Introduction: The Pulse of a Smart Nation’s Infrastructure

 

In the hyper-dense, meticulously planned urban landscape of Singapore, the health and longevity of its infrastructure are not merely engineering concerns; they are matters of national resilience, economic vitality, and public safety. Every bridge, skyscraper, tunnel, and public housing block is a critical asset in a complex, interconnected system. For decades, ensuring the integrity of these structures relied on periodic, manual inspections—a process that was often labour-intensive, subjective, and reactive. Today, a technological revolution is underway, transforming how the city-state watches over its built environment. This revolution is powered by the convergence of Structural Health Monitoring (SHM), Artificial Intelligence (AI), and Machine Learning (ML).

This report provides an exhaustive analysis of the application of AI and ML in SHM within the specific context of Singapore. It explores the foundational principles of SHM, delves into the advanced AI technologies being integrated into monitoring systems, and examines the unique ecosystem that makes Singapore a global leader in this domain. We will dissect the top-down government policies, from the ambitious Smart Nation initiative to the detailed frameworks of the Building and Construction Authority (BCA), that are catalysing this transformation. We will spotlight the cutting-edge research emerging from local powerhouses like the National University of Singapore (NUS) and Nanyang Technological University (NTU), and survey the commercial landscape of service providers bringing these innovations to market. Through local case studies—from the deep tunnels of the MRT to the automated inspection of building façades—this report will illustrate how AI-powered SHM is moving beyond theory and into practice.

Ultimately, this analysis will demonstrate that Singapore’s approach is more than just technological adoption; it is the deliberate creation of a self-reinforcing ecosystem where policy, research, and industry converge to build a safer, more efficient, and truly intelligent built environment. We begin by establishing the fundamentals of SHM, the bedrock upon which this new digital watchtower is built.

Section 1: The Foundation: Understanding Structural Health Monitoring in the Urban Landscape

 

Before delving into the complexities of artificial intelligence, it is essential to establish a firm understanding of Structural Health Monitoring (SHM) itself. SHM represents a paradigm shift in how we approach the lifecycle management of civil infrastructure. It moves away from the traditional, reactive model of “inspect and repair” towards a proactive, data-driven philosophy of continuous assessment and management.

 

1.1. Defining Structural Health Monitoring (SHM): From Concept to Necessity

 

At its core, Structural Health Monitoring is the process of observing and analyzing an engineering structure over time using periodically sampled response measurements to monitor changes to its material and geometric properties.1 This process involves the integration of sensors, data acquisition systems, and analysis models to provide a continuous or periodic diagnosis of the structure’s “state of health”.2 It is crucial to distinguish SHM from traditional Non-Destructive Evaluation (NDE). While NDE is typically a one-off, out-of-service inspection to find a specific flaw, SHM is an in-service, automated process designed to track the evolution of a structure’s condition throughout its operational life.4

The primary objective of SHM is to provide quantifiable performance data and early warnings of potential structural problems.6 This allows asset owners and engineers to intervene in a timely manner, which yields several critical benefits:

• Enhanced Safety: By detecting damage or degradation at its onset, SHM helps prevent catastrophic failures that could endanger lives and property.7 This is particularly vital in densely populated cities like Singapore, where the failure of a single piece of critical infrastructure could have devastating societal and economic consequences.9
• Economic Advantages: SHM facilitates a move from time-based maintenance (e.g., inspecting a bridge every five years) to more efficient condition-based maintenance (CBM), where repairs are performed only when necessary.3 This minimizes costly downtime, reduces unnecessary inspection and repair costs, and can significantly extend the operational lifespan of an asset.6
• Rapid Post-Event Assessment: After extreme events like earthquakes or blast loading, SHM systems provide near real-time information on a structure’s integrity, enabling rapid decisions about its safety and re-occupancy.5

The evolution of SHM has been remarkable. Its conceptual roots can be traced back to simple, qualitative methods, such as the 19th-century practice of railroad wheel-tappers using the sound of a hammer strike to evaluate a wheel’s integrity.1 Today, this has evolved into a highly sophisticated, multi-disciplinary field that integrates advanced sensing technologies, data science, and computational power directly into the fabric of our infrastructure.2

 

1.2. The Anatomy of an SHM System: Components and Workflow

 

A modern SHM system can be understood as a data pipeline, transforming raw physical responses into actionable intelligence. The system is comprised of several key integrated components, each playing a crucial role in this process.

• Sensors: These are the “sensory nerves” of the structure, responsible for measuring its physical responses. The selection and placement of sensors are critical for ensuring accurate and reliable data collection.6 A wide array of sensor types are used, each suited for different measurements:

• Accelerometers and Vibration Sensors: Measure dynamic movements and are fundamental to vibration-based monitoring techniques.7
• Strain Gauges: Measure the deformation or strain on a structural member’s surface, indicating stress levels.7
• Inclinometers and Tiltmeters: Measure the tilt or rotation of structural components, often used to monitor settlement or deformation in buildings and retaining walls.7
• Piezometers: Measure pore water pressure in the ground, crucial for geotechnical applications like deep excavations and dams.7
• Optical Fiber Sensors: Use light transmitted through fiber optic cables to measure strain and temperature with high precision, often over long distances, making them ideal for monitoring large structures like bridges and tunnels.13
• Piezoelectric (PZT) Transducers: Smart materials that can act as both sensors and actuators, forming the basis for high-frequency wave propagation and impedance-based methods.4

• Data Acquisition Systems (DAS): This is the hardware that collects the analog signals from the sensors and converts them into digital data. A typical DAS includes a digital recorder, often with its own internal battery backup to ensure operation during power outages, and an on-site server for initial data processing and storage.1
• Data Transmission and Storage: Once collected, the data must be transmitted to a central location for analysis. This can be achieved through wired connections or, more commonly, wireless telemetry using DSL or cellular modems.11 For security, the SHM network is often kept completely separate (“air-gapped”) from the facility’s main network to prevent unauthorized access.11 Data is then stored in databases, either on-premise or in the cloud, forming a historical record of the structure’s behavior.2
• Data Interpretation and Diagnosis: This is the “brain” of the SHM system, where sophisticated algorithms process the vast amounts of raw data to extract damage-sensitive features and convert them into practical, tangible information about the structure’s health.1 This stage involves statistical analysis, signal processing, and, increasingly, the application of AI and machine learning models. The ultimate goal is to detect the existence of damage, locate it, identify its type, and quantify its severity.1
• Alerting and Information Dissemination: The final step is to communicate the findings to the relevant stakeholders. Modern SHM systems provide automated alerts via email or SMS when predefined thresholds are exceeded or anomalies are detected.11 The information is often presented on user-friendly web-based dashboards with smart graphics and reports, enabling engineers and asset managers to make informed and timely decisions.8 A prime example of this complete workflow is the OpenSHM platform developed by the U.S. Geological Survey, which can notify users of potential structural damage within 3-5 minutes of an earthquake.11

This entire process underscores a fundamental characteristic of modern SHM: it is an inherently data-centric discipline. The integrity of the final diagnosis is wholly dependent on the quality of the data pipeline, from the precision of the initial sensor measurement to the robustness of the final analytical model. This data-centricity creates a “big data” challenge, as continuous monitoring can generate enormous datasets.17 However, it is precisely this challenge that makes SHM a perfect domain for the application of AI and machine learning, which thrive on large volumes of data to uncover hidden patterns and insights. The successful implementation of AI in SHM is therefore fundamentally a data engineering challenge before it is an AI challenge.

 

1.3. Core SHM Methodologies: A Comparative Overview

 

The data interpretation stage of an SHM system relies on a variety of scientific methods to diagnose a structure’s health. These methodologies operate on different physical principles and are sensitive to different types of damage.

