Evaluating the Feasibility of Digital Twin Adoption for MEP Systems

RIT
Evaluating the Feasibility of Digital Twin Adoption for MEP Systems

By

Tejas Adana

A Capstone Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Engineering in Engineering Management

Department of Industrial and Systems Engineering

Rochester Institute of Technology RIT Dubai
December 10, 2025

RIT
Master of Engineering in Engineering Management

Capstone Approval

Student Name: Tejas Adana
Capstone Title: Evaluating the Feasibility of Digital Twin Adoption for MEP Systems

Capstone Committee:

Name Designation Date

Name Designation Date

Acknowledgement
I am so thankful to my respected supervisor who guided me a lot to complete the dissertation. I am also grateful to my friends who helped me overcome the difficulties during the project. I would like to give thanks to my family, especially my parents, who supported me a lot during conducting this study.

Abstract
The mechanism of digital twin and the adoption of the concept in Mechanical, Electrical, and Plumbing (MEP) systems can be viewed as a revolutionary practice of building control that will resolve the existing inefficiencies in the functioning of modern facilities. This study has revealed the research gaps in the application of the digital twin in the MEP industry by a systematic literature review based on PRISMA 2020 guidelines, expert interviews with MEP specialists at various plants and facilities, and prioritization matrices of the identified operational issues. The research modifies a four-layer building architecture available in the literature of digital twins, and it is specifically applied to MEP systems in the context of a commercial building setting. This research prioritizes the use of the digital twin in MEP implementation in a systematic way through a systematic assessment of both business impact and technical feasibility. The modified framework shows considerable opportunities for performance increase in main operational measures, and the literature-based estimations suggest a decrease in operational costs of 42.1, in energy consumption of 18.3%, and in equipment reliability of 48.6%. The results would give useful implementation rules to the MEP industries and could offer evidence-based information to practitioners aimed at maximizing building performance as well as meeting the sustainability specifications in the built environment.
Keywords: Digital Twin, MEP Systems, Predictive Maintenance, Energy Optimization, Building Information Modeling (BIM), Internet of Things (IoT), Framework Adaptation.

1. Introduction
1.1 Motivation for Digital Twin Research in MEP Systems
The Architecture, Engineering, and Construction (AEC) industry continues to face increasing pressure to improve operational efficiency, reduce energy consumption, and extend asset lifecycles within complex building environments (Tang et al., 2019) in Figure 1.1. Modern MEP systems, which contribute nearly 60% of total building operational costs and represent a significant share of energy consumption in commercial facilities, have therefore become critical targets for performance optimisation (AlBalkhy et al., 2024). However, traditional management practices remain predominantly reactive, addressing equipment failures only after they occur instead of preventing disruptions through predictive strategies (Lu et al., 2020). This reactive model results in cascading inefficienciesincluding unplanned downtime, excessive maintenance expenditures, reduced occupant comfort, and persistent patterns of energy wastethat accumulate over building lifecycles (Khajavi et al., 2019).

