穆建华，夏志豪，卢兆强，梁源飞
（上汽通用五菱汽车股份有限公司，柳州 545007）
摘要：本文引入多目标优化方法对发动机性能仿真模型进行标定。应用GTpower和ModeFrontier进行多目标优化标定，通过合理选择优化输入变量，以发动机扭矩和发动机比油耗作为目标变量，克服了传统标定方法耗时长，误差大（通常控制在5%以内）的缺点。仿真结果表明，应用多目标优化对性能仿真模型标定不仅具有速度快、精度高、智能化等优点还能提供输入输出相关性分析，为发动机后续设计开发提供技术依据。
关键词：GTpower；ModeFrontier；多目标优化；模型标定
Abstract: This paper introduced a multi-objective optiomization to complete an
engine performance simulation model calibration. Based on reasonable input
parameters selected, the engine torque and BSFC are chosen as multi-objective
target outputs in the joint model based on GT-power and ModeFrontier. The new
method overcomes the disadvantages of time-consuming and big-error (5% error
range in tradition calibration method). Finally The result shows the model
calibration method based on multi-objective in this paper is time-saving and
has higher precision, higher automation and can get a high reliability results, especially provides
technical support for making decision or judgment later.
Keyword: GTpower; ModeFrontier;
Multi-objective Optimization; Model Calibration
|