• Vibration-Based Methods: This is one of the most widely used approaches for global damage detection. It is based on the principle that a structure’s dynamic characteristics—its natural frequencies, mode shapes (the patterns of vibration), and damping—are functions of its physical properties (mass, stiffness, and energy dissipation).4 When damage such as a crack occurs, it reduces the structure’s stiffness, which in turn causes measurable changes in these modal parameters. By monitoring for these changes over time, engineers can detect that damage has occurred somewhere in the structure.19
• Wave Propagation Methods (Acousto-Ultrasonics): These are active methods that use actuators, typically PZT patches, to generate high-frequency elastic waves (like Lamb waves) that travel through the structure.4 A network of sensors measures the response. If the wave encounters a discontinuity like a crack or delamination, it will be reflected, scattered, or its travel time will change.4 By analyzing these changes, it is possible to not only detect but also locate the damage with high precision.
• Acoustic Emission (AE): This is a passive monitoring technique. Instead of generating its own signal, it “listens” for the transient elastic waves (acoustic emissions) that are generated by the sudden release of energy within a material, such as when a crack grows or a fiber breaks in a composite material.4 AE is excellent for detecting the initiation and propagation of active damage in real-time.
• Electromechanical Impedance (EMI) Method: This is a high-frequency technique that leverages the electromechanical coupling property of PZT materials. A PZT patch is bonded to the structure and driven by a small alternating voltage over a range of frequencies, causing it to vibrate. The structure’s mechanical impedance (its resistance to motion) directly affects the PZT’s electrical impedance (the ratio of voltage to current).4 Because the sensing area is localized, the EMI method is highly sensitive to incipient, local damage like a loose bolt or a micro-crack, which will cause a measurable shift in the electrical impedance signature.4
• Optical Fiber Sensing: This method uses optical fibers as the sensing element. One common technique is Brillouin optical time-domain reflectometry (BOTDR), where pulses of light are sent down a fiber bonded to or embedded in a structure. Changes in strain and temperature along the fiber cause a shift in the frequency of the backscattered light.13 By analyzing this frequency shift, it is possible to obtain a continuous profile of strain and temperature along the entire length of the fiber, making it an extremely powerful tool for monitoring large-scale infrastructure like bridges, tunnels, and pipelines.13

The selection of a particular methodology depends on the type of structure, the damage to be detected, and the operational environment. The table below provides a comparative summary of these core techniques.

 

Methodology	Underlying Principle	Typical Frequency Range	Damage Detection Capability	Key Advantages	Limitations/Challenges
Vibration-Based	Damage alters global stiffness and mass, changing modal parameters (frequency, mode shapes). 4	Low (<100 Hz)	Global, sensitive to moderate-to-severe damage.	Well-established; can use ambient vibrations; good for overall structural assessment.	Low sensitivity to incipient/local damage; highly affected by environmental noise (temperature, traffic). 24
Wave Propagation	Damage (e.g., cracks) reflects or scatters high-frequency guided waves. 4	High (20 kHz – 1 MHz)	Local, high sensitivity to small-scale damage (cracks, delamination).	Good for damage localization; can inspect areas inaccessible to other methods.	Complex wave behavior; performance is material and geometry dependent. 4
Acoustic Emission (AE)	Passively detects stress waves released by active damage growth (e.g., cracking). 4	High (100 kHz – 1 MHz)	Local, real-time detection of active damage events.	Passive (low power); detects damage as it happens.	Cannot detect static/pre-existing damage; requires low operational noise.
Electromechanical Impedance (EMI)	Local damage alters the mechanical impedance of the structure, measured as a change in the electrical impedance of a PZT sensor. 4	High (30 kHz – 400 kHz)	Localized, extremely high sensitivity to incipient damage (e.g., loose bolts, micro-cracks).	Very sensitive to early-stage damage; insensitive to boundary conditions. 4	Small sensing area per sensor; requires baseline data.
Optical Fiber Sensing	Changes in strain and temperature along the fiber alter the properties (e.g., frequency) of backscattered light. 13	Static to Low	Distributed, continuous measurement of strain/temperature over long distances.	Can monitor large structures with a single sensor; immune to electromagnetic interference. 25	Higher initial cost; requires careful installation.

The availability of these diverse and powerful monitoring techniques lays the groundwork for a profound transformation in asset management. The ability to generate continuous streams of data about a structure’s health enables a fundamental shift from a reactive to a proactive maintenance philosophy. Instead of relying on fixed schedules and responding to failures after they occur, asset owners can now use real-time data to anticipate problems, schedule maintenance precisely when needed, and make informed decisions that optimize safety, cost, and performance over the entire lifecycle of the asset.6 This shift extends the role of the structural engineer from a designer and inspector to a long-term manager of structural performance and risk.5 This data-driven approach is the essential prerequisite for the next leap forward: the integration of artificial intelligence.

Section 2: The Intelligence Layer: Integrating AI and Machine Learning into SHM

 

While advanced sensors and data acquisition systems provide the raw data, the true value of modern SHM is unlocked through intelligent analysis. This is where Artificial Intelligence (AI) and Machine Learning (ML) have become indispensable. They provide the computational power to sift through massive, complex datasets and extract the subtle signatures of damage that are often invisible to the naked eye or traditional analysis methods.

 

2.1. The AI Catalyst: Why SHM is a Prime Candidate for Machine Learning

 

The core challenge in SHM is one of statistical pattern recognition.10 The goal is to distinguish the patterns in sensor data that indicate damage from those caused by benign operational and environmental variations, such as changes in temperature, humidity, or traffic loading.24 This is an incredibly difficult task for human analysts, as damage-induced changes can be very subtle and buried in noise.27

This is precisely the type of problem that AI and ML algorithms are designed to solve. They excel at learning complex, non-linear relationships directly from data, allowing them to automatically identify damage-sensitive features and build robust diagnostic and prognostic models.17 The increasing availability of low-cost sensors and IoT devices has led to an explosion in the volume of data collected from structures, creating a “big data” environment where AI’s ability to handle vast datasets provides a decisive advantage over traditional methods.17 Consequently, the SHM field is rapidly moving from being dominated by purely physics-based models to embracing data-driven and hybrid approaches that leverage the power of AI.26

 

2.2. AI for Enhanced Data Processing and Anomaly Detection

 

One of the first and most critical roles for AI in the SHM pipeline is to make sense of the raw sensor data. This involves two key tasks:

• Feature Extraction and Noise Reduction: Before a diagnosis can be made, the relevant information—or “features”—must be extracted from the raw signals. For example, in vibration-based monitoring, the key features are the structure’s modal frequencies. AI techniques can automatically identify these features while simultaneously filtering out the confounding effects of environmental noise, which is a persistent challenge in real-world deployments.17
• Anomaly Detection: This is often the first level in damage assessment: determining if the structure’s behavior has deviated from its normal, healthy state.10 Unsupervised learning algorithms are perfectly suited for this task. Models like Principal Component Analysis (PCA) or Autoencoder neural networks are trained exclusively on data from the structure in its healthy condition. These models learn to create a compact representation of “normalcy.” When new data from the structure is fed into the model, it attempts to reconstruct it. If the reconstruction error is high, it means the new data does not fit the learned model of health and is flagged as an anomaly, alerting engineers to a potential problem.26 This approach is powerful because it does not require prior data from damaged states, which is often unavailable.

 

2.3. Computer Vision: The Automated Eye for Structural Inspection

 

Perhaps the most visually impressive application of AI in SHM is in the domain of computer vision. This technology is revolutionizing the process of structural inspection, which has traditionally relied on manual visual checks that are slow, subjective, prone to human error, and can be dangerous for inspectors working at height or in confined spaces.30

• Automated Crack and Defect Detection: Deep learning models, especially Convolutional Neural Networks (CNNs), have proven to be exceptionally effective at this task.31 These models are trained on large datasets containing thousands of images of structural surfaces, some with defects and some without. The CNN learns to automatically identify the visual features associated with different types of damage, such as cracks, spalling (chipping of concrete), and corrosion staining.30 Advanced object detection architectures like YOLO (You Only Look Once) and Faster R-CNN can then be used to draw bounding boxes around detected cracks in new images with high speed and accuracy.30
• From Detection to Quantification: The technology has moved beyond simply flagging the presence of a crack. By incorporating reference frames or using techniques like binocular stereo vision, computer vision frameworks can now automatically measure the physical dimensions of the damage—such as crack length, maximum width, and total area—in real-world units like millimeters.34 This provides engineers with the quantitative data needed to assess the severity of the damage and plan repairs.
• Enabling Technologies: This automated analysis is made possible by new data capture methods. Unmanned Aerial Vehicles (UAVs), or drones, equipped with high-resolution visual and thermal cameras can be deployed to rapidly and safely capture detailed imagery of an entire building façade or bridge underside.28 This creates a seamless, end-to-end automated inspection workflow, from robotic data collection to AI-powered data analysis.

 

2.4. Predictive Maintenance and Structural Prognosis: The Holy Grail of SHM

 

The ultimate goal of SHM is not just to find existing damage, but to predict future problems before they occur. This is the domain of predictive maintenance and structural prognosis, and it is where AI’s predictive power is most valuable.

