Performance Analysis of Heat Exchanger System Using Deep Learning Controller

Abstract

Conventional PID controllers have utilised in most of the process industries. Despite being the most used controller, the traditional PID controller suffers from several disadvantages. Due to rapid development in the field of the process control system, various controllers have been developed that try to overcome the limitations of the PID controller. In this paper, a heat exchanger system has been simulated, and the generated data has been used to train a deep learning-based controller using Backpropagation. The obtained results are compared with the conventional controller on several metrics, including time response, performance indices, frequency response etc. The proposed model outperforms the conventional controller on all the evaluation metrics.

Performance analysis of various training algorithms of deep learning based controller

“…This approach allows for a more generalized assessment of model performance, as it can reveal how well the model can perform across different problems and scenarios. One benchmark problem is considered as a heat exchanger [7,46], and the second benchmark problem is a cascaded tank-level system [3,47]. A heat exchanger system was used to transfer heat from the source to the working fluid.…”

“…The most common types of classical controllers used in process control systems are proportional-integral-derivative (PID) controllers [5,6], which are based on a mathematical model of the process and adjust the control action based on the error between the setpoint and the process variable. However, they have several limitations that can make them unsuitable for controlling complex or highly nonlinear processes [3,7]. One of the main limitations is their reliance on a mathematical model of the process, which can be challenging to develop and may not accurately capture the dynamics of the system under all operating conditions [8].…”

“…Adaptive controllers, on the other hand, can struggle to maintain stability and performance when the process dynamics change significantly. To address these limitations, there has been a growing interest in developing more advanced control strategies, such as deep learning controllers [7,11,12]. Deep learning controllers are based on artificial neural networks that can learn to model the process dynamics directly from data, without requiring a priori knowledge of the system [13,14].…”

System Modelling and Identification for EEG Monitoring using Random Vector Functional Link Network

“…The main difference between them is the direct link between the input to the output layer [7] traditional models: it has better generalization performance, a simple architecture design, and a faster learning algorithm. Similarly, some other system identification techniques like Additive algorithms such as Least mean square (LMS) for identification and noise separation [8][9][10][11][12]. Various Nonlinear system identification and control related models are presented in [13][14][15][16][17][18][19][20][21][22].…”

Analysis of DC motor for process control application using neural network predictive controller

“…Various control techniques, such as PID control [15], fuzzy logic control [16], and Model Predictive Control [3], have been explored for effective speed and torque control in DC motor systems. Soft computing-based controllers, including genetic algorithms and neural networks, are gaining interest because of their potential to address complex and nonlinear relationships within the control system, enhancing adaptability and learning capabilities [17][18][19][20]. This research signifies the continual evolution of control strategies, aiming to further enhance the capabilities of DC motors in diverse applications.…”