Timber Drying Model Predictive Control Technology Based On Support Vector Machine

As one of the widespread concerned and used resources in today's society, timber resource is renewable, but what different from wind, solar, water and geothermal energy etc is that it possesses the characters of slow self-renewal speed and long reproductive cycle. Excessive depletion of timber resource aroused by the existing production mode that causes the ecological problems are more and more serious. People all over the world paid attention to these problems such as how to increase the utilization rate and reduce consumption of limited timber resources and improve the quality of wood products. It has been always a focus research in this field at home and abroad. Improving the performance of timber is the key to solve above problems, while improving the quality of timber drying is an important breakthrough.According to the characters of timber is porous permeable and moisture absorption material, the water exists in the timber with a variety of forms and the different timber with various moisture content, and combining the timber drying is a lagged, time-variable, coupling and complicated non- linear process. On the basis of the thorough research on the timber drying mechanism and controlling the internal moisture permeates outwards reasonably and orderly, two main researches, how to establish the optimal timber drying mechanism model and built a drying close-loop control system based on this model, were carried out. The drying mechanism model should have good dynamic approaching effect of moisture content variation curves. The closed-loop control system should have rapid and accurate response capability to the subtle changes of drying space, and then can establish the drying control strategy of the next time.About the establishment of timber drying model, existing timber drying models are of many constraints that lead to complex model and difficulty to practical application. While the models which were built up based on Neural Network usually met the questions of over learning and owe learning and the curse of dimensionality. In the dissertation, combing with the Statistical Learning Theory and the structural risk minimization principle, the nonlinear predictive regression method based on the least squares support vector machines is presented, which can improve the generalization ability of the models and simplify the training process. Mexican Hat wavelet function is introduced as kernel function of the least-square support vector machines and the good properties of the wavelet functions in signal capturing and signal analysis are fully display, it can improve the generalization ability of Support Vector Machines effectively by detecting transient signal through scale variations.In the control of drying process, traditional control methods were based on the accurate model of measured object. However, because of the lack of accurate models it usually adopts fixed control algorithm, which would lead to flexibility and adaptability lackness of whole system control as it was limited by model framework. Aiming at these problems, this paper introduces the predictive control method based on model; the drying model based on Support Vector Machine is used as prediction function, using roll optimizing strategy to realize the optimal control process of the limited moisture content interval in timber drying process and it will solve the optimization problem of object function in the design of prediction controller by means of a multivariable random iterative search method based on particle swarm optimization.To verify the actual effects of the prediction control system, a small timber drying kiln was used for experiments and the model numerical simulation was discussed. The results show that the timber drying model and predictive control system can meet the accuracy requirement and have good control effect. This dissertation is of great help for improving controlling level of wood drying and realizing controlling timber drying course fully automatically. It is the important basis for guaranteeing the drying quality, facilitating the continuous usages of the timber resource and the healthy growth of the relevant industry.