Research On Failure Warning Method Of Marine Diesel Engine

With the development and application of intelligent ships,the requirements for cabin intelligence are gradually improved,and marine diesel engine,as an important equipment in the cabin,is an important guarantee for the normal navigation of ships.In order to meet the demand of ship cabin intelligence,this paper takes marine diesel engine as the research object,researches on the fault warning method of marine diesel engine,and proposes the state monitoring and fault warning method of marine diesel engine based on deep learning.The method can identify the abnormal state of the equipment in time,and send out the warning alert before the occurrence of failure,so as to avoid the occurrence of failure as far as possible,thus avoiding major safety accidents,which is of great significance to guarantee the normal operation of diesel enginesThis paper firstly introduces the principle of marine diesel engine and typical faults in detail,analyzes the causes of the faults and gives solutions.Then,the deep learning models used in this paper are introduced in detail,focusing on the principles of the Convolutional Neural Networks(CNN),the Recurrent Neural Network(RNN)and their variants to lay the foundation for the establishment of the prediction model in the following.In the study of single parameter prediction and fault warning,a combined CNN-Bi GRU prediction model was established using convolutional neural network and Bidirectional Gated Recurrent Unit(Bi GRU)to predict the exhaust temperature of diesel engines,and the prediction results of the combined model were analyzed;based on the residual analysis method and sliding The upper and lower thresholds of residuals and standard deviation thresholds are determined based on the residuals analysis method and the sliding window analysis method;through experimental verification,the proposed fault warning method can realize real-time monitoring of diesel engine exhaust temperature and future state prediction,and can identify the abnormal state of the equipment,which can realize the fault warning of diesel engine.In order to study multi-parameter prediction and fault warning,a prediction method combining Principal Component Analysis(PCA)and two-way gated cyclic unit is proposed,and a PCA-Bi GRU multi-parameter prediction model is constructed to monitor the status of multiple operating parameters and predict the future status,and the standardized residual method combined with Euclidean Distance is used to set the multi-parameter prediction deviation,so as to determine the fault threshold,and at the same time,to prevent the false alarm phenomenon,the standard deviation value of the deviation is calculated by using sliding window,so as to set two fault thresholds;the effectiveness of the proposed multi-parameter fault warning method is proved by abnormal state analysis.Finally,the results of the marine diesel engine condition prediction and fault warning method studied in this paper are summarized,and the future research directions and contents are prospected.