【作者】 赵鑫；

【导师】 王伟杰；

【作者基本信息】 哈尔滨工业大学 ， 机械工程（专业学位）， 2018， 硕士

【摘要】 汽车发动机的气密性直接影响汽车的动力性能。发动机缸盖作为汽车发动机最重要的部件之一,其气密性检测是生产过程中的关键环节。目前,国内外对气密性检测的研究主要集中在三个层面,即检测方法研究、检测仪器研究以及检测装置研究。本文以HWS15型发动机缸盖作为研究对象,开发了一套具备自动化及智能化特点的气密性检测装置。首先,在充分了解国内外研究现状的基础上,本文对目前应用最为广泛的差压法和质量流量法进行建模研究和对比,选用后者作为被检缸盖的气密性检测方法,并选择MFL400作为气密性检测仪器。在此基础上,利用Simulink仿真的方法对检测过程中的关键参数进行了最优化设计。然后,为了使设备满足发动机缸盖气密性检测的精度要求,本文对检测过程中的缸盖孔道的封堵进行了研究。在建立孔道封堵泄漏模型的基础上,本文提出了一种封堵力设计和执行元件选用的方法,可以从提高封堵质量的角度提升气密性检测的整体精度。最后,根据实际检测要求设计了一套具备自动化及智能化特性的气密性检测设备。在提出检测装置总体方案的前提下,主要针对检测装置硬件系统部分进行了设备机械系统、气动系统、液压系统的设计以及基于PLC的控制系统方案设计。智能软件系统主要应用统计过程控制SPC技术实现了对检测过程稳定状态的质量监控,由于SPC控制图无法给出具体失控形式,本文建立了BP神经网络模型并利用由Monte Carlo法生成的数据样本进行训练,使该模型对控制图正常、趋势、阶跃和周期4种模式的有效识别率达到85%以上,在上述研究基础上开发的气密性检测装置可以实现对气密性检测过程的智能监控。更多还原

【Abstract】 The air tightness of an automotive engine cylinder is directly correlated with the vehicles dynamic performance.The automotive industry considers the cylinder as one of the chief components of the engine along with its air tightness detection procedure as a key step in the manufacturing process.At present,the research on leak detection including other relevant fields mainly focus on three levels which include detection method research,detection instrument research,and detection device research.Considering the HWS15 engine cylinder as the research subject,this research thesis develops a set of leak detection devices characterized with automation and a neural networking software system.Initially,the study examines and compares the most popularly used differential pressure method and mass flow method with the aim of primarily understanding the current research trend.After evaluation,the study selects the latter as the leak detection method for the cylinder cover,and chooses MFL400 as the detection instrument.Then by using Simulink,the simulation method is used to optimize the key parameters in the detection process.The second stage of the research involved designing a method to make the equipment meet the accuracy requirements of the cylinder leak test and to study the plugging of the cylinder during the detection process.Based on previous models a sealing model with modifications is developed.Finally,a design method for sealing force and the selection of actuators,which can improve the overall accuracy of leak detection from the perspective of improving the sealing quality is proposed.The next stage consisted of designing a set of automatic leak detection devices.Based on the previously obtained characterization and design results,the designing process of the hardware system,including the device fixture,pneumatic system,hydraulic system and the control system based on PLC is put forth.Finally,in the last stage the software system coupled with an intelligent artificial neural networking is described.The software primarily uses the SPC to monitor the quality and the stability of the process.However,the control chart could not generate a concrete failure form,so the BP neural network model is proposed to recognize the 4 modes of SPC data of the leak detection process.更多还原