Figure 1.1 The basic mechanism of DT (Pan & Guo, 2025).
Digital twin technology presents a transformative solution to these challenges by establishing dynamic virtual representations of physical assets that merge real-time operational data with predictive analytics (Grieves, 2022). Unlike conventional building management systems that offer limited monitoring capabilities, digital twins enable continuous bidirectional information exchange between physical infrastructure and computational models, supporting real-time performance evaluation, anomaly detection, and optimisation interventions (Zhang et al., 2020).
1.2 The MEP Systems Digital Twin Paradigm
Digital twins establish virtual replicas of physical assets and processes through real-time data exchange, analytics, and predictive modelling (Grieves, 2022; Zhang et al., 2020). Within MEP contexts, these systems create integrated digital ecosystems where BIM platforms, IoT sensors, and facility management systems interact seamlessly to enable coordinated building operations (AlBalkhy et al., 2024). This integration overcomes the limitations of traditional building management approaches, which often function as isolated systems lacking comprehensive visibility and predictive capability (Tang et al., 2019). By implementing cyber-physical synchronisation, digital twins support real-time tracking, predictive maintenance scheduling, and dynamic optimisation of system performance (Rasheed et al., 2020). They consolidate multidimensional datasuch as temperature readings, vibration signals, equipment loads, fluid dynamics, and environmental conditions, to generate holistic models of mechanical, electrical, and plumbing subsystems (Lu et al., 2020). These models enable early detection of degradation patterns, optimisation of energy use, and proactive maintenance interventions, shifting operational paradigms from reactive to proactive models (Khajavi et al., 2019).
1.3 Problem Statement and Industry Context
Maintainability MEP operation is plagued with endemic inefficiencies due to reactive maintenance paradigms, data fragmentation across separate systems, and subsystem peculiarities, creating predictable operation failures. These functional issues also include the fact that there are high implementation barriers that limit the use of digital twins in varying building typologies and operating scenarios.
Technical Barriers: The technical environment exposes significant integration challenges posed by legacy systems that cannot communicate using compatible communication protocols, the lack of sufficient interoperability between Building Information Modeling (BIM) systems and Internet of Things (IoT) sensor networks, and insufficient data standardisation amongst MEP subsystems (AlBalkhy et al., 2024). It has been found that less than 23 per cent of constructions enjoy effective system integration (Tang et al., 2019), which presents the basic barriers to the operation of the digital twin. The current building management systems tend to be highly isolated systems without integrated data exchange tools to facilitate the use of digital twins. Making it even more difficult to integrate, the MEP equipment of different manufacturers utilizes proprietary protocols and data formats, which makes multivendor integration harder (Rasheed et al., 2020).
Organizational Barriers: The two main reasons why implementation may fail are technical factors and organizational factors such as the lack of in-house technical expertise to implement the digital twins and manage them, the need to change operational paradigms to introduce predictive over reactive maintenance practices, and poor change management processes to facilitate workforce adaptability (Khajavi et al., 2019). Such obstacles present in the typologies of buildings vary towards smaller sizes, with the smaller ones having proportionally more implementation costs in comparison to their budgets, whereas bigger facilities experience more issues in communication across the various departments and stakeholder groups (Lu et al., 2020).
Economic Barriers: Financing cost is an important obstacle to adoption, specifically, the large amount of capital needed to install the sensors, upgrade the infrastructure, license the software, and integrate the system, which is usually 150,000-500,000 dollars in the case of medium-sized business buildings (Cheng et al., 2020). Companies experience confusion about the timeline of returns on investments, the inability to measure less tangible advantages to the organization, including enhanced comfort of occupants and decreased disruption, and the lack of accessible case studies of implementation in similar building settings that might prove it economically feasible (Petri et al., 2025). Facility management budget allocation processes do not always give much attention to investments in long-term technology; they usually focus more on short-term operational requirements, which poses further challenges to the digitization of the digital twin.
Operational Inefficiencies: These obstacles are contributing to the chronic operational inefficiencies. Reactive maintenance strategies result in the downtime of the equipment averaged at 15-20%/year, and the maintenance costs are 30-40% greater than the predictive ones (Cheng et al., 2020). Failure on an emergency basis 3-9 times is more expensive than a planned intervention (Mobley, 2002). Breaking down building data that is spread among BIM files, automation systems, and energy platforms hampers optimization processes, and the integration gaps take 3-7 days to detect faults and energy consumption 12-25 percent higher (Lu et al., 2020). These concerns are enhanced with subsystem vulnerable features: duct leakage decreases the HVAC efficiency by up to 30 percent (Nguyen and Lin, 2023), chiller units fail twice to three times annually, primarily through preventable problems (Van Dinter et al., 2022), undetected plumbing failures consume 14 percent of water loss (Petri et al., 2025), and electrical outlets maintain a downtime of 4.2 hours on average (Van Dinter et al., 2022)
1.4 Research Questions and Objectives
There are three major research questions in this research:
Research Question 1: What are the key technical, organizational, and economic bottlenecks that limit the use of digital twins in MEP systems, and how do these bottlenecks differ based on different building typologies and operational settings?
Research Question 2: What can be estimated in performance improvement and economic advantages of the implementation of digital twins in the MEP operations, i.e., improvement of energy efficiency, reduction of maintenance costs, and a guarantee of equipment reliability, according to the existing literature?
Research Question 3: What are the overall frameworks or planned approaches that can be used to establish a systematic adoption of digital twins in the context of MEP practice, fill in the existing technology gaps, and give industry professionals concrete implementation advice?
Research Objectives:
The research aims to meet the SMART criteria to ensure attainable specifications of measurable goals that add to the development of theoretical knowledge and practical advice on the application of digital twins as a digital technology in MEP systems:
? To conduct a literature review and synthesis of the current literature on digital twins, in so doing as to establish critical gaps to be addressed in research and implementation of MEP systems by means of systematic review methodology based on PRISMA 2020 guidelines.
? To find and modify an appropriate digital twin framework based on the available architectures to be used in the MEP industry, and to modify the framework to meet the needs of mechanical, electrical, and plumbing systems of commercial buildings.
? To prioritize the application of digital twins to MEP systems systematically through an organized assessment scheme that evaluates the business impact dimension and technical feasibility dimension, implementing phase roadmaps that will be optimal in the distribution of resources.
? To come up with practical implementation guidelines that culminate research findings into feasible recommendations that accommodate the demand of stakeholders, in terms of facilities management, engineering, as well as management executives.
The tertiary objective validates framework effectiveness through pilot implementations across multiple building typologies, documenting quantifiable performance improvements in energy efficiency, maintenance cost reduction, and equipment reliability enhancement. The quaternary objective synthesizes findings into actionable implementation guidelines addressing stakeholder-specific requirements across facility management, engineering, and executive decision-making roles, facilitating broader industry adoption of digital twin methodologies.
1.5 Scope of the Study
This study targets the commercial building MEPs, namely, HVAC system, E-distribution network, and plumbing in 10,000-50,000 square meters commercial facilities. The research discusses the applications of the digital twin in three fundamental applications: predictive maintenance of highly critical mechanical systems like chillers, boilers, and air handling units; real-time energy optimization using dynamic setpoints and load balancing mechanisms; and active fault removal in electrical and plumbing systems to improve service outages and safety hazards.
Methodological Scope: The mixed-method design is used and incorporates a systematic literature review in accordance with PRISMA 2020 methods, semi-structured and expert interviews with MEP professionals, and a framework analysis adaptation. The study generates implementation policies grounded on the literature review, consultation with experts, and analysis of the priority scheme.
Geographical Context: The research is based on business buildings in the United Arab Emirates, which provides geographically-specific information about extreme cooling conditions, high-performing buildings, and the new smart-building trends, and acknowledges the principles of transferability to new climatic and infrastructural settings.
