Implementation of an AI-driven predictive maintenance and asset lifecycle management system for university estate portfolios
Table Of Contents
Chapter ONE
INTRODUCTION
- 1.1Introduction
- 1.2Background of Study
- 1.3Problem Statement
- 1.4Objective of Study
- 1.5Limitation of Study
- 1.6Scope of Study
- 1.7Significance of Study
- 1.8Structure of the Research
- 1.9Definition of Terms
Chapter TWO
LITERATURE REVIEW
- 2.1Theoretical Foundations of Estate Management
- 2.2Asset Lifecycle Management Theories
- 2.3Predictive Maintenance Concepts in Facilities Management
- 2.4AI and Machine Learning in Real Estate and Estates Portfolio Management
- 2.5Data Retrieval and Data Quality in Estate Portfolios
- 2.6Urban Planning and Campus Estate Governance
- 2.7Risk Assessment and Compliance in Estate Management
- 2.8Sustainability and Energy Efficiency in Estate Portfolios
- 2.9Stakeholder Theory in University Estates
- 2.10Gaps in Current Estate Management Practices and Research
Chapter THREE
RESEARCH METHODOLOGY
- 3.1Research Design and Philosophical Underpinnings
- 3.2Study Population and Sampling Techniques
- 3.3Data Collection Methods and Tools
- 3.4System Architecture and Technology Stack
- 3.5Data Governance, Privacy, and Security
- 3.6Feature Engineering for Predictive Maintenance
- 3.7Model Development, Training, and Validation
- 3.8Evaluation Metrics and Benchmarking
- 3.9Ethical Considerations in Research
- 3.10Deployment Strategy and Change Management
Chapter FOUR
DATA PRESENTATION AND ANALYSIS
- 4.1Case Study: University Estate Portfolio Overview
- 4.2Documentation of Current Maintenance Practices
- 4.3Asset Inventory and Classification
- 4.4Data Acquisition and Cleaning Procedures
- 4.5Development of Predictive Maintenance Models
- 4.6Asset Lifecycle Simulation and Planning
- 4.7Cost-Benefit Analysis and ROI of the System
- 4.8System Prototyping, Testing, and User Feedback
- 4.9Impact on Sustainability and Energy Use
- 4.10Risk Mitigation and Compliance Outcomes
Chapter FIVE
SUMMARY, CONCLUSION AND RECOMMENDATIONS
- 5.1Summary of Findings
- 5.2Theoretical and Practical Implications
- 5.3Limitations and Recommendations for Future Research
- 5.4Policy and Governance Implications for University Estates
- 5.5Conclusion and Final Reflections
Project Abstract
This study presents the design, development, and evaluation of an AI-driven predictive maintenance and asset lifecycle management system tailored for university estate portfolios, aimed at optimizing facility performance, reducing downtime, and extending asset lifespans within constrained public funding environments. The research integrates data from multiple sources, including building management systems, IoT sensor networks, maintenance logs, capital plan documents, and energy consumption records, to create a unified, time-aware asset registry that supports proactive decision-making. A multilayer architecture is proposed, consisting of data ingestion and preprocessing, feature engineering, predictive analytics, optimization, and governance layers, each incorporating domain-specific constraints such as budget cycles, regulatory compliance, and safety requirements. The predictive analytics module leverages machine learning models for remaining useful life (RUL) estimation, anomaly detection, and fault diagnosis across mechanical, electrical, plumbing, and civil infrastructure components. These models are complemented by a digital twin framework that simulates asset behavior under varying usage scenarios and environmental conditions to predict performance degradation and schedule maintenance activities with minimal disruption to campus operations. The optimization component formulates maintenance scheduling, spare parts inventory, and capital expenditure planning as a multi-objective problem, balancing maintenance cost, asset availability, energy efficiency, and risk exposure, while accounting for limited technician capacity and vendor lead times. A governance layer enforces data quality, model transparency, and auditability, providing explainable AI outputs to facility managers and university administration for informed governance, policy alignment, and accountability. The research also addresses data interoperability challenges through standardized ontologies and interoperable APIs to ensure scalable integration with existing campus information systems. An empirical evaluation is conducted across several university facilities, spanning diverse building types and vintages, to validate model performance, forecast accuracy, and the impact on maintenance effectiveness. Key performance indicators include mean time between failures (MTBF), maintenance completion rates, mean time to repair (MTTR), asset downtime, total cost of ownership (TCO), energy intensity, and flexibility in budget adaptation. The results demonstrate that AI-driven maintenance planning substantially improves asset reliability and uptime while reducing unexpected maintenance costs by a measurable margin, enabling more strategic allocation of limited resources. Sensitivity analyses reveal the robustness of the system to data sparsity, sensor outages, and varying operational policies. The study also investigates change management considerations, including user adoption, stakeholder engagement, and training needs, to maximize the practical utility of the system in a real-world university setting. Finally, the research discusses implications for scalable deployment, governance under public procurement constraints, and potential extensions to integrate occupant comfort metrics and sustainability targets, positioning the proposed system as a transformative tool for modernizing estate management in higher education institutions.
Project Overview
What This Project Is About
A plain-language overview of how university estate portfolios can be kept in good condition using smart systems that predict when maintenance is needed and manage the assets over time. It combines data from building equipment, space usage, and costs to help facilities teams plan repairs and replacements more efficiently.
The Problem It Addresses
Universities own many buildings and assets that need regular upkeep. Maintenance is often reactive, costly, and disruptive. This project seeks a smarter way to predict failures, plan budgets, and extend asset life without unnecessary downtime. It helps campuses run more smoothly and sustainably.
Objectives of the Project
- Identify common maintenance needs across a portfolio of university buildings.
- Develop a simple model to forecast when equipment will require service.
- Create a plan for asset lifecycle management that balances cost and performance.
- Propose an easy-to-use dashboard for facilities staff to monitor health indicators.
- Assess potential cost savings and reliability improvements.
What You Will Do Step by Step
- Learn basic concepts of asset management and predictive maintenance.
- Collect or simulate data about building equipment, usage, and maintenance records.
- Develop a simple rule-based or basic machine-learning approach to predict failures.
- Design a lightweight lifecycle plan for major assets.
- Build a user-friendly dashboard mock-up for key metrics.
- Test the model with sample scenarios and refine it.
- Evaluate potential cost and reliability benefits.
Expected Outcome
A practical, easy-to-implement plan and a basic toolset that helps university facilities teams predict maintenance needs, schedule repairs, and manage assets over time, leading to lower costs and fewer disruptions.