Abstract: A method for monitoring of and fault detection in an industrial process, comprising at least a first sub-process and at least one second sub-process arranged in a process chain, comprising, for the at least one second sub-process the steps of collecting data and calculating a multivariate sub-model based on the collected data, said method being characterized by the steps of receiving in the first sub-process from the at least second sub-process information related to the multivariate sub-model calculated for the at least second sub-process, collecting data related to the first sub-process, and calculating a multivariate sub-model for the first sub-process based on collected data and received information.

Abstract: The present invention provides a method and system for controlling the quality of a product produced by an injection molding production process. The invention includes performing a multivariate analysis on injection molding process data collected real-time and determining whether the real-time data is within a predetermined production control limit. When the real-time production data exceeds the control limit, the process is considered out of control and product produced during the out of control condition is removed real-time from the injection molding production process.

Abstract: The invention provides a method and an arrangement for filtering or pre-processing most any type of multivariate data exemplified by NIR or NMR spectra measured on samples in order to remove systematic noise such as base-line variation and multiplicative scatter effects. This is accomplished by differentiating the spectra to first or second derivatives, by Multiplicative Signal Correction (MSC), or by similar filtering methods. The pre-processing may, however, also remove information from the spectra, as well as other multiple measurement arrays, regarding (Y) (the response variables). Provided is a variant of PLS that can be used to achieve a signal correction that is as close to orthogonal as possible to a given (y) vector or (Y) matrix. Hence, ensuring that the signal correction removes as little information as possible regarding (Y). A filter according to the present invention is named Orthogonal Partial Least Squares (OPLS).