By analyzing the historical data from a structure’s sensors, machine learning models can learn the patterns of degradation over time. Time-series forecasting models, such as Recurrent Neural Networks (RNNs) and their more sophisticated variants, Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) networks, are particularly well-suited for this.32 They can process sequential data to identify trends and predict the future state of the structure, including its Remaining Useful Life (RUL).10 This capability is the cornerstone of a true predictive maintenance strategy, allowing asset managers to move from reactive or preventative schedules to a proactive approach where interventions are optimized based on data-driven forecasts of future health.17 These AI-driven predictions can be integrated into comprehensive decision support systems, helping to prioritize maintenance actions, allocate resources more effectively, and optimize lifecycle costs.9

 

2.5. The Rise of the Digital Twin: Creating a Virtual Replica

 

The concept of the Digital Twin represents the pinnacle of SHM and AI integration. A Digital Twin is a dynamic, virtual model of a physical structure or system that is continuously updated with real-time data from its SHM sensor network.31 It is not a static BIM model; it is a living, breathing digital replica that mirrors the state and behavior of its physical counterpart.40

This fusion of the physical and digital worlds, powered by AI, unlocks unprecedented capabilities. Asset managers can use the Digital Twin to:

• Visualize Health: See the real-time status and performance of the structure on a detailed 3D model.
• Run Simulations: Conduct “what-if” analyses to simulate the impact of future loads, extreme events like earthquakes, or proposed retrofits without any risk to the actual structure.39
• Optimize Performance: Use AI and machine learning algorithms to analyze the integrated data within the twin, improving the accuracy of predictive models and optimizing maintenance strategies over the asset’s entire lifecycle.40

The application of these various AI models is not random; it systematically addresses the well-defined hierarchy of damage assessment in SHM.4 Unsupervised anomaly detection models tackle Level 1, answering “Is there damage?”. Computer vision models address Level 2 and 3, answering “Where is the damage?” and “What type of damage is it?”. Finally, predictive models like LSTMs tackle Level 4, prognosis, answering “How much useful life remains?”. This shows a structured and methodical approach by the research community to solving the SHM problem from the ground up.

The following table maps specific AI and ML algorithms to their primary applications within the SHM workflow, providing a clear guide to the technologies transforming the field.

 

AI/ML Category	Specific Algorithm	Primary SHM Application	Description
Unsupervised Learning	Principal Component Analysis (PCA), Autoencoders	Anomaly / Novelty Detection	Learns the pattern of “healthy” structural data. Flags new data that deviates from this pattern as a potential damage-induced anomaly. 26
Deep Learning (Supervised)	Convolutional Neural Network (CNN)	Image-Based Defect Detection	Trained on image datasets to automatically identify and classify visual defects like cracks, spalling, and corrosion from drone or camera footage. 30
Deep Learning (Supervised)	Artificial Neural Network (ANN)	Damage Classification & System Identification	A general-purpose classifier used to map sensor data inputs (e.g., vibration features) to damage state outputs (e.g., location, severity). 17
Deep Learning (Time-Series)	Recurrent Neural Network (RNN), LSTM, GRU	Predictive Maintenance & Prognosis	Analyzes time-series sensor data to model degradation trends and predict the future state and Remaining Useful Life (RUL) of a structure. 28
Hybrid Approaches	Physics-Informed Neural Networks (PINNs)	Data-Driven Modeling with Physical Constraints	Integrates the governing partial differential equations of structural mechanics into the AI model’s loss function, improving accuracy with less data. 26
Integrated Systems	Digital Twin	Holistic Asset Management & Simulation	A virtual replica of the structure, fed by real-time SHM data and analyzed by AI models to simulate behavior and optimize lifecycle management. 31

This technological advancement is not happening in isolation. It exists in a symbiotic relationship with hardware innovation. The development of high-resolution cameras on drones provides the rich datasets needed to train powerful computer vision models.35 The proliferation of low-cost, wireless IoT sensors makes it feasible to gather the continuous data streams required by predictive algorithms.28 This creates a powerful feedback loop: better sensors enable better AI, which in turn creates the business case for deploying more advanced sensors, accelerating the transformation of the entire field.

Section 3: The Singapore Mandate: Policy and National Initiatives Driving AI-Powered SHM

 

The rapid adoption of AI and ML in Structural Health Monitoring in Singapore is not merely an organic, industry-led trend. It is being actively and strategically driven by a comprehensive, multi-layered framework of national policies and government initiatives. This top-down mandate creates a unique and fertile ecosystem where technological innovation is not just encouraged but is becoming a fundamental requirement for operating in the nation’s built environment.

 

3.1. The Smart Nation Vision: A Top-Down Catalyst for Intelligent Infrastructure

 

The foundation of Singapore’s digital transformation is the Smart Nation initiative, launched in 2014. Its vision is to harness infocomm technologies, networks, and big data to create tech-enabled solutions that improve lives, create economic opportunities, and build stronger communities.43 This national-level ambition provides the political will and financial backing for large-scale investments in smart infrastructure. In 2017 alone, the government set aside S$2.4 billion to support the initiative.43

A key enabler of this vision is the Smart Nation Sensor Platform (SNSP). This initiative aims to deploy a nationwide network of sensors and data-sharing gateways to create a comprehensive, real-time picture of the city’s operations.43 A prominent example is the “Lamppost-as-a-Platform” (LaaP) project, which intends to fit Singapore’s 110,000 lampposts with various sensors to collect data on everything from weather and traffic to public safety.43 This ubiquitous data-gathering backbone provides a ready-made platform that specialized systems, such as those for SHM, can integrate with, enabling a move from monitoring individual structures to monitoring the health of the city as an interconnected system.

 

3.2. The BCA’s Transformation Agenda: Engineering a Digital Future for the Built Environment

 

While the Smart Nation initiative sets the broad vision, the Building and Construction Authority (BCA) is the key agency responsible for operationalizing this vision within the built environment sector. The BCA has implemented a suite of powerful policies and frameworks that directly mandate and incentivize the adoption of digital technologies like AI-powered SHM.

• The Built Environment Industry Transformation Map (BE ITM): Launched to unify previous roadmaps for the construction and facilities management industries, the BE ITM promotes a holistic, lifecycle approach to building transformation.46 It identifies three key transformation areas: Integrated Planning and Design (IPD), Advanced Manufacturing and Assembly (AMA), and Sustainable Urban Systems (SUS).46 The goals of IPD and SUS, in particular, are directly aligned with the capabilities of SHM, as they emphasize collaboration across the value chain and the use of technology to optimize building operations and maintenance.
• Integrated Digital Delivery (IDD): IDD is the central pillar of the BE ITM. It is defined as the use of digital technologies to integrate work processes and connect all project stakeholders throughout a building’s entire lifecycle—from design and fabrication to on-site construction and, crucially, operations and maintenance.49 Building on technologies like Building Information Modeling (BIM) and Virtual Design and Construction (VDC), IDD mandates the use of a
  Common Data Environment (CDE), a shared digital platform where all project information is stored and accessed.48 This CDE is the digital foundation upon which advanced SHM systems and Digital Twins are built. The BCA has set an aggressive target to increase IDD adoption in new developments from 34% to 70% by 2025, creating a powerful forcing function for the industry to digitalize.48
• Design for Maintainability (DfM): This regulatory guide fundamentally shifts the industry’s focus by embedding maintenance considerations into the earliest stages of design.52 The DfM guide, along with the voluntary
  Maintainable Design Appraisal System (MiDAS), requires designers to forecast downstream maintenance needs and make appropriate design provisions for access, material selection, and ease of repair.52 This proactive approach creates a clear regulatory and business imperative for installing SHM systems. These systems are needed to provide the data that validates the long-term performance and maintainability of the original design, effectively closing the loop between design intent and operational reality.

 

3.3. The Green Imperative: Sustainability as a Driver for Smart Monitoring

 

Another powerful driver for SHM adoption is Singapore’s commitment to sustainability, as outlined in the Singapore Green Building Masterplan (SGBMP). This plan, part of the broader Singapore Green Plan 2030, sets ambitious “80-80-80 in 2030” targets, which include greening 80% of the nation’s buildings by Gross Floor Area (GFA) and having 80% of new developments be Super Low Energy (SLE) buildings from 2030.55

To achieve these goals, the BCA’s Green Mark 2021 scheme has been updated to place greater emphasis not only on energy performance but also on designing for maintainability and creating healthier environments.55 SHM systems are critical tools in this context. They provide the continuous, granular data needed to monitor a building’s energy consumption, verify the performance of its systems (like HVAC), and ensure it maintains its Green Mark certification over its lifespan.

