基于内燃机振动信号时-频调谐特性的燃烧信息补偿及定量评价研究

Non-combustion excitation response signal is the bottleneck that restricts the extraction of combustion parameters based on vibration signals. Partial combustion information is eliminated when the combustion excitation response signal is removed, which further affects the quantitative evaluation of combustion parameters. To solve these problems, the dynamic migration rule of excitation response signal frequency band is firstly revealed. The level of different non-combustion incentives is then determined by the sensitivity analysis method. Secondly, the distortion characteristics of the reciprocating inertial force excitation response signal under the influence of the main non-combustion excitation are explored. And the influence of the separation of reciprocating inertial force excitation response on the combustion excitation response integrity is studied. A strategy based on "contribution degree" is proposed to compensate combustion excitation response signals. Thirdly, the relationship between the characteristics of vibration signals and the lag angle is studied. A multi-level description method is further proposed to describe the lag angle. And the recognition accuracy of phase combustion parameters is improved based on the "average and discrete distribution" method. Fourthly, the association expression between the vibration velocity signal and the pressure rise rate signal in the compression stroke is established based on the information entropy theory. And then the maximum pressure rise rate and peak pressure is obtained. Lastly, the online identification process of combustion parameters is developed. And the error range and influence factors of the identified combustion parameters are analyzed. Furthermore, the accuracy and efficiency are verified through the engine bench tests. The research results of this project provide academic reference for the evaluation of combustion status, and provide new idea and method for on-line control of new combustion technology.

非燃烧激励响应信号是制约基于振动信号提取燃烧特征参数的瓶颈，针对剔除非燃烧激励响应信号时部分燃烧信息损失并导致燃烧特征参数难以定量评价的基础问题，探索激励响应信号频带动态迁移规律，基于敏感性分析确定非燃烧激励影响级别；分析主要非燃烧激励影响下往复惯性力激励响应信号畸变特性，研究往复惯性力激励响应分离对燃烧激励响应完整性影响，提出基于“贡献度”补偿燃烧激励响应信号的策略；挖掘振动响应信号特征与相位滞后角关联性，提出多层次特征融合的相位滞后角描述方法，基于“平均及离散分布”提高相位燃烧特征参数识别精度；分析压缩冲程中振动速度信号和压力升高率信号对应关系，基于信息熵理论构建两者关联性表达式，探索获取最大压力升高率及峰值压力的方法；制定燃烧特征参数在线辨识流程，分析参数辨识误差范围及影响因素，并开展台架试验验证。本课题的研究为燃烧状态评价提供理论参考，为新型燃烧技术在线控制提供新的思路与方法。

项目围绕内燃机机体表面振动信号特性和基于振动信号提取燃烧信息进行研究，建立了单缸柴油机一维仿真模型和二缸柴油机缸盖-机体有限元模型，分析往复惯性力、活塞换向撞击及曲轴主轴颈负荷等非燃烧激励振动响应信号特性及其时、频耦合关系，确定对振动信号起主要作用的非燃烧激励；利用偏相干函数分析往复惯性力对振动响应信号贡献度的变化规律，研究缸内压力信号和实测振动信号中可用于表征高频谐波能量的特征参数；利用主成分分析法提取燃烧激励响应信号，分析不同工况下缸内压力对振动信号的贡献度，讨论测点位置和测点数目对贡献度的影响；探讨了贡献度与燃烧始点相位偏差间关系的变化规律，基于此规律修正了燃烧特征参数辨识误差。