AI-based Prognostic Analysis of NOx Emission from Industrial Diesel Engine for Feedback Control System

NOx emissions from diesel engines are critical issue because they have far-reaching consequences for public health, the environment, and the climate. NOx emission reduction is crucial for improving air quality, safeguarding human health, and minimizing climate change. The present study examines the extent to which two different machine learning (ML) models, Linear Regression (LR) and XGBoost, can forecast NOx emissions. The study utilizes a set statistical methods and Taylor diagrams to figure out how well the models can catch key statistical characteristics of measured NOx values. The Taylor graphs as well as statistical values show that XGBoost does better than LR. In particular, XGBoost gets R
²
values of 0.9967 and 0.9941 for training datasets and 0.9901 and 0.9902 for test datasets, while LR gets R
²
values of 0.9901 and 0.9902. Additionally, LR's MSE values are 1081.21 for training data and 580.48 for test data, while XGBoost's are 1.53 for training data and 3.59 for test data. Thus, while both ML models shows good forecasting performance, the XGBoost outperforms the LR in predicting the NOx emission from diesel engine.