Abstract:
Abstract: A rotary engine has broad application prospects in the plant protection of agricultural spraying, range extenders of an electric vehicle, and generator sets, due mainly to its light weight, small size, stable operation, low noise, and excellent high-speed performance. The current capability of gas distribution has a great challenge on the power, economy, and emissions in the rotary engine. Therefore, it is necessary to systematically design and optimize the valve system of a rotary engine for better performance of gas distribution. Many previous efforts have been made on the experiments and simulation of the intake model. However, the physical experiment is time-consuming and high cost, not suitable for the systematic optimization of an engine. The traditional optimization can consider only one or two variables at the same time. Alternatively, the recent parametric modeling has achieved the functional trade-off between the model complexity and computational efficiency. In this study, a parameter design of a fully variable gas distribution port system in a rotary engine was proposed to reduce the complexity of the valve model for the accurate description of the working process of the valves. The shape curve of the valve port was selected to obtain the opening area curves of gas distribution in the rotary engine. The working process of gas distribution was constructed using parameterized arrays in a one-dimensional simulation model, thereby efficiently realizing the continuous adjustment of parameters in the gas distribution system. The parameters of valve port (advance angle deviation of intake, speed of opening area at the intake port, duration coefficient of the intake port) were used as the response surface factor, and the monitoring indicators of performance (power and fuel consumption rate) were used as the response surface response. A single factor scanning value method was used to determine the optimal range of parameters. The response surface method was selected to regress the performance of the rotary engine. Finally, multi-objective optimization was utilized to optimize the valve parameters of the rotary engine. The results showed that the maximum error for the cylinder pressure and volume change curve of simulation and experiment was within 5%, indicating a high accuracy on the fully variable gas distribution parameter model of the rotary engine. There were a high accuracy and excellent predictive ability on all the quadratic response surface regression models. The determination coefficient of models, the adjusted and prediction determination coefficients were all above 0.95. and the maximum error between the test and the predicted value was less than 2.1%,so the model has good accuracy and predictive power. After optimization, the intake port is shifted forward by 19.21°CA, increasing the duration of the intake port by 78.40% and expanding the port width by 58.89%; the exhaust port is shifted backward by 19.92°CA, with no change in duration and an increase in port width of 46.60%.The optimized overlap angle of the valve was 39.3°CA higher than that in the original machine. The engine power increased by 1.07 kW, 25.36% higher than before. The fuel consumption of the engine was reduced by 104.46 g/(kW·h), 20.01% lower than that of the conventional engines. The maximum error between the optimization and the simulation was within 1%, indicating a high accuracy of the optimization model. The research findings can provide insightful suggestions to optimize the performance of diesel rotary engines.