Introduction

Construction projects are dynamic and complex, and hence, many building codes, standards, and guidelines were developed to maintain a certain quality in different aspects of the construction sector, such as the level of hazard protection, construction safety, occupant safety, and environmental protection [1,2,3,4,5]. Design engineers should comply with the codes and design guidelines [6]. Noncompliance with building design codes delays preconstruction activities (e.g., building permits) as noncompliant designs pose safety risks to building users.

Errors in building designs are commonly observed in construction projects [7]. Statistics show that approximately 30% of construction costs are associated with rework [8, 9]. This issue stems from traditional design approaches embedded in the construction sector, such as using 2D drawings to communicate with contractors. The use of 2D drawings forces design engineers to check code compliance manually despite the complexity of the constructed asset [10, 11]. Even though the construction industry has acknowledged 2D drawings as an error-prone and time-consuming process, reluctance to move from the traditional practices has been a roadblock [12, 13]. Building information modeling (BIM)-based rule checking can identify errors in a building model, minimizing the need for corrections and rework during the construction stage [10]. Furthermore, BIM-based rule checking enhances the accuracy and efficiency of compliance checking [10, 14].

BIM-based model checking is defined as a system that considers objects' configuration and relationships to assess a building design without any modifications [2]. BIM-based model checkers automatically evaluate a building design based on a set of functional requirements, which are denoted as rules [1, 15]. The result of the process can be "pass," "fail," "warning," or "unknown" [2]. BIM-based model checking is a four-step process: rule interpretation, building model preparation, rule execution, and rule reporting [16]. Applications and add-ons in the BIM suite, such as BIM Assure, Solibri Model Checker (SMC), Revit Model Review, and SmartReview APR, are used in this process.

Recently, BIM-based rule checking has received the attention of academics. For example, Eastman et al. [2] and Lee [17] applied a logical approach to check occupant circulation rules. Ding et al. [18] and Dimyadi and Amor [19] used an object-based approach for rule interpretation to check disabled access to buildings; Bouzidi et al. [20] applied an ontology-based approach for providing technical guides for tile roofs based on relevant codes; Nguyen and Kim [21] employed an Autodesk Revit model checker to check fire safety requirements, and Chahrour et al. [22] studied BIM-based rule checking for clash detection of a USD 75 million project and concluded that rework could be reduced by USD 15.2 million through rule checking procedure.

Most of the previous studies have developed BIM-based model checking tools using the ruleset from the Occupational Safety and Health Administration (OSHA) guidelines. Though OSHA is renowned throughout the world for safety guidelines, using such generic standards-based model checking tools may not address the specific safety requirements of a country or province. Hence, province or country-specific safety rule sets should be developed for building design review. However, to the best of the authors' knowledge, no approach has dealt with the development of a BIM-based occupant safety rule check for Ontario, Canada.

Hence, this study presents a BIM-based occupant safety ruleset to check institutional building designs for code compliance to prevent safety hazards for building users in Ontario, Canada. Ontario Building Code (OBC) and Canadian Standards Association (CSA) standards were reviewed to obtain safety-related guidelines. Accordingly, a ruleset was developed for use in building model checking. Regardless of the different approaches of this study, this research is more comprehensive in terms of the safety rule checking concept than recent similar research works in Canada [23] or other countries [24, 25] that utilized fire egress codes and focused only on fire-related safety. The outcomes of this research can aid building designers in model checking and ensuring the safety of building users. The rest of the paper's structure is as follows: Relevant literature was reviewed, and a research methodological framework was developed. Then, a case study was conducted, followed by an analysis of the results, discussion, and conclusions.

Literature review

BIM-based rule checking involves rule interpretation, model preparation, rule execution, and rule reporting. These aspects are explained in detail below.

Rule interpretation

Building codes and standards are typically in written texts, tables, and possibly equations. These rules should be translated into a machine-processable format to be used in a model checker [2]. This is performed in two ways. In the first method, rules are translated into computer codes by a programmer [23]. In the second method, rules are formally translated through the logic of human language [26]. Therefore, construction industry experts without any programming background can execute the second method of rule interpretation through software or plugin applications [27].

Programming-based rule interpretation

There are different techniques for interpreting the rules for rule checking development, such as an ontological approach, a logical approach, and an object-based approach [16]. Each method is explained below.

Object-based approach An object-based approach is a technique for organizing knowledge by defining attributes, procedures, rules, and machine learning [28]. Building rules are represented in a three-stage approach. In the first stage, building codes are classified. In the second stage, all relevant building objects and rules are identified. In the third stage, all data and values are stored and maintained in a tabular form to establish a knowledge base [29]. CoreNet e-PlanCheck is an example of an object-based rule checking system that is officially implemented in Singapore [30].

