The Research And Design Of Intelligent Control System For Tobacco Humidity

Conditioning cylinder is main equipment in re-drying tobacco produce process of tobacco industry. It is chiefly applied to warm and humidify the tobacco that is separated into leaf and stem so that its moisture and temperature can be controlled at the range of technology demand. The research and design of intelligent control strategy and system for tobacco humidity have much theoretic meaning and practical value to conditioning processes and other moistening equipment.In this paper, based on predictive PI and the hormone modulation of neuroendocrine system, an ultrashort feedback intelligent controller which is suitable for the tobacco moisture adjustment process with large time delay is built. Opinions and discussions are presented in many aspects such as model analysis, data preprocessing, control algorithm design, hardware and software configuration, etc.First, according to the knowledge on hygroscopicity & dehydration properties of tobacco leaves and air conditioning, a model of export tobacco's variational moisture content is built, after the discussion on the moisture relations of leaves and air on the mechanism, and the analysis of the main factors on the impact of tobacco humidity. The above model is based on the equilibrium moisture. Another model of the distribution of the tobacco humidity in the ordering-cylinder is also built by linearizing the relation between humidity margin and hygroscopicity & dehydration rate. These models provide the basis for the following frame of control strategy.Secondly, the suitable real-time filtering algorithms, meeting the system requirements, are confirmed, after comparing the denoising effect of filtering method. Then the characteristics of system and parameters of object model are analysed through denoising the locale sampling data. Thirdly, according to the hormone modulation of neuroendocrine system and modulation model, an ultrashort feedback intelligent controller which is suitable for the process with time delay is designed. It combined with predictive PI controller to get the final controller output. Furthermore, the controller parameters are identified and tuned quickly by the error function. Simulation results demonstrate that the control effect of the new controller is better than that of conventional PID and general predictive PI controller.Fourthly, the existent main problem and solutions in humidity automation is proposed according with the fact status. Combining the analysis and design, the control policy and frame are constructed, including the phased control, formula managing, self-learning and online optimization.Next, the thesis describes the hardware realization and the software logic process of the automatic control system of the conditioning cylinder. A real-time control demonstration system with abundant functions is built by Labview 7 at last.Finally, we conclude the thesis, and provide the direction for further studies.