Abstract
JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 2021, VOL. 20, NO. 3, 314–325 https://doi.org/10.1080/13467581.2020.1796675 CONSTRUCTION MANAGEMENT A novel method for occupational safety risk analysis of high-altitude fall accident in architecture construction engineering Xiao-Ping Bai and Yu-Hong Zhao School of Management, Xi’an University of Architecture & Technology, Xi’an, China ABSTRACT ARTICLE HISTORY Received 27 December 2019 The occupational safety risk of high-altitude fall accident in architecture construction engineer- Accepted 7 July 2020 ing is dynamically changeable. Considering these dynamic changes, how to make a qualitative and quantitative analysis of the changing trends of the occupational safety risk, and providing KEYWORDS the basis for the latter occupational safety risk control measures, become the main target of the Architecture engineering; dynamic risk management of construction projects. From the point of view of dynamic change, construction engineering; this paper modularizing analyzes the presented model, quantitative analyzes the impact of safety; accident; risk occupational safety risk of the whole construction system, proposes some methods about transforming the dynamic fault-tree module into the Bayesian network model, and puts forward a novel occupational safety risk analysis method of high-altitude fall accident in construction project management based on integrated Dynamic Fault Tree Analysis (DFTA) and Bayesian network method. The presented method and detailed realization process can offer a useful reference for occupational safety risk management of high-altitude fall accident in architecture construction engineering. 1. Introduction are combined with a lot of different construction stages according to a certain time and logical order, The architecture construction engineering has the so the former construction activity does not attain the characteristics of long duration, complexity, change- standard, the next construction activities will have able environment, multi participation and so on. These potential occupational safety hazards. Smoothly realiz- characteristics determine that architecture construc- ing the goals of progress, occupational safety, and the tion engineering is dynamic and interrelated; the occu- period of the entire construction system needs the pational safety risk of high-altitude fall accident in mutual coordination of many construction stages. architecture construction engineering is also dynami- The traditional occupational safety risk analysis meth- cally changeable. ods are not specifically aimed at the dynamic, relevant The occupational safety risk of the project in differ - and time sequence and using the separation, local, ent stages of construction is different; generally, with static attitude to look at occupational safety risk. the increase of the construction height, the occupa- Methodological approach: considering the dynamic tional safety risk will increase, and the occupational changes of occupational safety risk factors of architec- safety risk of the main stage is much more than that ture construction engineering, how to make at the decoration stage. Due to construction projects CONTACT Xiao-ping Bai 1507742734@qq.com School of Management, Xi’an University of Architecture & Technology, Xi’an 710055, China © 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group on behalf of the Architectural Institute of Japan, Architectural Institute of Korea and Architectural Society of China. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 315 a qualitative and quantitative analysis of the changing as state space explosion and transformation program trends of the occupational safety risk factors, and pro- error, with respect to the traditional fault tree transfor- viding the basis for the latter occupational safety risk mation to the Markov chain method. It also benefits control measures, become the main target of the from the inherent advantages of Bayesian networks, dynamic risk management of architecture construction such as probability updating, etc. (Xin Ge, Tang, and engineering. From the point of view of dynamic Zhe 2009). Lee, Park, and Shin (2009) studied the appli- change, this paper modularizing analyzes the pre- cation of the Bayesian network model in risk manage- sented model, quantitative analyzes the impact of ment of large engineering projects. Manno et al. (2012) occupational safety risk of the whole construction sys- presented a high-level modeling framework that tem, presents some methods about transforming the exceeds all these limitations, based on Monte Carlo dynamic fault tree module into the Bayesian network simulation. Chiacchio et al. (2013) tested the feasibility model, and presents a novel occupational safety risk of a composition algorithm based on a Weibull distribu- analysis method of high-altitude fall accident in con- tion. In order to overcome the limitation of the tradi- struction project management based on integrated tional method, Leu and Chang (2015) proposed a supply Dynamic Fault Tree Analysis (DFTA) and Bayesian net- chain