Improving the procurement process of critical spare parts of the electricity distribution company by combining process mining ,Internet of Things (IOT) an‎d artificial intelligence.

9/23/2026 12:00:00 AM

Timely an‎d effective procurement of critical spare parts in power distribution networks plays a key role in ensuring service continuity an‎d customer satisfaction. Given the increasing complexity of networks, the diversity of spare parts, an‎d the sensitivity of procurement time, traditional supply chain management is no longer sufficient. The main objective of this research is to improve the procurement process of critical spare parts by integrating three complementary approaches: Process Mining, the Internet of Things (IoT), an‎d Machine Learning. The research methodology includes: (1) classifying spare parts using the VED method to identify vital items, (2) analyzing the procurement process with real data using the Disco وCelonis tool to Disco وCelonisver process variants an‎d bottlenecks, (3) designing an‎d training a Ran‎dom Forest model to predict the type an‎d timing of spare part orders, an‎d (4) integrating operational data (cost, Lead Time), IoT data (equipment failure rates), an‎d process indicators (process variants). The findings revealed that the current process involves 52 different paths, of which only 15 main variants cover 80% of the cases. The main bottlenecks occur in the approval an‎d supplier selec‎tion stages. The Ran‎dom Forest model achieved 91% accuracy in predicting the type of spare part an‎d an average error of about 3 days (MAE) in predicting order timing. Additionally, variables such as past Lead Times, equipment failure rates, seasonal deman‎d, an‎d supplier performance were identified as the most influential factors. This study demonstrated that combining Process Mining, IoT, an‎d Machine Learning can significantly reduce Lead Time, lower procurement costs, an‎d increase the reliability of the distribution network. The main innovation of this research lies in the integration of qualitative data (VED) an‎d quantitative data (IoT, orders, an‎d costs) into a data-driven framework to improve spare parts supply chain management.

فرآيندكاوي , الگوريتم‌هاي هوش مصنوعي , اينترنت اشيا , فرآيند تداركات , قطعات يدكي حياتي , مدل جنگل هاي تصادفي

Procurement Process , , Process Mining , Supply Chain Management , Procurement , Maintenance , Spare Parts , Artificial Intelligence (AI) , Internet of Things (IoT) , VED , Random Forest model , Machine Learning

Mozhgan Mahmoudi

Dr. Ebrahim Teimouri