A sub-health online identification and diagnosis method based on performance monitoring data

Abstract

The invention discloses an online sub-health identification and diagnosis method based on performance monitoring data, belonging to the technical field of fault diagnosis. First, establish the initial model of probabilistic neural network state classification and calculate the threshold standard deviation, use the current model to carry out online monitoring and diagnosis classification of monitoring equipment, and further identify and extract sub-health state data, and put them into the sub-health state data group; if When the sub-health data group to be identified reaches the storage limit or has a known state, the storage work is suspended, and all elements in the group are subjected to K-means cluster analysis to obtain the classification result and clear the storage space of the sub-health data group. Then merge the sub-health state data set after cluster analysis with the previous training samples, and update it to the initial model to obtain a new classification model; finally, repeat the above steps to identify the sub-health state and repair it in time when a fault occurs. The invention takes timely and effective measures according to the state of the equipment to reduce losses caused by failures.

Description

technical field [0001] The invention belongs to the technical field of fault diagnosis and relates to an online sub-health identification and diagnosis method based on performance monitoring data. Background technique [0002] Traditional fault diagnosis methods include two states, normal and fault, and use this as the basis for division to construct a diagnosis model. However, during the operation of the equipment, it does not always work efficiently and with high accuracy, and there is still a sub-health state; therefore, the two-state model based on normal and fault will misdiagnose the sub-health state as a normal state or a fault state, and cannot reflect The real state of the device. [0003] At present, in engineering applications, the data of equipment failure status is generally obtained through FMECA report, so as to identify the equipment failure status; however, it is difficult to obtain the sub-health status of the equipment through the hardware structure and w...

AI Technical Summary

Problems solved by technology

However, during the operation of the equipment, it does not always work efficiently and with high accuracy, and there is still a sub-health state; therefore, the two-state model based on normal and fault will misdiagnose the sub-health state as a normal state or a fault state, and cannot reflect real state of the device

[0003] At present, in engineering applications, the data of equipment failure status is generally obtained through FMECA report, so as to identify the equipment failure status; however, it is difficult to obtain the sub-health status of the equipment through the hardware structure and working mechanism of the equipment. Over time, the status data of the equipment is gradually obtained by monitoring the operating status of the equipment and the data changes of the monitoring points, and then the sub-health status is identified based on the status data

[0004] However, this method of identifying and diagnosing the sub-health state of the equipment based on the state data of the equipment operation is currently mainly used offline, that is, the sub-health state is diagnosed by analyzing the historical state data of the equipment, so the working status of the equipment cannot be diagnosed in real time.

Images