Data Driven Analytics (Machine Learning) for System Characterization, Diagnostics and Control Optimization

Abstract

This presentation discusses the potential use of machine learning techniques to build data-driven models to characterize an engineering system for performance assessment, diagnostic analysis and control optimization. Focusing on the Gaussian Process modeling approach, engineering applications on constructing predictive models for energy consumption analysis and tool condition monitoring of a milling machine tool are presented. Furthermore, a cooperative control optimization approach for maximizing wind farm power production by combining Gaussian Process modeling with Bayesian Optimization is discussed.