Fault-tolerant method of redundant sensor measurement system

Abstract

The present invention provides a fault-tolerant method of a redundant sensor measurement system. The method comprises the steps of obtaining a measurement equation of the redundant sensor measurement system; determining a parity vector and a voting matrix; establishing a state equation and a measurement equation of a fault-tolerant system; obtaining the drift state estimated values of the sensors of the redundant sensor measurement system; according to the drift state estimated values, determining the drift factors; carrying out the feedback compensation on the drift state estimated values of the faulted sensors; taking the compensated sensor output as the system output. The fault-tolerant method can compensate the drifts of the faulted sensors real-timely, guarantees the redundancy characteristic of the system, at the same time, can realize the fault-tolerant operation without needing the threshold values and the observation cycles any more, enables the system to operate stably even though all sensors generate the fault drift, and has a better robustness.

Description

technical field [0001] The invention relates to the technical field of sensor measurement, in particular to a fault-tolerant method for a redundant sensor measurement system. Background technique [0002] Fault-tolerant design is an important means to improve the reliability of the sensor measurement system. The most important method is to use the geometric redundancy design, that is, to use the redundant configuration of the sensor to improve the overall reliability of the sensor measurement system. This fault-tolerant design requires the sensor measurement system to have reasonable fault detection and elimination techniques. When there is a sensor fault, the faulty sensor can be detected in time, and its fault can be eliminated from the measurement system where it is located to ensure that the measurement system can work normally. At present, there are three main detection and isolation technologies for fault sensors: the first is a fault detection and isolation method ba...

AI Technical Summary

Problems solved by technology

There are the following three problems in the detection and isolation techniques of these three types of fault sensors: (1) The first and third methods all need to set a threshold value, and the value of the threshold value is an empirical value, which leads to artificial subjective The factor is relatively large; (2) The initial error of the sensor in the second method has a great influence on its performance; (3) These three methods all need an observation period, when the sensor output in this period is not normal , it can be judged that the sensor is faulty, and the existence of this observation period makes it impossible to perform fault diagnosis in real time

[0003] However, none of the above three sensor fault detection and isolation methods can make the redundant sensor measurement system maintain its redundant characteristics, because drift is an inherent characteristic of sensors, and each sensor may drift at any moment. Drifting sensors can be correctly diagnosed and isolated by the redundant inertial navigation system, but when the number of sensors reaches the minimum configuration of the system, the system cannot correctly diagnose and isolate the faulty sensor due to the loss of its redundancy characteristics. Therefore, in view of The inherent drift characteristics of sensors require a fault-tolerant method for redundant sensor measurement systems, which can enable redundant sensor systems to have self-repair and compensation functions, thereby ensuring the redundancy and accuracy of sensor measurement systems

Images