Multi-dimensional model and interactive simulation of intelligent construction based on digital twins

Integrating heterogeneous information from multiple systems and delivering the necessary data to key stakeholders along the phases are essential for efficient building management. Integrating the latest building data with the 3D model improves understanding of intricate building systems, with digital twins as essential support pillars. Created for the industrial sector, digital twin technology now finds applications in many others. Digital twins are quickly gaining traction in the smart factory industry. More efficient operations with reduced waste can be achieved with the help of DT, which enables a virtual world to mimic the physical one and replicate real-world data in a virtual setting. It isn’t easy to generalize from the current research, as much of it is devoted to developing the foundation and potential of DTs and suggesting architectures personalized to specific use cases. Hence, this study proposes the Multi-Dimensional Digital Twin Technology-assisted Building Information Modeling (MD-DTT-BIM) to create the assembly of the construction operations and maintenance systems. The study utilizes DT technology and BIM to examine building modelling in light of current research and issues to make the building construction process more efficient. One new aspect is how quickly environmental factors like weather and air pollution may impact the construction process in the real-world. The construction’s real-life scenario is mapped to the virtual area using DT technology, allowing additional modelling study. Research on prefabricated building construction using BIM and DTs technology has been conducted.

The construction model’s initial state can be mathematically represented as a function of various parameters such as dimensions, material properties, and environmental factors for real-time monitoring.

As shown in Eq. (1), where \(\:\text{N}\left({\text{t}}_{0}\right)\) denotes the initial model at time \(\:{\text{t}}_{0}\), \(\:D\:\)signifies the dimensional data (e.g., 3D model), \(\:\text{P}\) indicates the material properties, \(\:\text{E}\) denotes environmental conditions.

Equation (2) shows that a six-dimensional DT model for the construction management of buildings can be built by following the rules, adding the construction mentioned above elements for buildings, and referencing model elements, objectives, and criteria from existing DT systems.

The model includes VG, PG, DG, SG, and MG, where VG signifies the virtual entity containing construction components, PG denotes the physical entity of construction, DG comprises all twin data and fuses virtual and physical entities, SG is the construction process service system, and MG indicates the existing project management systems covering engineering data. The interrelation between VG, PG, DG, SG, and MG is signified by CG. The model’s principal data source and the cornerstone of the service process are the green components in both real and virtual entities. ‘People, machine, material, technique, and environment.’ Green information is gathered at four points in the real world: transit, storage, lifting, and installation. The virtual model conducts modelling, simulation, optimization, interaction, and visualization at the ‘geometry-physics-behaviour-rule’ level. Collecting and analyzing a wide variety of data is necessary to effectively manage a project’s timeline, cost, quality, and safety. Optimization of green construction processes, monitoring energy usage, optimization of equipment operation, recycling of materials, and deployment are all tasks offered by the management service layer.

Figure 1 shows the sequence diagram. To start the simulation, the user uses the User Interface (UI) to submit a request to the Digital Twin System. This system functions as the primary coordinator in charge of overseeing the whole simulation. As soon as the request comes in, the DT System starts the required building models by communicating with the Multi-dimensional Model Engine. After that, the Multi-dimensional Model Engine gets the necessary model data like 3D building structures, environmental data, and construction schedules from the Data Management System. After retrieval and processing, it is returned to the DT System to include the processed model data in the simulation environment. After the models are ready, the DT System tells the Simulation Engine how to customize the simulation. Environmental factors, the building schedule, and the layout of the necessary equipment are all examples of such characteristics. To gather any extra data that may be needed during this setup step, the Simulation Engine communicates with the Data Management System. The Simulation Engine will simulate after the necessary setup is finished. To do this, you must communicate with the Construction Equipment via the simulation and provide it with precise instructions. The machinery reports to the sim on the building operations’ location, status, and real-time advancement. The Feedback System receives this real-time data from the Simulation Engine for analysis. Digital Twin System receives processed data, such as simulation performance, possible problems, and required modifications, from the Feedback System, which updates it continually. Afterwards, the Digital Twin System communicates this information to the user interface, allowing the user to see the building process in real-time and engage with the simulation while running. Lastly, whether the simulation is running or has finished, the Digital Twin System will produce a comprehensive report and transmit it to the user interface for evaluation. Information on the effectiveness of the building process, the functioning of the machinery, and any schedule deviations may be included in this report. Throughout this whole process, every component is vital to the success of the intelligent construction simulation in delivering an interactive and accurate platform for optimizing and making decisions on building projects. The research assumes that all sensors and IoT devices are functioning as intended, ignoring that problems like data loss, errors, and communication failures may significantly impact system performance. For example, misreading of thermal conditions may occur if temperature sensors used in smart buildings wander by around 1.5 °C over six months. Data accuracy may be reduced by as much as 10% when using LiDAR-based structural monitoring systems because of occlusions caused by dust or fog. Zigbee and LoRaWAN are two wireless transmission systems that may have packet loss rates higher than 5% in highly interference settings. This can result in incomplete data logs and inaccurate analytics. These problems may reduce the efficacy of real-time monitoring and decision-making without strong error-handling procedures. Internet of Things (IoT) digital twin applications may be more reliable using adaptive error correction techniques, redundant multi-sensor networks, and Kalman filtering for sensor fusion to reduce data discrepancies.

