The article studies the fault diagnosis of the first-stage power system of a carrier rocket made up of multiple engines. It establishes the simplified 6-DOF nonlinear model and uses the extended Kalman filter to generate residual errors. The quadratic receding time-domain algorithm is used to detect faults, and estimation vector fault characteristics are used to estimate precisely the thrust loss degree. Faults are located with the direction of abrupt change in overload fault. The simulation results show that the method can quickly detect engine faults, estimate the thrust loss degree with fault estimation values, and locate precisely the serial number of the faulty engine of the power system made up of multiple engines. The method boosts the development of the online health management technology of the power system of a carrier rocket.

研究了多发运载火箭一级动力系统的故障诊断。它建立了简化的 6 自由度非线性模型，并使用扩展卡尔曼滤波器生成残差。采用二次后退时域算法进行故障检测，利用估计向量故障特征精确估计推力损失程度。故障定位在过载故障突变方向。仿真结果表明，该方法能够快速检测发动机故障，利用故障估计值估计推力损失程度，准确定位多发动机组成的动力系统故障发动机的序列号。本发明促进了运载火箭动力系统在线健康管理技术的发展。