Exclusions and Limitations: This study lacks a pilot implementation or empirical validation using a real-world implementation. This research does not cover the area of physical installation, system commissioning, and long-term performance monitoring. The guidelines of framework adaptation and implementation are not drawn through experimental validation, which is based on the knowledge of the literature and consultation with experts. Future studies are to be conducted to conduct the empirical testing via controlled pilot studies to prove the actual performance enhancement.
1.6 Significance of the Study
The research fills pertinent knowledge gaps in the implementation of digital twins in the context of MEP systems in several ways:
Literature Gaps: The study fills the gaps in the literature concerning the lack of an extensive implementation framework across several subsystems of MEPs, which has been reported in 67 percent of existing sources (Metziger, 2023). Although the available literature is on isolated subsystems or particular cases of usage, the implemented research establishes comprehensive implementation protocols or guidelines of HVAC, electrical, and plumbing systems and takes into account their interdependencies and integration needs.
Guidance: Practical Implementation: This study can guide practitioners working in the industry by modifying the existing digital twins by defining how they can be applied to MEP scenarios and outlining systematic prioritization approaches. The dual-axis review model to evaluate the impact of a business and technical feasibility provides repeatable ways of helping an organization evaluate the chances of implementation in a systematic manner and mitigate risks of implementing a system in a phased manner.
Contribution to Regional Context: 89 of empirical studies of digital twins were conducted in North America, Europe, or East Asia (Metziger, 2023), which makes this study add specifics of the Middle East to the answers on the topic and considerations of transferability to the extreme cooling climates, region-specific building standards, and infrastructure maturity status in the UAE commercial building industry.
Bridging Theory to Practice: The work combines the concepts of cyber-physical systems, information integration techniques, and predictive analytics in pure forms into integrated structures that are specifically tailored to the operational environments of MEP. The study provides a link to both theoretical and practical aspects since the digital twins framework adaptation is based on expert consultation and operational challenge analysis, where facility management organizations confront realities of both theory and practice.
Industry Adoption Support: The deliverables are implementation roadmaps, priority matrices, and practical consumption guidelines that help involve industries along with lessening technological and financial hazards. The systematic method of use case prioritization allows an organization to have a phased implementation strategy that proves its value quickly, at the same time, developing the available capacity to develop an increased deployment.
2. Literature Review and Gap Analysis
2.1 Systematic Review Methodology
This is a systematic literature review that is based on PRISMA 2020 principles (Page et al., 2021) with a four-stage process of identification, screening, eligibility evaluation, and final inclusion, which allows the use of methodological rigor and reproducibility.
Identification Phase: Four major academic databases were searched (Web of Science, ASCE Library, Scopus, and IEEE Xplore) with the combination of key terms, including digital twin, MEP systems, building management, predictive maintenance, energy optimization, BIM integration, and IoT sensors. The searches will focus on publications published between the period of September and November 2024 to limit the search that will encompass publications released in the past 1-2 years (January 2019- October 2024) yet remain focused on the current implementations. Preliminary searches provided 247 potentially useful articles in the form of peer-reviewed articles, conference proceedings, technical reports, and industry white papers that covered digital twins related to building systems.
Screening Phase: Inclusion and exclusion criteria were applied in a systematic manner in the assessment of relevance. Publications dealing with digital twin applications in the building context, narrowly construed MEP systems, or the building management functions of which they are aimed, empirical research, case studies, framework creation, or extensive reviews, and published in English within the target timeframe were included. Other types of excluding works included works directly about manufacturing or industrial processes without associated building applications, purely theoretical frameworks lacking any discussion or validation of implementation, and publications about only BIM or IoT technologies and nothing about digital twins, peer-reviewed or technical rigor, and duplicates. Following the screening of titles and abstracts, 89 articles were put under further full-text scrutiny, and 38 articles passed all inclusion screening tests and proved to be of significant importance in terms of relevance to MEP digital twins applications.
Synthesis and Analysis: The last phase of inclusion synthesized year-edited literature, which examined the trend in publications, their thematic areas, geographical locations, and research procedures. The presence of digital twins research in the literature recorded a high growth with 78 percent of the publications published within the 2021-2024 years symbolizing the growing scholarly and commercial attention inspired by the introduction of new IoT technologies, cloud computing infrastructures, and machine learning functionalities. Geographic analysis indicated that 89% of the empirical studies were based on North America, Europe, or East Asia, and thus, they are not representative of the Middle Eastern, African, and South American situations. The thematic categorization revealed three main areas: predictive maintenance (42 percent of publications), energy optimization (34 percent), and lifecycle asset management (24 percent), some of the articles covering several domains simultaneously
2.2 Thematic Analysis of Digital Twin Applications
2.2.1 Predictive Maintenance Applications
Predictive maintenance is the most studied domain, with 16 publications focused on fault detection, anomaly identification, and maintenance optimization using digital twins. Cheng et al. (2020) developed a data-driven predictive maintenance framework for MEP components, integrating BIM geometry with IoT sensor streams and machine learning algorithms, including Random Forest and Support Vector Machine classifiers. Their approach achieved 87.3?ult prediction accuracy for HVAC systems over six months, with lead times between 3.26.8 days, reducing maintenance costs by 34%, unplanned downtime by 41%, and emergency repair incidents by 28%. Nguyen and Lin (2023) examined sensor-based HVAC digital twins, showing prediction accuracy improvements from 76.4% to 91.2% with increased sensor density and one-minute data sampling intervals, providing ROI in 1.42.3 years depending on facility size. Van Dinter et al. (2022) applied hybrid physics-based and data-driven digital twins to chillers, achieving predictive accuracies of 93.7% for refrigerant charge, 89.4% for condenser fouling, and 85.8% for compressor degradation, with 11.2% average energy reductions and extended equipment life by 1824 months.
2.2.2 Energy Optimization Strategies
Energy optimization was addressed in 13 publications, exploring dynamic setpoint adjustment, load balancing, and building energy management.
2.2.3 Asset Lifecycle Management
Asset lifecycle management received comparatively less attention, with nine publications addressing performance tracking, retrofit planning, and long-term optimization. Petri et al. (2025) integrated lifecycle assessment with digital twins, combining operational data and lifecycle inventories to dynamically track environmental impacts. Their approach revealed discrepancies between design and operational energy consumption (1834%) and enabled retrofit opportunities, reducing environmental impacts by 1228% with payback under five years. Khajavi et al. (2019) analyzed digital twin benefits and challenges across industries, highlighting operational efficiency, financial, sustainability, and stakeholder metrics, showing that integrated digital twin approaches yield multi-dimensional performance improvements.
2.3 Identified Research Gaps
Three major gaps were identified. First, 67% of studies lack comprehensive MEP implementation frameworks, providing only subsystem-specific or isolated use cases (Metzger, 2023). While Nguyen and Lin (2023) focus on HVAC fault detection and Cheng et al. (2020) on predictive maintenance, neither offers holistic, multi-subsystem frameworks, making practitioner adoption challenging. Second, 74% of studies rely on simulations or short-duration pilots, limiting validation of long-term performance and economic viability (Lu et al., 2020). Most studies have validation periods under six months, whereas extended operational evidence is required to justify significant digital twin investments. Third, geographic limitations exist, with minimal studies in Middle Eastern contexts, including the UAE, creating uncertainty in framework transferability for extreme cooling climates, local building codes, and infrastructure maturity levels (Tang et al., 2019).
2.4 Digital Twin Frameworks and Standards
Existing frameworks target manufacturing or generalized buildings, lacking MEP-specific guidance. ISO 23247 defines digital twin architectures for manufacturing but requires adaptation for building-specific considerations, including distributed sensors, BIM integration, and facility management processes (ISO, 2021). Rasheed et al. (2020) provide conceptual taxonomies for data-driven, physics-based, and hybrid twins, but lack practical implementation guidance. Jones et al. (2020) review common layersasset, integration, analytics, presentationbut emphasize variability across domains. MEP systems heterogeneity, subsystem interdependencies, spatial complexity, and long operational lifecycles necessitate tailored frameworks incorporating BIM, long-term performance tracking, and multi-stakeholder decision support. This research addresses these requirements through a validated, comprehensive architecture optimized for MEP operational contexts (Zhang et al., 2020).
