No abstract

The intelligent approaches emerge as leading techniques in providing of stable and high performance control of industrial plants with nonlinearity, model uncertainty, variables coupling and disturbances. In the present research a novel approach for the design of a nonlinear model-free fuzzy logic controller (FLC) with two inputs – the system error and the main measurable disturbance and a rule base for disturbance compensation is suggested. It is based on off-line parameter optimisation via genetic algorithms. The approach is applied for the development of a FLC for the control of the level of ammonia brine solution in a carbonisation column with compensation of the changes in the inflow pressure. The control algorithm is implemented in a general purpose industrial programmable logic controller in ”Solvay Sodi” SA – Devnya, Bulgaria. The FLC system with disturbance compensation outperforms in an increased dynamic accuracy the FLC with the system error as a single input even when linear feedforward disturbance compensation is added. The performance of all systems is assessed from the real time control and the simulations based on a derived TSK plant model.