Furthermore, the government has put in place direct financial incentives to spur adoption. The Green Mark Incentive Scheme for Existing Buildings 2.0 (GMIS-EB 2.0) is a S$63 million fund to help building owners offset the upfront capital costs of energy-efficient retrofits.55 Additionally, developers can receive a bonus GFA of up to 3% for projects that achieve Green Mark Platinum Super Low Energy certification

with a Maintainability Badge, directly linking sustainability goals with the need for long-term monitoring and maintenance.55

These policies do not exist in isolation; they are designed to be a cohesive and self-reinforcing ecosystem. The BE ITM provides the overarching vision. IDD creates the digital backbone and collaborative processes. DfM and the SGBMP generate the regulatory and market demand for buildings that are designed to be high-performing and easily maintained. AI-powered SHM emerges as the critical enabling technology that allows the industry to meet this demand by providing the necessary monitoring, analysis, and predictive capabilities. The data generated by these SHM systems then feeds back into the IDD process, creating a digital twin that informs and improves the design of future buildings. This creates a virtuous, closed-loop system that is strategically designed to accelerate the transformation of the entire sector.

The following table summarizes this integrated policy framework and its direct impact on the adoption of AI-powered SHM in Singapore.

 

Policy/Initiative	Responsible Agency	Key Objective	Effect/Outcome for the BE Sector	Direct Impact on AI/SHM Adoption
Smart Nation	PMO (SNDGO, GovTech)	Harness technology to improve urban living, economy, and public services. 43	Creates a national mandate for digitalization and provides the physical data infrastructure (e.g., SNSP). 43	Provides the high-level justification and data backbone for city-scale intelligent infrastructure monitoring.
Built Environment ITM	BCA	Create an advanced, integrated, and sustainable built environment sector through a lifecycle approach. 46	Unifies transformation efforts for construction and facilities management, promoting collaboration across the value chain.	Sets the strategic direction that necessitates advanced operational technologies like SHM.
Integrated Digital Delivery (IDD)	BCA	Integrate workflows and connect stakeholders via digital technologies throughout the building lifecycle. 49	Mandates the use of digital platforms and a Common Data Environment (CDE) for all project data. 48	Creates the essential digital foundation (the CDE) required for data-intensive applications like SHM and Digital Twins.
Design for Maintainability (DfM)	BCA	Integrate operations and maintenance considerations into the initial design phase of a project. 52	Requires designers to plan for future maintenance, access, and material performance. 53	Creates a clear regulatory and business need for SHM systems to monitor and validate the long-term maintainability of assets.
Green Building Masterplan (SGBMP)	BCA, SGBC	Achieve ambitious energy efficiency and sustainability targets for buildings (“80-80-80 in 2030”). 55	Drives demand for high-performance buildings and provides financial incentives (e.g., GMIS-EB 2.0, bonus GFA). 55	Necessitates SHM for monitoring energy performance and system health to achieve and maintain Green Mark certification.

This framework reveals a sophisticated approach where the government acts not just as a regulator, but as an active market-maker. By mandating IDD adoption, providing grants and incentives to de-risk investment for companies 57, and funding foundational R&D 59, the government is systematically overcoming the primary barriers of cost, risk, and inertia that typically hinder technology adoption in the conservative construction industry.61

Section 4: The Innovation Engine: Research and Development in Singapore

 

Singapore’s leadership in AI-powered SHM is not solely the result of top-down policy. It is equally driven by a vibrant and collaborative research and development (R&D) ecosystem. The nation’s universities and research institutes are at the forefront of developing the next generation of SHM technologies, from advanced sensors to sophisticated AI frameworks. This innovation engine is fueled by significant government investment, such as the S28billionallocatedundertheResearch,InnovationandEnterprise(RIE)2025planandadedicatedinvestmentofoverS1 billion into AI compute, talent, and industry development.59

 

4.1. National University of Singapore (NUS): From Advanced Sensors to Robotic Systems

 

The National University of Singapore’s Department of Civil & Environmental Engineering is a key contributor to SHM research, with a notable focus on both the fundamental hardware and innovative deployment systems. The department’s work is supported by extensive facilities, including a state-of-the-art Structural Engineering Laboratory.62

• Advanced Sensor Development: NUS researchers are pushing the boundaries of sensor technology. One area of focus has been on developing novel optical fiber sensors, such as lens-based plastic optical fiber (LPOF) sensors, which have demonstrated excellent potential for real-time monitoring of fatigue cracks and strain with high sensitivity.63
• Piezoelectric Applications: The university has conducted extensive reviews and research into the application of piezoelectric (PZT) materials. This work covers not only their use as sensors and actuators for damage detection in beams, plates, and pipes but also their potential for active structural repair, showcasing a holistic approach to asset lifecycle management.15
• Robotics and Automation for SHM: Recognizing the practical challenges of deploying sensors on large structures, NUS has pioneered innovative solutions to make SHM more efficient. One notable project involves a robot-assisted system for the modal analysis of bridges. This approach uses programmable wheeled robots, which can be remotely controlled, to carry a minimal number of accelerometers to various points on a bridge. This allows for the identification of structural frequencies and high-resolution mode shapes without the need for a dense, permanently installed sensor array, offering a cost-effective and flexible solution for structural assessment.64
• Institutional Support: This research is fostered within dedicated research clusters like “Built Environment 4.0” and “Resilient Infrastructures,” and supported by facilities like the Centre for Advanced Materials and Structures, ensuring a focused and well-resourced R&D effort.62

 

4.2. Nanyang Technological University (NTU): Pioneering Unified AI Frameworks

 

Nanyang Technological University has established itself as a leader in applying artificial intelligence and systems integration to solve complex SHM problems. Their research often focuses on developing sophisticated, data-driven frameworks for real-world infrastructure challenges.

• Unified AI-Enabled SHM Framework: A flagship technology developed at NTU addresses the complexities of vision-based SHM. Instead of simplifying damage assessment to a binary “damaged/undamaged” classification, this framework treats it as a multi-attribute, multi-task problem. It uses a comprehensive, open-sourced image dataset and advanced deep learning techniques to recognize multiple key attributes in structural images simultaneously.65 The research involved extensive benchmarking of various CNNs (VGG16, VGG19, ResNet50) and demonstrated the superiority of a
  Multi-Task Learning (MTL) framework, which trains all tasks concurrently to leverage inter-task relationships and improve performance and computational efficiency.65
• Advanced Sensing and Analysis: NTU’s research also covers other SHM methodologies. Projects include the design of real-time SHM systems for underground caverns using distributed fiber optic sensors (BOTDR) 14 and the investigation of Lamb wave techniques with PZT sensors for monitoring load and damage in structural components.22
• Focus on Real-World Impact: NTU’s work is consistently framed by the need to ensure the safety and durability of civil infrastructure in densely populated urban environments like Singapore.9 Their research into combining deep neural networks (DNNs) with random forests (RF) to strategically place a limited number of strain gauges for maximum monitoring effectiveness is a prime example of this practical, problem-solving approach.9

 

4.3. A*STAR: The Collaborative Research Powerhouse

 

The Agency for Science, Technology and Research (A*STAR) plays a crucial role as a national research institute that collaborates with academia and industry to translate scientific discoveries into impactful technologies. While not a university, its institutes are deeply involved in SHM-related R&D.

• Bridging Research and Industry: ASTAR frequently partners with NUS and NTU on research projects. For example, ASTAR’s Institute of Materials Research and Engineering (IMRE) and Institute of High-Performance Computing (IHPC) collaborated with NUS and NTU on developing a new generation of wearable hydrogel sensors for health monitoring, a technology with clear parallels and potential applications in flexible sensors for SHM.67
• Direct Research Contributions: ASTAR researchers are active contributors to the global SHM community. At the 9th Asia-Pacific Workshop on Structural Health Monitoring (APWSHM), researchers from ASTAR presented work on “Estimating the Probability of Detection of Cracks in Metal Plates Using Lamb Waves” and “Damage Detection in Hybrid Metal-Composite Plates Using Ultrasonic Guided Waves,” demonstrating direct expertise in core SHM techniques.68
• Enabling Technology Development: A*STAR’s focus on areas like underwater SHM 69 and developing prognostic software tools for assessing damage and predicting residual life 70 helps build the foundational technologies that the entire ecosystem relies upon.