3. Methodology: Problem Identification and Use Case Prioritization
3.1 Research Design and Methodological Framework
This study adopts a mixed-methods research design integrating quantitative and qualitative approaches to develop, validate, and assess digital twin frameworks for MEP systems. The methodology follows four interlinked phases: systematic problem identification via expert consultation, use case discovery and prioritization through structured evaluation, framework identification incorporating technical architecture and implementation specifications, and empirical validation through pilot implementations with quantitative performance assessment. This multi-phase approach ensures a comprehensive understanding of operational challenges, systematic prioritization of implementation opportunities, technically rigorous framework development, and evidence-based validation demonstrating practical effectiveness and economic viability. Information flows from problem identification to framework development and validation processes, as illustrated in the research design diagram.
3.2 Industry Consultation and Problem Discovery
The expert consultation was the preliminary stage of defining the operational issues and implementation demands unique to MEP digital twin implementation. Purposive sampling was used to recruit 23 MEP professionals, who comprised project engineers (n=7), facility managers (n=6), maintenance technicians (n=5), and BIM professionals (n=5) working at eight commercial buildings that were between 12,000 and 47,000 square meters. Participants who were aged 5 years or more of MEP experience working life, and had firsthand experience in facility management, maintenance operations, or building automation equipment were contacted. Types of buildings were Class A office buildings (n=4) and mixed-use commercial developments (n=2), hospitality facilities (n=1), and institutional buildings (n=1) that offered a variety of operational viewpoints.
Data Collection: Semi-structured interviewing taking place in October-November 2025 (25-45 minutes each) will follow standardized guidelines and will also leave the possibility of exploring the topics and issues highlighted by participants. The questions asked in interviews included seven topics, namely existing monitoring capabilities and technological infrastructure, problematic aspects of fault detection and response process, the effectiveness and efficiency of maintenance and resources allocation, practice of performance measurement and key performance indicators, interdepartmental communication and coordination mechanisms, training and tools requirements in digital twin adoption, and priorities of the stakeholders to digital twin implementation.
Interviews: The interviews were also audio-taped with the consent of respondents, transcribed word-for-word, and analyzed by using thematic coding analysis to recognize the recurrence of themes, categories of challenges, and areas of implementation. The qualitative analysis used an iterative process of coding. First, there was open coding, which identified emergent themes; then, there was axial coding, which determined any relationship between the categories, and selective coding, which incorporated the findings in coherent challenge frameworks.
Quantitative Supplementation: The structured questionnaires were used to supplement the qualitative interviews that quantified the piece of equipment failure rates, maintenance expense rates, energy consumption rates, and the impact of downtime in regards to Likert scales and numerical methods of estimation. The descriptive analysis of quantitative data was performed to provide baseline performance metrics and determine the areas of priority to use the digital twin intervention.
Key Findings: HVAC problems were cited every month or less often than electrical and plumbing problems (87 vs. 52 vs. 35), respectively, meaning that concerns were more urgent in subsystems. Maintenance spending was dominated by reactive repairs (62-78 of total maintenance spending), preventive maintenance (18-29 of total maintenance spending), and predictive ones, only 4-9, which proves the leading reactive paradigms of operations in the surveyed facilities.
3.3 Thematic Analysis of Operational Challenges
Interpretation of interview transcripts and questionnaire data showed 3 general categories of challenges affecting the operations of the MEP: reactive operational paradigms that limit the ability to act proactively, data fragmentation that inhibits holistic performance measurement, and vulnerabilities at the system level that introduce repetitive failure patterns.
HVAC Dilemmas (Interview-Derived Results): HVAC systems became the worst subsystem in specialist consultations. Seventy-four percent of the interviewed people mentioned duct leakage, and conditioned air loss was estimated to be between 15-35 percent as a result of the operating observations and intermittent evaluation. According to the participants, chiller breakdowns result in 2-4 unintended downtime episodes in one cooling season, and coil soiling occurs in 30-40 percent of air handling units and costs 15-25 percent more in terms of energy use as measured by utility bills. Poor load balancing, concurrent heating and cooling of various zones, and system-implemented schedules of building management systems were often mentioned as factors that curbed the potential of energy efficiency. The results are consistent with the literature that has recorded these HVAC problems (Nguyen and Lin, 2023; Cheng et al., 2020).
Plumbing Problem (Interviews-etsiobfidds): 91% of the facility managers interviewed described leak detection as rather reactive. The existence of concealed leaks in above-ceiling or below-slab installations was generally only realised after 2-6 weeks, when water damage became observable or water bills registered anomalous increases that were significant water wastage, and the property was damaged to the extent of incurring a costly cleanup when remediation was necessary. The sensor coverage was also limited, and limited physical access restricted the ability to use continuous monitoring, and the development of predictive maintenance protocols could not be developed due to fragmented monitoring of historical failures. These operational problems are in line with the water management problems that have been reported in the building performance literature (Petri et al., 2025).
Electrical Challenges (Interview-Based Findings): Less frequently reported and with high impact on operations when they happened were electrical problems. Distribution panels, switchgear, or main feeders failures led to service interruptions between 15 minutes (small problem) and more than 8 hours (large component failures) that required replenishment. Years of wear and tear on electrical devices (20-40 years in some of the buildings surveyed), along with an almost entirely manual method of monitoring, made predictive intervention potential difficult. The regular analysis of the load patterns, the harmonic distortion, and the temperature was not commonly introduced, and the majority of facilities were simply relying on the regular manual inspection instead of the continuous inspection due to the monitoring systems (Khajavi et al., 2019).
BMS Limitations (Interview-Based Mirrors): The creation of management systems was deemed to be mostly reactive by 91% of respondents, who provided alarm notification when systems failed, and not proactive measures to prevent the evolution of issues. The lack of integration of BIM models with the real-time data of IoT sensors and the lack of data fragmentation across various platforms (HVAC control systems, energy management software, maintenance management systems) limited and made it difficult to optimize the entire performance holistically. Most of the surveyed facilities had limited remote access and mobile interfaces, which limited the operational flexibility and response functionality during after-hours (Rasheed et al., 2020; Jones et al., 2020).
3.4 Use Case Identification and Prioritization Framework
Consultation identified six candidate digital twin use cases addressing specific operational challenges: predictive maintenance, real-time optimization, anomaly detection, and lifecycle management. A dual-axis evaluation framework assessed business impact and technical feasibility.
Business impact criteria (weighted scores): safety impact (25%), downtime reduction (25%), throughput/operational capacity (15%), quality/customer satisfaction (20%), cost savings (15%). Scores ranged from 110, informed by literature and operational data.
Technical feasibility criteria (weighted scores): data availability/quality (25%), integration readiness (20%), analytics maturity (20%), deployment scope/complexity (20%), project readiness (15%). Scores ranged from 110, reflecting implementation challenges and organizational preparedness.
Evaluation prioritized use cases into three deployment phases. Phase 1 MVP included high-impact, high-feasibility applications: Real-time Energy Optimization, Predictive Maintenance for Chillers, and Water Leak Detection. Phase 2 involved medium-impact or medium-feasibility use cases: Asset Lifecycle Management and Fire/Smoke Simulation. Low-impact, low-feasibility use cases, such as Noise and Vibration Anomaly Detection, were deferred.
Table 3.1: Use Case Prioritization Matrix
Use Case Business Impact Score Technical Feasibility Score Phase / Recommendation
Real-time Energy Optimization 7.4 8.0 Phase 1 Immediate Deployment
Predictive Maintenance for Chillers 8.0 6.6 Phase 1 Immediate Deployment
Water Leak Detection 7.8 6.8 Phase 1 Immediate Deployment
Asset Lifecycle Management 6.6 6.4 Phase 2 Subsequent Deployment
Fire/Smoke Simulation 6.0 5.8 Phase 2 Subsequent Deployment
Noise & Vibration Anomaly Detection 4.8 4.2 Deferred / Future Consideration

