Construction Research Congress 2012 © ASCE 2012
A Framework for Automatic Safety Checking of Building Information Models Sijie ZHANG1, Jin-Kook LEE2, Manu VENUGOPAL3, Jochen TEIZER4, and Charles M. EASTMAN5 1
PhD Student, School of Civil and Environmental Engineering, Georgia Institute of Technology, 790 Atlantic Dr. N. W., Atlanta, GA, 30332; PH (404)-901-2441; email:
[email protected] 2 Research Scientist, College of Architect, Georgia Institute of Technology, Atlanta, GA; email:
[email protected] 3 PhD Candidate, School of Civil and Environmental Engineering, Georgia Institute of Technology, 790 Atlantic Dr. N. W., Atlanta, GA, 30332; PH (510)-579-8656; email:
[email protected] 4 Assistant Professor, School of Civil and Environmental Engineering, Georgia Institute of Technology, 790 Atlantic Dr. N. W., Atlanta, GA, 30332; PH (404)-8948269; email:
[email protected] 5 Professor, College of Computing and College of Architecture, Georgia Institute of Technology, Atlanta, GA, 30332; email:
[email protected] ABSTRACT In the past two decades more than 26,000 U.S. construction workers have died at work. Of these fatalities, 40% involved falls. It is indicated that safety in construction remains a big problem. Safe construction requires planning throughout the project lifecycle, from design, planning, through construction execution and extending into operations and maintenance. The literature and past research show that there is a lack of tools and resources to assist designers and engineers with addressing construction safety. Despite the implementation of safety practices, most of them applied in the field are primarily text-based checklist. Further improvements can be gained in construction safety through the use of technology. This paper contributes in solving this problem by developing a framework of automatic safety checking to Building Information Models (BIM). The presented framework and case study extends BIM to include automated hazard identification and correction during construction planning and in certain cases, during design. As a result, the developed automated safety checking platform informs construction engineers and managers in knowing, why, where, when, and what safety measures are needed for preventing fallrelated accidents before construction starts. Keywords: Building Information Model (BIM), Fall Protection, Construction Safety, Planning, Rule Checking, Simulation. INTRODUCTION A key for successful projects in all industries is safety. As good safety practices and records create a positive, hazard free, and productive work environment, planning for safety at the front-end of a project is not only the first but also a fundamental step for managing safety (Waly 2002). Typical safety planning consists of the identification of all potential hazards as well as the decision on choosing
574
Construction Research Congress 2012 © ASCE 2012
correspond safety measurement (Bansal 2011). Precisely and accurately identifying the potential safety hazards is critical to the safety planning process. Traditional safety planning relies on manual observation, is labor-intensive, time-consuming, and thus inefficient. Safety manager/superintendents can hardly be aware of all potential hazards on job sites at all times. The existing safety rules and regulations provided by OHSA and industrial best practices have been successful in lowering incident rates, such as Total Recordable Incident Rates (TRIR) according to the Construction Industry Institute (CII) (BMM2010-2). However, by applying experience based knowledge, it is hard for them to learn, compare, combine, and/or implement for safety planning. From another aspect, it is difficult for workers to acquire the knowledge in a short time and stick to the rules accordingly when performing work tasks (Cox and Cheyne 2000). Another main problem in current safety planning system is that design choices often determine construction methods and schedule while limited attention is given to safety during the design phase (Hinze and Wiegan 1992). Construction site safety often remains the sole responsibility of the contractor. NEED FOR AN AUTOMATED RULE-BASED SAFETY CHECKING SYSTEM The planning and design phases provide a vital opportunity to eliminate hazards before they appear on a job site. Previous research indicates, there is a lack of responsive tools and resources to assist designers with addressing construction safety (Ku et al. 2010). Technology is believed (Teizer et al. 2007) can play a key role in reducing incident rates further once it positively influences current practices in safety planning. The growing implementation of BIM in the A/E/C industry is changing the way safety is approached today. In an attempt to take advantage of the potential of BIM for safety and health in construction (building) design and planning and to further facilitate the integration of construction safety and health in BIM, automated hazard recognition and prevention method generation is proposed. We limited the research scope and focused on developing the rule implementation for fall protection. Falls are the most frequently cited type of fatality in construction. Falls are the cause of many serious injuries and fatalities. More than half the falls in construction are related to environmental factors involving somewhat the working surface or facility layout conditions. Inadequate or inappropriate use of fall protection Personal Protective Equipment (PPE) and removed or inoperative safety equipment contributed to more than 30% of the falls (Huang and Hinze 2003). FRAMEWORK AND METHODOLOGY The proposed framework of a rule-based safety checking system is illustrated in Figure 1. The first step is to collect and analyze construction data including work breakdown structure, schedule, and cost. These typically provide some of the most important technical aspects of a project by defining the project in terms of hierarchically-related, product-oriented elements and the work processes required for each element's completion. Each element of the Work Breakdown Structure (WBS)
