A Fault Diagnosis Method for Satellite Temperature Control System Based on Temperature Curve Characteristics

Abstract

The invention discloses a temperature-curve-characteristic-based fault diagnosis method of a satellite temperature control system. The method comprises: step one, telemetering temperature data are collected according to a time sequence and telemetering temperature data information is stored according to a cycle; step two, a cycle i is initialized; step three, difference operation is carried out onNi data of the i-th cycle to obtain a differential sequence corresponding to the i-th cycle; step four, a threshold alpha is set and statistics of the number Qi of numerical values with absolute values larger than alpha in the differential sequence is carried out; step five, 1 is added to the i; if the i is larger than a total cycle number M, a step six is performed; and if the i is less than orequal to the M, the step three is carried out; step six, calculation is carried out at the step four to obtain calculated fluctuation intensity characteristic values Q1, Q2, ..., QM of the M cycles; step seven, for newly-received to-be-detected data, statistics of the number with the difference values exceeding a threshold is carried out based on the first difference and the number is used as a to-be-detected characteristic; and step eight, a confidence interval and a rejection domain in normal distribution are calculated at the significance level and determination is carried out. Therefore, the diagnostic accuracy is improved by analyzing historical telemetering temperature data.

Description

technical field [0001] The invention relates to a fault diagnosis method of a satellite temperature control system based on temperature curve characteristics, which belongs to the field of computers. Background technique [0002] The temperature control system of the satellite is an important part of the satellite. It controls the heat exchange between the internal and external environment of the satellite and is the guarantee for the normal operation of each system in the satellite. Its reliability directly affects the work of itself, other systems, and even the entire satellite. state. The electronic components inside the satellite have a clear range of temperature conditions for the working environment. Once the temperature control system fails, when the ambient temperature changes but cannot be effectively regulated, the temperature of the components themselves will change accordingly. If the temperature exceeds the design range, their performance parameters will also c...

AI Technical Summary

Problems solved by technology

The fault diagnosis system based on expert knowledge can make full use of database technology, store expert diagnosis knowledge rules and diagnosis results in the database in an orderly manner, and can browse, update and maintain at any time, but the quality of knowledge expression has an impact on the application range and efficiency of knowledge Large, but also affects the acquisition of knowledge and the learning mechanism of fault solutions

The model-based fault diagnosis method first establishes the temperature control system model, then gives the solution method according to the characteristics of the model, and finally gives an example of fault diagnosis, but the accurate construction of the model is often very difficult

In recent years, the diagnostic method based on neural network has been developed, and its advantages are highly nonlinear, highly fault-tolerant and associative memory, etc. However, there are still many shortcomings in the application of artificial neural network to fault diagnosis. Reveals some potential relationships within the system, unable to give a clear explanation of the diagnostic process

The network training takes a long time, and it has no ability to diagnose faults that do not appear in the training samples, and even draws wrong diagnosis conclusions, which increase the difficulty of neural networks in practical applications.

[0005] In addition, the above methods are only based on the residual analysis, and the satellite is diagnosed according to the fault status word and the normal range of the data, or the degree of similarity to the existing fault mode is judged, which is likely to cause misjudgment or missed judgment of the abnormal situation, and has great limitations

On the one hand, if the temperature control system fails in the initial stage, there is only a small fluctuation but does not exceed the threshold. In this case, it is difficult to detect potential failures with existing methods; As a result, the current temperature change is very different from the historical data. In fact, it is normal operation, but it is easy to be judged as a fault.

The fundamental reason is that the original method is too focused on the value of the data at a fixed moment, which makes the universality of the diagnostic rules poor.

Images