Study On Fault Diagnosis Technology Of Marine Medium Speed Diesel Engine Based On Thermal Parameters

With the increasing development of global trade,the safety of marine transportation has become the focus of various countries.The normal operation of diesel engines is the basis for ensuring the safe navigation of ships.The high-intensity work of marine diesel engines and the harsh working environment have led to frequent malfunction of marine diesel engines.Therefore,it is very necessary to vigorously develop the fault diagnosis technology for diesel engines.This paper mainly conducts research on onboard fault diagnosis technology of marine medium-speed diesel engines based on thermal parameters.This research is of great significance for improving the economy and safety of marine diesel engines.The main research contents and conclusions of this article are as follows:1.In this research,various mathematical models and calculation principles in simulation software are introduced.simulation software is used to model and simulate Dafa 6DE-18 medium-speed diesel engine,the original 6DE-18 diesel engine model is tested through bench.The error of the laboratory and simulation values of the cylinder pressure curve,fuel consumption rate and output power under the rated speed and various working conditions are within the allowable range,and the accuracy of the 6DE-18 diesel engine model is verified.2.The common thermal faults of 6DE-18 diesel engine are selected,the fault mechanism is analyzed,a simulation model of typical faults is designed,and a large number of representative thermal parameters are obtained by running the diesel fault model,and these data are normalized and pre-processing of dimension reduction,build a fault diagnosis model to provide sample database.3.The performance of fault diagnosis based on BP neural network and support vector machine for marine medium-speed diesel engines is compared and analyzed.The analysis shows that for the specific research object and less training samples,the performance of SVM fault diagnosis is better than that of BP neural network.Because the random selection of SVM model parameters seriously affects the accuracy rate,the particle swarm optimization algorithm is used to optimize it,resulting in high accuracy and high response speed PSO-SVM fault diagnosis model.4.The bench test verification of the onboard fault diagnosis technology for marine medium-speed diesel engines based on thermal parameters is carried out.Subject to the test conditions,this test only carried out the diesel engine intake pipe leakage fault test and the fuel injector blockage fault test.The collected cylinder pressure curves are compared with the cylinder pressure curves of the fault model corresponding to simulation software to verify the effectiveness of the fault model.The data collected by the fault test is input into the fault diagnosis PSO-SVM splitter model after normalization and dimensionality reduction to verify the accuracy of the PSO-SVM fault diagnosis model.And according to the PSO-SVM fault diagnosis model,the original machine-side controller is updated and embedded into the fault diagnosis module to integrate monitoring,alarm and fault diagnosis into one.In summary,the PSO-SVM mathematical model fault diagnosis method based on thermal parameters has a good fault diagnosis performance,at the same time,the faults of diesel engines under different working conditions can be accurately identified,with stability and universal applicability,which can meet the requirements of practical engineering applications.