A Fuzzy Mathematics Based Fault Auto-diagnosis System for Vacuum Resin Shot Dosing EquipmentHE Zheng-wen , XU yu , WU JunSchool of Management, Xian Jiaotong univervsity, Xian 710049, P.R. ChinaAbstract: On the basic of the analysis of faults and their causes of vacuum resin shot dosing equipment, the fuzzy model of fault diagnosis for the equipment is constructed, and the fuzzy relationship matrix, the symptom fuzzy vector, the fuzzy compound arithmetic operator, and the diagnosis principle of the model are determined. Then the fault auto-diagnosis system for the equipment is designed, and the functions for real-time monitoring its operation condition and for fault auto-diagnosis are realized. Finally, the experiments of fault auto-diagnosis are conducted in practical production and the veracity of the system is verified.Key words: fuzzy model, fault auto-diagnosis system, vacuum resin shot dosing equipment1 IntroductionVacuum Resin Shot Dosing Equipment (VRSDE) is a key special equipment used to perfuse various kinds of electrical components, such as striking windings of vehicles and motorcycles, transformers, sensors, capacitors and so on (hereafter referred to as “workpieces” ) with epoxy resin. Its function is to purify and to mix epoxy resin (liquid A) with solidifying catalyst (liquid B) under a vacuum condition firstly, and then to perfuse them into workpieces in accurate proportion and quantum, thus finishing epoxy resin perfusion.Similar to most complicated equipment, the fault information environment of VRSDE is a fuzzy one, basically. Beginning with analyzing the relationship of faults and their causes, we construct the mathematical model of faults and their causes, we construct the mathematical model of fault diagnosis for the VRSDE based on fuzzy theory. On the basis of this model we build a fault auto-diagnosis system for the equipment that fulfills the functions of real-time monitoring on the operation condition and fault auto-diagnosis of the equipment. The veracity of the system is verified in practical productions.2 The main faults and their causesAccording to the impacts on the quality and the process of perfusion, the main faults of the VRSDE can be classified as the following kinds: proportion fluctuation, insufficient mixture, inadequate purification, leaking from shot mouth, quantum fluctuation, quantum reduction and no liquid shooting out from shot mouth. Our research on the formation mechanism of faults shows that these faults have relations on the following reasons to different extents: uni-direction valve being in malfunction state ,shutting-valve being in malfunction state, the pipe out of the perfusion chamber seeping, the pipe within the perfusion chamber seeping, liquid level in purifying canister being too low, vacuum degree of purifying canister being too low, liquid temperature fluctuation, moving down speed of the cylinder being too low, air pressure being too low, measuring pump wearing and tearing seriously, machining precision of measuring system being too low, proportion being too big or too small and resin containing arenaceous quartz. The correlative degree of faults and their causes can be described by Table 1 qualitatively.Table 1 The correlative degree of main faults and their causes of VRSDEProportionfluctuationInsufficientmixtureInadequatepurificationLeakageFrom shotmouthQuantumfluctuationQuantumreductionNo liquid shooting out from the shoot mouthUni-direction valve being in malfunction state/Shutting-valve being in malfunction state/The pipe out of perfusion chamber seeping/The pipe out within perfusion chamber seeping/Liquid level in purifying canister being too low/Vacuum degree of purifying canister being too low/Liquid temperature fluctuation/Moving down speed of the cylinder being too low/Air pressure being too low/Measuring pump wearing and tearing seriously/Machining precision of measuring system being too low/Proportion being too big or too small/Resin containing arenaceous quartz/Note: The number of “ “ indicates the correlative degree of causes and faults.3 construction of fuzzy model for fault diagnosisWe define that the operation condition set of the VRSDE consists of all kinds of faults and is expressed by V: V1 Proportion fluctuation V2 Insufficient mixtureV3 Inadequate purificationV = V4 = Leakage from shot mouthV5 Quantum fluctuation V6 Quantum reductionV7 No liquid shooting out from shot mouth