复杂行驶工况下智能汽车稳定系统复合故障诊断方法研究

The complex fault diagnosis problems of intelligent vehicle stability system with signal transmission delay, input constraints and uncertain nonlinear interference are investigated in this project. First, in view of the possible disturbance of the vehicle stability system under complex driving conditions, a nonlinear dynamic mathematical model with higher accuracy and wider adaptability is established by focusing on factors such as side wind and friction. Then, a more accurate mathematical model of fault diagnosis is obtained for the complex faults of actuators and sensors in the actual system, the complex faults are separated and identified using signal processing technology. Finally, the compound fault diagnosis control strategies of actuators and sensors are designed based on sliding mode control and artificial intelligence method, give full play to the robustness of sliding mode technology and well adaptive performance of intelligent control, the fault estimation of actuators and sensors in automotive stability system are realized, and make the system can still normally work in the event of a failure, the feasible intervention measures are determined in advance to restore the vehicle to a stable driving state and realize the safe, accurate and stable control of the vehicle. The research work to be carried out based on the problems existing in the practical application of vehicles, it is a helpful attempt and exploration for the large-scale application of fault diagnosis technology in the intelligent vehicle stability control system.

本项目针对一类含有信号传输时延、输入受限、不确定非线性干扰的智能汽车稳定系统的复合故障诊断问题展开研究。首先，针对复杂行驶工况下汽车稳定系统可能出现的扰动情况，重点考虑侧向风和摩擦力等因素，建立准确率更高、适应性更广的非线性动力学数学模型。其次，针对实际系统中执行器和传感器发生复合故障问题，探求利用信号处理技术对复合故障进行有效的识别、定位和分类，建立更准确的系统故障诊断数学模型。最后，基于滑模控制和人工智能方法设计复合故障诊断控制策略，充分发挥滑模技术的鲁棒性和智能控制的良好适应性能，实现汽车稳定系统执行器和传感器的故障估计，使系统在发生故障时仍能够正常工作，提前裁决出可行的干预措施使车辆恢复到稳定行驶状态，保证车辆的安全、准确、稳定的控制。本项目拟展开的研究工作，完全从汽车实际应用中存在的问题出发，为故障诊断技术在智能汽车稳定控制系统中的大范围应用做出有益的尝试和探索。

本项目主要以车辆电子稳定控制系统为研究对象，针对系统长时间高强度和高负荷工作导致的执行器和传感器故障问题，综合考虑系统中存在外部干扰、信号传输延时、不确定等因素，设计相应的智能鲁棒故障诊断策略。具体研究内容如下：首先将车辆稳定系统中的执行器和传感器故障信息作为附加状态变量，建立与故障相关的增广数学模型，针对不确定、时延系统设计几种基于观测器的鲁棒诊断策略，然后考虑系统中存在时变的非线性不确定干扰，设计基于等效输入扰动的滑模观测器，对集总扰动进行实时估计，以提高故障辨识准确度，最后通过大量联合仿真实验验证了算法在不同复杂工况下的有效性，实现了汽车稳定系统执行器和传感器的故障估计，提高系统抗干扰能力，使系统在发生故障时仍能够正常工作，提前裁决出可行的干预措施使车辆恢复到稳定行驶状态，保证车辆的安全、准确、稳定的控制。本项目故障诊断系统为车辆安全运行提出一种有效方法，为提高车辆稳定系统的故障诊断提供进一步的理论和技术支持。