The paper presents some approaches in modeling and control of unmeasurable or hardly measurable technological plants. Hybrid modeling is described using First Principles (FP) models, Fuzzy Logic - based (FL) models and Neural Network (NN) models. Different kinds of aggregations are examined - linear combination, Hammerstein-like models, Gain Scheduling models. The results of comparative analyses of the behaviour of different models are presented, which show higher accuracy of best hybrid models with 6-10 % according to Mean Square Error (MSE). Some problems of advanced Inference Control (IC) implementation using hybrid models, are discussed. Applications of hybrid modeling as well as Inference Control of industrial plants are described - steam boiler mill-fan and dust preparation system, hot strip mill runout table cooling system modeling and control.
All articles published by Studies in Informatics and Control (SIC) are published in accordance with the terms and conditions of the Creative Commons Attribution License (CC BY-NC-SA 4.0).