تصميم وتنفيذ نظام تحكم ذكي بالماتالب لمطحنة فحم

Abstract

Abstract This research addresses -via the matlab software environment- the intelligent control of a coal mill; due to the importance of coal mills in heavy industries and the energy industry, and the problem in traditional industrial control systems where early fault diagnosis is not performed, non-ideal precision in evaluation, poor estimation occurs, and human errors exist. Additionally, the detection of instability conditions and the prediction of recurring patterns represent a characteristic of human limitations. Due to the human factor in traditional industrial control systems, an intelligent model of a Nonlinear Autoregressive Neural Network with Exogenous Inputs was designed to predict mechanical vibration behavior. Based on the analytical methodology, a time-series database comprising 3,000 real samples was created, which were collected from the operation of a coal mill operating at Berber Cement Factory, and used for training and testing the model to ensure the realism of the results. Mechanical vibration behavior was accurately predicted at the level of the neural network model because it is a vital indicator directly linked to stability and all operating parameters, such as the mill feed rate per hour, mill outlet temperature, differential pressure as a prim indicator of the actual filling degree relative to nominal value, and other operating parameters. It was found that the mean square error reached 0.007 within only 11 iterations, coupled with the positioning of all error correlations within the confidence limits, representing conclusive evidence of the model's ability to capture the complex nonlinear and dynamic relationships inherent in industrial operational data.