One advantage of this approach is that contextual concept lattice classes can cover the standard scope; therefore, rules can be encapsulated within the classes [31]. Creating and maintaining a complex concept lattice is a time-consuming and error-prone task [32].

Logic-based approach First-order predicate logic is the most common language and is applied to interpret natural language. This method is common because humans define rules, and interpreting rules can be made by translating logic into human language statements [33]. In this approach, explicit and implicit logical conditions such as “and,” “or,” and “If–Then” are combined and applied to validate and check a building element [34]. In order to interpret rules into basic logic, a conceptual graph can be employed by rule experts without programming knowledge [35]. There are four stages for translating rules into the conceptual graph:

  1. I.

    Identification of the main concept of rules

  2. II.

    Identification of independent sub-rules of each rule

  3. III.

    Identification of constraints and restrictions

  4. IV.

    Identification of connections of all elements to develop the proper conceptual graph.

Then, to support the process of compliance code checking in the BIM environment, BIM Rule Language (BIMRL) is developed through Structured Query Language (SQL) [36]. BMRL defines a simplified schema that uses the relational database to facilitate access to the building data. This approach is useful for unidirectional data processing where Industry Foundation Classes (IFC) models are loaded into the simplified schema [37]. The advantage of this approach is that predicate calculus constructions provide a sufficient means for constructing any expression [32]. However, they cannot efficiently represent the different evaluation strategies of data items for object classes of different designs. Thus, the final model of the predicate logic standard is complicated and arduous to create and maintain [38].

Ontology-based approach This approach is based on using a graph that displays objects and their logical relations [39]. The ontological approach involves four aspects, as follows:

  1. I.

    Developing a construction regulation database

  2. II.

    Creating a reasoning model

  3. III.

    Capitalizing the database

  4. IV.

    Validating the framework by integrating all aspects into a prototype [40].

The advantage of this approach is the effective use of computers in the design process. However, this approach is unable to provide formal language for describing models [41]. Hence, it is difficult to understand how design classifications are achieved. Furthermore, this approach does not explain how evaluation provisions are represented, leading to problems in understanding the compliance checking process [32, 41].

BIM application-based rule interpretation

There are several model checkers currently being used in the market. Solibri Model Checker (SMC) is a popular BIM application used to check the integrity, quality, and safety of BIM models [42]. SMC, which contains built-in geometry-oriented rules, automatically analyzes and checks virtual building models to detect design deficiencies and missing elements [43]. SMC strictly checks model structure requirements [16]. Furthermore, it provides a user-friendly visualization of the building model and permits a virtual walk-through. This feature has been used for checking safety rules related to fall hazard prevention for construction workers [44]. The SMC-based SMART Code system was developed by the International Code Council in North America to check the reliability of rule checking [2]. The advantages of SMC are (i) its ability to read and correctly interpret IFC files exported from a wide range of BIM authoring tools, (ii) a complete set of spatial operations, and (iii) a high level of maturity (widely used by designers) [45]. On the other hand, SMC's disadvantages include a lack of formal and standard schema for the rule definition and rule templates with hard-coded rules [45].

BIM Assure gives users visibility into building data. The features of BIM Assure are cloud-based model viewing (2D/3D), the support for IFC file format, the ability for element classification and data viewing, and the ability for editing and rule analysis [45]. The advantages of BIM Assure are ease of defining rules, low requirements for programming constructs, minimum domain knowledge required, and proper performance indication [46]. On the other hand, the disadvantages of this tool have no standard schema for the rule definition, low support for complex rules, no integrated geometry engine, and low interface to other languages and systems [45, 46].

There are plug-ins for BIM applications, such as the Autodesk Revit plug-in, that enable rule checking. Nguyen and Kim used this possibility to check openings in firewalls, fire resistance ratings, and horizontal continuity of the firewalls based on the International Building Code (IBC) in the USA [21]. Although the Revit Model Review requires a low level of logic or programming constructs, it has no integrated geometry engine, no standard schema for the rule definition, low support for complex rules, and low interface to other languages or rule checking systems.

Table 1 shows the BIM-based approaches and techniques for developing automated compliance checking. As shown in this table, different rule interpretation techniques have been used by previous researchers. However, no previous research has interpreted BIM-based safety rules according to OBC and CSA standards.

