Prediction of Equipment Effectiveness using Hybrid Moving Average-Adaptive Neuro Fuzzy Inference System (MA-ANFIS)...

Area of Science:

• Operations Management
• Artificial Intelligence in Manufacturing
• Predictive Maintenance

Background:

• Dynamic production environments necessitate higher operational performance.
• Accurate prediction of machine performance is crucial for productivity and resource allocation.
• Existing research inadequately addresses reliable equipment effectiveness assessment for managers.

Purpose of the Study:

• To introduce a hybrid moving average - adaptive neuro-fuzzy inference system (MA-ANFIS) for predicting equipment effectiveness.
• To develop and evaluate distinct equipment effectiveness prediction models using real-world manufacturing data.
• To provide a decision support system for production managers to enhance equipment performance.

Main Methods:

• Development of a hybrid moving average - adaptive neuro-fuzzy inference system (MA-ANFIS).
• Application and evaluation of three distinct prediction models based on real-world manufacturing problems.
• Comparative analysis of model performance, focusing on the Gaussian membership function.

Main Results:

• The hybrid MA-ANFIS model utilizing a Gaussian membership function achieved superior prediction accuracy compared to other developed models.
• The developed models were integrated into a decision support system.
• Demonstrated potential for reducing time and cost in bus body building through improved equipment performance.

Conclusions:

• The MA-ANFIS model offers a reliable method for predicting equipment effectiveness in dynamic manufacturing settings.
• The decision support system aids managers in optimizing resource allocation and improving operational efficiency.
• This approach contributes to enhanced productivity and cost reduction in manufacturing processes.