering the characteristics and requirements of fault forecast,the deficiencies of traditional forecast methods were pointed out.A dynamic fuzzy comprehensive evaluation method was put forward,and a fault forecast model based on dynamic fuzzy comprehensive evaluation and multi-Agent synchronous inference was proposed.Forward inference and backward inference,different knowledge expression methods and inference mechanisms were integrated in the same framework.A candidate fault set including the probability of each fault symptom was provided by forward inference based on dynamic fuzzy comprehensive evaluation.Then,the factual fault symptoms,the corresponding fault causations and components could be confirmed by backward inference process based on multi-Agent synchronous inference.