Collectively, Singapore’s R&D institutions are pursuing a comprehensive, “full-stack” approach to SHM innovation. NUS and A*STAR are conducting fundamental research on the hardware layer, developing new sensor materials and technologies.63 NTU is pioneering the

software and systems layer, creating integrated AI frameworks and advanced analytical models.65 Both universities are innovating at the

deployment layer, using robotics and automation to make data collection more practical and efficient.64 This holistic strategy ensures that progress in one area, like AI algorithms, is supported by corresponding advances in sensor capabilities and deployment methods, leading to more robust and commercially viable solutions.

This R&D effort is also deeply integrated with the international scientific community. Researchers in Singapore benchmark their AI models against global standards like the ImageNet dataset 65, actively publish in top-tier international journals like

Structural Health Monitoring 71, and participate in major global conferences.68 This outward-looking posture ensures that local research remains at the cutting edge and benefits from global knowledge exchange, further solidifying Singapore’s position as a hub of innovation.

Section 5: On the Ground: Applications, Case Studies, and Commercial Solutions in Singapore

 

The convergence of forward-thinking policy and cutting-edge research has cultivated a dynamic market for AI-powered SHM in Singapore. A mature ecosystem of both global and local technology providers is now actively deploying these solutions across the nation’s infrastructure. This section examines the commercial landscape and highlights specific case studies that demonstrate how intelligent monitoring is being applied to solve real-world challenges in Singapore.

 

5.1. The Commercial Ecosystem: SHM Service Providers in Singapore

 

The presence of a diverse range of commercial SHM providers is a clear indicator of a healthy market. These companies offer a spectrum of services, from specialized sensor deployment to comprehensive, IoT-ready monitoring platforms.

• Global Leaders with a Local Presence:

• SGS: A world leader in inspection and certification, SGS offers a remote, sensor-based SHM solution in Singapore that acts as a “24/7 NDT inspector.” Their IoT-ready system provides continuous, real-time assessment of defects, helping to reduce asset failure risk and optimize maintenance costs for critical assets like bridges and cranes.73
• Sixense Asia: Leveraging its experience from major international projects like the Bosphorus bridges, Sixense provides SHM services in Asia with specialist expertise in monitoring cable-stayed and prestressed concrete structures. Their solutions include robust measurement tools and acoustic monitoring for the real-time detection of cable strand failures due to corrosion or fatigue.23
• DYWIDAG: This company offers a versatile, sensor-agnostic SHM platform designed for long-term use. Their product portfolio includes specific solutions for monitoring rail temperature, the force in post-tensioned tendons, and structural movement with smart tilt sensors, catering to a wide range of infrastructure markets.74

• Local Specialists and Innovators:

• Sofotec Singapore: A Singapore-founded firm that has established itself as a leading specialist in SHM using Fiber Optic Sensors. They have a strong track record of monitoring critical infrastructure like pipelines and tunnels for major local clients including the Housing & Development Board (HDB), SP Group, and Surbana Jurong, demonstrating deep expertise in a key technology niche.25
• Ackcio: A Singapore-based company that has developed a patented long-range wireless mesh network for industrial data acquisition. Their Ackcio Beam solution is specifically designed to provide reliable, real-time data from sensors in challenging environments, particularly underground, overcoming the limitations of other wireless technologies.75

 

5.2. Case Study 1: Protecting Critical Lifelines – AI-Powered Monitoring of MRT Tunnels

 

The construction and operation of Singapore’s Mass Rapid Transit (MRT) system, much of which runs underground, presents a significant SHM challenge. Tunneling for new lines, such as the Thomson-East Coast Line, creates potential risks for adjacent structures and existing utility tunnels.75

• The Challenge: Traditional manual monitoring of these tunnels is fraught with difficulties. Access is often restricted, requiring special permits and limiting the frequency of readings. This manual approach is labor-intensive, costly, and provides only intermittent snapshots of the tunnel’s condition, which is insufficient for managing the dynamic risks associated with ongoing construction.75
• The Solution: To overcome these challenges, project contractors have deployed automated, real-time monitoring systems. In one project monitoring a common services tunnel near an MRT construction site, the contractor Soil Investigation deployed the Ackcio Beam wireless monitoring system. This involved placing biaxial tiltmeters deep underground, connected to wireless nodes that could reliably transmit data through four stories of earth and concrete to a gateway at the surface.75
• The AI and Digital Integration: This automated system provides a continuous stream of real-time data, which is fed to servers for analysis, enabling far more effective risk management than manual methods.75 This forms the basis for AI integration. More advanced systems used in Singapore’s MRT tunnels employ the SmartSensys RSCS, a vehicle-mounted system with 16 high-speed cameras that scans the tunnel walls as it drives, using computer vision to create an ultra-fine digital map of any cracking patterns.76 This monitoring data is further integrated into the larger digital ecosystem. The Land Transport Authority (LTA) utilizes
  BIM + ERP (Enterprise Resource Planning) systems on major projects, creating a digital twin foundation where real-time SHM data can be integrated to provide a comprehensive view of the project’s health and progress.77

 

5.3. Case Study 2: Intelligent Buildings – From Façade Inspection to Predictive Maintenance

 

The application of AI and SHM to buildings is a direct response to the regulatory drivers established by the BCA.

• Automated Façade Inspection: Singapore’s Periodic Façade Inspection (PFI) regime mandates that buildings over 13 meters high and 20 years old undergo a 100% visual inspection of their façades at least once every seven years.78 This has created a significant market for automated solutions. Instead of relying on costly and slow manual inspections using gondolas, firms are now deploying drones equipped with high-resolution visual and thermal cameras. The captured images are then processed by AI algorithms trained to automatically detect defects like cracks and spalling.78 This approach is not only faster and more cost-effective but also significantly safer for inspection personnel.36
• Predictive Maintenance in Smart Buildings: Aligning with the goals of the Green Building Masterplan and Design for Maintainability, companies in Singapore are offering predictive maintenance services for building systems.37 These solutions use a network of IoT sensors to monitor the performance of critical equipment, such as HVAC systems. AI-powered analytics then process this data to identify anomalies, predict potential failures before they occur, and optimize maintenance schedules.38 This proactive approach minimizes disruptive downtime, reduces energy consumption, and lowers overall operating costs, directly contributing to a building’s sustainability and maintainability credentials. While there is no specific public case study on SHM for an iconic building like Marina Bay Sands, its extensive use of robotics and automation for its operations signals a high level of technological readiness for such systems.81

 

5.4. Case Study 3: The Ultimate Vision – Virtual Singapore, the Nation’s Digital Twin

 

Virtual Singapore represents the culmination of all these trends, scaling the concept of SHM from a single asset to an entire nation.

• The Project: Initiated by the Singapore Land Authority (SLA), Virtual Singapore is a dynamic, data-rich, and highly detailed 3D digital twin of the entire country.39 It was built using data from laser-scanning aircraft and vehicles and encompasses both above-ground features (buildings, roads) and underground infrastructure.39
• The Capability: This is far more than a static map. It is a collaborative platform that is continuously updated with real-time data streams from thousands of IoT sensors across the city, covering traffic flow, environmental conditions, and utility usage.40 This allows government agencies, planners, and researchers to run complex simulations for a huge range of applications, including optimizing urban planning, improving infrastructure management, and enhancing disaster preparedness by modeling scenarios like flooding or fires.39
• The Role of AI: Artificial intelligence is the engine that drives Virtual Singapore. AI and ML algorithms are essential for processing and analyzing the immense volumes of data flowing into the twin, running the predictive simulations, and extracting the actionable insights that inform national-level decision-making.39 It is the ultimate SHM and asset management tool, providing a holistic, system-level view of the nation’s built environment.

These case studies reveal a clear maturity curve in Singapore’s application of SHM. The journey begins with tackling specific, high-risk problems at the component level, such as monitoring a section of an MRT tunnel.75 It then broadens to encompass the lifecycle management of entire assets, like a building’s façade or its mechanical systems.37 The final, most ambitious stage, embodied by Virtual Singapore, is the management of the entire city as an integrated system of systems.39 This strategic progression from tactical problem-solving to holistic, city-scale planning demonstrates a sophisticated and long-term vision for intelligent infrastructure management.

Section 6: Navigating the Future: Challenges and the Road Ahead for AI in SHM

 

Despite the remarkable progress and strong systemic support in Singapore, the path to ubiquitous, fully autonomous structural health monitoring is not without its challenges. Widespread adoption requires overcoming significant technical, financial, and regulatory hurdles. This final section examines these challenges and explores the future trends that will define the next frontier of intelligent infrastructure.

