Long‐term statistical stability of industrial plants: Performance indicators and monitoring of an industrial pet plant

Abstract

In the present work, usual statistical process control (SPC) definitions are reformulated to allow for long‐term analyses, giving rise to extended statistical process control (ESPC) procedures. In order to allow for implementation of the ESPC approach, t‐ and F‐control charts and monitoring indexes (NEPM, EPY and OEP) are designed in this work and are used to monitor the performance of a real industrial poly(ethylene terephthalate) (PET) site based on six process outputs (intrinsic viscosity, crystallinity, acetaldehyde concentration and colours a, b and L). As shown in this manuscript, ESPC tools allow for proper and systematic analysis of huge amounts of industrial data, using simple, fast and efficient statistical techniques, which can be used to improve the quality of the process operation in most industrial polymerisation plants.

Use of variance spectra for in‐line validation of process measurements in continuous processes

“…Considering the possible use of the spectrum of variances for surveillance purposes, the amount of variability of a particular signal is important for the definition of acceptable boundaries for the signal. This is a key point for the implementation of process monitoring schemes, product quality control, equipment health management, and measurement validation …”

“…This is a key point for the implementation of process monitoring schemes, product quality control, equipment health management, and measurement validation. [3,12,14] Figure 2 illustrates the small and large signal boundaries in a typical industrial process. These systems possess measured and unmeasured signals (disturbances) that can be diagnosed with the help of appropriate boundaries (window lengths).…”

“…One of the main reasons for this slow transition concerns the compliance to regulatory issues, as the appraisal of an individual result (that can be either approved or disapproved) is in a certain sense less complex than the analysis of continuous process data, which can provide contradictory information about the current product quality in a short period of time. In the polymerization industry, the applications are so vast and the products are so diversified that operation changes and grade transitions are very common at plant sites, making the proper implementation of quality control procedures a key issue for product acceptance in the market …”

“…In the polymerization industry, the applications are so vast and the products are so diversified that operation changes and grade transitions are very common at plant sites, making the proper implementation of quality control procedures a key issue for product acceptance in the market. [3] The previous paragraph illustrates that chemical industries are demanding improved reliability of process data and technical evolution of measuring systems. As an example, process analytical technology is evolving to new miniaturized, modular, and smart pieces of equipments.…”