【作者】 刘金刚；

【导师】 周云山；

【作者基本信息】 湖南大学 ， 车辆工程， 2008， 博士

【摘要】 金属带式无级变速器(CVT)是汽车的理想传动装置,它具有结构紧凑、性能优良、易于自动控制等优点,发展无级变速器成为我国自动变速器工业的共识。电液控制系统是无级变速器最重要的部分,包括TCU、液压控制装置以及执行机构3部分。其关键技术为:1、液压控制系统的设计与试验方法;2、比例电磁阀试验与开发技术;3、速比控制技术;4、产品一致性及可靠性的技术。本文围绕上述关键技术进行了深入的研究,主要成果和创新点为:1)根据液压控制系统的压力等级,将整个系统分为:变速与夹紧力控制、离合变矩控制、润滑冷却3个子系统。把一个复杂的系统逐一简化,在功能等效的基础上完成了液压系统的集成,并制作液压控制模块的样品。通过台架试验的结果表明:液压控制模块的稳态压力、动态压力和瞬态流量等主要性能指标,都能达到汽车不同行驶工况所需的目标值;整车道路试验表明:所设计的液压控制模块完全满足汽车任意行驶工况下对夹紧力与速比控制、起步离合器、液力变矩器控制的性能要求。根据润滑冷却流量分配关系,在设计时充分考虑了润滑阀流量—压力特性对流量分配产生的影响,利用优化算法对润滑冷却系统进行了优化设计。试验结果表明:在任意工况下,优化后润滑冷却系统可以满足整机和重要部件(如金属带)的润滑冷却要求。2)系统研究了CVT专用电液比例阀的设计方法,从液压阀的工作原理入手,建立了阀的理论模型,该模型的仿真结果与液压阀的试验结果的完好一致性验证了该理论模型的正确性。基于所建立的理论模型,利用遗传算法对专用阀的敏感结构参数进行了优化。优化后的电液比例阀的试验结果表明:各项动态性能指标明显优于原型阀,更能满足无级变速器的使用要求。3)根据CVT夹紧力控制阀性能测试的数据,采用粗糙集和最小二乘支持向量机相结合的方法,完成了产品质量分类器的设计;该方法利用粗糙集理论约简原始条件属性集,去除大量冗余信息,对最小二乘支持向量机分类器进行了有效的简化,降低了分类器的复杂程度,减少了训练时间和测试时间。试验结果表明,该分类器的精度高,对保证产品性能一致性具有良好的应用价值。4)针对CVT速比控制中的波动问题,将广义预测控制理论应用于CVT速比控制,根据实际应用情况,还引入人工智能技术,对常规广义预测控制器进行了改造,开发了专用的CVT速比智能预测控制器。该控制器利用支持向量机建立非线性模型作为预测模型,通过混沌寻优算法完成在线最佳控制量的计算。台架试验结果表明,所设计的智能速比预测控制器具有良好的实时性,能有效抑制速比跟踪中的波动问题,具有更好的稳态控制品质;在阶跃响应过程中,可有效降低超调量和过渡时间,具有更好的动态品质。5)不增加硬件成本和大计算量,对CVT控制系统夹紧力的在线故障诊断进行了探讨,提出了基于模型的故障诊断方法。基本策略:把系统全局的和局部的预测模型的输出与实际输出比较,根据残差就得出相应的诊断结果。通过人为制造的故障进行测试:结果表明所建立的故障自诊断系统具有较高的精度和反应速度,对CVT的备用安全是非常有效的。更多还原

【Abstract】 The metal pushing V-belt type continuously variable transmission (CVT) has the advantages of simple and compact structure, excellent performance and easy operation, which is an ideal transmission form for economical automobiles. Developing continuously variable transmission becomes a common standpoint in automatic transmission industry domestically. The electro-hydraulic control system is an important part for the whole continuously variable transmission, which is composed of TCU, hydraulic control system and mechanism. The key technologies of the electro-hydraulic control system include the following sections: 1, the design and test method for the hydraulic control system; 2, the test and development technology for proportional electromagnetic valve; 3, the design technology for speed ratio controller; 4, the technology of improving quality stability and coherence. This paper mainly focuses on the above-mentioned key technologies research, and acquires a few of achievement and innovation as follows.1) The whole hydraulic control system is divided into three sub-systems in terms of their unattached pressure levels, which are speed shifting mechanism control subsystem, clutch and torque converter control subsystem, and lubrication and cooling flux distribution subsystem. A complicated system is de-escalated, which is integrated based on the principle of function matching, and patterns of the hydraulic control module are manufactured. It is shown from the table-board test results that the primary performance indices such as steady pressure, dynamic pressure and transient flux of the hydraulic control module are adequate for the requirements of vehicle normally running. It is shown from the vehicle road test results that the hydraulic control module is adequate for the control requirements of clamping force, clutch and speed ratio. The influence of flux distribution result from the lubrication valve flux-pressure down characteristic is considered when designing the lubrication and cooling flux distribution subsystem. The lubrication and cooling flux distribution subsystem is optimized by means of reasonable optimal algorithm and it is shown from the test results that the optimized lubrication and cooling flux distribution subsystem could satisfy lubrication and cooling requirement of the whole CVT and local components under random working conditions. 2) The design method for CVT appropriative proportional valves is systemically studied. The structure and basic working principle of the valve are analyzed firstly, and then a dynamic model is set up by means of mechanism analysis. For the purpose of checking the validity of the modeling method, prototype workpieces of the valve is manufactured for comparison test, and its simulation result follows the experimental result quite well. An associated performance index is founded considering the response time, overshoot and saving energy, and several sensitive structural parameters are selected to adjust for deriving the optimal associated performance index. The optimization problem is solved by the genetic algorithm (GA) with necessary constraints. Finally, the properties of the optimized valves are compared with those of the prototype workpieces, and the results prove that the dynamic performance indexes of the optimized valves are much better than those of the prototype workpieces, which are more suitable for the automatic transmission.3) A new hybrid scheme of rough sets and support vector machine for quality classification is proposed based on the performance testing characteristics of the CVT clamping force control valve. The proposed quality classification method reduces the redundant feature parameters based on the rough sets theory. And then, based on the reduction result, the least squares support vector machine (LS-SVM) classifier is used to set up the quality classifier. The experiments show that the accordance rate of the proposed method is very accurate, which has great project application value in engineering.4) Aiming at the problem of the CVT speed ratio tracking fluctuation, the concept of generalized predictive control is introduced to the CVT speed ratio control. The artificial intelligent technology is introduced to design an intelligent predictive controller according to the requirements of engineering practices. The proposed controller utilizes the support vector machine (SVM) predictive model to forecast the future control effect; the introduction of this SVM predictive model improves the precision of forecast for its nonlinear behaviors. The optimal control is determined by using a simple chaos optimal algorithm (COA); the introduction of the chaos optimal algorithm improves the optimal control searching efficiency effectively and solves the problem of real-time in control process. It is shown from the bench test that the proposed control algorithm accomplishes the better steady and dynamic performances than those of general PID: the fluctuation reduces in static tracking; the overshoot and the transition time decrease obviously in step-response. 5) At last, fault diagnosis of the clamping force control system is explored with no additional hardware cost and no additional computation. The diagnosis strategy based on model is proposed, which realized fault diagnosis in terms of the errors of real-time measured signal and the outputs of the global and local diagnosis model. The experiment results indicated that the proposed self-diagnosis system of the clamping control system had higher precision and response speed, which is very important for CVT security protection initiatively.更多还原