Rotary Kiln's Control Based On Fuzzy Neural Network And Its Realization In Siemens PCS7

Rotary kiln is the main production equipment in cement production, thus the automatic control of it, especially of the temperature inside the kiln, a vital controlled variable, is of great importance to guarantee the quality and stabilize the production. Due to the nonlinearity, instability and time lag in controlling, satisfactory result is quite hardly to be obtained by the traditional controlling system. Since fuzzy system can reason logically with human experimental knowledge and neural network has a strong computing capability, if a fuzzy system is constructed with neural network, a fuzzy inference system with learning self-adapting function will be realized by modulate and optimize the system parameters automatically according to the sample.The fuzzy system utilizes an average fuzzy inference model which consists of hierarchical neural network. Considering that the steepest descent method is easily trapped into the minimal value for the neural network learning algorithm, this paper introduces a two-part optimization solution. Firstly search the network parameters for optimization with the global searching ability of genetic algorithm to roughly modulate the system parameters. Then search the network parameters locally with steepest descent method to precisely modulate the parameters, thus enhance the control ability of fuzzy neural network. The efficiency of the solution is proved through simulation experiment with hybrid learning algorithm.Siemens PCS7 is a process control system widely used in industry control, which integrated the solution of the rotary kiln control system, combined fuzzy neural network learning algorithm. utilized SCL to design adaptive fuzzy control, and realized the intelligent control of the rotary kiln temperature.