On The Modeling & Advanced Control Of Pulp Refining Process

Pulp refining process plays an important role in pulp & paper plant. Automatic control for the process is necessary to get refined-pulp with consistent quality. Based on modern pulp-refining theory and advanced control technologies, this dissertation focuses on the solution to problems in modeling and control of pulp refining process in a large-scale pulp & paper plant. Related topics include first principle based and genetic programming (GP) based modeling of refining process, predictive control, predictive inferential control, etc.Main contents in the dissertation include:1 .Based on the analyses of current status in the research of modeling and control of pulp refining process, some opening problems for developing its computer control system are pointed out.2.The modeling of high-consistence pulp refining process is studied. A dynamic mathematical model for high-consistence pulp refining process, with flow rate of pulp, flow rate of white water, power of disc refiner as the model inputs, and specific energy, specific edge load, concentrator level as the model outputs, is presented based on specific edge load (SEL) theory. Also a refined-pulp quality predictive model, with specific energy, specific load as the model inputs, and beating degree, wet weight as the model outputs, is presented. Simulation results show that the models can be acted as theoretical bases for the automatic control of high-consistence pulp refining process.3.GP algorithm was investigated. Simulation results show that GP with fitness stimulating mechanism could greatly improve evaluation efficiency compared withconventional GP. A GP based dynamic model of low-consistence pulp refining process is presented.4.The predictive inferential control of SISO and MIMO cascade systems is studied. An algorithm is developed, which verifies the estimations for unmeasurable primary outputs with errors between real-measured values and predicted values of secondary variables. Application results show that the algorithm is effective for refined-pulp quality control.5.To solve the problem in control of concentrator level, predictive control for integrating plant is studied. A dynamic matrix control (DMC) algorithm and a feedforward DMC algorithm for integrating plant are presented.6.The key technologies for the implementation of pulp refining process control system are studied. Since high reliability and friendly human-computer interface are very important hi an industrial computer control system, the software and hardware in our system are constructed consummately. With the soft sensor we developed, stable control of high-consistence refined-pulp quality is achieved by adjusting concentrator level, pulp consistence and disc refiner power properly. To insure safety in the refining process, sequential control system and interlock protection system are designed and realized.Finally, some problems to be solved are pointed out based on the summarization of the works done in the dissertation.