OPTIMAL CONTROL SIMULATION OF DIESEL INJECTION SYSTEM USING NEURO MODEL PREDICTIVE-BASED ELECTRONIC CONTROL UNIT

Subject Area: Control system engineering

This paper presents an optimal control simulation of a diesel injection system using a Neuro model predictive-based Electronic Control Unit (ECU). The research aimsto improve diesel engine efficiency while maintaining fuel economy. This was achieved using a model predictive control system and a neural network trained with data collected from diesel engine vehicles online during a 54km to model a Neuro-MPC algorithm which was used to improve the ECU. The Neuro-MPC-based ECU was implemented on a Diesel Control System (DCS) using Simulink. The resultof the Neuro-MPC when evaluated with mean square error is 0.047964Mu and Regression of 0.9773. The results implied good training performance at tolerable error and the ability to predict engine behaviour and supply desired fuel. The result of DCS when tested at a distance of 54Km showed that the Air to Fuel Ratio (AFR) is 14.53: 1 which gives a tolerable error of + (0.03) error when compared with the ideal Stoichiometric standard for DCS which is 14.5: 1, thus giving a mechanical efficiency of 94.3%.