Aeroengine Performance Modeling Method Based on Self-tuning Wiener Model

Abstract

The invention discloses a self-adjusting Wiener model-based aero-engine performance model building method. The method comprises the steps of firstly generating training and testing data of a model; secondly transmitting an initial combination of characteristic parameters of the model and the training data to a nuclear extreme learning machine to obtain corresponding optimal model structure parameters, nuclear parameters and output weight, and building a self-adjusting Wiener model; and finally forming a turbofan engine overall performance model by a self-adjusting Wiener model cluster. The method has the advantages that a time constant value of a linear dynamic part can be self-adjusted by utilizing the learning machine, so that the dynamic performance estimation is more accurate; a block Wiener model has a nonlinear static module, and the static estimation precision is superior to that of a conventional machine learning method; and in addition, globally optimal model characteristic parameters and weighted vectors can be directly obtained in a training stage, and an optimization algorithm is not adopted for adjusting the model parameters, so that the calculation time can be remarkably shortened.

Description

technical field [0001] The invention relates to the field of aero-engine model parameter identification, in particular to an aero-engine performance model modeling method based on a self-adjusting Wiener model. Background technique [0002] The structure of the aero-engine is complex, and it has been working in harsh environments such as high temperature and high pressure for a long time, and it is a failure-prone system. The traditional regular maintenance method is not only resource-intensive, inefficient, but also expensive. Due to its obvious advantages such as small scale, high efficiency, good economic affordability, and avoidance of major catastrophic accidents, condition-based maintenance is very suitable for maintenance support of large and complex systems. One of the prerequisites for implementing condition-based maintenance is to require the system to have the ability to predict its own faults and manage its health status, and the concept of health management has...

AI Technical Summary

Problems solved by technology

In the traditional Wiener model, the linear dynamic part simplifies the time constant of the first-order inertia link, which affects the dynamic performance of the model, and the group optimization algorithm is inefficient in adjusting the structural parameters and output weights of the Wiener model. Therefore, it is necessary to design a a new modeling method

Images