Development of energy efficient drive for ventilation system using recurrent neural network

• Published in: Neural computing & applications Vol. 33; no. 14; pp. 8659 - 8668
• Main Authors: Prince, Hati, Ananda Shankar, Chakrabarti, Prasun, Abawajy, Jemal H., Keong, Ng Wee
• Format: Journal Article
• Language: English
• Published: London Springer London 01.07.2021; Springer Nature B.V
• Subjects: Adaptive control; Algorithms; Artificial Intelligence; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Control methods; Control stability; Controllers; Data Mining and Knowledge Discovery; Derivatives; Energy dissipation; Image Processing and Computer Vision; Low speed; Model reference adaptive control; Neural networks; Original Article; Probability and Statistics in Computer Science; Proportional integral; Recurrent neural networks; Rotor speed; Speed control; Ventilation; Sensor-less speed control; FOPI; Recurrent neural network; Ventilation system
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-020-05615-x

This research article corroborates the working of a model reference adaptive model (MRAS) with fractional-order proportional-integral (FOPI λ )-based encoderless speed control approach for ventilation system drive using recurrent neural network (RNN) for the low- and medium-range operation. The purpose of this study is to minimize the energy loss due to fluctuations and variation in the rotor speed and also find the optimum values of FOPI λ by using a recurrent neural network to enhance the overall implementation of the system. In this perspective, the low-speed execution of MRAS is poor due to the existence of a pure integral and derivative parameter. Towards enhancement of the performance at speed region, a MRAS method with RNN is used. The network is trained using the Levenberg–Marquardt (LM) algorithm, and FOPI λ control method is used for tuning the gain of proportional-integral of speed and current controller of the encoderless speed control of the ventilation drive. The presented RNN speed estimator with FOPI λ controller has shown better performance and stability in transitory and stable operation as well as it also provides an enhancement in the overall efficiency of the ventilation drive. The validation of the presented algorithm is detailed experiments on a fully digitized 5.5 kW ventilation system using the Lab VIEW interface.