Based On LabVIEW And Support Vector Machine For Fault Diagnosis Study Of Diesel Engine

With the development of science and technology, the diesel engine as a commonmachinery and equipment, are widely used in engineering machinery, mining, shipbuildingand other fields. With the complex structure of the engine, resulting in a variety of fault proneduring operation, not only affect the normal operation of the diesel engine and workflow, butalso may cause serious injury or death. Therefore, diesel engine fault diagnosis has importantpractical significance.First, by reading a lot of domestic and foreign references, commonly used diesel enginefault diagnosis method with in-depth understanding. On this basis, this paper ant colonyalgorithm and support vector machine for diesel engine fault diagnosis. And R6105AZLDdiesel engine as the research object, through in-depth study of two of its five major systemsand structures, known as the presence of some of the typical diesel engine fault and faultsources, and set up a normal, single-cylinder off the oil, the oil off the cylinder, supplyadvance angle increases, the fuel supply advance angle decreases, the air filter clogged sixkinds of conditions.Second, in order to efficiently implement data acquisition, display and storage, this paperuses LabVIEW virtual instruments and PCI8620data acquisition card to develop a dieseldynamic data acquisition system, including systems acquisition interface design, systemhardware selection and commissioning.Again, for the diesel six kinds of conditions, select the appropriate acceleration sensorsand eddy current displacement sensor, and arranged the eleven measuring point.Finally, the diesel engine vibration signal characteristics, using wavelet transform forsignal noise reduction pretreatment, and then the signal time frequency analysis and waveletpacket energy spectrum analysis, and extracting the characteristic parameters. For diesel engine fault diagnosis using ant colony algorithm to optimize the parameters of support vectormachine and selection of K-Ford as a cross-validation method to determine the parameters ofthe ant colony optimization criteria, and then fault classification, and with non-optimizedSVM classification results in contrast, the results show that the ant colony optimizationsupport vector machine has a good classification results.