基于模糊方向规则的火箭发动机传感器故障检测与分离

提出了一种直接从训练样本中获取模糊方向规则的学习算法，并应用于火箭发动机的传感器故障检测与分离。每种传感器故障模式由一些模糊方向规则聚集形成，模糊方向规则的全隶属区是一个由单位方向、夹角和两个半径确定的方向超体。模糊方向规则一次循环学习形成，在学习中能不断融合新样本信息。液体火箭发动机传感器故障检测与分离的仿真研究验证了模糊方向规则系统的优越性能。

SENSOR FAULT DETECTION AND ISOLATION OF ROCKET ENGINE BASED ON FUZZY DIRECTION RULES

n this paper we discuss a new method for extracting fuzzy direction rules directly from training input-output data for sensor fault detection and isolation in a liquidpropellant rocket engine. Each sensor fault class is an aggregate of fuzzy direction rules The full membership region of a fuzzy direction rule is a direction hyperbody that is described by a direction vector an included angle and two radii The fuzzy direction rule system can learn nonlinear sensor fault boundares in a single pass through the training data and provide the ability to refine existing fuzzy direction rules in the whole learning process The sensor fault detection and isolation simulation of liquid rocket engine demonstrates the strong qualities of the fuzzy direction rule system.