Micro sensor fault detection and separation method based on statistical magnitude Mahalanobis distance

Abstract

The invention discloses a micro sensor fault detection and separation method based on statistical magnitude Mahalanobis distance, and belongs to the field of industrial process monitoring and fault diagnosis, and the method comprises the steps: collecting normal sensor measurement as a training data set; giving a window width, and sequentially calculating the sample statistics of the measurement variables in each window to obtain a data matrix formed by the statistics; calculating a mean value and a covariance of the data matrix, and giving a control limit of the Mahalanobis distance; acquiring real-time sensor measurement as test data; calculating statistics by using test data in a window, calculating a Mahalanobis distance by using the mean value and the covariance, and comparing the Mahalanobis distance with a control limit to judge whether a fault exists or not; and if there is a fault, solving the statistical magnitude reconstruction contribution of each sensor by using fault reconstruction, and setting the sensor with the maximum contribution value as a fault sensor to complete fault separation. Compared with the prior art, a process mathematical model is not needed, and detection and separation of faults of various types of micro sensors can be achieved.

Description

technical field [0001] The invention belongs to the field of industrial process monitoring and fault diagnosis, in particular to a small sensor fault detection and separation method based on statistical Mahalanobis distance. Background technique [0002] Distributed control systems are widely used in modern industrial production processes, and a large number of industrial sensors such as pressure, temperature and flow sensors are equipped to collect important measurement information. On the one hand, the operating environment of industrial systems is relatively complex, and sensors are usually affected by harsh factors such as vibration, high temperature, and humidity, which may lead to sensor performance degradation or even measurement failure; probability. Real-time monitoring of sensor measurements, timely detection and location of abnormalities will contribute to the safe, reliable and efficient operation of the system. In recent years, data-driven process monitoring a...

AI Technical Summary

Problems solved by technology

However, fault characterization is not obvious in the early stages, and the small anomalies caused are also easily masked by normal variations and noise in process data, making it challenging to diagnose

At present, a small amount of data-driven fault diagnosis work has considered the problem of micro-fault detection in industrial processes, but the solution to the problem of micro-fault separation is relatively rare

Moreover, the methods given in the existing work are usually effective for a fixed type of small sensor faults, and lack a certain degree of universality

Images