Fault Diagnosis And Optimal Control Of Hydrogen Engine Based On Information Fusion

Nowadays,global fossil fuels have been reduced sharply,and the task of development of new energy sources is imminent.Hydrogen has been favored by all walks of life for its cleanliness and permanent renew ability.Hydrogen fuel has many advantages when it is used to replace conventional gasoline and diesel for engines,but it also has certain problems.For example,hydrogen engines have lower output power than gasoline engines and are prone to abnormal combustion such as flashback and pre-ignition.The abnormal combustion phenomenon of hydrogen engines will affect their normal operation.Based on the information fusion technology,the research carried out in this paper includes: the diagnosis research of hydrogen engine pre-ignition based on the fusion of SOM neural network and Multi-Agent system,the research of optimal control of hydrogen engine pre-ignition based on Ant Colony-Fletcher-Reeves Conjugate Gradient Method fusion and the research on the optimization of hydrogen engine overall performance based on PSO-BP information fusion.Specifically,the advantages of the hydrogen engine and the mechanism of abnormal combustion are described.The premature combustion of the hydrogen engine is diagnosed and its performance is optimized based on the pressure increase rate.The research contents of this paper include:1.The SOM neural network method was used to diagnose the phenomenon of hydrogen engine pre-ignition.For the diagnosis results of mild pre-ignition,the Multi-Agent system was used to track and verify the topology change of hydrogen engine.According to the verification results,it can be found that the fault line that is trained to be visually consistent with the fault location where the provided sampledata matches.This method improved the accuracy of the diagnosis result and the limitation of diagnosing only a single fault.2.The ACO-CGF information fusion was used to optimize the ignition timing of hydrogen engine under all operating conditions.The prediction of the optimal ignition advance angle was obtained through experiments.The optimal ignition MAP of hydrogen engine was constructed successfully.Under the circumstances,the optimal ignition advance angle obtained can avoid the occurrence of pre-ignition of the hydrogen engine,and the error accuracy of ACO-CGF reaches 0.99921,which meets the expected requirements.3.Using PSO-BP information fusion by changing the method of hydrogen injection,the effects of two injection modes(single injection and split injection)on the performance of hydrogen engines at different speeds were explored.The actual training and the predicted value were as well as 0.95592.Comparing the two injection methods,it can be concluded that the hydrogen injection method using split injection can increase the temperature and pressure in the cylinder,increase the heat release rate of the mixture in the cylinder,and shorten the combustion duration,thereby making the engine’s power and economy Performance has been improved,but nitrogen oxide emissions will increase,that is,emissions performance will decrease.