Robust model selection decision-making using a fuzzy supervisory Approach

Industrial machinery/assets are usually operated under different operating conditions or modes. Local empirical models can be built within specified operating condition boundaries to represent system dynamics more accurately than a global model, which is generally applicable over the entire operating regime. However, when the change of system operating regime occurs, none of local models can capture the characteristics of the system outside of the operating condition boundaries they were built upon. This paper presents a novel approach to model selection decision-making based on a fuzzy supervisory approach. The supervisory method selects local models and fuses these models to represent system dynamics as system transits from one operating regime to another. Through this fuzzy supervisory approach, the modeling errors caused by an operating regime switch can be significantly reduced. We present experimental results from the application of this approach to high bypass commercial aircraft engine.