Knowledge Modeling Method Based On Rough Set Theory And Its Application In Methanol Synthesis

Rough set theory is a valid mathematical theory dealing with uncertainties, which provides a new tool for modeling complex process. This thesis discusses the systematic steps of the knowledge modeling method based on rough set theory, from raw data processing to the resulted model. For the step of decision table attribute reduction, an attribute reduction method based on MQPSO optimization is proposed. This method combines MQPSO algorithm with attribute reduction of rough set, the optimal reduction results can be obtained by the use of the algorithm, and the effectiveness of the method is verifed by experiment data.Methanol is a very common chemical raw material. In recent years, its new usages have been well developed, and the demand is rapidly increased. The methanol synthesis unit in a chemical plant is considered in this thesis. The knowledge modeling for the industrial process is studied according to the production data. On the basis of its investigation, the knowledge modeling method based on rough set theory is applied to model the methanol synthesis process.Two models are established in the thesis:one is hot temperature model of synthesis reaction, the other one is the temperature model of boiler drum. The modeling steps and their implement flowcharts are given in details, including data discretization based on adaptive discretization algorithm, attribute reduction based on MQPSO, attribute value reduction based on and rule reduction based on RRSEA algorithm. Both models are tested with practical production data, and the results showed that the model can meet the requirements of accuracy.An on-line operation optimization software is designed according to the practical condition of methanol synthesis section of the plant, and its real-time communication with the field devices is realized by the use of OPC technology. The developed software consists of the function modules of real-time data acquisition, database access, modeling and optimization etc., and its use can provide information technology services to industrial production.