Normative Visual Patterns for Hazard Recognition: A Crisp-Set Qualitative Comparative Analysis Approach

Heap-Yih Chong,Mingxuan Liang,Pin-Chao Liao 대한토목학회 2021 KSCE JOURNAL OF CIVIL ENGINEERING Vol.25 No.5

Understanding the mental representations used for hazard recognition would help the development of inspection strategies for effective safety management. This study is part of ongoing research to identify hazard recognition patterns based on users’ mental representations. Hence, this study explored normative visual patterns for improving hazard recognition performance using a crisp-set qualitative comparative analysis (cs-QCA). Eye-tracking data and visual trajectories were collected using an eye-tracking device in a structural laboratory. A cs-QCA approach was adopted to analyze and summarize normative visual patterns that were used to successfully detect hazards by all participants, namely, potential electrical contact, a large machine with no guardrails, and steel bars dump. The results show that object identification should suffice as the basis for identifying electricity-related hazards, while struck-by hazards should focus on the objects and their pivot points or potential movement trajectories. The experimental design and analytical approach provide new insights into visual analytics in hazard recognition. The research extends and supplements recognition by component theory in the context of construction hazard recognition. The results also provide new and practical references for hazard inspection training, as well as for future development of automated hazard recognition systems.

A Novel Approach Based on Fluid Dynamics for On-Site Safety Assessment

Mei Liu,Heap-Yih Chong,Pin-Chao Liao 대한토목학회 2021 KSCE JOURNAL OF CIVIL ENGINEERING Vol.25 No.5

Construction safety assessment is a major component of safety management in construction projects. However, thus far, most assessment studies have focused on either cross-sectional or longitudinal performance. This study aims to develop a proactive and preliminary safety assessment method using a fluid dynamics (FD) approach by integrating longitudinal contractor performance with cross-sectional safety conditions. To this end, an FD framework was first developed using three processes: (a) identifying the connections between FD principles and safety analogs, (b) incorporating the Darcy–Weisbach equation into the framework, and (c) modifying FD equations for construction safety assessment. Subsequently, two case studies were investigated, and the results obtained were compared and verified with one existing assessment results. The results indicate that the state of safety at a construction site incorporates both the occurrence of hazards and conservation of energy. Thus, the proposed FD approach can be used to preliminarily assess and predict safety conditions by combining safety-related measures with the dynamic characteristics of construction processes. The approach considers a comprehensive range of indicators and parameters, which enables the comparison of safety performance between projects or assessment periods by independently changing parameters.

A System Dynamics Model of Prevention through Design towards Eliminating Human Error

Xinlu Sun,Heap-Yih Chong,Pin-Chao Liao,Dongping Fang,Yanqing Wang 대한토목학회 2019 KSCE JOURNAL OF CIVIL ENGINEERING Vol.23 No.5

Theoretically, prevention through design (PtD) is the most effective approach for eliminating human error in the field of construction. However, there is a lack of rigorous empirical analysis regarding the focal point of PtD. Hence, this research aims to develop a system dynamics model that depicts a mechanism for identifying the focal point of PtD and relevant design factors with respect to human error. Elevator installation serves as the scope of the model, and three sub-models of “cognition-behavior,” “environment-cognition,” and “design-environment” are developed. Their quantitative relationships are analyzed based on empirical data prior to the development of the system dynamics model. Subsequently, safety performance is simulated based on an eight-hour work day, and sensitivity analyses are conducted to determine the impact of each design factor on human error. This study provides insights into PtD using various design factors that can be generalized and used to develop safer designs under complex working environments.

Latent Provisions for Building Information Modeling (BIM) Contracts: A Social Network Analysis Approach

Su-Ling Fan,Heap-Yih Chong,Pin-Chao Liao,Cen-Ying Lee 대한토목학회 2019 KSCE JOURNAL OF CIVIL ENGINEERING Vol.23 No.4

The effective adoption and use of Building Information Modeling (BIM) require appropriate contract design to fairly allocate the contracting parties’ rights and responsibilities. Several standards for BIM protocols and contracts have been developed for the industry. However, the awareness and the use of these are rather limited, leading to unclear provisions in BIM contracts. Therefore, the research aims to identify the influential legal aspects that serve as the latent contract provisions in BIM contracts. A questionnaire survey was conducted to survey experts and active BIM users in construction projects. The data were analyzed using Social Network Analysis (SNA) by assuming interdependent relationships among various the legal aspects in BIM contacts. The key legal aspects associated with BIM contracts pertain to the roles and responsibilities of the project participants. The results also reveal that data security is the center of all latent legal aspects in the contracts. The study provides significant new insights into clarifying the required contract provisions in BIM contracts.

Distinctive Judicial-Tailored Causation References of Construction Accidents

Jiaming Wang,Rui Cheng,Heap-Yih Chong,Pin-Chao Liao 대한토목학회 2022 KSCE JOURNAL OF CIVIL ENGINEERING Vol.26 No.8

Accident liability allocation is significant for social justice and industry safety. The allocation of liability between individuals and organizations depends on the judgment of their negligence in accidents. However, current judicial practice does not show a clear distinction between different accident types when allocating liabilities, and there is a lack of scientific evidence to guide the legal decision-making process in liability allocation. Unfortunately, current accident causation research does not sufficiently meet the requirements as scientific evidence to support the realization of more strategic and targeted liability allocation. This study introduces the qualitative comparative analysis (QCA) method into the field of accident analysis using crisp-set QCA (csQCA) to explore the causal distinctions between fall and non-fall accidents based on the human factor analysis and classification system (HFACS). The results show that organizational influence and unsafe acts distinctively signify fall and non-fall accidents, and that the precondition of unsafe acts is their junction. Theoretically, this study 1) furnishes a QCA tool that allows academic research with uncertainty to be used in the body of evidence that meets the requirements of law, thereby extending the application scope of accident research to the legal practice of liability allocation; and 2) provides meaningful references for the judgment of negligence and accident liability for judicial practice.

Identifying Effective Management Factors Across Human Errors – A Case in Elevator Installation

Pin-Chao Liao,Zhangming Ma,Heap-Yih Chong 대한토목학회 2018 KSCE JOURNAL OF CIVIL ENGINEERING Vol.22 No.9

Human errors are recognized as the main factor in causing construction accidents. Previous studies mainly focused on justifying theories associated with human errors and hardly quantifying the causal relations between external stimuli and human errors. Hence, the aim of this research is to develop comprehensive management measurements for addressing human errors in construction projects. Deductive reasoning was used to describe the methodological process. Firstly, we constructed a theoretical model of human errors based on Cognitive Reliability and Error Analysis Method (CREAM) and knowledge-combined structure learning algorithm. Then, Bootstrap method was adopted to verify the reliability of network topology, while the similarity-flooding algorithm was used to analyze similarity of factors across various causal models for their commonalities. Subsequently, Bayesian parameter estimation was to analyze the sensitivity of the nodes. The results show that inadequate quality control, design failure and inattention are the most fundamental causes of human errors in promoting safety management. This research has proposed an analytical approach that consolidated the influential mechanics to reflect the overall influence of a root cause in the human error. Ultimately, the research lays an analytical foundation for safety management research in the future.