A 2025 automated system, and Joo’s 2017 BIM risk model

Despite institutional efforts and new tools, accidents on construction sites have not fallen as hoped, the notes say, in part because safety work is still manual and unstandardized, and because people on site often cannot yet use BIM well. In a 2025 Sungkyunkwan University master’s thesis, Geum Dae-yeon automates BIM safety-object placement and risk assessment with RPA, leaving room for an engineer’s judgment, and records about six times the time efficiency in risk assessment, three times in writing a work plan and in education, five times in approval, and fifteen times in prior management of unsafe conditions — figures left as that study places them.

Joo Hyeong-woo’s 2017 Sungkyunkwan master’s thesis joined BIM and risk-assessment methods so that hazards could be modeled and visualized in advance, rather than managed mainly after the fact. A model that systematized accident types, kinds of work, and spatial information was checked twice by Delphi and set out in field scenarios with a model viewer; it was found to allow more precise, consistent safety planning, with positive expert feedback on usefulness.

Lee’s 2023 fire-safety tool, and Park’s 2018 reverse design

Lee Byeong-heun’s 2023 Hoseo University doctoral thesis starts from the limits of an insulation-centered design, which can hardly reflect ejected flame, windows coming away, or damage on upper floors. A BIM simulation and design guidance were built to predict fire spread through exterior openings, using fire-load density, flashover limits, and heat-release rates, and refined with physical tests, including heat-stress tests on aluminum and PVC frames. Reliability was also checked against Grenfell Tower, where, as the 2023 thesis places it, the path of ejected flame and the predicted range of damage were similar to the actual case.

Park Hye-jin’s 2018 Kyungpook National University master’s thesis applies a BIM reverse-design survey to aging buildings with no drawings, or incomplete ones, combining 3D measurement, photographs, and restored plans so that connections, deformation, and defects can be read in the model and turned into a diagnosis report. Field cases, as that study places them, showed time saved and quality raised, more markedly in older, more complex buildings, and administrative time shortened. This page stays with the 2025 figures and with these earlier theses, and does not add later official construction-safety or fire-casualty tallies. The same Korean report, at fuller length, is on CampusDown.