 

6.1. Implementation Hurdles: Beyond the Technology

 

While the technology itself is advancing rapidly, several practical barriers can slow its implementation.

• Data Quality and Availability (The “Sensor Data Dilemma”): The adage “garbage in, garbage out” is especially true for AI. The performance of ML models is critically dependent on access to large volumes of high-quality, well-labeled training data.18 In SHM, this presents a unique problem: data corresponding to structural damage is, by design, rare. Furthermore, real-world data is often “unclean,” degraded by anomalies arising from sensor noise, network instability, or equipment malfunctions, all of which can compromise the reliability of AI-driven monitoring systems.82
• Cost and ROI (The “SHM Cost Constraint”): The initial capital investment for a comprehensive SHM system—including sensors, data acquisition hardware, software licenses, and installation—can be substantial.18 While the long-term benefits in terms of extended asset life and reduced maintenance costs are significant, demonstrating a clear and compelling short-term Return on Investment (ROI) can be a challenge for asset owners, particularly for smaller firms or on projects with tight budgets.18
• System Integration and Interoperability: Real-world structures are complex ecosystems with diverse components. A key technical hurdle is integrating data from various sensor types, legacy Building Management Systems (BMS), and different proprietary software platforms into a single, cohesive system. A lack of standardization can lead to data silos, preventing a truly holistic view of the structure’s health.37
• Talent and Skills Gap: The convergence of civil engineering and data science has created a demand for a new type of professional—one who is “bilingual” and proficient in both structural mechanics and AI technologies.83 There is a significant need to upskill the existing workforce and train a new generation of engineers and data scientists to develop, deploy, and manage these complex systems effectively.46

 

6.2. The Digital Fortress: Cybersecurity and Data Governance in a Connected Infrastructure

 

As Singapore’s infrastructure becomes increasingly “smart” and interconnected, it also becomes a more attractive target for cyber threats. Securing these systems is a paramount concern.

• The Threat Landscape: Every connected sensor, actuator, and controller in an SHM or smart building system represents a potential entry point for malicious actors.85 The threat is not just theoretical; state-sponsored groups are known to target Critical Information Infrastructure (CII), which includes sectors like energy, transport, and water.86 A successful attack could have devastating consequences, ranging from data theft to the physical compromise of a structure’s safety systems. The infamous 2013 Target data breach, which was initiated through a vulnerability in their HVAC vendor’s network, serves as a stark reminder of how operational technology (OT) systems can be a gateway to catastrophic IT breaches.85
• Singapore’s Regulatory Framework: Singapore has established a robust legal framework to address these risks.

• The Cybersecurity Act: This legislation imposes strict obligations on the designated owners of CII. These organizations must comply with codes of practice, conduct regular cybersecurity audits and risk assessments, and report prescribed cybersecurity incidents to the authorities, often within two hours of detection.87
• The Personal Data Protection Act (PDPA): This act governs the collection, use, and disclosure of personal data. Its requirements for making reasonable security arrangements to protect data and for mandatory notification in the event of a data breach are highly relevant for any SHM system that may handle data linked to individuals, such as in a smart residential building or office.87

This creates a dual challenge for any SHM implementation: it must protect the integrity of the OT systems to ensure physical safety and operational reliability, while also protecting the confidentiality and integrity of the IT data the system generates. Successfully navigating this complex intersection of engineering, data science, and law is no longer optional; it is a fundamental requirement for deploying intelligent infrastructure in Singapore.

 

6.3. The Next Frontier: Future Trends in Intelligent SHM

 

Looking ahead, several key trends are poised to shape the future of AI in SHM, pushing the boundaries of what is possible.

• Edge AI and Real-Time On-Site Inference: To overcome challenges with data transmission latency and bandwidth, there is a growing trend towards Edge AI. This involves deploying lightweight, optimized AI models to run directly on edge devices, such as smart sensors or on-site gateway computers equipped with specialized hardware like Neural Processing Units (NPUs).42 This allows for real-time analysis and decision-making at the source, reducing reliance on cloud connectivity and enabling faster response times.
• Physics-Informed AI: To address the “black box” problem of some deep learning models and to improve performance when training data is scarce, researchers are developing hybrid models. Physics-Informed Neural Networks (PINNs) integrate the known physical laws of structural mechanics (expressed as partial differential equations) directly into the AI’s learning process.26 This helps to constrain the model’s predictions, making them more accurate, robust, and physically plausible, thereby increasing trust in the AI’s output.17
• Robotics and Full Automation: The convergence of AI with advanced robotics, drones, and autonomous ground vehicles will continue to drive towards fully automated inspection and monitoring workflows. This will further reduce the need for human intervention in hazardous environments, improve the consistency and quality of data collection, and enhance overall safety and efficiency.35
• AI-Driven Materials Science: One of the most transformative future trends is the application of AI to the very materials of construction. Researchers are using AI to accelerate the discovery and design of novel materials with enhanced properties, such as greater durability, sustainability, or even self-healing capabilities.59 This could fundamentally change the nature of SHM, shifting the focus from simply detecting damage to monitoring the performance of these advanced, adaptive materials.

 

Conclusion: Building a Resilient Future on a Foundation of Trust

 

The integration of Artificial Intelligence and Machine Learning into Structural Health Monitoring is not a futuristic concept in Singapore; it is a present-day reality that is fundamentally reshaping the nation’s approach to infrastructure management. The analysis reveals a uniquely powerful and strategically designed ecosystem where top-down policy, world-class R&D, and a responsive commercial market work in concert. National mandates like the Smart Nation vision and the BCA’s BE ITM have created the imperative for transformation, while specific frameworks like IDD and DfM have laid the digital and procedural groundwork. This has cultivated a fertile environment for the cutting-edge research from NUS and NTU and the practical solutions from providers like SGS and Sofotec to take root and flourish.

The result is a clear trajectory of increasing sophistication, moving from the monitoring of individual components to the holistic management of entire assets, and culminating in the ambitious vision of a national-scale digital twin. However, this journey is not merely a technological one. As infrastructure becomes more intelligent and autonomous, the critical challenges shift towards data governance, cybersecurity, and talent development. The future of the built environment in Singapore will be defined not only by the intelligence of its systems but by the trust that society places in them.

Successfully navigating this future requires a new, multi-disciplinary mindset. The structural engineer of tomorrow must be conversant in data science, the AI developer must understand the principles of structural mechanics, and both must operate within a robust legal and ethical framework. As Singapore continues to build its digital watchtower, its greatest success will lie in its ability to foster this human capital and to build a resilient future founded not just on smart technology, but on unwavering public and regulatory trust.

