Title: Non-linear hydraulic system (MIMO) plant for optimal control and identification by hybrid BCMO-RERNN strategy
Authors: Rathinam Muniraj; Ganapathiapillai Kannayeram; Ragavan Saravanan
Addresses: Department of Electrical and Electronics Engineering, P.S.R. Engineering College, Sivakasi, Tamil Nadu, 626 140, India ' Department of Electrical and Electronics Engineering, Sri Ramakrishna Institute of Technology, Perur Chettipalayam, Pachapalayam, Coimbatore, Tamil Nadu, 641 010, India ' Department of Electrical and Electronics Engineering, Balaji Institute of Technology and Science, Warangal, Telangana, 506 331, India
Abstract: This manuscript proposes a non-linear hydraulic system multiple input multiple output (MIMO) plant for optimal control and identification by hybrid BCMO-RERNN strategy. The BCMO-RERNN hybrid strategy is a combination of balancing composite motion optimisation (BCMO) and recalling enhanced recurrent neural network (RERNN), therefore it is called BCMO-RERNN. The proposed adaptation process is executed by utilising data occupied through the operation of control system, which plant is recognised and model is used for improving the controller. Plant is corresponding to MIMO hydraulic system, which containing 2 tanks fed through the pump and the 3 path valve. The first aspects are achieved by plant identification and controller training. Plant identification is performed in two ways: first one is offline by using BCMO technique and collecting data in an open loop, second one is online by RERNN, where training data is collected through control system operation in a close loop.
Keywords: BCMO; balancing composite motion optimisation; RERNN; recalling enhanced recurrent neural network; MIMO; multiple input multiple output; hydraulic system; MATLAB.
DOI: 10.1504/IJHVS.2024.138404
International Journal of Heavy Vehicle Systems, 2024 Vol.31 No.3, pp.413 - 442
Received: 29 Jul 2023
Accepted: 25 Sep 2023
Published online: 02 May 2024 *