基于[火用]补偿策略的某型增压锅炉故障定位研究

Research on Fault Localization of a Supercharged Boiler based on Exergy Compensation Strategy

为了解决某船用增压锅炉缓变类故障难以诊断的问题,开展基于热经济学的诊断方法研究。首先,在GSE仿真平台上构建增压锅炉的仿真模型,参考常见故障机理及判断依据,对8类单一热力故障进行模拟,采集故障工况数据建立多程度缓变故障样本库;其次,根据热经济学结构理论,将增压锅炉按照“产品”的定义划分为4个过程组元,建立锅炉的生产结构图,并计算故障过程中每个组元额外的不可逆损失,作为故障定位及量化的依据;最后,针对故障引发的扩散效应,提出基于神经网络的[火用]补偿策略。结果表明:经过叠加故障的对比,该策略能够有效增强结构理论的故障定位能力,允许设备的独立诊断,检测叠加故障与外部冲击情况下锅炉内部设备的异常状况,在12组典型故障组合的检测中精度达到了91.7%,能够有效找到真实故障组元。

In order to solve the problem of difficult diagnosis of gradual faults in a certain marine super-charged boiler,research on diagnostic methods based on thermoeconomics was carried out.Firstly,a simulation model of the supercharged boiler was constructed on the GSE simulation platform,and 8 types of single thermal faults were simulated based on common fault mechanisms and judgment criteria.Fault condition data were collected to establish a sample library for multi-level gradual faults.Secondly,ac-cording to the theory of thermoeconomic structure,the supercharged boiler was divided into four process components based on the definition of"product",and a production structure diagram of the boiler was es-tablished.The additional irreversible losses of each component during the fault process were calculated as the basis for fault localization and quantification.Finally,in response to the diffusion effect caused by faults,a neural network-based exergy compensation strategy was proposed.The results show that through the comparison of superimposed faults,this strategy can effectively enhance the fault localization ability of structural theory,allow for independent diagnosis of equipment,and detect abnormal conditions of internal equipment of boiler under superimposed faults and external impact conditions.In 12 typical fault groups,the detection accuracy reaches 91.7%,which can effectively find the real fault components.