Comparative Study Of Fuel Consumption Rate Optimization Methods For Atkinson Engine

As a complex nonlinear system,the design variables and operation variables of automobile engine are highly coupled and interacted with each other.For example,the Atkinson cycle engine requires greater LIVC(Late Intake Valve Closed)and greater angle of ignition delay to avoid detonation,if the geometric compression ratio is too large.However,it will greatly reduce the engine WOT(Widely Open Throttle,WOT)torque at the same time,and larger ignition angle will in turn offset advantages of fuel economy bringing by large expansion ratio.In this paper,research has been completed with the help of the school enterprise cooperation project,which was to modify an Otto cycle gasoline engine to the Atkinson cycle gasoline engine.The main contents and innovations of this paper are as follows:The simulation model of Atkinson cycle engine was built with GT-Power,and used to research the effect of engine operating parameters on fuel consumption.The calibration results verified the correctness of the model that meant the model was prepared for the optimization with genetic algorithm.The influence of main operation parameters on the performance of Atkinson cycle engine was investigated,and the optimization range was determined.At the same time,the load control method of Atkinson cycle engine was put forward which was well prepared for the next step.In order to reduce the workload of bench test and theoretical calculation,the DOE(Design Of Experiment)method was used to select the engine operating point,and the used GT-Power model to calculate the points.The artificial neural network was used to simplify the simulation model of the Atkinson cycle engine.The two optimization schemes are different that GT-Power united genetic algorithm and artificial neural network combined with genetic algorithm.The optimization results of the two methods showed that the neural network model has more advantages,and the optimal time could be saved by 322 times.Compared with the optimization results of program two,the average fuel consumption was about 3%decrease.The test calibration results under rated speed showed that the ANN optimization method could be used for the rapid and accurate optimization of Atkinson cycle engine,and the fuel economy of Atkinson cycle engine has been improved better.The maximum error of the ANN model was 8.2%,the average error was 3.5%relative to the optimization test results of the platform.Compared with the original test results,the maximum improvement was 5.4%at 5200rpm and 52.4Nm,and the average improvement was 3.6%.