575
Construction Research Congress 2012 © ASCE 2012
provides logical summary points for measuring cost and schedule performance. Afterwards, this information can be represented in BIM applying the corresponding construction schedule. Compared to the traditional process, the proposed system takes the existing safety rules, guidelines, and best practices (e.g., construction safety standards from Occupational Safety and Health Administration (OSHA) or industry best practices) and links them with the building model. Through the data fusion, potential safety hazards as well as suggested solutions can be automatically generated and provided by mapping the rules and building objects. Since the building model is usually updated during the project design and operation phases, the safety checking system is connected to the system and can be re-run after each model or schedule update to ensure the planning for safety at front-end of the project.
Figure 1. Framework for Implementing an Automated Rule-based Safety Checking System After the developed safety rule checking system has identified the hazards in the BIM, design for safety and safety planning can be conducted. The goal of the rule checking system is to assist human decision makers in the safety planning and scheduling task by proposing realistic solutions to resolve the identified issues. The rule checking results first can be communicated to the designer along with corresponding safety requirements from the contractor to facilitate safer design since the designer is able to modify the model directly after the checking results. Especially in Design-Build project delivery method, the use of the design for construction safety concept may gain momentum (Behm 2005). The safety knowledge is transferred from contractor to designer based on the checking results and specific design-forsafety suggestions. Then, after model update and safety re-check, the system is used by the contractor for safety planning. Thirdly, safety reports are generated by the rule checking system. Since the prevention methods are visualized they engage human decision makers through a three-dimensional immersive environment. Such view can enable better decision making and increase the awareness of project participants, including workers, e.g., in pre-task planning or daily meetings. At the operational and monitoring stage of a project, automatic reports in table form can guide the installation of protective equipment correctly and in a timely manner. They can also be used to document safety issues and become part of legal project obligations. Rather than performing safety inspection based on experience, a safety manager who needs to conduct site inspection everyday can use the reporting feature to ensure the safety of the site is following strictly the designed safety plan. The
576
Construction Research Congress 2012 © ASCE 2012
effectiveness of safety education and training for construction workers on site can be improved since interactive modules can be prepared that are much more engaging than paper- or lecture-based techniques. Trends in weekly or daily safety training can be recorded and potential safety hazards along with location and prevention methods can be visualized. Workers can become much more aware of these hazards before work is started. One of the promising directions of BIM applications in AEC industry is to facilitate various rule checking and simulations for evaluating building designs in earlier phase of project (Eastman et al. 2008). Rule-based systems apply rules, execute them to a given model, and return the reports which basically consist of “pass” or “fail”. The process of rule checking is composed of four major stages (Eastman et al. 2009): 1) Rule interpretation: The interpretation of safety rules from safety regulation or best practices is logic-based mapping from human language to machine readable form. The name, type, and other properties in the rule can be analyzed and extracted from the rule. They can then be classified in differing site conditions using IF-THEN context to determine the corresponding measures. Rule translation typically has two aspects: (a) the condition or context where the rule applies and (b) the properties upon which the rule applies. The first step might identify the target building object for example a slab, and the second step would then check the width, length, location, etc. of the identified slab. 2) Building model preparation: In object-based modeling, all building objects have type and properties. This information can be used in addition to checking geometric features. Thus the requirements of a rule checker for building models are stricter than existing 2D drawing or 3D modeling requirements. The basic requirement for the rule-based checking system is that the building objects need to be modeled correctly in terms of name, type, attributes. 3) Rule execution: The rule execution phase brings together the translated rule sets with prepared building model. Since the rules have been transformed to machine readable code, their executions are straightforward. The building objects can be mapped to the rule sets by name, type or other attributes. Providing the variety of both building project and safety protection method, the rule execution is designed to have two steps: (a) automatically check the model and apply safety measures according to default settings/suggested solution, and (b) provide all possible solutions which can be selected or changed according to an individual best practice after automated checking. 4) Rule checking reporting: The checking results will be reported in two different forms: (a) visualization of applied safety protective equipment, and (b) table-based check of the results showing detailed information from model and applied solution. In addition, quantity-take-off information for resource leveling of safety equipment and importing the generated information into project schedules is also possible.
