Process control structures: theory and practice-a review

The synthesis of good control is the primary objective of a control engineer. Whilst there are numerous control synthesis methods, the same cannot be said for choosing the system model and controller structure. The main reasons for the difficulty in structure selection are the multivariable nature of most process plants and the interactions among various plant variables. This necessitates looking at a number of issues while selecting a simple structure for control synthesis. These include: interaction analysis, decoupling, observability, controllability, input-output/loop pairing, robustness, implementation considerations and economy. Measurement, too, creates its own problems. Frequently there are too few variables for direct or indirect measurement. The question is which combination gives the optimum results in terms of good control performance and economics of implementation and operation. The paper examines the selection of system and controller structure for process control and presents avenues which need to be further explored