Learning and Analytics in Intelligent Systems, 2020
This paper continues the presentation of new contributions to improve significantly the results d... more This paper continues the presentation of new contributions to improve significantly the results disseminated in the recent paper presented for FedCSIS’ 2018 International Conference on single-input single-output autoregressive moving average with exogenous input models in terms of accuracy, simplicity and real time fast implementation. The improvement is performed by developing accurate multi-inputs single-output adaptive neuro-fuzzy inference system models, suitable to be tailored also for modelling the nonlinear dynamics of any process or control systems. Compared to the conference paper the novelty is a further investigation on design and implementation of hybrid closed-loop control strategy structures consisting of the proposed improved models in combination with digital PI controllers that enhance significantly the tracking accuracy performance, overshoot and number of oscillations, transient convergence speed and robustness to consecutive step changes in the input setpoints. The effectiveness of this new control approach is demonstrated by extensive simulations conducted in MATLAB software environment followed by a rigorous performance analysis by comparison to MATLAB simulation results obtained for a same case study by a standard two input and a single output digital PI controller.
Learning and Analytics in Intelligent Systems, 2020
This paper continues the presentation of new contributions to improve significantly the results d... more This paper continues the presentation of new contributions to improve significantly the results disseminated in the recent paper presented for FedCSIS’ 2018 International Conference on single-input single-output autoregressive moving average with exogenous input models in terms of accuracy, simplicity and real time fast implementation. The improvement is performed by developing accurate multi-inputs single-output adaptive neuro-fuzzy inference system models, suitable to be tailored also for modelling the nonlinear dynamics of any process or control systems. Compared to the conference paper the novelty is a further investigation on design and implementation of hybrid closed-loop control strategy structures consisting of the proposed improved models in combination with digital PI controllers that enhance significantly the tracking accuracy performance, overshoot and number of oscillations, transient convergence speed and robustness to consecutive step changes in the input setpoints. The effectiveness of this new control approach is demonstrated by extensive simulations conducted in MATLAB software environment followed by a rigorous performance analysis by comparison to MATLAB simulation results obtained for a same case study by a standard two input and a single output digital PI controller.
Uploads
Papers by Sorin Radu