【摘要】：針對渦輪增壓汽油機氣路系統(tǒng)中節(jié)氣門與廢氣旁通閥動力學耦合、機理建模復雜的問題,本文提出基于神經網絡模型的氣路系統(tǒng)預測控制方法,實現(xiàn)了節(jié)氣門與廢氣旁通閥的協(xié)調控制.首先,針對渦輪增壓汽油機氣路系統(tǒng)map與機理混合描述的特性,利用系統(tǒng)的輸入輸出數據,采用反向傳播神經網絡(back propagation neural network,BPNN)訓練得到一個非線性氣路模型;其次,基于泰勒展開式對預測模型進行線性化,并對模型的精度進行了驗證,進而利用該模型預測系統(tǒng)的未來動態(tài);然后,在考慮系統(tǒng)存在輸入約束的條件下,設計了一個線性模型預測控制器對節(jié)氣門與廢氣旁通閥進行協(xié)調控制,實現(xiàn)了進氣歧管壓力和升壓的跟蹤控制進而滿足發(fā)動機的扭矩需求;最后,通過離線仿真和基于d SPACE的快速原型實驗(rapid control prototyping,RCP)驗證了控制系統(tǒng)的有效性和實時性.
[Abstract]:Aiming at the problem of dynamic coupling of throttle and exhaust gas bypass valve in turbocharged gasoline engine gas path system and complicated mechanism modeling, a predictive control method of gas path system based on neural network model is presented in this paper. The coordinated control of throttle and exhaust gas bypass valve is realized. Firstly, a nonlinear gas path model is obtained by using the input and output data of a turbocharged gasoline engine based on the hybrid description of map and mechanism. A nonlinear gas path model is obtained by using the backpropagation neural network (back propagation neural network BPNN). The prediction model is linearized based on Taylor expansion, and the accuracy of the model is verified, and then the future dynamics of the system are predicted by using the model. A linear model predictive controller is designed for coordinated control of throttle and exhaust gas bypass valve. The tracking control of intake manifold pressure and boost pressure is realized to meet the torque requirement of engine. The effectiveness and real-time performance of the control system are verified by off-line simulation and (rapid control prototyping experiment based on d SPACE.
【作者單位】： 吉林大學汽車仿真與控制重點實驗室;吉林大學通信工程學院;中國一汽集團公司研究設計中心;
【基金】：Supported by National Natural Science Foundation of China(61703177,61520106008) Jilin Provincial Science and Technology Department Project(20170520067JH) Jilin Provincial Education Department Project(JJKH20170801KJ)
【分類號】：TK411;TP183
，

本文編號：2147964