Search

This paper reports on the development of a first principle model of a plastic film manufacturing process, which is implemented in the Matlab/Simulink dynamic simulation environment. The real-life cross directional controller that is incorporated into the model is also described with a few performance results. Development of the model and implementation of the controller aim at developing models for performance monitoring and diagnosis of the film manufacturing process.

In a steel plant, reheating furnaces are used for heating the steel slabs to a temperature of approximately 1200degC before rolling. Reheating furnaces consumes a lot of energy in steel plants. For competitive advantage it is important to improve the heating quality of slab and reduce the energy consumption as much as possible. This paper explores the potential of nonlinear model predictive control techniques to improve the temperature control of the metal slabs in a hot mill reheat furnace, and particularly whether or not these control techniques can be exploited to reduce the energy consumption. An example of actual furnace operation is presented to show the effectiveness of the proposed scheme.

This paper investigates the application of multivariable PID controllers to a wastewater treatment process. Four multivariable PID control schemes are investigated. The methods considered are those suggested by Davison, Penttinen-Koivo, Maciejowski and a new method proposed in this paper. All of the methods are suitable for multi-input multi-output (MIMO) control loops that experience loop interaction. The methods, furthermore, requires only simplistic plant models. The performance of the methods is assessed using a nonlinear benchmark model, and the optimal tuning values are determined using an optimisation method. The simulation results show the significance of the study and lead to the conclusion that the proposed method yields somewhat better results than other methods with respect to decoupling capabilities and closed-loop performance.

This paper explores the use of a multivariate statistics approach to monitor and diagnose the performance of a steam heated drying process. The results presented in this paper are a first stage investigation into the capabilities and limitations of these methods when applied to this particular process. The results show that, despite the non-linear behaviour of the drier process, the model predictions correspond well with the process measurements. The results also indicate ability to detect process failures, however, accurate diagnosis of the failures remains a subject for future studies

A toy helicopter makes for an impressive classroom control demonstration. However the challenge to obtain a robust, stable, multivariable controller is nontrivial given the nonlinear nature of the helicopter, the stochastic nature of the disturbances, and the performance limitations of PCs. This paper applies a model predictive controller using the Mathworks xPC target to obtain significantly improved results over that achievable using classical control algorithms.

The online measurement of basis weight in paper production is hampered by excessive noise, and the zig-zagging motion of the scanner transversing the sheet. Decoupling variations in the direction of production form variations across the sheet and visa-versa. An algorithm is presented that attempts to reconstruct this two-dimensional data from the single scanner.