Relative Gain Array and Singular Value Analysis to Improve the Control in a Biomass Pyrolysis Process

Abstract

The pairing of actuators and controlled variables for process control is non-trivial due to couplings among process quantities and further disturbing effects. Here, we consider the product gas stage in a biomass pyrolysis process. Based on a dynamic model of the latter, several analyses, namely Relative Gain Array (RGA), Dynamic Relative Gain Array (DRGA) and Singular Value Analysis (SVA) are performed to solve the pairing problem. In particular, we emphasize the practical benefit from DRGA's results by showing that challenges, like non-measurable disturbances or time delays, can be mitigated by simple changes in an implemented control concept.

Fuzzy Automatic Control of the Pyrolysis Process for the Municipal Solid Waste of Variable Composition

“…Furthermore, sufficient attention has been paid to the development and research of mathematical and simulation models of various pyrolysis plants and their separate components (reactors, circulation systems, heat exchangers, etc.) [25][26][27][28], including the use of fuzzy logic, soft computing and other intelligent techniques [29,30]. Also, currently, a sufficient number of examples of the successful application of intelligent and other advanced algorithms and devices for the implementation of effective control of certain technological variables of waste utilizing complexes have been presented [31,32].…”

Control Loop Pairing and Interaction Analyses of the Falling Film Evaporator Process

“…Some advice on determining the relevant frequency domain of a plant is given in Sec. IV and in [8]. To practically evaluate R(iω), it is common to consider the magnitude of its elements |r ij (iω)| over ω.…”

“…The latter approach is referenced as dynamic relative gain array (DRGA) throughout the present paper. Although the steady-state RGA and DRGA are often applied to various processes [8]- [10], Question 3) cannot be answered by these methods. To remedy this issue, there exist the notions of decentralized integral controllability (DIC) [2], [11] and the Niederlinski index (NI) [12].…”