Works cited

1. en.wikipedia.org, accessed July 8, 2025, https://en.wikipedia.org/wiki/Structural_health_monitoring
2. Introduction to Structural Health Monitoring – ISTE, accessed July 8, 2025, http://www.iste.co.uk/data/doc_xqjujdlhnfls.pdf
3. (PDF) Introduction to Structural Health Monitoring – ResearchGate, accessed July 8, 2025, https://www.researchgate.net/publication/273059503_Introduction_to_Structural_Health_Monitoring
4. Overview of Structural Health Monitoring – Smart Material, accessed July 8, 2025, https://smart-material.com/wp-content/uploads/2025/01/SHM-Overview-Inman.pdf
5. FEM1:, accessed July 8, 2025, https://www.ltrc.lsu.edu/nsf/files/FEM1%20An%20Introduction%20to%20Structural%20Health%20Monitoring.pptx
6. SHM – Structural health monitoring – Sisgeo, accessed July 8, 2025, https://sisgeo.com/geotechnical-monitoring-applications/structural-health-monitoring/
7. A Complete Guide to Structural Health Monitoring: AECbytes Tips and Tricks, accessed July 8, 2025, https://aecbytes.com/tipsandtricks/2024/issue109-StructuralHealthMonitoring.html
8. Structural Health Monitoring – Kinemetrics, accessed July 8, 2025, https://kinemetrics.com/application/structural-health-monitoring/
9. Applications of structural health monitoring technology in Asia – ResearchGate, accessed July 8, 2025, https://www.researchgate.net/publication/304344956_Applications_of_structural_health_monitoring_technology_in_Asia
10. An introduction to structural health monitoring | Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, accessed July 8, 2025, https://royalsocietypublishing.org/doi/10.1098/rsta.2006.1928
11. Structural Health Monitoring (SHM) Frequently Asked Questions, accessed July 8, 2025, https://earthquake.usgs.gov/monitoring/nsmp/buildings/faq.php
12. Effects of Environmental and Operational Conditions on Structural Health Monitoring and Non-Destructive Testing: A Systematic Review – MDPI, accessed July 8, 2025, https://www.mdpi.com/2075-5309/13/4/918
13. Structural Health Monitoring with Application to Offshore Structures, accessed July 8, 2025, https://www.worldscientific.com/worldscibooks/10.1142/11302
14. Structural health monitoring using distributed fiber optic sensors – DR-NTU, accessed July 8, 2025, https://dr.ntu.edu.sg/entities/publication/440a7959-a3af-4f79-a7af-63df8abb43f0
15. Applications of Piezoelectric Materials in Structural Health Monitoring and Repair: Selected Research Examples – PubMed, accessed July 8, 2025, https://pubmed.ncbi.nlm.nih.gov/28883375/
16. Structural Health Monitoring (SHM) – Sixense Group, accessed July 8, 2025, https://www.sixense-group.com/en/offer/monitoring/structural-health-monitoring-shm
17. AI in Structural Health Monitoring for Infrastructure Maintenance and Safety – ResearchGate, accessed July 8, 2025, https://www.researchgate.net/publication/386532412_AI_in_Structural_Health_Monitoring_for_Infrastructure_Maintenance_and_Safety
18. Use of Artificial Intelligence in Structural Health Monitoring for Improving the Resilience and Sustainability of Concrete Structures – White Rose eTheses Online, accessed July 8, 2025, https://etheses.whiterose.ac.uk/id/eprint/35594/
19. Structural Damage Detection Based on One-Dimensional Convolutional Neural Network, accessed July 8, 2025, https://www.mdpi.com/2076-3417/13/1/140
20. Structural Health Monitoring – Wiley, accessed July 8, 2025, https://www.wiley.com/en-us/Structural+Health+Monitoring-p-9780470394403
21. Deep Learning Empowered Structural Health Monitoring and Damage Diagnostics for Structures with Weldment via Decoding Ultrasonic Guided Wave – MDPI, accessed July 8, 2025, https://www.mdpi.com/1424-8220/22/14/5390
22. Structural health monitoring using piezoelectric materials – DR-NTU, accessed July 8, 2025, https://dr.ntu.edu.sg/handle/10356/61229
23. Structural Health Monitoring (SHM) – Extending life of structures …, accessed July 8, 2025, https://asia.sixense-group.com/offer/instrumentation-monitoring/structural-health-monitoring-shm
24. (PDF) Railway bridge structural health monitoring and fault detection: State-of-the-art methods and future challenges – ResearchGate, accessed July 8, 2025, https://www.researchgate.net/publication/319280857_Railway_bridge_structural_health_monitoring_and_fault_detection_State-of-the-art_methods_and_future_challenges
25. Sofotec Singapore | Smart Cities | Fiber Optic Sensors, accessed July 8, 2025, https://www.sofotec.com.sg/
26. Machine learning for structural health monitoring: challenges and opportunities – SPIE Digital Library, accessed July 8, 2025, https://www.spiedigitallibrary.org/conference-proceedings-of-spie/11379/1137903/Machine-learning-for-structural-health-monitoring-challenges-and-opportunities/10.1117/12.2561610.full
27. Enhancing structural health monitoring with machine learning for accurate prediction of retrofitting effects – AIMS Press, accessed July 8, 2025, https://www.aimspress.com/article/doi/10.3934/math.20241472
28. AI in Structural Health Monitoring for Infrastructure Maintenance and Safety – MDPI, accessed July 8, 2025, https://www.mdpi.com/2412-3811/9/12/225
29. Structural Damage Detection Based on the Correlation of Variational Autoencoder Neural Networks Using Limited Sensors, accessed July 8, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC11054692/
30. Surface Crack Detection using Computer Vision – Wipro, accessed July 8, 2025, https://www.wipro.com/engineering/surface-crack-detection-using-computer-vision/
31. Deep Learning for Structural Health Monitoring: Data, Algorithms, Applications, Challenges, and Trends – MDPI, accessed July 8, 2025, https://www.mdpi.com/1424-8220/23/21/8824
32. Deep Learning Algorithms for Structural Health Monitoring | Encyclopedia MDPI, accessed July 8, 2025, https://encyclopedia.pub/entry/51824
33. (PDF) Neural Network Approaches for Structural Damage Detection and Classification, accessed July 8, 2025, https://www.researchgate.net/publication/391280329_Neural_Network_Approaches_for_Structural_Damage_Detection_and_Classification
34. A Three-Step Computer Vision-Based Framework for Concrete Crack Detection and Dimensions Identification – MDPI, accessed July 8, 2025, https://www.mdpi.com/2075-5309/14/8/2360
35. Improving the Concrete Crack Detection Process via a Hybrid Visual Transformer Algorithm, accessed July 8, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC11125435/
36. Resilient Infrastructure – Ministry of National Development (MND), accessed July 8, 2025, https://www.mnd.gov.sg/urbansustainability/resource/resilient-infrastructure
37. Smart Buildings: Transforming facilities management in Singapore – Infinity Wave, accessed July 8, 2025, https://infinitywave.io/blog/singapore-facilities-management/
38. Predictive Maintenance in Singapore | Nanoprecise Sci Corp, accessed July 8, 2025, https://nanoprecise.io/predictive-maintenance-in-singapore/
39. Virtual Singapore – Singapore’s virtual twin – Observatory of Public Sector Innovation, accessed July 8, 2025, https://oecd-opsi.org/innovations/virtual-twin-singapore/
40. Digital twins and the future of Singapore – Intelligent CIO APAC, accessed July 8, 2025, https://www.intelligentcio.com/apac/2024/06/10/digital-twins-and-the-future-of-singapore/
41. How Singapore Leads the Way in Urban Planning with Digital Twin Technology, accessed July 8, 2025, https://www.mingothings.com/post/how-singapore-leads-the-way-in-urban-planning-with-digital-twin-technology
42. integrating edge-ai in structural health monitoring domain – arXiv, accessed July 8, 2025, https://arxiv.org/pdf/2304.03718
43. Smart Nation – Wikipedia, accessed July 8, 2025, https://en.wikipedia.org/wiki/Smart_Nation
44. Singapore’s Smart Nation Initiative – A Policy and Organisational Perspective, accessed July 8, 2025, https://lkyspp.nus.edu.sg/docs/default-source/case-studies/singapores-smart-nation-initiative-final_112018.pdf?sfvrsn=354e720a_2
45. Smart Nation Singapore, accessed July 8, 2025, https://www.smartnation.gov.sg/
46. Built Environment Industry Digital Plan – Singapore – IMDA, accessed July 8, 2025, https://services2.imda.gov.sg/ctoaas/builtenvironmentindustrydigitalplan
47. 7 Sep 2022 [for support, pls] Launch of Built Environment Industry Transformation Map (BE ITM) ACES is pleased to support the re – Association of Consulting Engineers Singapore, accessed July 8, 2025, http://aces.org.sg/wp-content/pdf/2022/312_2022-BCA_BE%20ITM%20for%20Member%20Support.pdf
48. Built Environment Industry Transformation Map to Facilitate Integration and Collaborative Breakthrough Across the Entire Value Chain – Building and Construction Authority (BCA), accessed July 8, 2025, https://www1.bca.gov.sg/about-us/news-and-publications/media-releases/2022/09/06/built-environment-industry-transformation-map-to-facilitate-integration-and-collaborative-breakthrough-across-the-entire-value-chain
