The application of kernel density estimates to condition monitoring for process industries

Discusses the application of kernel extraction for estimating the non-parametric density function of a multivariate process system for condition monitoring purposes. In particular, the paper concentrates on a real industrial case study to demonstrate the differences and practical capability of three different estimators. It is shown that the kernel density estimate has the potential to be an important technique of obtaining real nonparametric empirical density function of the process population as an aid to more effective intelligent condition monitoring.