Table 1 Approaches and techniques for a rule interpretation
Full size table

Model preparation

The steps that should be undertaken in preparing the building model for rule checking include (i) creating a model view to derive the required data for a specific type of rule checking and to extract subsets of an overall building model, (ii) extracting implicit properties using enhanced objects to derive new information and compute complex properties, (iii) obtaining the new model to make an evaluation of certain implicit relations and properties easier, (iv) performance-based analysis to check rule adequacy, and (v) checking layout rule parameters to check the building model layout types [71].

Rule execution

Rule execution brings the BIM model and the corresponding rules together [2]. The BIM model should be validated before performing the rule check to ensure properties, names, and objects are available [72]. Moreover, in the rule execution stage, a management system is required to coordinate the application of rule modules and their results. This system checks the completeness of the rule checking and model version consistency [73].

Rule reporting

Rule reporting validates whether design conditions are satisfactory according to the ruleset. A rule checking report also provides a graphical representation of each element of the BIM model in reference to the source rule [7].

Methodology

Figure 1 illustrates the methodological framework of this research. The four-step process proposed by Eastman et al. [2] for BIM-based rule checking was considered in this research. This research was performed in 5 interrelated steps, which are explained in detail below:

Fig. 1
figure 1

Research methodological framework

Full size image

Step 1. Building safety rule identification

OBC and CSA Standards were reviewed to identify safety-related guidelines. The approach used by Solihin [36] and Nguyen and Kim [21] in finding safety-related was used in this study. The list of used codes and standards in this study are as follows: OBC and CSA A500-16 (building guards), guidelines in CSA Z1002-12 (occupational health and safety, hazard identification and elimination, and risk assessment and control), CSA Z259.2.4:15 (fall arresters and rigid vertical rails), CSA Z259.13–16 (manufactured horizontal lifeline systems), CSA Z1000:14 (occupational health and safety management), and CSA Z45001:19 (occupational health and safety management system, requirements with guidance for use included safety-related guidelines).

The safety-based rules and standards used in this study are classified as follows: (i) egress analysis, which deals with issues such as the number and location of means of egress, (ii) stairs and steps, to check stairs and step dimensions and configuration, (iii) ramps, to check the slope and dimension of ramps, (iv) landings, to deals with the location and dimension of landings in a building model, and (v) handrails and guards, to check the existence and height of building guards (if they exist).

Step 2. Decision table preparation

The object-based decision table was developed to show the geometric information of objects based on the relevant rules from the OBC and CSA standards (Table 2). Overall, there are 64 parameters, which can be organized into five categories: (1) stairs, (2) ramps, (3) guards, (4) Egress analysis, and (5) ceiling height.

Table 2 Summary of object-based decisions
Full size table

Step 3. Ruleset preparation

The customized parametric ruleset file (BIM-based safety rules) was created using a model checking application (SMC). First, available rule templates in SMC were reviewed to identify the appropriate template related to each rule in the decision table (Table 2). Second, based on the rules identified in step 1, safety classes for rule checking were defined in SMC. Finally, the ruleset classes were customized based on the 64 rules in the decision table.

Step 4. BIM object checking

BIM model objects were validated by focusing on the level of detail and specifications required by decision tables. Object validations include the correctness of the required attribute value, the correctness of embedding an attribute within a BIM object, and the definition of relationships between objects within a building model.

Step 5. Building model diagnosis

The building model was checked in this step, and the design errors were determined. The graphical report (i.e., output) indicated building elements that did not comply with standards and building codes after the rule execution. Based on the model checking, each element was defined as accepted, rejected, critical severity (major), moderate severity (normal), and low severity (minor).

Case study: the method implementation

The proposed rule set was implemented as a case study for the Center for Automotive Research and Education (CARE) building of the University of Windsor located in Windsor, Ontario, Canada. This 2-story building was renovated in 2014. The building is classified as an educational building. The BIM model of the CARE building was created based on the 2D construction drawings (Fig. 2). Autodesk Revit was used to create the building model.

Fig. 2
figure 2

The CARE building used for the case study

Full size image

An Industry Foundation Class (IFC) file was used to transfer building data to the SMC. The exported IFC file from the BIM model includes all details with a high level of specificity. In this transfer, it is important to ensure the required elements for safety rule checking are available in the IFC file. This measure was performed to verify that the IFC file contains all required elements for the rule checking process.

The BIM-based safety ruleset file was coded in SMC. As presented in Table 1, SMC has been used by previous researchers to check safety rules in building evaluation, such as building evacuation, fall protection for workers, building safety design, and accessibility. SMC is capable of creating specific rules identified in Table 2. This file was imported into the rule checking section of SMC to execute rules and report the result of code checking. Table 3 provides a summary of the model checking report. This report determines whether each component meets the rules identified in Table 2. The detailed report informs designers of each problematic element and compares the designed value with the standard value of that parameter identified in Table 2. In this case study, there is no "rejected" or "major" error. Therefore, it shows that this case study does not need to be entirely redesigned. However, several "normal" errors indicate the need for a major revision in the parameters of the "means of egress" section.