49. BCA Integrated Digital Delivery Website – Global BIM Network, accessed July 8, 2025, https://globalbim.org/info-collection/bca-integrated-digital-delivery-website/
50. Integrated Digital Delivery (IDD) – Building and Construction Authority (BCA), accessed July 8, 2025, https://www1.bca.gov.sg/buildsg/digitalisation/integrated-digital-delivery-idd
51. Principles of Integrated Digital Delivery – BCA Academy, accessed July 8, 2025, https://www.bcaa.edu.sg/what-we-offer/courses/online-courses/Details/principles-of-integrated-digital-delivery
52. DESIGN FOR MAINTAINABILITY GUIDE: – Building and …, accessed July 8, 2025, https://www1.bca.gov.sg/docs/default-source/dfm/design-for-maintainability-guide—residential-(version-2-0).pdf?sfvrsn=dc1a298f_0
53. Design for Maintainability Guide (Non-residential) – Building and Construction Authority (BCA), accessed July 8, 2025, https://www1.bca.gov.sg/docs/default-source/docs-corp-buildsg/sustainability/dfm-guide-non-residential.pdf
54. BCA Singapore Piloting Voluntary Maintainable Design Appraisal System | Grow Your Business – Distributors in Southeast Asia, accessed July 8, 2025, https://www.growyourbusiness.org/business-news/bca-singapore-piloting-voluntary-maintainable-design-appraisal-system/
55. Green Building Masterplans | Building and Construction Authority …, accessed July 8, 2025, https://www1.bca.gov.sg/buildsg/sustainability/green-building-masterplans
56. Singapore Green Building Masterplan, accessed July 8, 2025, https://www.cop-pavilion.gov.sg/about/green-initiatives/mitigation/singapores-green-building-masterplan
57. BCA sets out initiatives to support transformation in built environment sector, accessed July 8, 2025, https://www.allenandgledhill.com/sg/publication/articles/27682/bca-sets-out-initiatives-to-support-transformation-in-built-environment-sector
58. Unlocking Government Grants for AI Projects in Singapore: A Guide for SMEs – Business+AI, accessed July 8, 2025, https://www.businessplusai.com/blog/unlocking-government-grants-for-ai-projects-in-singapore-a-guide-for-smes
59. Singapore sees a material science focus in the use of AI funds for research – GovInsider, accessed July 8, 2025, https://govinsider.asia/intl-en/article/singapore-sees-a-material-science-focus-in-the-use-of-ai-funds-for-research
60. Singapore Budget 2024: Govt announces S$1 Bn investment over the next five years into AI, accessed July 8, 2025, https://ciosea.economictimes.indiatimes.com/news/policy/singapore-budget-2024-govt-announces-s1-bn-investment-over-the-next-five-years-into-ai/107751263
61. Integrated Digital Delivery: Implementation Status, Barriers, and Strategies – ASCE Library, accessed July 8, 2025, https://ascelibrary.org/doi/10.1061/9780784482865.078
62. Structural Engineering Laboratory – Civil and Environmental …, accessed July 8, 2025, https://cde.nus.edu.sg/cee/research/laboratories/structural-engineering-laboratory/
63. KSC Kuang PhD, BEng(Hons), MBA Associate Professor at National University of Singapore – ResearchGate, accessed July 8, 2025, https://www.researchgate.net/profile/K-Kuang
64. Chan Ghee Koh’s research works | National University of Singapore and other places, accessed July 8, 2025, https://www.researchgate.net/scientific-contributions/Chan-Ghee-Koh-2195492021
65. Structural Health Monitoring (SHM) Utilising Unified AI technology | Nanyang Technological University | Innovation and Entrepreneurship | NTU Singapore, accessed July 8, 2025, https://www.ntu.edu.sg/innovates/tech-portal/tech-offers/detail/structural-health-monitoring-utilising-unified-ai-technology
66. Study on Structural Health Monitoring (SHM) applications – DR-NTU, accessed July 8, 2025, https://dr.ntu.edu.sg/entities/publication/058dddbf-b02d-485f-b123-88fe8f225813
67. Stretching the limits of health monitoring – A*STAR Research, accessed July 8, 2025, https://research.a-star.edu.sg/articles/highlights/stretching-the-limits-of-health-monitoring/
68. 9th Asia-Pacific Workshop on Structural Health Monitoring | Monash University, accessed July 8, 2025, https://www.monash.edu/engineering/shm
69. underwater structural health monitoring Archives – A*STAR Research, accessed July 8, 2025, https://research.a-star.edu.sg/tag/underwater-structural-health-monitoring/
70. Structural Health Monitoring – AlphaSTAR Corporation, accessed July 8, 2025, https://alphastarcorp.com/structural-health-monitoring/
71. Structural Health Monitoring | SAGE Publications Inc, accessed July 8, 2025, https://us.sagepub.com/en-us/nam/structural-health-monitoring/journal201592
72. Structural Health Monitoring – Impact Factor & Score 2025 – Research.com, accessed July 8, 2025, https://research.com/journal/structural-health-monitoring
73. Structural Health Monitoring | SGS Singapore, accessed July 8, 2025, https://www.sgs.com/en-sg/services/structural-health-monitoring
74. Structural Health Monitoring (SHM) Systems – DYWIDAG, accessed July 8, 2025, https://dywidag.com/products/shm-systems
75. CASE STUDY – Real-Time Monitoring of an MRT Tunnelling Project in Singapore – Ackcio, accessed July 8, 2025, https://www.ackcio.com/wp-content/uploads/2023/02/case-study-real-time-monitoring-of-an-mrt-tunnelling-project-in-singapore.pdf
76. The SmartSensys RSCS for Singapore MRT tunnel scans. – ResearchGate, accessed July 8, 2025, https://www.researchgate.net/figure/The-SmartSensys-RSCS-for-Singapore-MRT-tunnel-scans_fig4_292337584
77. Knowledge management practices in Singapore’s tunnel industry – RealKM, accessed July 8, 2025, https://realkm.com/2025/05/21/knowledge-management-practices-in-singapores-tunnel-industry/
78. Building Façade Inspection Through AI Thermography and UAV Technology – Voltin, accessed July 8, 2025, https://www.voltin.com.au/facade-inspection-tech/
79. From reaction to prevention: BE safe and productive with technology | BuildSG Magazine, accessed July 8, 2025, https://www1.bca.gov.sg/buildsg-emag/articles/from-reaction-to-prevention-be-safe-and-productive-with-technology
80. Predictive Maintenance – AIRSYS, accessed July 8, 2025, https://www.air-sys.sg/PredictiveMaintenance
81. Robots at Singapore’s Marina Bay Sands cut labour dependency by 30 per cent, accessed July 8, 2025, https://www.techgoondu.com/2025/06/18/robots-at-singapores-marina-bay-sands-cut-labour-dependency-by-30-per-cent/
82. RETRACTED: Enhancing Data Quality Management in Structural Health Monitoring through Irregular Time-Series Data Anomaly Detection Using IoT Sensors – MDPI, accessed July 8, 2025, https://www.mdpi.com/2075-5309/14/7/2223
83. Singapore’s Smart Nation 2.0 policy focuses on AI and building resilience – GovInsider, accessed July 8, 2025, https://govinsider.asia/intl-en/article/singapores-smart-nation-20-policy-focuses-on-ai-and-building-resilience
84. Impact of AI on Singapore’s Labor Market – IMF eLibrary, accessed July 8, 2025, https://www.elibrary.imf.org/view/journals/018/2024/040/article-A001-en.xml
85. Cybersecurity in Smart Buildings: From Afterthought to Strategic Investment by Vardaan Arora on Siemens Blog, accessed July 8, 2025, https://blog.siemens.com/2025/03/cybersecurity-in-smart-buildings-from-afterthought-to-strategic-investment/
86. Cybersecurity in Singapore: Strengthening Resilience Amid Evolving Cyber Threats, accessed July 8, 2025, https://www.factsasia.org/blog/cybersecurity-in-singapore-strengthening-resilience-amid-evolving-cyber-threats
87. Singapore: Data Protection & Cybersecurity – Country Comparative Guides – Legal 500, accessed July 8, 2025, https://www.legal500.com/guides/chapter/singapore-data-protection-cybersecurity/
88. Security Requirements and Breach Notification | Singapore | Global Data and Cyber Handbook | Baker McKenzie Resource Hub, accessed July 8, 2025, https://resourcehub.bakermckenzie.com/en/resources/global-data-and-cyber-handbook/asia-pacific/singapore/topics/security-requirements-and-breach-notification
89. Cybersecurity 2025 – Singapore | Global Practice Guides – Chambers and Partners, accessed July 8, 2025, https://practiceguides.chambers.com/practice-guides/cybersecurity-2025/singapore/trends-and-developments
90. PDPA Overview – Singapore – PDPC, accessed July 8, 2025, https://www.pdpc.gov.sg/overview-of-pdpa/the-legislation/personal-data-protection-act
91. Data protection and cybersecurity laws in Singapore | CMS Expert Guide, accessed July 8, 2025, https://cms.law/en/int/expert-guides/cms-expert-guide-to-data-protection-and-cyber-security-laws/singapore

Singapore is innovating with construction 2.0 – Saint-Gobain, accessed July 8, 2025, https://www.saint-gobain.com/en/news/singapore-innovating-construction-20