【作者】 陈华；

【导师】 韩震宇；

【作者基本信息】 四川大学 ， 机械制造及其自动化， 2002， 硕士

【摘要】 本论文对发动机的普通电喷系统的组成进行了分析，研究ECU的基本控制理论和算法。为开发新型的发动机电控单元，设计并制作了一种ECU测试装置，可从国外ECU中获取MAP数据，获取设计ECU控制算法的必要数据。该平台不仅可以用于检测国外已制作好的ECU，同时也可以用于制作新型ECU时的参数优化。PC机从测试台中获取数据，并保存在数据库中，以备编写ECU控制程序时使用。 在研究过程中，还进行了发动机台架实验。实验中，调节节气门在特定位置，并让发动机的速度依次从最低变化到最高，同时记录水温、油温、进气压力和转速等数据，还可以测出相应的喷油量、点火提前角。在此基础上构造稳态控制所需的MAP图，这些数据不仅作为ECU控制算法研究的样本，而且为测试装置的调节、测试和验证提供参考。 另外，在论文中也讨论单点发动机的怠速控制。用模糊控制理论、PID调节、BP神经网络对发动机怠速的速度稳定性进行了深入分析和研究。给出了模糊控制和传统PID调节在怠速控制上的对比。BP神经网络用于学习模糊控制器的输入与输出之间的函数关系，以取代传统模糊控制规则表。在BP网络的训练过程中，采用了自调整的学习算子以加速收敛得到较好的学习效果，最后给出了仿真结果。更多还原

【Abstract】 This paper describes the components in the common electronic control system of engine and presents the basic control theory and algorithm in the electronic control unit. To develop new-style engine ECU, the studies designed and fabricated an ECU measuring device to get the MAP data from foreign ECU. The ECU measuring device can not only detect the foreign ECU, but also be used to optimize the control parameters in the ECU developed. The PC gets the data from the ECU measuring device and stores them in the database, which can be used to write ECU’s control program.In our studies, engine experiment is carried out. In this experiment, adjusting the throttle to the specific position, and letting engine rotary speed change from the lowest to the highest, at the same time, the data of the water temperature, the fuel temperature,the air press and the rotary speed can be noted. Moreover, we can measure the fuel quantity and the ignition angle. On the base of the data, the MAP of the steady state control can be constructed. The data not only become specimen for studying ECU control algorithm but also become reference for adjusting, measuring and verifying the measuring device.This paper also discusses SI engine idle speed control. The rotary speed stability of engine in idle speed state is analyzed and studied further by using fuzzy control theory, BP neural network and PID theory. The simulation results are given though fuzzy control compared with PID control. The BP neural network has been used to learn the relationship of input and output function in fuzzy control to replaced the fuzzy rule table. The advanced and efficient algorithm variable scale method for learning is used in BP neural network training. Finally the control simulation results are given.更多还原