Abb Research Ltd.

Eduardo Alvarez Gallestey

Alec Stothert

Dario Castagnoli

Modelling of industrial processes is simplified with the use of Mixed Logical Dynamic (MLD) framework. Optimal control problems can be generated for application to industrial processes. For example, two arbitrarily connected MLD blocks are automatically merged to obtain one composite MLD block. Via a repeated use of the procedure, any arbitrarily complex system containing the complete description of an industrial process can be generated from the simplest MLD building blocks. The optimal control problem is generated via adding an MLD block whose unique output becomes the cost functional of the problem. In a graphical environment, any specific industrial process may be reproduced by instantiating blocks from a library of basic MLD elements or atomic MLD blocks and by properly connecting them. In case an appropriate library is available, this process will not require any expert knowledge from the end user apart from the ability to build the graphical interconnections mentioned.

Method of generating optimal control problems for industrial processes

The disclosed process control system makes use of Optimal control (OC) and model predictive control (MPC) techniques for selection of the Expert Systems (ES) targets values U. The ES target values U are selected to minimize the performance criterion J. A mathematical model of an extended system given by the process P and the ES is developed. This hybrid mathematical model has both continuous dynamics and logical relationships. Controlled variables of the mathematical model are the ES target values U and inputs are the measurements y and the performance criterion J. The OC and/or MPC techniques are used to compute values U. An optimizer of the OC/MPC selects values of the ES target values U only. This activity has lower sampling rates than selection of controller values, which simplifies the design of the OC/MPC controller.

Growing community of inventors

Tägerig, Switzerland

Eduardo Alvarez Gallestey