A Decision Risk Assessment and Alleviation Framework under Data Quality Challenges in Manufacturing

Area of Science:

• Industrial Engineering
• Manufacturing Systems
• Data Science

Background:

• The proliferation of Internet of Things (IoT) and artificial intelligence (AI) enables real-time control in manufacturing.
• Data quality issues pose a significant challenge to the reliability of automated decision-making models.

Purpose of the Study:

• To propose a risk assessment framework for systematically evaluating data quality issues in continuous production workshops.
• To address challenges in parameter identification under uncertainty, sensor accuracy selection, and fault-tolerant control.

Main Methods:

• Developed a framework transforming data quality issues into data deviation problems.
• Utilized Monte Carlo simulation to quantify the impact of data quality on decision risk.
• Established a direct link between sensor quality and operational risks.

Main Results:

• The framework provides specific steps to address data quality challenges.
• A case study in the steel industry validated the framework's effectiveness.
• Demonstrated a quantifiable relationship between sensor quality and decision-making risks.

Conclusions:

• The proposed framework offers a novel approach to assessing safety in industrial automated decision-making.
• It enables risk reduction for real-time unmanned control systems.
• Enhances the robustness and reliability of AI-driven manufacturing processes.