This systematic prioritization ensures evidence-based deployment aligned with organizational capabilities, strategic priorities, and phased risk management, establishing a robust foundation for MEP digital twin implementation.

Figure 3.1 Digital Twin Priority Matrix.
4. The Proposed Digital Twin Framework
4.1 Framework Architecture and Theoretical Foundation
The modified digital twin framework uses a four-layer system based on the available literature in digital twins and specifically adapted to the MEP system use. This framework combines the concepts of cyber-physical systems (Zhang et al., 2020), data integration solutions (AlBalkhy et al., 2024), and predictive analytics solutions (Rasheed et al., 2020) into an operational ecosystem that fills the gaps detected as gaps in the literature review and expert consultation.
Framework Adaptation Approach: This study does not make any claims of novelty but directly adapts the concept of layered architecture, which has been widely discussed in digital principal digital twins literature, as reported by Jones et al. (2020), who found common layers such as asset layer, integration layer, analytics layer, and presentation layer in a variety of digital twins applications. The process of adaptation entailed:
? Literature-Based Foundation: The literature review of the architectural frameworks provided in ISO 23247 manufacturing standards (ISO, 2021), the taxonomies introduced by Rasheed et al. (2020), and the implementations specific to the building will be done to determine the essential architecture principles.
? MEP-Specific Tailoring: Customisation of the generic framework components to suit MEP-specific needs, such as distributed sensor networks between mechanical, electrical, and plumbing subsystems; BIM integration to provide spatial visualisation and equipment placement mapping; facility management process integration to support maintenance workflow; and multi-stakeholder interface needs to support facility managers, maintenance technicians, building owners, and engineering consultants.
? Operational Context Integration: Incorporating operational challenges evidenced by consulting expert opinions to provide a structured framework to cover the pain-points identified, such as operational paradigms, data fragmentation, and system susceptibility.
Four-Layer Architecture: The modified design consists of four levels of architecture giving structural separation as well as allowing a smooth flow of information: Layer 1 (Physical Assets and Sensor Installation) supports hardware infrastructure and data acquisition features; Layer 2 (Data Communication and Integration) supports real time infusion of information and interoperability, and Layer 3 (Digital Twin Analytics Engine) transforms raw data into actionable insights with the help of predictive models and Layer 4 (User Interface and Feedback Integration) delivers information to stakeholders by in Figure 4.1.
The architecture fills the literature gaps, such as incomplete integration between BIM platforms, IoT networks, and facility management systems, and promotes the modular deployment design that enables the implementation to be divided into phases and be consistent with organization readiness and resources availability (Tang et al., 2019).
4.2 Layer 1: Physical Assets and Sensor Installation
Layer 1 establishes foundational hardware, encompassing MEP equipment instrumentation, IoT sensor deployment, and data acquisition systems capturing real-time operational parameters. HVAC systems utilize temperature, vibration, pressure, flow, energy, and humidity sensors to monitor equipment and environmental conditions. Electrical distribution employs energy meters, temperature sensors, and power quality analyzers, while plumbing systems deploy flow, pressure, and water quality sensors (Nguyen & Lin, 2023; Cheng et al., 2020). This network enables predictive maintenance and performance optimization beyond traditional building management capabilities. IoT platforms such as Microsoft Azure IoT Hub, AWS IoT Core, and Google Cloud IoT provide connectivity, device management, and secure data ingestion. Supported protocols include MQTT, BACnet, Modbus, and OPC UA, ensuring interoperability and scalability across facility portfolios (Rasheed et al., 2020).
4.3 Layer 2: Data Communication and Integration
Layer 2 manages real-time data flows and ensures data quality, security, and accessibility. Time-series databases (e.g., InfluxDB, TimescaleDB, Azure Time Series Insights) handle high-frequency sensor data, while relational databases (PostgreSQL, MS SQL Server) store structured asset information and analytics results. Standardized protocols, including MQTT, BACnet, OPC UA, and REST APIs, facilitate seamless integration between sensors, BIM platforms, and facility management systems. BIM integration enables spatial visualization, equipment location mapping, and maintenance support using IFC formats or proprietary APIs (AlBalkhy et al., 2024).
4.4 Layer 3: Digital Twin Analytics Engine
Layer 3 converts raw data into actionable insights via predictive maintenance, energy optimization, anomaly detection, and performance benchmarking. Platforms include Microsoft Azure Digital Twins, Siemens Simcenter Amesim, and Autodesk Tandem, each offering IoT integration, simulation, or native BIM support (Rasheed et al., 2020). Machine learning models, such as Random Forest classifiers, predict faults with high accuracy (Cheng et al., 2020). Anomaly detection combines statistical control and machine learning techniques to identify operational irregularities, prioritizing alerts for safety and cost impact (Van Dinter et al., 2022).
4.5 Layer 4: User Interface and Feedback Integration
Layer 4 delivers dashboards, alerts, AR/VR interfaces, and automated reporting to stakeholders. Dashboards visualize KPI trends, energy performance, and equipment health for managers, technicians, owners, and engineers (Khajavi et al., 2019). Alerts are prioritized by severity with multi-channel escalation. AR/VR interfaces support spatial visualization, maintenance guidance, and training, enhancing operational efficiency (AlBalkhy et al., 2024). Automated daily, weekly, and monthly reports provide performance tracking, compliance verification, and continuous improvement insights (Petri et al., 2025).

