Collaborative project with Prof. Dr. Scheuring,
Institute for Automation & Industrial IT - Cologne University of Applied Sciences
The optimum setting for temperature and air humidity controllers in climatic chambers, nitriding furnaces and other devices represents a constant problem. Increasingly stricter demands are being placed on such control systems with regard to the standard and the quality of the control. These types of control are becoming increasingly difficult using conventional adjustment processes.
Climatic chambers, for example, require to be controlled by two parameters that act against each other, namely according to the temperature and the air humidity. It is a case of a classical multiparameter system. If one acts on the temperature control parameter, then one unintentionally has an adverse effect on the air humidity control parameter and vice versa.
On the other hand, nitriding furnace control systems require to be scanned according to specific set value graphs in relation to temperature, pressure and gas concentration. In this case, specific tolerances must be observed, that is the actual value of the control parameter may only deviate by a certain percentage from the set value as currently prescribed. In addition, there is a problem in that this must also be precisely observed for different loads, above all as the control section also changes in practice with the size and the composition of the load placed in the furnace and, accordingly, also the optimum control parameters.
In practice, traditional PID controllers have been used up until now, which are parametrised in a one-off operation when starting up operations, according to the well-known adjustment rules for specific types of furnace and load concentrations. Further adjustments are carried out manually, as a result of which the potential for optimisation that is available is not exploited. The number of possible application scenarios as well as peripheral conditions is, in practice, as high as one wants, which makes the development of a universally applicable solution difficult.
Progress has been made through research in the last few years, particularly in the control of climatic chambers, and successes notched up, above all in the reduction of the consumption of energy. In this connection, model-predictive control, for example, may be mentioned as one of these possible methodical approaches to solving the problem.
This approach was explored by the GECO►C team in collaboration with the company Stange Elektronik GmbH which is based in Gummersbach and which, inter alia, develops control systems for nitriding furnaces on behalf of its customers. The aim was to exploit the potential available for optimisation.
In the case of model-predictive control systems the ideal course of the regulating variable is calculated mathematically using an existing section model. Such systems are suitable, inter alia, for the control of multiparameter systems and systems with a conspicuous level of downtime. A high degree of quality on the part of the existing section model is of great significance for the trouble-free operation of this process.
As requested, not only was the current set value taken into account in this connection, but entire future set value trends were also included in the calculation, which might be of benefit in the case of the present problem.