【作者】 周能辉；

【导师】 赵华； 谢辉；

【作者基本信息】 天津大学 ， 动力机械及工程， 2007， 博士

【摘要】 均质压缩燃烧(Homogeneous Charge Compression Ignition,HCCI)方式具有可以同时降低油耗和排放的优势,得到了内燃机界的广泛研究,并日趋成为一个新的燃烧方式。但其还存在着火时刻控制、变工况控制及其控制系统等方面的难点。本文主要是围绕着HCCI发动机的这一系列控制问题展开了研究。针对基于全可变气门机构的汽油HCCI发动机,为满足HCCI发动机的闭环反馈控制的研究需要,提出了基于单片机和CAN总线的分布式HCCI控制体系结构。同时为了保证控制策略的实时性和简化控制参量之间的耦合关系,提出了HCCI分层闭环管理的思想。将控制系统划分为负荷控制、燃烧相位控制和机构控制等三个层次,三个层次均采用相应的闭环控制。系统结构清晰,易于实现,而且具有较高的鲁棒性。本文开发了基于分布式控制系统的HCCI发动机管理单元、喷油点火控制单元、全可变气门机构控制单元以及缸压采集分析单元等基本功能模块。采用基于神经元自适应PID控制策略,实现了对全可变气门机构的精确、有效控制,为实现HCCI发动机运行控制奠定了基础。开发了缸压采集及实时分析算法,解决了燃烧信息在线反馈问题。开发了基于动态递归神经网络的进气量预测算法,有效解决了空燃比准确、快速控制问题。采用基于模型的优化方法和试验相结合,研究了HCCI发动机稳态运行时的气门控制规律。提出了利用经典PID控制和前馈控制相结合的办法,实现了HCCI燃烧下的IMEP和燃烧相位CA50的闭环控制,该策略实时性强,可以实现基于循环的控制。研究开发了基于混合放热率管理的HCCI/SI模式过渡闭环控制策略。试验表明,所提出建立的分层闭环控制策略可以实现有效的HCCI运行控制。建立了HCCI虚拟车辆仿真平台,结合发动机工作循环仿真和道路运行仿真,在整车的NEDC工作循环下评价了控制策略,并预测了HCCI发动机在NEDC循环下的性能。更多还原

【Abstract】 Homogeneous Charge Compression Ignition combustion has the potential of providing better fuel economy and emission characteristics. Therefore, it has been widely researched and is gradually becoming a new combustion mode apart from the original ones. However, it is still confronted with problems in ignition timing control, transient operation and control system. In this paper, the aforementioned problems that occur in the application of an HCCI engine are systematically studied in detail.A CAN-bus-based distributed HCCI control system was designed to obtain close loop control on the fully variable valve actuation (FVVA) HCCI gasoline engine. Meanwhile, a layered management strategy was developed to achieve highly real-time control as well as to simplify the couplings between the inputs and the outputs. The entire control system was stratified into three layers, responsible for load management, combustion phase control and mechanical system control respectively, each with its own specified close loop control strategy. The system is outstanding for its explicit configuration, easy actualization and robust performance.Based on the distributed control system, software and hardware were developed for the primary functioning modules of HCCI engine management unit, injection and ignition control unit, FVVA control unit and in-cylinder pressure acquisition unit. With the single-neural-network-based self-adaptive PID control strategy employed, accurate and effective control of FVVA was achieved, which contributed to the operation control of the HCCI engine. Also, an algorithm of in-cylinder pressure acquisition and real-time analysis was developed, realizing the on-board feedback of combustion condition. Another algorithm to predict the amount of inlet flow was developed based on dynamic recurrent neural network, leading to accurate and rapid control of A/F ratio.Model-based optimization and experiments were employed to investigate the general valve control law when the HCCI engine was operated in steady state. The strategy of combining classical PID and feedforward control was proposed to realize the control of IMEP and CA50 in HCCI mode, and cycle-to-cycle control was achieved due to the high real time of this strategy. Meanwhile, HCCI/SI mode transition control strategy based on managing hybrid heat relese curve was studied. Experimental results show that, with the layered management strategy applied, effective control over HCCI operation becomes possible.A virtual HCCI vehicle simulation platform was established, on which engine operation simulation and road simulation were combined to evaluate the HCCI control strategy in NEDC drive cycle as well as to predict the HCCI engine performance throughout the entire NEDC cycle.更多还原