Sources

1. This page reads prior research gathered for a graduation project on automating highly variable safety-management work on Seoul construction sites. Despite institutional efforts and technical progress aimed at lowering the accident rate, accidents have still not decreased, and one cause named is a manual, unstandardized way of managing safety. Risk-assessment work is inefficient because information is limited and dependence on hand work is high, and even after BIM technology was introduced, a further barrier is that people on site often lack the ability to use it. The work defines the problem from a practical understanding of everyday safety tasks on site, then develops and verifies a technology and a system to address it, on the ground that a lasting effect can be expected only when obstacles on the actual site are clearly seen and improvement is fitted to them. The flow of site safety work and the methods of practitioners were first surveyed, and related prior research was analyzed so that existing problems and earlier attempts at improvement could be set in order. A system using automation technology (RPA) and BIM was then designed on the basis of the problems analyzed, applied on site, and measured for actual effect. The main study is Geum Dae-yeon’s 2025 master’s thesis at Sungkyunkwan University, in Seoul, a study in developing a system to automate highly variable safety-management work on Seoul construction sites. It automates BIM-based placement of safety objects and the risk-assessment process by an RPA method, and presents a way of raising practical applicability through an intuitive user interface and a parametric BIM library. The core is a system structure that efficiently automates repetitive, highly variable site safety work while still allowing an engineer’s professional judgment and a responsible decision. BIM-based digital information was linked with RPA technology, and a striking point is that complicated procedures such as writing, reviewing, and approving safety documents were made more efficient. According to figures in the body of that research, risk-assessment work recorded about six times the time efficiency, writing a work plan three times, education three times, and approval five times. Prior management of unsafe conditions showed a fifteen-fold gain in efficiency, which is taken to suggest that automation is highly effective for variable, repetitive safety work. Those 2025 figures are left as the study places them, and later official construction-safety tallies are not added. Joo Hyeong-woo’s 2017 master’s thesis at Sungkyunkwan University, in Seoul, tried to build a more systematic safety-management system by integrating BIM technology and risk-assessment methods into safety work on a construction site. It is a study centered on the need for a system that, leaving a qualitative and subjective way of managing safety, models and visualizes in advance the possibility of an accident and hazard factors, so that objective decisions become possible. The researcher structured hazard factors in relation to spatial information through BIM, and designed a model so that this could be used directly in safety-management procedures. The study concentrates on empirically analyzing the limits of existing management methods and on verifying whether a BIM-based alternative can actually be applied. The method first carried out a theoretical review of risk assessment, and on that basis developed a BIM model that systematizes accident types, kinds of work, and spatial information, because existing safety management was often more centered on response than on prevention, and a safety plan based on prior prediction and visualization was judged necessary. The model’s effectiveness and possibility of use were then verified twice through a Delphi method among experts. In order to present a practical way of applying the BIM model, field-centered working scenarios were also included, using existing software and a Model Viewer. The thesis is of interest for extending BIM from a mere design tool to a safety-management platform for preventing accidents. In the course of modeling, hazard factors were coded and structured by being distinguished according to accident type and process. The thesis presented a result that a risk-assessment model built on BIM makes possible more precise and consistent establishment of a safety plan than existing safety-management methods. Through experts’ Delphi evaluation, positive feedback on the system’s possibility of use and effectiveness was confirmed, a result that also showed a high possibility of practical application. These research results suggest that BIM technology need not remain only at the design stage, but can operate as a practical risk-management tool on a construction site. The system indicates that it can make repetitive, complex hazard factors visually clear, and can also be usefully applied in worker education and in the process of decision-making. It also suggested that later research would need a strategy for expanding field application that considers feasibility and economy. Lee Byeong-heun’s 2023 doctoral thesis at the graduate school of Hoseo University, in Chungcheongnam-do, is a study that tries to predict, through BIM technology, the risk of fire spread through exterior-wall windows, and to propose a fire-safety design process on that basis. It starts from the awareness that an existing design method centered on insulation performance can hardly reflect various actual hazard factors such as flame ejected to the outside, windows coming away, and damage on upper floors. The researcher therefore focused on developing a BIM-based simulation system and design guidance that can quantitatively predict fire spread through openings in a building’s exterior wall. The work did not stop at modeling, but treated comprehensively the distribution of combustibles, fire-load density, limits at which flashover occurs, heat-release rate, and structural analysis. The tool was refined on the basis of physical experiments and simulation data, and a process was built so that it could actually be applied at the design stage, a study that can be assessed as having made a connecting link among technology, policy, and the field. A process was designed that fuses data of various layers and precisely predicts an actual fire situation. Fire loads were first classified according to the use of the building, and fire-growth rates and conditions for the occurrence of flashover that reflect the combustion character of each combustible were turned into numbers, because it is hard to explain properly