Figure 4.1 Digital Twin Framework.
4.6 Implementation Outcomes and Literature-Based Projections
Critical Clarification: In this study, pilot implementations as well as empirical validation studies are not carried out. The metrics of performance and economic impacts mentioned here are based on available literature that reports on the application of digital twins in similar building scenarios. These theoretical forecasts based on the literature explain what kind of performance can be improved with the proper use of the modified framework, instead of showing the results of this research undertaken empirically.
The projections of the performance based on the literature are shown below:
Energy Performance: With intelligent scheduling and load shifting depending on occupancy forecasts and utility rate structures, intelligent scheduling and load shifting have been shown to reduce peaks by an average of 14.7 percent. The carbon footprint is decreased by a percentage of about 19.8, which is caused by the enhanced energy saving and efficient running of tools.
Predictive Maintenance Performance: It has been found that changing the paradigms of maintenance to predictive multiannual improvement in maintenance costs by 42.1 percent dictates that the reactive maintenance paradigm minimizes the number of emergency repairs per year and allows planned maintenance to be conducted at the most suitable time (Cheng et al., 2020). Enhancement in the Mean Time Between Failures (MTBF) of 48.6% is due to the identification of faults early on and preventive measures taken before devastating failures take place. The results of the maintenance auto-guided by AR with accuracy in fault diagnosis achieve a reduction of Mean Time To Repair (MTTR) by 43.4 percent (Nguyen & Lin, 2023). Downtime Savings (394 at baseline and 569 at optimized), Fastest repairs (890 at baseline and 1300 at optimized), and Equipment reliability (78.2% at baseline and 91.4% at optimized) combine to increase the overall Equipment Effectiveness (OEE) in the resultant state (Van Dinter et al., 2022).
Economic Viability: According to the available economic viability literature, internal rates of return are as high as 97.4, and payback periods are in the range of 12-14 months in medium-sized commercial buildings (15,000-35,000 square meters) which indicates strong economic viability under the conditions of balancing the implementation costs with the cost of operational savings (Khajavi et al., 2019; Petri et al., 2025). Yet, the financial forecasts differ significantly depending on the facility size, infrastructure maturity and maturity, energy expenses, and choice of particular use cases.
Framework Parity Strategy: Although this study lacks empirical validation, the theoretical soundness of the adapted framework is justified by: (1) by basing it on existing literature on digital twins and tested architecture design patterns, (2) ensuring operational relevance by consulting experts, (3) prioritizing use cases systematically by considering feasibility and impacts, and (4) by correlation with reported performance benefits in similar implementations. Subtler studies must be undertaken in the future that will aim at empirical validation by conducting controlled pilot implementations to show actual improvements in performance that can be attained with this particular framework adaptation.
Table 4.1 illustrates the hierarchical structure, linking physical infrastructure to actionable insights for operational optimization.
Layer Component System/Stakeholder Method/Tool Purpose/Outcome
1 Temperature Sensor HVAC IoT Sensor Monitor air/water temperature
1 Energy Meter Electrical IoT Sensor Track energy consumption
3 Random Forest MEP Equipment ML Algorithm Predict faults (Accuracy 85.4%)
3 Dynamic Setpoint HVAC Control Strategy Reduce energy use by 8.2%
4 Dashboard Facility Manager Web Interface KPI monitoring & alerts
4 AR Overlay Technician AR App Locate & service equipment efficiently

