| Decoupling theory was brought in at the same time as the control subject, which is one of the oldest control issues. Regardless of how the development of decoupling, the main content is the two questions: one is how to choose the amount charged matches the amount; one is what kind of decoupling network is used.Hot air reflow machine system is a complex multi-zone temperature industrial system, with some characteristics not benefit to study control algorithm, as strong coupling and nonlinear. During debugging at the scene there had been the temperature inter - string situation. Paper [2] ,from the study of hot mechanism, set up the hot air reflow machine model based on the heat balance mechanism. For its strong coupling characteristic, the paper presents two methods decoupling control algorithms to compare, which is benefit for the project practice.After discussing the decoupling situations at home and abroad, the author obtains the system relative amplification coefficient matrix from the model, analyzes the coupling level, and conforms the system original amount's matches are right in theory. For the model's high order, first simplifies it to the classic second-order system using Pade approximation method, then designs the feed-forward decoupling compensation with the invariance principle. After adding the PID controller, simulating in the MATLAB, results show that the method is an effective one. Taking into account that the classical method relies on the accurate model too much, and there is so many uncertainties in the industrial scene, the paper designs a fuzzy neural network decoupling controller, which integrates the fuzzy control and neural network. The characteristics are not dependent on precise mathematical model so much and it can be adjusted the controller parameters online. From the simulation results , see that the above-mentioned decoupling method can effectively solve the coupling problem of the system, which provides a good theoretical basis to engineering practice. |
PDF Full Text Request