基于新理念的缸盖表面振动信号表征内燃机燃烧过程特征参数的研究

The situation is that the vibration response signals excited by the in-cylinder pressure and the reciprocating inertia force are overlapped in time domain and in frequency domain and so are hard to be separated by the conventional signal processing effectively. This project has proposed the new concept on the research and analysis of the vibration signals of cylinder head according to the characteristics of ICE's combustion process and deduced the variable rule of the reciprocating inertia force versus the crank angle in internal combustion engines; calibrated the amplitude of the response signals excited by the reciprocating inertia force using the nature that the vibration response signals excited by in-cylinder pressure are coupled with those excited by reciprocating inertia force, accordingly, to remove the response signals excited by the reciprocating inertia force and to increase the signal-to-noise ratio of the response signal excited by the in-cylinder pressure; provided the method to extract the response signal excited by in-cylinder pressure within the range of the peak pressure using improved phase-shifting filter and blind source separation algorithm; studied the method to calibrate the amplitude parameters of the vibration signal by analyzing the relative relationship of the vibration response signal in the compression stroke versus the variation of the in-cylinder pressure so as to achieve the assessment for the combustion parameters related to the amplitude of the cylinder pressure; and analyzed the errors and the influencing factors of the in-cylinder pressure-related combustion characteristic parameters depending on the cylinder pressure information reflected by the surface vibration signals from the cylinder heads. The research in this project has set a theoretical basis for the assessment and the close-loop control to the in-cylinder combustion process via the surface vibration signals from the cylinder heads.

针对缸内压力和往复惯性力激励的振动响应信号在时域及频域都出现重叠，常规信号处理技术难以有效地对其进行分离的问题，提出了从内燃机工作过程角度提高缸压激励响应信号信噪比的新理念，在曲柄转角域研究内燃机往复惯性力的变化规律，利用缸压激励和往复惯性力激励响应信号的耦合特点，对往复惯性力激励响应信号幅值进行标定，去除往复惯性力激励响应信号；提出了采用改进移相滤波和盲源分离算法分段提取缸内峰值压力点前后的缸压激励响应信号的方法；通过对压缩冲程中振动响应信号与缸压变化规律的对应关系分析，研究对振动信号幅值参数进行标定的方法，实现对与缸压幅值相关的燃烧特征参数的评价；基于缸盖表面振动信号所能提供的缸压信息，对缸内燃烧过程特征参数的评价误差及其影响因素进行分析。课题的研究将为利用缸盖表面振动信号评判缸内燃烧过程或进行缸内工作过程闭环控制奠定理论基础。

针对缸盖表面振动信号的低频扰动与燃烧激励响应信号在时频域存在耦合，常规处理技术难以有效进行分离的问题，依据仿真分析和试验分析，揭示了往复惯性力、倾覆力矩、活塞侧压力产生的摩擦力等非燃烧激励源与燃烧激励源的耦合关系，针对缸盖表面振动速度信号和振动加速度信号分别提出了高信噪比燃烧激励响应信号的提取方法；仿真分析结果表明，缸盖表面振动速度信号中，非燃烧激励响应信号的激励源主要是往复惯性力和活塞侧压力产生的摩擦力，燃烧激励响应信号淹没在非燃烧激励响应信号中，后者存在于整个工作循环中，为此，提出了以往复惯性力激励为主体的非燃烧激励响应信号描述模型，并采用模式识别技术获取了模型有关参数，利用预测模型从实测振动速度信号中剔除非燃烧激励响应信号的影响，有效地提高了燃烧激励响应信号的信噪比；仿真分析指出，缸盖表面振动加速度信号中的非燃烧激励源主要为往复惯性力激励，其响应信号与缸压激励响应信号呈线性叠加，为此，基于经验模式分解技术，对实测振动加速度信号进行了分解，选取与燃烧激励相关性强的本征模函数，重构得到了燃烧激励响应信号；基于仿真分析，得到了激励信号频谱成分在响应信号中的反映，提出了燃烧激励响应信号滞后于激励信号相位角的描述方法；基于得到的高信噪比的燃烧激励振动速度、振动加速度响应信号与压力升高率、缸内压力二次导数的对应关系，提出了角度相关的缸内燃烧过程特征参数的评估方法；利用提取的燃烧激励振动速度响应信号与压升率信号在陡升段的线性对应关系，标定得到了最大压升率值；提取燃烧激励响应信号上燃烧始点至峰值压力点区域内的数据，采用积分后得到的峰值位移表征峰值压力，利用改进的平均位移算法得到的平均位移表征平均指示压力；讨论了基于振动信号的燃烧过程特征参数评估方法的误差范围及其影响因素。