577
Construction Research Congress 2012 © ASCE 2012
DEVELOPMENT OF AN AUTOMATED RULE-BASED SAFETY CHECKING SYSTEM FOR FALL PROTECTION According to OSHA regulations (OSHA), fall protection rules can be classified into three parts: (1) definition, (2) general requirement, and (3) prevention criteria. Definitions specify the unsafe area; general requirements show the protection methods which should be applied in a specific scenario; and prevention criteria relates to the detailed information of the prevention system to be used. Safety checking rules need at least three components: (1) the objects, attributes and relations needed to represent a safety condition, and (2) the logic for carrying out the assessment. Once a safety condition is identified, a third aspect comes into play: (3) how to resolve the safety issue. An initial set of rules was generated using a set of fall protection rules from OSHA. A more comprehensive open source repository for organization-based safety rules and regulations can be extended in the future. The research focused on fall prevention, such as openings in slabs, edges on floor, and openings in walls. According to OSHA, a “hole” means a gap or void of two inches (5.1 cm) or more in its least dimension. A hole or opening can exist in a floor (e.g., concrete slab), roof (e.g., skylight), wall, (e.g., window), or any other walking/working surface. Regardless of its length, we implemented a default fall arrest system (e.g., guardrail system for edges on slabs or for openings in walls), if the location of the object was elevated more than 1.8 meters (six feet). For holes on a floor measuring more than a pre-defined value, for example 1.5 meter (59 inches), in its least dimension, we applied also a guardrail system. Holes were “covered” if an opening measured less than one meter but more than five centimeters in its least dimension. Holes with less than five centimeters (two inches according to OSHA) in its least dimension were ignored (due to the small size of the hole and lower likelihood objects falling through). The default table-based safety rule translation for fall protection is shown in Table 1. Least Dimension (x) of a Slab Opening Prevention Method < 5 cm “Not considered” 5 cm < x < 1.5 m “Cover” > 1.5 m “Guardrail system” Table 1. Example of Table-based Rule Interpretation Then different scenarios are determined by acquiring the corresponding geometric information to each object: (1) exterior walls are detected to determine where edge protection is needed; (2) openings in slabs/roofs are detected to prevent fall through openings; (3) openings in exterior walls are detected to determine where additional wall opening protection is required; (4) interior walls around slab openings are detected for fall protection from wall openings. The different cases of fall protection scenarios are listed in Table 2, edge in red represent edges need to be protected.
578
Construction Research Congress 2012 © ASCE 2012
Case
Case1
579
Case2
Case3
Diagram
*edges that need protection are highlighted in red. Table 2. Different cases of fall protection After applying and visualizing an automated version of rule checking, human input is optional to assist in the final decision making process. Finally, the checking results and visualization are updated in the BIM. Each hazard is detected and the proper protection method is shown. After the checking and visualization process ends, the safety rule checkers provides additional information with high relevance to decision makers: (a) quantity take-off of the required safety protective equipment, and (b) scheduling of the installation of safety protective system. CASE STUDY A building model was created showing a construction site in progress for test purpose. The BIM includes different types of openings that could be a potential fall hazard. The identified openings have different sizes and geometric shapes. The holes are located on walls and floor slabs. The model includes a four-story building with walls on the first two floors and emergency staircases.
