Research On Optimization Of Sensor Locating For Diesel Engine Fault Diagnosis

Diesel engine plays an important role in the ship transportation and power machinery industries due to its excellent power performance and extremely high combustion efficiency.The structure of the diesel engine is complex,and most of its components work in the harsh environment of high temperature,high pressure and severe vibration for a long time,which will lead to the failure of the diesel engine.Failure of diesel engine will cause huge economic losses and even seriously threaten the safety of people.It is very important to ensure the safety of the diesel engine by monitoring the status information of diesel engine to diagnose the potential fault.The status information of the diesel engine is obtained through sensors.Generally speaking,the information will become more comprehensive and complete with the increase of the number and types of sensors,so that it is easier to diagnose diesel engine faults.Placing too many sensors aimlessly will not only destroy the strength of the diesel engine,but also cause information redundancy,which makes it difficult to extract core information.On the contrary,a small number of sensors will lead to incomplete information,especially when the sensor fail,which will lead to serious loss of information in the monitoring system.Therefore,it is necessary to study an optimization method of sensor placement for diesel engine fault diagnosis tasks,which is of great significance to the practical application of diesel engine fault diagnosis technology.This paper studies the causality model of diesel engine,sensor layout evaluation system,sensor layout optimization selection method and sensor fault detection method from four aspects.First,establish a diesel engine simulation model based on the structure and working principle of the diesel engine.Extracting the causal relationship model of diesel engine failure through simulation experiment data.Based on the above model,an evaluation system for sensor layout schemes is designed in this paper.It includes three evaluation indexes: fault isolation ability,fault isolation ability after sensor failure and the number of sensors.Secondly,the optimal sensor placement scheme is selected through the Pareto optimal strategy,in which the improved binary particle swarm algorithm and the banker’s rule are used to optimize the optimization speed of the Pareto population and the construction speed of the non-dominated solution set.Finally,sensor failures will seriously affect the authenticity of the information obtained.To ensure that the sensor monitoring information is true and effective,this paper uses clustering and principal component analysis methods to monitor sensor failures based on Q statistics and contribution Data isolation measures are adopted to implement efficient and accurate fault diagnosis for diesel engines.In this paper,taking MTU 8V 396 diesel engine lubrication system as an example,the proposed sensor optimization method and sensor fault detection method are validated.The results show that the sensor arrangement scheme can accurately identify the working state of the sensor using the proposed approaches,which completely isolate all faults under the condition of sensor failure.