the degree of risk of various building types by a single standard. An algorithm was then built, reflecting both laboratory conditions and actual cases, that can predict the moment when flame is ejected, the moment when an exterior-wall window comes away, and the intensity and range of heat release. A simulation tool that can be linked with BIM software was then developed, making it possible to visualize, at the design stage, the trajectory of ejected flame, temperature change, the range of radiant-heat spread, and so on. The tool’s reliability was verified by applying a large actual fire case such as Grenfell Tower, and in that process it was also experimentally demonstrated how much the moment of dropout and the range of fire spread change according to differences in window material. The core is that an analysis-based design environment was built, not a mere tool. The thesis made it possible to predict numerically fire-spread scenarios at exterior openings that existing design had not considered, such as the path of fire spread, the moment flashover occurs, and the possibility of a window coming away. The process by which flame ejected to the outside spreads to an adjacent building or an upper floor had until then been hard to interpret quantitatively, but through a BIM-linked tool, spatial analysis and risk prediction became possible already at the pre-construction stage, which is placed as the greatest achievement: a preemptive, digitally predictive design rather than a merely manual, drawing-based defense. In this research, actual heat-stress experiments were carried out on aluminum and PVC frames, and on the basis of these data the moment of a window’s structural dropout was quantified. It became possible to predict when an exterior-wall window would collapse in a fire, and to simulate in BIM the height of the flame that would be ejected at that moment, the axial temperature, the range of radiant heat, and so on. In a simulation that applied the actual Grenfell Tower case, the trajectory of ejected flame and the prediction of the range of damage showed results similar to the actual case, as the 2023 thesis places that check, demonstrating that a BIM-based tool can be a realistic alternative in building fire-safety design. The notes also describe a method whose core is to compose a BIM model that reflects actual elements such as structure, materials, occupant composition, and the location of combustibles, then to simulate fire spread, heat transfer, and changes in the concentration of toxic gas through FDS. Park Hye-jin’s 2018 master’s thesis at the graduate school of Kyungpook National University, in Daegu, is a study that tries, by applying a BIM-based reverse-design technique to the safety-diagnosis process of aging buildings that have no drawings or incomplete ones, to raise survey accuracy and to achieve standardization of data. In order to overcome the limits of existing survey methods — breaks in information, dependence on the surveyor’s experience, errors in handwritten records, and so on — it seeks a way of using BIM structurally from the stage before the survey. The study organizes the whole process of survey, record, and diagnosis on a digital basis, so that information can be used consistently through subsequent diagnosis and the establishment of a repair plan. The thesis therefore focuses on practically showing that BIM can be used not only at the new-construction stage, but also in the maintenance of aging buildings and in diagnosis based on life-cycle cost (LCC). The researcher first analyzed the basic components of BIM, and designed a field-centered survey system to make up for the difficulty of obtaining information from an aging building. Various data such as 3D measurement, photography, and restoration of existing drawings were integrated, and on that basis the structure of the actual building was realized inside a BIM environment, because an existing handwritten survey method has limits in accuracy and consistency. Through the BIM model generated, connection relations among members, the deformed state of the structure, defect factors, and so on were visually analyzed, and a template that can convert this into a diagnosis report was also developed. Document output in a format that can be submitted to a public agency, and a data-conversion function, were also examined, with emphasis on securing practical applicability. Through field cases it was verified that this system can contribute to time saving and quality improvement in actual safety-diagnosis work, as the 2018 thesis places those cases. Even in a building that has no drawings or has aged, a BIM model can be reconstructed through measurement and reverse design. As a result, by applying BIM-based reverse design to the safety-diagnosis process of an aging building, the thesis empirically demonstrated that precise information gathering and spatial analysis are possible. Existing survey methods had large limits such as missing drawings, errors in records, and insufficient field information, but by reconstructing the state of the actual structure as a digital model through a reverse-design technique, this problem could be resolved to a considerable extent. The more complex and old the building, the more marked the efficiency of this method. Investigators, instead of handwritten drawing work, integrated various information such as 3D scanning, measurement, photographs, and existing documents to build a BIM model, and this model served as reference information for subsequent safety diagnosis. Through this process it was emphasized that, going beyond a mere visual reproduction, prior detection of practical hazard factors such as interference among members, deterioration factors, and structural cracks became possible. As the integration and accuracy of the data rose, the reliability of the diagnosis report rose with it. BIM-based survey results were also used as reports for submission to the agencies concerned, and an effect of shortening administrative processing time was also confirmed. This shows that a BIM model can sufficiently perform a function not only as an information tool for internal technicians, but also as a means of communication with the outside. A 2019 master’s thesis at Changwon University, in Gyeongsangnam-do, appears only as a bibliography line, without a title recovered here, and is not filled in. This page stays with the 2025 time-efficiency figures and with the 2017, 2018, and 2023 theses as they are placed, and does not add later official construction-safety tallies or later official fire-casualty figures.

Last updated: 14 November 2026