摘要
发动机故障诊断的一种主要方法是根据故障方程和发动机性能参数的测量值确定故障的类别和故障程度,故障方程组通常是亚定的。主状态量模型是求解亚定的故障方程的有效方法。主状态量模型的主要内容是根据最少故障原理,利用最优化方法求解然后根据合理性准则选择合理最优解。本文详细地讨论了主状态量模型可能采用的数学模型,并且指出:(1)主状态量模型可以采用不同的数学模型求解;(2)并非所有的数学模型都能给出合理的结果。本文的研究结果表明,约束或无约束超定最小二乘法和单个优选的散度法是发动机故障诊断主状态量模型的理想算法,而本文中所讨论的其他算法都或多或少存在一定的问题。
Fault diagnosis is a matter of resolving a limited set of measurements in terms of a large set of possible but improbable malfunctions.The relation between the measurements and the malfunctions is modeled in part as a set of linear equations △y=A△x. Engine fault equations are usually underdetermined and can be resolved by the MCPM(Main Condition Parameter Model),which is,to put it in a nutshell,to find the minimizing solution of the fault equations based on the Least Fault Principle and by the reasonableness criteria. Mathematical models which can be used for MCPM are discussed in the paper and the fol- lowing viewpoints have been obtained:1)Different mathematical models can be used for MCPM,but the information provided by the mathematical model is limited by the physical model of MCPM.Different mathematical models,so long as they conform to the physical model of MCPM and are correct in mathematics,would give about the same results.In other words,different mathematical models are only different approaches to resolving MCPM and are not different new methods of fault diagnosis which can give new fault diagnosis informa- tion,and 2)Not all the mathematical models can give reasonable results.In fact,many mathematical models would not conform to the physical model of MCPM and some of them would not conform to mathematical principles.It is shown that the constrained or unconstrained superdetermined least square method and the dispersity method based on indi- vidual minimization are ideal algorithm for engine fault diagnosis,and all of the other meth- ods discussed in the paper are somewhat defective.
出处
《中国民航学院学报》
1991年第3期1-9,共9页
Journal of Civil Aviation University of China
关键词
航空发动机
故障诊断
数学模型
jet engine
fault diagnosis
main condition parameter model minimization
