FPGA-based Fault Diagnosis Algorithm Simulation And Its Application

The cost of traditional fault diagnosis systems generally use the computer monitor, but isnot conducive to real-time operation of the computer is bulky, and expensive, increase thediagnostic system. FPGA implementation has the advantages of good real-time, Strongspecificity, high reliability, and low power consumption and so on, and use FPGA achievefault diagnosis and fault diagnosis technology will become the mainstream direction.However FPGA has its undesirable aspects: limited internal resources to realize the difficultyof complex algorithms, is currently facing a problem.This article uses the FPGA to achieve network clustering algorithm based on geneticalgorithm optimization process fault diagnosis, fault diagnosis object reciprocatingcompressors experiment. Network clustering algorithm for fault diagnosis model using thegeodesic distance instead of Euclidean distance to measure the relationship between the data,the similarity function by the genetic algorithm to optimize network clustering algorithm, tomake the network clustering algorithm has a better environment adaptability. FPGAimplementation network clustering algorithm based on genetic algorithm optimization onFPGA network clustering algorithm and genetic algorithm, the design of internalcommunication module to pass the fitness function. Top-down design approach to achieve thealgorithm is divided into several modules, respectively, to achieve the function of eachmodule, integrated into a complete algorithm based on the relationship between the module.Altera Corporation’s Cyclone III series FPGA use in the process of realization of the pipelinedesign, finite state machine, speed and area interchange design method to improve theoperating speed of the algorithm. Using VC++6.0preparation of test software on the hostcomputer, the existing data were tested on FPGA systems and computer systems, analyze thefault diagnosis accuracy and speed of execution, the results show that: the correct rate ofcomputer fault diagnosis system slightly higher than the FPGA system, but the execution ratethan FPGA fault diagnosis system is much lower, FPGA Guzang diagnostic system to meetthe real-time requirements.The last analog reciprocating compressor valve leakage experiment platform, artificialdrilling way analog valve leaked, analog valve four common faults: a cylinder intake valveslight leak, the intake valve of a cylinder serious leaked, two cylinder exhaust valve slightleak, two cylinder exhaust valve serious leak, as normal operating conditions the data were collected. Using the temperature and pressure sensors were tested at the temperature andpressure data of the fault, this data as an object for testing, testing the effect of fault diagnosis.Experimental results show that the FPGA-based fault diagnosis system has a higher accuracyrate and high speed, good performance, to meet the real-time requirements.This article uses the FPGA system fault diagnosis, implemented on FPGA networkclustering algorithm based on genetic optimization proposed a set of high-accuracy,high-speed, fault diagnosis system, high-speed and reliable fault diagnosis becomes flexiblediagnostic methods to reduce diagnostic cost, is a feasible design.