Guardrail System
Cover
Guardrail System
Figure 2. Examples of slab edge, slab openings, wall opening with different shapes/dimensions and protection Figure 2 show examples of slab edge, opening and wall opening which were detected by the system. Safety protective equipment was installed automatically. According to the interpreted rules, small holes are protected by “cover”; larger ones are protected by installing guardrail system. All potential fall hazards were 100% detected and corresponding safety protective equipment were installed.
Construction Research Congress 2012 © ASCE 2012
CONCLUSION AND DISCUSSION This research outlined a framework for a rule-based checking system for safety planning and simulation integrating BIM and safety. It is helpful to identify potential safety hazards and provide corresponding prevention methods in an automated approach. The design-for-safety concept can be implemented for early hazard identification in design model and communication between designers and contractors. The table-based safety rule translation prototype was developed based on fall protection from OSHA rules and construction safety best practices. Preliminary results demonstrate the feasibility of the developed safety rule-checker on the BIM tool. The developed automated safety-rule model checker shows good capability of practical applications in building information modeling and planning of work tasks. Once applied in construction design and execution phase, it may possess large potential for reducing errors and waste in safety planning for construction site work sequences and activities. From a safety management perspective, time and effort of safety staff/engineers can be saved through an automated safety code checking and simulation tool that assists labor-intensive safety tasks. The next steps of the research may focus on research that studies the applicability and performance of the safety-rule checker in simple to complex building information models. The analysis of parametric and complex rules customized to the specific type of project might also be explored. It is envisioned that the implementation of the safety prevention methods might need to be adapted to the scope of the project, type of firm, and the design process. Research will also need to focus on surveying additional cases studies, scenarios, guidelines, and best practices to explore system applicability and to convince practitioners of its usefulness. REFERENCES Bansal, V.K. (2011). “Application of geographic information systems in construction safety planning.” International Journal of Project Management, 29 (1) 66–77. Behm, M. (2005). “Linking construction fatalities to the design for construction safety concept.” Safety Science, 43(8) 589-611. BMM2010-2, CII Safety Report (2010), Construction Industry Institute, The University of Texas at Austin. Cox, S. J., and Cheyne, A. J. T. (2000). ‘‘Assessing safety culture in offshore environments.’’ Safety Sci., 34, 111–129. Eastman, C.M., Teicholz, P., Sacks, R., Liston, K. (2008) “BIM Handbook: A Guide to Building Information Modeling for Owners, Managers, Architects, Engineers, Contractors, and Fabricators.” John Wiley and Sons, Hoboken, NJ. Eastman, C.M., Lee, J., Jeong, Y., and Lee, J. (2009). “Automatic rule-based checking of building designs.” Automation in Construction, 18(8) 1011-1033. Hinze, J. and Wiegand, F. (1992). “Role of designers in construction worker safety.” J. Construction Eng. Manage., 118 (4) 677–684. Huang, X., Hinze, J. (2003). “Analysis of construction worker fall accidents.” Journal of Construction Engineering Management, 129 (3) 262–271. Ku, K. (2010). “Research needs for Building Information Modeling for Construction Safety”, Ph.D. dissertation, Virginia Tech.
580
Construction Research Congress 2012 © ASCE 2012
Safety and Health Topics: Fall Protection. Occupational Safety and Health Administration Home. Accessed May 2011. . Teizer, J., Caldas, C.H., Haas, C.T. (2007). “Real-time three-dimensional occupancy grid modeling for the detection and tracking of construction resources.” Journal of Construction Engineering Management, 133 (11) 880–888. Waly, A.F., Thabet, W.Y. (2002). “A virtual construction environment for preconstruction planning.” Automation in Construction, 12 139–154.
581
