Search

[Objectives] In view of the limitations of support vector machine (SVM) in the application of tractor diesel engine fault diagnosis, a hybrid population optimization-support vector machine (HPO-SVM) tractor diesel engine fault diagnosis model was proposed in this paper. [Methods]SVM was used as the base of fault diagnosis model and population algorithm was used to optimize c and g of SVM parameters. Based on particle swarm optimization (PSO)and grey wolf optimization (GWO)algorithm, hybrid population optimization (HPO) algorithm was put forward to optimize c and g of SVM parameters. The fault mechanism of diesel engine was analyzed, and four common faults of diesel engine and eight sensor data signals reflecting the faults were determined. Sensor signal data of Weichai WP6 tractor diesel engine were collected by the CAN (controller area network) bus and the Arduino UNO-MCP 2551 combined module to test the performance of HPO-SVM. The test results of SVM, PSO-SVM, GWO-SVM, GWOPSO-SVM, and LWD-QPSO-SOMBP (linear weight decrease-quantum particle swarm optimization-self organizing maps back propagation) neural network were compared. [Results] Compared with the other four SVM models, HPO-SVM took full advantage of the advantages of GWO algorithm and PSO algorithm in SVM parameter optimization, and the fault diagnosis accuracy was greatly improved. Compared with SVM, the diagnostic accuracy increases from 80% to 100%, up by 20%, which proved that HPO-SVM model had good diagnostic performance. HPO algorithm improved the optimization performance of single-population optimization algorithm. Compared with the other four models, the average fitness level of HPO algorithm was the highest. Compared with PSO algorithm, the optimum fitness of HPO algorithm was improved from 70 to 90, which increased by 22.22%. When the optimum fitness was reached, the number of iterations decreased from 105 to 27, which decreased by 74.29%. This showed that HPO algorithm had a good optimization efficiency. In order to avoid accidental six repeated tests on five SVM models, the test results showed that the performance of HPO-SVM models was more stable than that of the other four models, and the total accuracy of six diagnoses of HPO-SVM was 100%. Compared with LWD-QPSO-SOMBP neural network, the running time of HPO-SVM model decreased from 45 s to 15 s, which decreased by 66.67%, indicating that the diagnosis efficiency of optimized SVM model was due to the optimization of BP neural network model. [Conclusions]This study results can provide reference for fault diagnosis of high-efficiency tractor diesel engine. [ABSTRACT FROM AUTHOR]

With the rapid development of agricultural economy and the continuous improvement of modern agricultural mechanization level, agricultural machinery gradually tends to be precise and complicated, which puts forward higher requirements on the timeliness and accuracy of fault diagnosis technology of agricultural machinery. Mechanical fault diagnosis is a key technology in the field of agricultural machinery. In this paper, based on the existing theoretical knowledge, and combined with domestic and foreign literature, the fault diagnosis technology of agricultural machinery was comprehensively expounded. The research status of the four major mechanical fault diagnosis technologies of vibration monitoring, oil sample analysis, noise monitoring and infrared temperature measurement, and the time-frequency domain signal analysis methods of mechanical fault diagnosis technology were introduced in detail. Finally, the development trend of agricultural machinery fault diagnosis technologies was pointed out from different aspects, including the introduction of general machinery diagnosis technology, the continuous improvement of the application of intelligent diagnosis technology in the diagnosis of agricultural machinery faults, and the coordinated development of multiple diagnostic technologies. This paper can provide a reference for the further development of fault diagnosis technology of agricultural machinery. [ABSTRACT FROM AUTHOR]