TAILIEUCHUNG - Artificial Neural Networks Industrial and Control Engineering Applications Part 12

Tham khảo tài liệu 'artificial neural networks industrial and control engineering applications part 12', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | 374 Artificial Neural Networks - Industrial and Control Engineering Applications time sec time sec 0 1000 2000 3000 4000 5000 time sec Fig. 13. Piecewise Linearized Model Predictive control response Discrete controller tuning online Control loop of this technique is connected in a way introduced briefly in section . Differential evolution is chosen as search technique. After some experiments eligible parameters are chosen this way NP 30 CR F N 20. Cost function is selected according to Eq. 19 where h1 h2 . Control response is depicted in Fig. 14. There is no exact alternative in classical control theory to this technique. However in a certain way it is close to predictive control therefore it can be compared to Fig. 13. It is remarkable that control response shown in Fig. 14 provides the most suitable performance of all experiments. But on the other hand it is highly computationally demanding technique. Artificial Neural Network - Possible Approach to Nonlinear System Control 375 time sec Fig. 14. Discrete controller tuned online 5. Conclusion The aim of this work was to design a controller which provides control performance with control error less than 10 C. Because of the nonlinearity of the plant two groups of advanced control techniques were used. The first group is based on artificial neural networks usage while the second one combines their alternatives in modern control theory. Generally speaking neural networks are recommended to use when plant is strongly nonlinear and or stochastic. Although reactor furnace is indispensably nonlinear it is evident that control techniques without neural networks can control the plant sufficiently and in some cases especially predictive control and internal model control even better. Thus neural network usage is not strictly necessary here although especially Discrete Controller Tuning Online brings extra good performance. 6. Acknowledgement This work was supported by the 6th Framework Programme of .

TỪ KHÓA LIÊN QUAN
Đã phát hiện trình chặn quảng cáo AdBlock
Trang web này phụ thuộc vào doanh thu từ số lần hiển thị quảng cáo để tồn tại. Vui lòng tắt trình chặn quảng cáo của bạn hoặc tạm dừng tính năng chặn quảng cáo cho trang web này.