Table 3 Summary of safety code checking report
Full size table

SMC graphically illustrates the errors in the building design. Figure 3 indicates three instances of the graphical report. After the rule execution step, SMC highlights elements with errors. SMC categorizes errors, and users can click on the intended element in the report section of SMC to observe the element in the BIM model.

Fig. 3
figure 3

Graphical report

Full size image

Discussion

This study proposed a BIM-based occupant safety ruleset for institutional buildings in Ontario, Canada. A BIM ruleset enables accurate and fast evaluation of building designs to ensure compliance with building codes, standards, and guidelines [74]. The application of the proposed ruleset was demonstrated by using an existing building. The rule check returned no elements rejected or with major errors. A building design should comply with building codes and guidelines to receive construction approval. Hence, this validates the proposed rule set and its application. The theoretical and practical implications are discussed below.

Theoretical implications

The study developed a comprehensive database comprising occupant safety-based rules for buildings in Ontario, Canada. The guidelines available in the OBC, CSA A500-16 (building guards), CSA Z1002-12 (occupational health and safety—hazard identification and elimination and risk assessment and control), CSA Z259.2.4:15 (fall arresters and rigid vertical rails), CSA Z259.13–16 (manufactured horizontal lifeline systems), CSA Z1000:14 (occupational health and safety management), and CSA Z45001:19 (occupational health and safety management system—requirements with guidance for use included safety-related guidelines) were used in developing the aforementioned database. SMC was used to define the ruleset.

SMC generates a detailed report with the causes of errors, enabling building designers to revise the building model accordingly. The severity parameters defined in step 5 of the proposed framework were used to interpret the report generated by SMC. Therefore, the integration of the proposed ruleset and the rule result severity tables provided by SMC (refer to Table A1) interprets the generated report for designers.

Practical implications

The present study aids in preventing safety incidents for building occupants by identifying elements that don't comply with safety regulations in Ontario, Canada. The proposed ruleset can be extended to enhance the efficiency and accuracy of the building permit review and approval. It is important to point out that the proposed method requires the building design to be developed as a BIM model, which requires additional effort. The graphical and detailed reports enable designers to know the standard and correct value to edit the BIM model. Thus, the proposed framework allows designers to reduce model errors and prevent safety hazards for building occupants.

The proposed rule set was developed for occupant safety in Ontario, Canada. However, this ruleset can be customized to suit building codes in other provinces. Furthermore, additional safety rules can be identified and programmed into the present ruleset using the same approach or by using an application programming interface. The Architectural Engineering and Construction (AEC) industry should spearhead this cause by standardizing further safety.

Limitations

Several limitations have been identified in the proposed model checking tool. Since SMC does not allow users to freely manipulate data for generating new properties, using the proposed ruleset in SMC prevents the system from adequately covering some elements, such as passive fire protection. Furthermore, SMC software utilizes hard-coded rules. Hence, the individual steps of the checking processes are not visible and consequently limit the users' ability to alter them.

Conclusions

This study presented a BIM-based rule checking approach to reduce the safety risk for building occupants in Ontario, Canada. A comprehensive review was first conducted to provide up-to-date information regarding available rule checking techniques and systems. Then, OBC and CSA guidelines were used to develop a methodological framework that checks compliance with safety guidelines. The CARE building at the University of Windsor in Ontario, Canada, was chosen to test the framework. The goal of the case study was to see if the established ruleset was error-free and appropriately reflected the research's content. The case study's findings revealed that the intended rulesets were all included without errors and were easily utilized for rule checking, confirming the validity of the proposed framework. The case study further demonstrated the ability of this system to identify safety errors in a building design, informing the designers whether the building design requires a redesign, major revision, or minor revision. Moreover, due to the object-based classification of the ruleset, graphical and detailed reports can be generated that indicate the specific model elements with errors.

Overall, the outcomes of this study aid in preventing safety incidents for building occupants by identifying design elements that do not comply with standard safety regulations in Ontario, Canada. The outcomes of this research can be further extended to enhance the efficiency of the building permit review and approval. This study also promotes BIM adaptation in the Canadian construction industry.

Model checkers cannot make modifications to a design. Artificial intelligence (AI) can be used to modify the model based on identified errors. Hence, future research should look at developing an AI-based model checking interface for BIM applications. Moreover, this study focused on the architectural aspect of a building. Future studies can develop rulesets and automated safety rule checking systems for structural, plumbing, or mechanical sections.