Applied Machine Learning in Industry 4.0: Case-Study Research in Predictive Models for Black Carbon Emissions

Area of Science:

• Industrial Engineering
• Environmental Science
• Data Science

Background:

• Industrial furnaces (IFs) are critical in manufacturing, requiring precise heat treatment.
• Emission of black carbon (EoBC) from IFs is a significant operational challenge.
• EoBC is influenced by fuel quality, furnace efficiency, operational practices, and process conditions.

Purpose of the Study:

• To present a methodological approach for predicting EoBC in IFs using machine learning (ML).
• To identify the most suitable ML model for EoBC prediction based on real-world data and implementation constraints.

Main Methods:

• Utilized a real-world dataset of historical IF operation data.
• Trained and evaluated various ML models for EoBC prediction.
• Selected the optimal ML model through rigorous evaluation against operational data.

Main Results:

• Confirmed the feasibility of accurately predicting undesirable EoBC well in advance.
• Identified a specific ML approach that best fits the dataset and industrial constraints.
• Demonstrated the potential for proactive management of black carbon emissions.

Conclusions:

• Machine learning offers a viable solution for predicting EoBC in industrial furnace operations.
• This research pioneers the application of ML for EoBC prediction in the IF industry.
• Predictive modeling can enhance environmental compliance and operational efficiency in industrial settings.