A vast amount of production data are being collected in modern production processes. Although, the historians offer representative production data, there is still little or no idea how to efficiently exploit the data. Article discusses the problem of finding an appropriate empirical model of production performance, which could be employed to adjust manipulative variables to enhance production performance. The main steps of production performance modelling are described. Special attention is given to neural network identification and representation of the developed assisting tool, which ease the identification of a production performance model. In the article, a simulation of a complex production process is used as a case study. Identification of the production performance models and their practical application are illustrated on a case study, where models for three main production performance indicators (i.e. costs, quality and production rate) are identified.
COBISS.SI-ID: 26263079
This article addresses some features of process manufacturing that have to be taken into account during the design of a production control system in the process industries. The description of a model of a case study polymerization production plant is presented. Based on this model, a control structure framework is proposed, which makes it possible to automate part of the manager’s work. In this study, the model-based controller is introduced to the control structure.
COBISS.SI-ID: 22934823
A mathematical model was designed for an industrial, semibatch polymerization reactor, which describes the chemical reactions and heat balances in polymerization process. The model predicts the course of temperature in the reactor as a function of adding reagents, and key output parameters of the final product, such as conversion, solids content and viscosity. The main contributions are the integration of the two models the chemical reaction and the energy balance model, the validation of the model on realplant data from industrial operation, and the analysis and design of control algorithms for online dosing of reacting chemicals, which preserve reactor temperature close to the desired setpoint and so contribute to uniform product quality. The designed reactants dosing control represents an original solution for polymerization reactors, where reactor cooling is performed only through evaporative cooling. The desired control performance was proved by simulation based on realplant data and also experimentally on an industrial polymerization reactor.
COBISS.SI-ID: 24978727
Data gathered in the production process represent a useful source of information. But in production plants these data are still rarely exploited as the production operators often do not recognise their practical value. This article demonstrates how to effectively employ production data in order to identify the production variables with the strongest influence on the quality of the final product. Moreover, it shows how databased models of the production process can be used for advanced control techniques. The practical use of the data analysis is shown for a case study of the company Kolektor KFH.
COBISS.SI-ID: 12548379
The use of nonparametric Gaussian process model for fault detection is proposed. Detection relies on statistical tests applied to model prediction error. A novelty is model validity index which assesses whether the model is being used in the region in which it was learned and validated. In case this is not true caution is needed prior to releasing an alarm. Hence false alarm rate is reduced.
COBISS.SI-ID: 25017127