面向抗干扰性的HCCI发动机点火正时控制策略研究

发布时间：2018-06-09 11:14

本文选题：点火正时 + 抗干扰性　；参考：《重庆邮电大学》2016年硕士论文

【摘要】：在能源和污染问题日益严重的今天,HCCI(Homogeneous Charge Compression Ignition,均质充气压缩燃烧)发动机由于具有高效和低排放的特性而备受欢迎。HCCI发动机的正常运行需要有效的点火正时控制作为基础。考虑到发动机存在多种形式的外部干扰和内部参数扰动,其点火正时控制策略应具备良好的抗干扰性。然而,由于未充分考虑干扰的影响,或由于算法的限制,现有的控制策略均不能有效抑制内外部干扰的影响,其抗干扰能力仍有待提升。因此,本文将围绕控制策略的抗干扰性,对HCCI发动机的点火正时控制展开研究。论文主要工作包括:1.HCCI发动机模型建立为了方便控制策略的设计,首先,选取进气门开启时刻和排气门关闭时刻的曲轴转角为模型输入变量,点火正时为输出变量,进气门关闭时刻的混合气温度为状态变量,在充分考虑残余废气对点火正时的影响,而且在保证精度的同时确保较小的计算量的前提下,通过分析缸内的化学动力学,建立了HCCI发动机的状态空间模型。2.基于BP神经网络的点火正时PI控制策略研究针对传统固定参数的PI(Proportion Integration)控制器无法有效抑制外部干扰的问题,研究了基于BP(Back Propagation)神经网络的点火正时PI控制策略。以期望点火正时和实际点火正时信号之间的误差为输入,以进气门开启时刻曲轴转角的增量为输出,设计了点火正时PI控制器。为了使PI控制器的参数适应实时的运行情况,以期望点火正时、实际点火正时以及两者之间的误差为输入,以PI控制器的参数为输出,建立了BP神经网络,对点火正时PI控制器的参数进行整定。仿真结果显示:在跟踪发生阶跃变化的点火正时信号时,基于BP神经网络的点火正时PI控制策略具有比传统PI控制器更快速平稳的瞬态响应过程,并且稳态误差为零,显示出良好的跟踪性能。在出现多种形式的外部干扰后,相比固定参数的PI控制器,基于BP神经网络的PI控制策略能够将偏离期望值的点火正时更加快速平滑地调节至期望值,显示出更强的抗干扰能力。3.基于干扰观测器的点火正时前馈控制策略研究针对2中所提策略瞬态响应速度稍显缓慢,以及抗干扰性能有待提升的问题,研究了基于干扰观测器的点火正时前馈控制策略。为了提高点火正时控制的瞬态响应速度,以发动机的实测状态值和期望状态值为输入,以进气门开启时刻的曲轴转角为输出,设计了前馈控制器以控制点火正时。为了更好地抑制内部参数扰动和外部干扰,该策略将内外部干扰统一视为复合干扰,设计干扰观测器对该复合干扰进行观测,并在前馈控制律中加入复合干扰的观测值以抵消复合干扰的影响。仿真结果表示:在跟踪发生阶跃变化的点火正时信号时,基于干扰观测器的前馈控制器在一个循环之内即可跟踪到期望值,证明了其瞬态响应速度较2中策略有显著提升。当出现由内部参数扰动和外部干扰组成的复合干扰时,基于干扰观测器的点火正时前馈控制策略能够将点火正时控制在期望值的一个更小的邻域内,证明了其具备比2中所提策略更强的抵抗内外部干扰的能力。
[Abstract]:Today, with the increasingly serious energy and pollution problems, the HCCI (Homogeneous Charge Compression Ignition, homogeneous inflatable compression combustion) engine has been popular for the high efficiency and low emission characteristics, and the normal operation of.HCCI engine needs effective timing control as the basis. Interference and internal parameter disturbance, the control strategy of point fire should have good anti-interference. However, because of the lack of consideration of the influence of interference, or due to the limitation of the algorithm, the existing control strategy can not effectively suppress the influence of internal and external interference, and its anti-interference ability still needs to be improved. Therefore, this paper will revolve around the anti dry of control strategy. The main work of this paper is as follows: the main work of this paper is that the 1.HCCI engine model is built for the design of the control strategy. First, the crankshaft rotation angle of the inlet valve opening time and the exhaust valve closing time is selected as the model input variable, the ignition timing is the output variable and the intake gate closes the mixture gas. When the temperature is the state variable, the effect of the residual gas on the ignition timing is fully taken into account, and on the premise of ensuring the accuracy while ensuring the small amount of calculation while analyzing the chemical dynamics in the cylinder, the state space model of the HCCI engine,.2. based on the BP neural network, is set up for the traditional fixed time PI control strategy. The parameter PI (Proportion Integration) controller can not effectively suppress the external interference. The ignition timing PI control strategy based on the BP (Back Propagation) neural network is studied. The error between the desired ignition timing and the actual ignition timing signal is input, and the increment of the crankshaft angle increment at the opening time of the intake gate is designed. In order to make the parameters of the ignition timing PI controller, in order to make the parameters of the PI controller adapt to the real time operation, in order to expect the ignition timing, the actual ignition timing and the error between the two are input, the BP neural network is set up with the parameters of the PI controller as the output, and the parameters of the PI controller are set in the ignition timing. The simulation results show that the tracking occurs When the step change ignition timing signal is changed, the ignition timing PI control strategy based on BP neural network has a faster and stable transient response process than the traditional PI controller, and the steady-state error is zero, showing good tracking performance. After various forms of external interference, compared with the fixed parameter PI controller, the BP neural network is based on the BP neural network. The PI control strategy can adjust the departure time from the expected value more quickly and smoothly to the expected value, and show a stronger anti-interference ability,.3. based on the interference observer, the ignition timing feedforward control strategy studies the problem that the transient response speed of the strategy proposed in 2 is slightly slower and the anti-interference performance needs to be improved. In order to improve the transient response speed of ignition timing control, a feedforward controller is designed to control the ignition timing in order to improve the internal parameter disturbance in order to improve the transient response speed of the ignition timing control, in order to improve the transient response speed of the ignition timing control, the measured state value and the expected state value of the engine are input, and the crankshaft angle of the inlet valve opening moment is the output. And external interference, the strategy considers the internal and external interference as a compound interference. The interference observer is designed to observe the complex interference, and the observation value of the compound interference is added to the feedforward control law to counteract the influence of the compound interference. The feedforward controller can track the expiration value in a cycle. It is proved that the transient response speed is significantly higher than that in the two strategy. When there is a complex interference consisting of internal and external disturbances, the ignition timing feedforward control strategy based on the disturbance observer can control the ignition timing to a more desired value. Within small neighborhood, it is proved that it has stronger ability to resist internal and external interference than the strategy mentioned in 2.
【学位授予单位】：重庆邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TK411

【参考文献】