Corresponding behaviour patterns indicate a construction system’s capabilities; the system works when all these patterns are present. Analyzing the system’s behaviour mode, which is the product of multiple features interacting, might provide information on the system’s dependability. However, determining the internal behaviour mode becomes more challenging when dealing with large building systems since they often include more qualities with coupling interactions. It is worth noting that data acquired in real-time on the job site may accurately portray the current state of the resource equipment. Consequently, the digital twin rule model’s primary focus is on assessing the state of the construction system via the analysis of on-site real-time data and the quantitative determination of the relationships between various parameters.

The data sequence \(\:{x}_{1},\:{x}_{2},\:{x}_{3},\:…,\:{x}_{m}\:\)is transmitted unceasingly at fixed speed as the digital layer’s vital point tracking information of the construction lines in real-time. One can use the sliding window models to assess this data or to examine gathered information that has been updated over time. It is possible to use a matrix to represent the data that has to be handled in the digital layer.

In the Eq. (3), \(\:{C}_{DT}\:\)signifies the digital tie layer information matrices, composed of \(\:m\) components signifying various characteristics, which can define the force, temperature, and other characteristics of the vital tool of construction lines, and values respective to the attribute are signified by \(\:{d}_{nn}\).

As a means of communication, measuring and regulating building things is crucial for DT technology. Further, DTs are driven by monitoring sensors that complete the data collection for the prefabricated structure. Using this method, the displacement and rotating torque of the construction components may be properly monitored while assembled. This study successfully transfers and analyzes the data and manages every part of the assembly process. The following objective function represents the model algorithm’s acquired building picture data. One way to state the objective function principle for energy consumption reduction is:

In Eq. (4), where \(\:s\) signifies the weight matrices, \(\:{s}^{T}\:\)denotes the reversal of the weight matrices, \(\:{y}^{j}\:\)symbolizes the input, \(\:{x}_{j}\:\)indicates the output, \(\:a\) specifies the deviations, \(\:{\lambda\:}_{1}\:\)designates the eigenvalues of matrices and \(\:{v}_{jj}\:\)signifies the component values of the \(\:jth\) row and the \(\:j\)-column. Various security systems, including perimeter security personnel, access controls, video surveillance, and smoke detectors, may identify when an unsafe situation has occurred inside the structure. In addition to intelligent security systems, the building’s facilities systems, including ventilation, smoke prevention and exhaust, and automated fire extinguishing, adjust the conditions inside the building by acting as actuators for indoor environment quality prediction.

In the Eq. (5), \(\:{N}_{SEQ}\:\)signifies the physical model of building smart security equipment; \(\:ID\) is identification codes of smart security devices, corresponding to the smart security device individually; \(\:type\) is the type and model of the device; \(\:production\) is the manufacturer of the device; \(\:date\) is the data that the device has installed; \(\:location\) is the location of the device; \(\:device\) is the parameter set of devices, which is a set containing several variables. There is a wide variety of parameter settings. Device states such as shutdown and running, as well as the symbol state, are abnormal. Building equipment uses a lot of power and produces a lot of pollution. Mainly, power and fuel are used up by the machinery throughout the building assembly process. Typically, while equipment runs, a smart energy consumption and emission acquisition tool is utilized to gather energy consumption and emissions data. Integrating heterogeneous sensors into mechanical equipment allows for collecting machine location data in real-time, and installing an embedded terminal allows for storing and transmitting this data. Using these strategies, one may accomplish green machinery and equipment management from the beginning to the end of the building process. The model assumes that the sensor accuracy is always within ± 0.5 °C for temperature readings and ± 2% RH for humidity. However, in actual deployments, the drift may reach 1.5 °C and 5% RH over long periods. Likewise, it is presumed that data transmission latency is less than 10 ms, but network-induced delays might approach 50–100 ms under high-load situations, which affects real-time monitoring. Furthermore, optimization tactics might be impacted by energy consumption projections of 120 kWh, which could have an unexplained variation of ± 8 kWh due to the lack of error margins in the predictive outputs. To improve accuracy, a more thorough evaluation of the model’s performance might be achieved by using an extended Kalman filter (EKF) to account for sensor drift or by conducting Monte Carlo simulations with 95% confidence intervals.

