Research On Key Technology Of Fault Diagnosis And Condition Evaluation For Flight Vehicle

With the increasing complexity of flight vehicle,its reliability and safety are paid more and more attention.Fault diagnosis is a valid technique to monitor and maintain the reliability of flight vehicle,which is widely studied in both industry and academia.Testability is the design characteristic of fault diagnosis,and it can offer the diagnosis capability in a timely and accurate manner.Missile is a type of flight vehicle.Its flight environment includes both inside and outside the atmosphere,and the environmental profile the missile experiences is diverse and complex.In this article,the missile is taken as the research object and its key technologies for fault diagnosis and condition evaluation are studied,which has realistic engineering and military significance,as well as representativeness and promotion values.In order to improve the missile’s fault diagnosis and condition evaluation capabilities,the overall plan is studied considering the function and structure characteristics of the missile weapon,and a series of design methods are formed,including the multi-level built-in test design,rapid fault diagnosis,false alarm suppression,and condition evaluation.The key equipment of the missile is taken to validate the effectiveness of the proposed methods.The approaches formed in this article can provide scientific and effective technical support for the overall design of flight vehicle,and lay a solid foundation for improving the operational readiness.The main contents are as follows.(1)Design of fault diagnosis and condition evaluation schemeIn view of the missile’s function and mission characteristics,the design items of for testability and diagnosis of the missile are planned,and the testability engineering process is established,in which the design for fault diagnosis and condition evaluation as the core part is proposed.By analyzing the fault characteristics of each key equipment of the missile,the diagnosis requirements of each key equipment are clarified,and the corresponding technical research overall plans are formed,which are taken as the input of the subsequent researches.(2)Design method for multi-level built-in test systemBased on the engineering requirements of built-in test,the design method for built-in test of flight vehicle is systematically studied,and the design scheme of built-in test in system-level,equipment-level,and board-level are formed,which effectively integrates the test and diagnostic resources,and provides carrier and support for fault diagnosis and condition evaluation implementation.(3)Knowledge-based fault diagnosis methodOn the basis of the diagnosis knowledge,a new entropy value based on the information entropy theory,the irrational entropy,is proposed,which realizes the effective compression and unique expression of the diagnosis knowledge.Based on the irrational entropy,the hash search method is utilized to realize the fast fault diagnosis.Besides,through the test selection of the dependency matrix,the diagnosis knowledge is described in the form of a binary tree or a table,and then the binary tree search algorithm is used to realize the fast fault diagnosis.(4)False alarm suppression method considering its degradation characteristicsBy analyzing the impact of product degradation on false alarm,the occurrence of false alarm is regarded as an evolutionary process.Based on this,the support vector machine is utilized to identify the stage boundaries of false alarm evolution process,and then the hidden Markov model is introduced to construct the time series model for false alarm recognition.(5)Condition evaluation method based on grey clusteringDrawing lessons from the sub-health theory in modern medicine,three states of flight vehicle are proposed as the input of condition evaluation.Considering the various status parameters and different index results,the grey clustering theory is introduced to effectively integrate a variety of heterogeneous data to achieve accurate assessment of flight vehicle,which supports the realization of condition-based management.