A Momentum Wheel Success Envelope Analysis and Application Method

Abstract

The invention discloses a success envelopment analysis and application method of a momentum wheel, comprising: firstly, obtaining product performance index parameters of the momentum wheel, sorting the parameters into terminal value judge type parameters, process analysis type parameters and trend prediction type parameters; then forming success envelopes of the three parameter types; finally, judging whether a current parameter is in the envelopment range according to the success envelopes of the three parameter types, further obtaining whether the current parameter is usable. Compared with the prior art, various parameters of the momentum wheel in the invention are sorted into the terminal value judge type parameters, the process analysis type parameters and the trend prediction type parameters, statistic analysis is carried out to the parameters of the momentum wheel more accurately, further the parameters of the momentum wheel are judged and predicted better; and the accuracy of the success envelopment is improved.

Description

technical field [0001] The invention relates to a product data detection and prediction method, in particular to a momentum wheel success envelope analysis and application method. Background technique [0002] The traditional data envelopment analysis (Data Envelopment Analysis, DEA) is an efficiency evaluation method proposed by operations research scientists A. Chames, W.W. Cooper et al. In the effectiveness evaluation of the unit, it has the advantages of avoiding subjective factors, simplifying the algorithm, and reducing errors. The above method is based on the concept of relative efficiency and is a non-parametric statistical method for evaluating whether the decision-making units with the same type of multi-input and multi-output are technically effective. The basic idea is to regard each evaluated unit as a decision-making unit ( Decision Making Units, DMU), and then many DMUs constitute the evaluated group, through the comprehensive analysis of the input and output...

AI Technical Summary

Problems solved by technology

[0003] Due to the characteristics of small batches of spacecraft, the amount of successful historical parameter data is usually not large. If it is assumed that the parameter data is normally distributed or the distribution is known, it will usually bring about two problems: (1) the controllable time envelope The running length of the normal distribution will seriously deviate from the target value, so that the judgment of whether the process is out of control or not is unfounded; (2) The test statistics obtained by the normal distribution are not necessarily sensitive to the fluctuation of the non-normal process, so when the process is out of control , it is usually difficult to quickly call the police

This method has the following disadvantages: For multivariate non-parametric control charts, it is not easy to perform online control due to the inability to define sign statistics and rank statistics in multivariate situations

Images