Study On The Intelligent Fault Diagnosis Of The Parameter Measure System For High-power Laser Device

Condition monitoring and fault diagnosis of the parameter measurement systemfor high-power laser device plays a vital role in guaranteeing the reliability and thesafety of the device. At present, the traditional threshold monitoring method is themain method to monitor the important parameters of the parameter measurementsystem for high-power Laser device. The threshold monitoring method only canprovide the deviation of the important parameters of the parameter measurementsystem for high-power laser device, but it is difficult to find the intrinsic fault causeand the developing trend of abnormalities. Therefore the development of theintelligent fault diagnosis methods have the vital significance to improve thereliability, the security, the maintainability and the validity of High-power Laserdevice.The thesis is sponsored by the big science project named “High-power laserdriver”. The research is carried out on the electric control system of the parametermeasurement system for high-power laser device and puts an emphasis on the basicstructure, the technical principle and methodology of the intelligent fault diagnosissystem. The main work and innovation points can be summarized as follows:1The perfect intelligent fault diagnosis system was proposed based on theelectric control system of the parameter measurement system for high-power laserdevice. Its main functions include the real-time condition monitoring, diagnostic ruleacquisition, fault diagnosis, fault maintenance functions and so on.2A reduction algorithm based on the Core Searching algorithm was proposed,which meets the IND (B)=IND (A)(A is the attribute set and B is the minimumknowledge reduction set). From the mathematical point of view, the reductionalgorithm unifies the attribute reduction and the attribute value reduction. By makingthe best of the condition of IND (B)=IND (A), the algorithm can ensure the reductionset is minimal and at the same time get the most minimal attributes and attributevalues. The algorithm greatly reduces the amount of calculation.3The fault diagnosis method based on RBR, CBR and RBF cloud neuralnetwork was proposed. The intelligent fault diagnosis method not only has the cloudtheory advantages of randomness, fuzziness, and has the RBF advantages ofself-learning, adaptive ability, but also has the advantages of association, analogy and inference. By the intelligent fault diagnosis method we can achieve the fast andaccurate fault location. The efficiency of fault diagnosis is improved greatly.4In view of the breakdown maintenance decision problem, the RCM theory wasresearched. The RCM method is used to analysis maintenance pattern and calculatemaintenance cycle, and on this basis to determine whether a system or equipmentneed for examination, repair, and maintenance pattern, etc5The intelligent breakdown diagnosis system successfully realized theestimate functions. The confirmation result shows that the intelligent faultdiagnosis system is feasible and effective.