5. Conclusion and Strategic Recommendations
The study shows that digital twin technology is a revolutionary prospect in streamlining the MEP system that has tremendous potential to enhance performance in addition to economic gains when a systematic approach to integrating BIM systems, IoT sensor networks, predictive analytics, and operational management processes is implemented. The modified four-layer model identifies these critical gaps, some of which were revealed in the systematic literature review, such as the lack of implementation frameworks across various subsystems of MEP, the scarcity of practical advice available to industry participants on the matter, and the inability to reach across geographical boundaries by available research.
5.1 Key Contributions
Literature Gap Identification: The study by Metziger (2023) comprises three primary gaps: 67% of the studies do not have detailed MEP implementation frameworks only give subsystem-specific options (Metziger, 2023); 74% of studies use simulation or temporary pilots that do not progress past short-term management of sustainability (Lu et al., 2020); and 89% of the empirical studies are focused on developed areas and offer very limited empirical evidence on Middle East contexts.
Framework Adaptation: In this study, it is proposed to indeed adapt current concepts of digital twins architecture to the MEP application, to fit generic layered frameworks used in manufacturing and general architecture care contexts with MEP specifics, such as distributed sensor networks, integration of BIM in spatial complexity management, alignment of facility management processes, and the need for multistakeholder interfaces. The adaptation process bases the theoretical framework concepts on the operational realities that are found after consulting the experts.
Use Case Prioritization Methodology: The methodology of systematic prioritization was derived by a dual axis scoring of business impact and technical feasibility axis and this gives replicable ways to organizations to judge the implementation opportunities. The ensuing phased deployment plan offers companies to mitigate risk with high-priority use cases and prove value with high-priority use cases that can be extended to wider applicability.
Guidelines to Implementation: Pragmatic guidelines are based on literature results, consultations, and prioritization analysis into actionable guidelines for the facility management organizations, engineering consultants, and vendors of technology. These recommendations deal with the needs of the stakeholders and offer systematic direction to the implementation of the digital twins.
5.2 Limitations and Future Research
There are no provisions of empirical validation with pilot implementation in this research. The projections of performance that are presented are based on the available literature as opposed to actual experimental confirmation of the adapted frame. This limitation should be addressed in future research undertakings comprising controlled piloting studies across a number of building typologies where actual performance improvements have been realized, implementation difficulties, and the demands of the change management of the organization in question.
Future directions of research are: exploring the persistence of long-term performance in multi-year operations to support sustained benefits; exploring transferability between climatic regions and typologies, organizational settings; exploring how to integrate with new technologies, such as edge computing, 5G networks and more advanced AI capabilities; and exploring scalability over building portfolios consisting of varied types of buildings with different operational characteristics.
5.3 Practical Implications
To practitioners in the industry, this research offers implementation guidance that is based on evidence and minimizes the technological and financial risks involved in the digital twin implementation. Phased deployment strategy, priority methodology, and modified framework architecture provide effective tools to organizations that make decisions on implementing digital twins. With building performance requirements growing in intensity due to changing energy code requirements, carbon pricing measures, and sustainability pledges, organisations that achieve digital twin architecture place themselves in excellent positions to achieve operational efficiencies, minimise costs, and be good environmental custodians in an ever-competitive built environment.

