自走式采棉机HMCVT负载自适应速比匹配传动鲁棒性及控制策略研究

Cotton production as a backbone industry of agriculture and economic of Xinjiang, cotton harvest mechanization has become a weak link in Xinjiang. Due to many problems of transmission system in homebred cotton picker,the process of cotton harvest mechanization in Xinjiang has been hindered in a long period.Although transmission characteristics of HMCVT suits the operating mode of cotton picker well,HMCVT hasn’t been applied in cotton picker dues to the technical barriers and the relative lag in the research of HMCVT. Thus, to achieve the point control of HMCVT load adaptively in a large range ,the study construct a mathematical model matching cotton picker power plant, a control strategy also has been builded to achieve self-tuning of PID parameters based on BP neural network .In reagrd to robustness of load control ,speed ratio controller will be activated that engine speed and the rotational speed rate of change beyond a certain range to ensure self-tuning of small scale fluctuation of speed ratio matching,to relize the reasonable matching control of the best fuel economy and power performance between cotton picker and engine.In the study, The optimum composing of key influences factors of HMCVT shift quality will be determined .The research team not only has some research basis,but also has thorough experimental conditions.The conclusion of project results has great significance for cotton harvest mechanization.

棉花生产作为新疆农业及经济发展的支柱性产业，实现棉花机械化收获已成为我区亟待解决的难题。国产采棉机由于存在无法自适应负载和阻力变化、换挡冲击大及速比匹配不合理等问题，严重影响着我区棉花生产全程机械化的进程。由于HMCVT无级变速换档平稳、能够实现传动比大范围连续无级变化，因此本研究根据采棉机不同作业工况，通过建立采棉机多动力需求匹配数学模型、优化液压机械无级变速传动方案、确定变速箱各区段传动比速比，完成采棉机HMCVT结构和控制系统的设计，并基于BP神经网络PID参数自整定算法，通过控制器调节变速箱速比，实现变速箱负载自适应点位控制；同时，研究负载自适应控制的鲁棒性，避免小范围的负载随机波动对变速箱速比的频繁动作，进行速比的自整定，实现采棉机发动机的最佳燃油经济性或动力性匹配；通过仿真模拟及试验研究确定影响HMCVT换段品质的关键因素及最优组合。该研究对新疆棉花种植全程机械化具有重大意义。

国产采棉机虽已具备一定的研发、制造能力，但采棉机的变速箱、控制器等关键部件仍依赖进口，并已形成采棉机技术壁垒，国产采棉机产业化发展迫在眉睫。本项目研究通过采棉机液压机械耦合传动的行星齿轮排汇流结构参数优化，及行星齿轮排耦合传动特性构建了动力HMCVT模型，完成了样机的试制、相关台架实验及传动检测电子控制系统的设计，通过对传动系统液压元件的匹配、优化及仿真，完成了传动电子控制系统的设计，构建采棉机传动系统控制模型，并优化了BP 神经网络PID 算法，通过人工神经网络的自学习能力实现PID 参数的在线自整定，实现自适应的负载自整定鲁棒性控制，并基于BP神经网络的PID参数自整定控制策略对变速箱控制算法进行了应用测试，优化了负载自适应条件下的采棉机变速箱控制策略。校验了发动机与变速器速比匹配多元控制逻辑鲁棒性控制算法。通过构建HMCVT 换段过程多体动力学模型及HMCVT 换段品质的工况因素、设计因素、可控因素的正交多因素实验，确定了控制因素HMCVT影响传动性能的影响规律，掌握了HMCVT输出的载荷波动、冲击度特性，采棉机操作的便利性及经济性提高，研究成果可强化对国产采棉机科技创新，对我国自主知识产权的采棉机应用开发具有重要的理论指导和现实意义。