Expert system, fuzzy logic, and CMPC control of nonlinear multivariable industrial processes

A practical advanced control scheme has been developed to control nonlinear multivariable industrial processes. An Expert System, Fuzzy Logic, and Constrained Model Predictive Control (CMPC) strategies have been utilized in developing the advanced control scheme. The developed control scheme performance has been demonstrated and tested in real time on a simulated cement mill process.; The cement production industry suffers from decades of wasting energy by using low efficiency processes and poor control strategies. The most common control strategy used in cement production industries today is supervisory control, which is based on using single input/single output traditional PID controllers. Supervisory control strategy depends mainly on operator experience in handling different operation scenarios with the help of traditional PID controllers.; A cement production plant consists of two major processes. In the first process, a kiln is used to cook the raw material and produce clinker. In the second process, a cement mill circuit is used to grind the clinker to the right grade. This work focuses on the control of the cement mill process. The mill process is a multivariable nonlinear open loop unstable process. The process dynamic characteristics continuously change due to the changes in the fresh feed hardness, and the need to operate at different operating conditions to produce different cement grades. All these factors together, make the design of an efficient control scheme for the cement mill process a challenging problem. The currently used supervisory control strategy is very often unable to keep the process stable, the product quality on target, and ensure smooth and quick grade transition.; In this dissertation, a cement mill process simulator is developed and used to demonstrate the performance of the developed control scheme in real time fashion. The proposed control scheme is designed based on industrially available process control and artificial intelligence tools. The developed control scheme consists of fuzzy logic, expert system, and CMPC.; The fuzzy logic is used for online tuning of PI controllers to compensate for the effect of the process nonlinearity and to linearize the overall process. The expert system is used to infer the unmeasured hardness of the fresh feed and guide the process to the optimum operating conditions. The CMPC control strategy is implemented to control and optimize the overall linearized process. The developed controller components (i.e. fuzzy logic, expert system, and CMPC) showed great capabilities to achieve their objectives.; The developed controller demonstrated excellent ability to control and stabilize the cement mill process, at different operating conditions and scenarios including grade transition and disturbance rejection.