Research And Application On Electrical Heating Furnace Based On Neural Network Internal Model Control

The control of Temperature in industrial countries of the world today is facing an important problem, through more sophisticated controling strategies to achieve a better temperature controling has become a consensus. In recent years, from the industrial characteristics of the process, put forward by the internal model control methods do not ask for much on the model of precision and on-line calculation is simple, the process of the environment and the uncertainty of a certain degree of adaptability. In this paper, the characteristics of electric furnace, application of model control for heating water temperature of the internal model control, but the furnace temperature control process is a complex nonlinear physics of the process, which has lagged behind the dynamic properties of large inertia, interference Strong and non-linear, and other typical characteristics, it is difficult to establish precise mathematical model. In order to solve this problem, this paper, a neural network based on BP's internal model furnace temperature control system design, combined with our control over the laboratory testing equipment in industrial electric furnace to carry out targeted research and design .First of all, this article on a variety of methods to control the temperature has done a comparative study on the furnace system to do the analysis and to identify targets and strategies for the control. For the establishment of furnace temperature control system model of neural network algorithm, using BP algorithm and the model control strategy combining the new method, this algorithm and simulation, analysis, to verify this algorithm and the application of electric furnace possible Sex.Then, combined with electric furnace temperature control analysis of the characteristics of the course, the above method successfully used in the process of temperature control, set up the system's neural network models. The model can be used in the process of state-variable temperature control of the estimates and projections.Again, this article in accordance with the parameters of BPNN linear features of the design of the structure of the neural network based on BP's non-linear model controller, the controller applied to the temperature control process in order to drive heater was charged for the amount of current variables, To carry out simulations and hardware given design.Finally, according to BP neural network algorithm for computing large, complex algorithm, and other characteristics of this article is also innovative use of digital signal processor (DSP) hardware as part of the microprocessor.Through the study for the method based on neural network internal model controling for eletric heating stove temperature ,try a neural network to intelligent control technology into eletric heating stove temperature control in the area. The simulation results show that based on neural network model control system has a strong self-learning and adaptive, as well as robust, non-linear process of biochemical reactions can be accurately controlled, to achieve good results.