Digital twin model in intelligent construction management.

Figure 2 shows the digital twin model in intelligent construction management. Digital Twin technology potentially streamlines comfort assessment by combining BIM with sensors and real-time building data. This research could not locate previous work that improved data collection, occupant feedback visualization, or understanding of building discomfort triggers using Digital Twin technologies and risk assessment models. A Bayesian network model for occupant comfort level evaluation and an HVAC automated defect detection system with innovative approaches for leveraging BIM as a visualization platform and predictive maintenance are presented in our article. One unique aspect of our work is integrating JSON data for better integration and presenting a BIM plug-in that can handle data every 5 min. These features were not accessible before. Additionally, 10 variables are examined together for the first time to tackle the issue of space adequacy. The performance of the nine machine learning algorithms was evaluated utilizing different matrices, necessitating substantial data processing and real-time training and prediction. Fresh issues, such as compressor failure, were identified using new algorithms, and a novel approach to visualize outside building system problems is shown. Several criteria are used to propose a new way of estimating the remaining usable life of HVAC systems, which may extend their lifespan by at least 10% and lead to substantial savings. This article presents several new methods that improve HVAC defect detection and occupant comfort assessment. Compared to existing models, the proposed MD-DTT-BIM model enhances the operational efficiency ratio, real-time monitoring ratio, energy consumption reduction, occupant satisfaction ratio, and indoor environment quality prediction ratio.

Proposed MD-DTT-BIM model.

Figure 3 shows the proposed MD-DTT-BIM model. Historical data from completed projects may be used for benchmarking and comparative analysis. Timelines, expenditures, materials, and building techniques could all be part of it. To offer live updates on factors like temperature, humidity, and equipment condition, sensors and Internet of Things devices installed on the building site gather data in real-time. Information gathered from outside the construction project, such as weather predictions, supply chain updates, regulatory details, and other relevant elements, is known as external data. Physical features such as temperature, pressure, or structural strain may be measured and recorded in real-time using sensors. The Internet of Things (IoT) may supplement existing data on construction equipment, worker actions, and site conditions. With these devices, data collecting and monitoring may go on indefinitely. Integrating data from sensors, historical records, and external inputs are all examples of data sources that may be included in a single structure via data aggregation. Statistical and computational approaches are used in data analysis to understand the collected information, identify trends, and derive valid conclusions. This may require looking at past data for strategic planning or current data for fast decision-making. BIM Integration creates a comprehensive, organized digital model of a building project that incorporates structural, architectural, and MEP (Mechanical, Electrical, and Plumbing) details. BIM data is merged with it to make the digital twin more accurate and complete. Using 3D/4D modelling, one may represent the building site’s physical structure in three dimensions, and one can generate a representation of the construction progress and future states in four dimensions by including time. An engine component employs the digital twin to test hypotheses, make predictions, and examine the impact of alternative building techniques or environmental factors. Spatial linkages and geometric physical dimensions in the building project, including room layouts, structural components, and the overall structure design. The dimensions of the materials used in the building and details on their qualities, amounts, and where they were located. Dimension of time-building schedules, project milestones, and phase-specific timetables are all examples of time-related data. Energy Measurement Information about energy use and efficiency, including the effects of various building techniques and materials on energy consumption. Dimension of Cost Estimates, actual and projected costs are all parts of the project’s financial data. Information on safety procedures, risk evaluations, and incident monitoring is provided by Safety Dimension to guarantee a safe working environment on construction sites. The graphical user interface (GUI) dashboard facilitates user interaction with the model, data viewing, and simulation control. To determine the effect on the project, Scenario Analysis involves designing and testing several scenarios, such as alternative building techniques, materials, or timetables. Decision Support System tools that examine the outcomes of simulations and provide suggestions or insights to aid users in making well-informed decisions. Automation On the job site, automated technologies like robotics, drones, and autonomous vehicles carry out tasks with greater accuracy and efficiency.

The suggested approach considers the impact of regulatory requirements, such as safety laws, environmental compliance standards, and building rules, on project execution schedules and resource allocation. Project efficiency and the capacity to make real-time decisions are affected by the unpredictability introduced by variables like labour shortages, severe weather, supply chain interruptions, and unanticipated delays caused by permits and approvals. Due to the ever-changing and frequently unpredictable nature of external restrictions, the framework’s flexibility to real-world applications is crucial. Intelligent construction systems might benefit from improved decision-making capabilities by adding constraint-aware optimization methods and risk assessment models, increasing the model’s practicality.