Automated learning and control using speedup and neural networks
Само за регистроване кориснике
2000
Чланак у часопису (Објављена верзија)
Метаподаци
Приказ свих података о документуАпстракт
This paper studies complex dynamic neural network learning models. Backpropagation was used to train a neural network for dynamic simulation and control of a chemical stirred tank. The generalized delta rule algorithm was used to train the network minimizing the sum of squares of the residual. It was assumed that a historical database of plant inputs and outputs is available. A Pseudo Random Binary Sequence-PRBS was used as a disturbance. For training the database the 1% PRBS signal was superimposed upon its steady state value from SPEEDUP simulation. Once a trained neural network model was available, it then used in real time learning and pH control. The examined inverse and standard neural network models controllers achieved better performance than a conventional PI controller. Complex Internal Model Control - IMC achieved the best results in control and local stability. The obtained models in this paper improve noisy handling and reduce process variability. Some of these systems can... be used for self - maintaining non-linear, multivariable models and day to day troubleshooting.
Извор:
Chemical Industry and Chemical Engineering Quarterly, 2000, 6, 3, 377-384Издавач:
- CI and CEQ
Институција/група
Tehnološko-metalurški fakultetTY - JOUR AU - Savković-Stevanović, Jelenka B. PY - 2000 UR - http://TechnoRep.tmf.bg.ac.rs/handle/123456789/5309 AB - This paper studies complex dynamic neural network learning models. Backpropagation was used to train a neural network for dynamic simulation and control of a chemical stirred tank. The generalized delta rule algorithm was used to train the network minimizing the sum of squares of the residual. It was assumed that a historical database of plant inputs and outputs is available. A Pseudo Random Binary Sequence-PRBS was used as a disturbance. For training the database the 1% PRBS signal was superimposed upon its steady state value from SPEEDUP simulation. Once a trained neural network model was available, it then used in real time learning and pH control. The examined inverse and standard neural network models controllers achieved better performance than a conventional PI controller. Complex Internal Model Control - IMC achieved the best results in control and local stability. The obtained models in this paper improve noisy handling and reduce process variability. Some of these systems can be used for self - maintaining non-linear, multivariable models and day to day troubleshooting. PB - CI and CEQ T2 - Chemical Industry and Chemical Engineering Quarterly T1 - Automated learning and control using speedup and neural networks EP - 384 IS - 3 SP - 377 VL - 6 UR - https://hdl.handle.net/21.15107/rcub_technorep_5309 ER -
@article{ author = "Savković-Stevanović, Jelenka B.", year = "2000", abstract = "This paper studies complex dynamic neural network learning models. Backpropagation was used to train a neural network for dynamic simulation and control of a chemical stirred tank. The generalized delta rule algorithm was used to train the network minimizing the sum of squares of the residual. It was assumed that a historical database of plant inputs and outputs is available. A Pseudo Random Binary Sequence-PRBS was used as a disturbance. For training the database the 1% PRBS signal was superimposed upon its steady state value from SPEEDUP simulation. Once a trained neural network model was available, it then used in real time learning and pH control. The examined inverse and standard neural network models controllers achieved better performance than a conventional PI controller. Complex Internal Model Control - IMC achieved the best results in control and local stability. The obtained models in this paper improve noisy handling and reduce process variability. Some of these systems can be used for self - maintaining non-linear, multivariable models and day to day troubleshooting.", publisher = "CI and CEQ", journal = "Chemical Industry and Chemical Engineering Quarterly", title = "Automated learning and control using speedup and neural networks", pages = "384-377", number = "3", volume = "6", url = "https://hdl.handle.net/21.15107/rcub_technorep_5309" }
Savković-Stevanović, J. B.. (2000). Automated learning and control using speedup and neural networks. in Chemical Industry and Chemical Engineering Quarterly CI and CEQ., 6(3), 377-384. https://hdl.handle.net/21.15107/rcub_technorep_5309
Savković-Stevanović JB. Automated learning and control using speedup and neural networks. in Chemical Industry and Chemical Engineering Quarterly. 2000;6(3):377-384. https://hdl.handle.net/21.15107/rcub_technorep_5309 .
Savković-Stevanović, Jelenka B., "Automated learning and control using speedup and neural networks" in Chemical Industry and Chemical Engineering Quarterly, 6, no. 3 (2000):377-384, https://hdl.handle.net/21.15107/rcub_technorep_5309 .