PID Based Artificial Neural Network For Rotary Dryer In The Fertilizer Industry

Abstract

Rotary dryers are often used to dry materials in fertilizer industry. The drying process using a rotary dryer may occasionally encounter issues, such as an uneven distribution of hot air and overheating, which can be caused by operator input errors or control parameter adjustments that are not optimal. The conventional proportional-integral-derivative (PID) controller is a common method for regulating temperature in rotary dryers. However, it is not particularly effective in dealing with changes that occur in real time, which can result in extending the system’s transient period before equilibrium and overshoot for temperature output in rotary dryers. One potential solution that can be offered is to integrate PID control with an Artificial Neural Network (ANN) which would enable the system to adapt to the dynamics of the operational environment without operator assistance. The methodology used is backpropagation neural networks, trained with empirical data gathered during the operation of a rotary dryer. The output from the ANN model is then used to adjust the PID controller within the system in order to prevent extending the system’s transient period before equilibrium and overshoot. Backpropagation is used because this algorithm can effectively reduce errors when recognizing data patterns. The control design aims to improve the efficiency of the drying process, optimizing it, and reducing costs associated with production