舰用燃机系统的故障诊断与容错控制研究

The fault diagnosis and fault tolerant control of naval gas turbine with poor data and small sample information is the key for optimizing the allocation of ship combat and training missions, as well as the arrangement of ship maintenance and support activities.To deal with the poor data and insufficient training sample problems of fault diagnosis and fault tolerant control in naval gas turbine, the fault symptom and its evolution trend model, fault and its evolution trend model is establishmented, the diagnosis and prognosis model suitable for poor data and insufficient training sample is constructed, the fault tolerant controller design method and operating status assessment method are studied. More focus is placed on three technical issues: how to build the coupling relationship between fault symptom characteristic variables and fault characteristic variables; how to implement a fault diagnosis model for device with poor data and small samples; how to execute the hierarchical decomposition and weight distribution in the running state evaluation.With the innovative achievements,a naval gas turbine fault diagnosis and fault tolerant control system is established, which contributes to the necessary theoretical and engineering knowledge base of the practical application of gas turbine fault diagnosis and fault tolerant control.

在贫信息、小样本条件下完成舰用燃机故障诊断与容错控制是舰船开展作战、训练任务、维修保障活动得以优化配置与合理安排的关键所在。针对舰用燃机故障诊断与容错控制中监测数据少、训练样本不足的问题，通过建立故障征兆及其演化趋势模型、故障及其演化趋势模型，建立适于贫信息、小样本条件下的故障诊断与趋势预测模型，研究复杂冗余系统的容错控制器设计与运行状态评估方法。重点突破故障征兆特征变量与故障特征变量的耦合关系建立、贫信息、小样本情况下的故障诊断模型、运行状态评估中的层次分解与权重分配等3个技术难题，以期获得一些创新性成果，为舰用燃机故障诊断与容错控制的实际应用建立必要的理论与工程知识储备库。

在贫信息、小样本条件下完成舰用燃机故障诊断与容错控制是舰船开展使用、维修保障活动得以优化配置与合理安排的关键所在。针对舰用燃机故障诊断与容错控制中监测数据少、训练样本不足，以及复杂冗余系统的运行状态评估中的层次分解与权重分配等问题，开展舰用燃机典型故障模式分析，建立舰用燃机典型故障征兆模型、故障机理及演化机理模型，将GAP分析与知识相结合，研究适于贫信息、小样本条件下的故障诊断及演化趋势分析方法；将层次分析与变权理论相结合，研究面向冗余系统的运行状态评估方法，并提出舰用燃机各运行状态情况的保护性健康管理策略、预测容错控制策略、主动容错控制策略、维修健康管理策略，能够为舰用燃机使用、维修保障阶段的控制策略与保障措施设计提供支撑。