提高数控系统可靠性的解析余度研究

Research on Analytical Redundancy to Improve the Reliability of Digital Controlling System

为提高航空发动机数字控制系统的可靠性，本文用解析余度技术，对系统传感器故障检测、隔离、调整的方法进行仿真研究。仿真系统中的发动机数学模型是用一系列小偏差线化模型去逼进某一实际发动机的试车数据而得到的。这里建立的传感器故障模型模拟各种故障的发生，设计分离的卡尔曼滤波器产生的残差秩序构造一决策函数，对故障进行检测及隔离调整。本文在ＭＡＴＬＡＢ语言环境，对可能发生的故障进行了仿真研究，所得结果说明，这种算法是可行的。

Computer simulation of sensor failure detection, isolation and accommodation is studied by an analytical redundancy technique to improve the reliability of digital control system of aero-engine. The engine mathematical model in the simulation system is obtained by approximation of testing data of an actual engine with a series of small deviation and linearized models. The established model of sensor failure can simulate various failures. The sequence of residual differences generated by the designed separating Kalman filter constructs a decision function. It detects, isolates and accommodates the failure. Simulation research on probable ocurrance failures is conducted with MATLAB Language. The obtained results show that this algorithm is feasible practically.