Research On Diesel Engine Control Technology Based On Neural Networks

The quantity of fuel injected into the Cylinders is determined by the rack position withthe control of the Electronic Governing System most of which use the double closed-loopPID control strategy。The diesel engine is a complicated system with the characteristic ofnonlinear and time-varying. But the best parameters of the PID controller are different in thevarious working conditions and the parameters can’t be regulated online once put into use.The control strategy combined PID with neural networks is an important solution whichbelongs to intelligent control to overcome the shortcomings of PID controller.Back Propagation(BP) neural network with capacity of self-learning and adaptive canoptimize the PID controller parameters online. Based on MCU of MC9S12XEP100, thecontroller was experimentally evaluated on rack position actuator of diesel engine. Theexperimental result demonstrated that the controller is of better performance inadaptability and capability of anti-jamming but more calculating trouble compared to PID.As an effective tool in the development diesel engine control system,Hardware-In-Loop (HIL) simulation and off-line simulation can verify and debug a newalgorithm controller in a low cost and process-safe way. The core of the HIL andoff-line simulation is engine model which must be a high real-time and precise model. So aparametric model is built with with the the steady and dynamic state data of D6114dieselengine based on the software of Automotive Simulation Models(ASM) whose theoreticalbasis is the filling and emptying model and the quasi-steady state model in this paper. Andthe model which is verificated and calibrated can be used as the real-time simulation model.To verify and debug a new algorithm controller in a low cost and process-safe way, aHIL system based on the dSPACE is built with a real actuator and the D6114diesel enginemodel. A controller combined the Cerebellar Model Articulation Controller(CMAC) neuralnetwork with PID is designed and implemented to the speed control of diesel engine and theHIL and off-line simulation show that the diesel engine can have a better performance withthe controller.After the simulation, the CMAC-PID controller is experimentally evaluated which isthe most convincing part in the development of the control strategy on the diesel generatingset of D6114. The experiment on the speed adjustment, steady-state operation andmutation load of the D6114diesel demonstrated that the controller is of good performanceand it have the similar control effect to the product-level controller...