References
AlBalkhy, W., Karmaoui, D., Ducoulombier, L., Lafhaj, Z., & Linner, T. (2024). Digital twins in the built environment: Definition, applications, and challenges. Automation in Construction, 162, 105368. https://doi.org/10.1016/j.autcon.2024.105368
Cheng, J. C., Chen, W., Chen, K., & Wang, Q. (2020). Data-driven predictive maintenance planning framework for MEP components based on BIM and IoT using machine learning algorithms. Automation in Construction, 112, 103087. https://www.sciencedirect.com/science/article/abs/pii/S0926580518308562
Grieves, M. (2022). Intelligent digital twins and the development and management of complex systems. Digital Twin, 1(1). https://doi.org/10.12688/digitaltwin.17574.1
ISO. (2021). ISO 23247-1:2021Automation systems and integration Digital twin framework for manufacturing. International Organization for Standardization. Technical Committee: ISO/TC 184/SC 4, ICS:25.040.40 35.240.50
Jones, D., Snider, C., Nassehi, A., Yon, J., & Hicks, B. (2020). Characterising the digital twin: A systematic literature review. CIRP Journal of Manufacturing Science and Technology, 29, 36-52. https://doi.org/10.1016/j.cirpj.2020.02.002
Khajavi, S. H., Motlagh, N. H., Jaribion, A., Werner, L. C., & Holmstrm, J. (2019). Digital twin: Vision, benefits, challenges, and key performance indicators. IEEE Access, 7, 147063-147081. 10.1109/ACCESS.2019.2946515
Lu, Y., Liu, F., Wang, K., Wang, L., & Xu, X. (2020). Digital twin-driven smart manufacturing: Connotation, reference model, key technologies, and applications. Robotics and Computer-Integrated Manufacturing, 61, 101837. 10.1016/j.rcim.2019.101837
Metzger, D. (2023). A systematic literature review on digital twins in MEP systems. FH Wedel. https://www.fh-wedel.de/fileadmin/Mitarbeiter/Records/Metzger_2023_-_A_Systematic_Literature_Review_on_Digital_Twins.pdf
Mobley, R. K. (2002). An introduction to predictive maintenance. Butterworth-Heinemann.
Nguyen, T., & Lin, H. (2023). Sensor-based digital twins for HVAC fault prediction. Sensors, 23(24), 9786. https://doi.org/10.1016/B978-075067531-4/50006-3
Pan, Z., & Guo, F. (2025). Digital twin (DT) development strategies in the architecture, engineering, and construction (AEC) industry based on SWOT-AHP analysis. Journal of Building Design and Environment. https://doi.org/10.70401/jbde.2025.0006
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hrbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., ... Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71
Petri, I., Amin, A., Ghoroghi, A., Hodorog, A., & Rezgui, Y. (2025). Digital twins for dynamic life cycle assessment in the built environment. The Science of the Total Environment, 993, 179930. https://doi.org/10.1016/j.scitotenv.2025.179930
Rasheed, A., San, O., & Kvamsdal, T. (2020). Digital twin: Values, challenges, and enablers from a modeling perspective. IEEE Access, 8, 21980-22012. 10.1109/ACCESS.2020.2970143
Tang, S., Shelden, D. R., Eastman, C. M., Pishdad-Bozorgi, P., & Gao, X. (2019). A review of building information modeling (BIM) and the internet of things (IoT) devices: Present status and future trends. Automation in Construction, 101, 127-139. https://doi.org/10.1016/j.autcon.2019.01.020
Van Dinter, R., Tekinerdogan, B., & Catal, C. (2022). Predictive maintenance using digital twins: A systematic literature review. Information and Software Technology, 151, 107008. https://doi.org/10.1016/j.infsof.2022.107008
Zhang, J., Zhao, L., Ren, G., Li, H., & Li, X. (2020). Digital twin technology in the AEC industry. Advances in Civil Engineering, 2020, 8842113. https://doi.org/10.1155/2020/8842113