AI-Driven Predictive Maintenance System for Manufacturing Equipment Optimization

Thesis Abstract

The persistent challenge of unplanned equipment downtime and maintenance costs significantly impedes operational efficiency in manufacturing industries, necessitating innovative solutions that leverage advancements in information and communication technology. This study aims to develop and validate an Artificial Intelligence (AI)-driven predictive maintenance system designed to optimize manufacturing equipment performance, thereby reducing downtime, extending machinery lifespan, and lowering maintenance expenses. The specific objectives include identifying relevant sensor data and features indicative of equipment failure, designing a machine learning framework capable of accurately predicting maintenance needs, evaluating the system's predictive accuracy and operational impact in real-world settings, and providing actionable insights for maintenance scheduling. Employing a mixed-method research design, the study integrates quantitative data analysis with qualitative insights to comprehensively evaluate system performance and usability. The population comprises manufacturing equipment within three mid-sized electronics manufacturing plants in a metropolitan industrial zone, with a combined total of 150 machines. A stratified random sampling technique was used to select a representative sample of 60 critical equipment units displaying high operational variability. Data collection involved deploying IoT sensors to continuously monitor operational parameters such as vibration, temperature, sound levels, and power consumption over a six-month period, resulting in a dataset of approximately 10 million sensor entries. Complementary semi-structured interviews with maintenance personnel and plant managers provided contextual insights into maintenance workflows and system integration challenges. Data analysis incorporated several techniques feature extraction and selection were conducted using Principal Component Analysis (PCA) to identify the most predictive sensor indicators. Supervised machine learning algorithms—specifically, Random Forest and Support Vector Machines (SVM)—were trained on labeled historical failure data to develop predictive models. Model performance was assessed through metrics such as accuracy, precision, recall, F1-score, and the Area Under the Receiver Operating Characteristic Curve (AUC-ROC). Additionally, the study applied the Theory of Planned Behavior to understand maintenance personnel's acceptance of AI-driven recommendations and used time-series analysis to evaluate predictive lead times and maintenance scheduling efficiencies. The anticipated findings demonstrate that the AI-based predictive maintenance system significantly enhances failure prediction accuracy—achieving over 85% accuracy and an AUC-ROC exceeding 0.9—compared to traditional calendar-based or reactive maintenance strategies. The system's deployment is expected to reduce unplanned downtime by at least 30%, lower maintenance costs by 20%, and improve overall equipment effectiveness (OEE). Qualitative insights suggest high acceptance levels among maintenance staff, provided adequate training and system integration support are provided. This research contributes to the existing body of knowledge by bridging the gap between theoretical predictive analytics and practical maintenance management in manufacturing contexts, particularly in developing an integrated AI framework tailored for sensor-rich industrial environments. It extends current understanding of how machine learning models can be optimized for real-time failure prediction and operational decision-making. The findings affirm the efficacy of combining sensor technology with advanced algorithms under the lens of established behavioral theories to facilitate technological adoption. The study concludes that AI-driven predictive maintenance systems present a viable pathway to industrial digital transformation, capable of delivering substantial operational improvements. It recommends strategic investments in sensor infrastructure, staff training on AI tools, and continuous system performance evaluation to sustain benefits. Future research should explore scaling such systems across different manufacturing sectors and integrating them with enterprise resource planning (ERP) systems for holistic operational intelligence. This study thereby provides a comprehensive framework for manufacturing firms aiming to harness AI for equipment reliability and production excellence.