risk assessment model based on the fault tree work method. transformation. Chena and Leub (2014) have established a risk assessment model of bridge construction based on fault tree transformation. Rauzy and Blériot-Fabre 2. Analyzing related references (2015) proposed a direct translation of dynamic fault trees into guarded transition systems. Abdo and Flaus Fault tree analysis (FTA) is widely used for the relia- (2016) provided a simulation technique based on MCS bility analysis of systems (Vesely et al. 2002) (Kabir (Monte Carlo Simulations) to solve dynamic fault tree 2017). Although fault tree models are well structured taking into account epistemic uncertainty in the deter- and easily understood, the classical FTA models are mination of the failure rate of basic events. Merle, et al. generally unable to model some aspects of system (2016) focused on the quantitative analysis of Dynamic behavior, e.g., priorities or functional and stochastic Fault Trees (DFTs) by means of Monte Carlo simulation. dependencies between events (Kabir et al. 2018). Yevkin (2016) used Approximate Markov chain method These can be looked as shortages of classical FTA. for dynamic fault tree analysis for both reparable and The modeling capability of classical fault trees has non-reparable systems. been enhanced through several extensions, such as The Bayesian network (simply referred to as BN), also dynamic fault trees (DFTs) (Dugan, Bavuso, and Boyd known as Belief (probability distribution) network, is 1992) and Pandora Temporal Fault Trees (TFTs) a directed acyclic graph (DAG), which is formed by the (Walker 2009). Merle et al. (2010) focused on a sub- nodes representing variables and directed edges con- class of Dynamic Fault Trees (DFTs). Merle, Roussel, necting these nodes. Leu and Chang (2015) presented and Lesage (2011) presented an algebraic framework the development of a fall risk assessment model for SC allowing algebraically modeling dynamic gates and (steel construction) projects by establishing a Bayesian determining the structure–function of any Dynamic network (BN) based on fault tree (FT) transformation. Fault Tree (DFT). Rahme and Xu (2015) adopted the Sutrisnowati, Bae, and Song (2015) proposed a method dynamic fault tree (DFT) formalism to model the sys- to generate a Bayesian network from a process model tem reliability before and during a software rejuvena- which can be discovered from event logs in port infor- tion process in an aging cloud-based system. For mation systems. Miyauchi and Nishimura (2016) con- instance, in DFTs, dynamic gates like Functional structed Bayesian networks connecting the findings Dependency (FDEP), Priority-AND (PAND), and SPARE from a physical examination and questionnaire on gates are introduced to model the dynamic failure daily lifestyle choices. Laitila and Virtanen (2016) elabo- behavior of systems. DFTs are primarily analyzed rated on the ranked nodes method (RNM) that is used quantitatively, and for the analysis of fault trees, espe- for constructing conditional probability tables (CPTs) for cially the DFTs, different approaches like algebraic Bayesian networks consisting of a class of nodes called (Merle et al. 2016; Merle et al. 2010), Markov chain- ranked nodes. Mohammadfam et al. (2017) presented based (Dugan, Bavuso, and Boyd 1993; Boudali, a model for managing and improving occupational Crouzen, and Stoelinga 2010), stochastic (Zhu et al. safety behavior of employees using the Bayesian net- 2017; Zhu et al. 2016), Bayesian network-based works approach. Chebila (2018) applied the Bayesian (Codetta-Raiteri and Portinale 2017; Boudali and networks method to estimate the occurrence frequency Bechta Dugan 2006), Sequential Binary Decision of failures and consequences in a straightforward way. Diagram (SBDD)-based (Xing, Tannous, and Dugan Haddawy et al. (2018) explored the use of Bayesian 2012; Ge et al. 2015) approaches have been devel- networks to model malaria. Tabar and Elahi (2018) com- oped (kabir et al. 2020). bined the Bayesian networks (BNs) and the simulated Dynamic fault tree transforming to Bayesian network annealing algorithm Dąbrowski and Górski. (2019) was model can effectively avoid some disadvantages, such 316 X.-P. BAI AND Y.-H. ZHAO to determine the influence of the milling tool setup on status that we mostly do not want to happen as the occupational safety hazards associated with a kickback goal of fault analysis system, then look for factors in furniture making. that lead to the occurrence of the fault, and looks The dynamic risk factors for the entire construction for these next-level factors, until the original factors project will have different effects. When dynamic that we do not want to further process. In this changes of construction project risk factors, how to method, the top-down analysis approach is used qualitative and quantitative analysis of the change to find the logical relationship of the top event trend of risk sources, as well as the influence on the and all kinds of intermediate events, bottom events, reliability of the whole system, and to provide a basis etc. (embodying internal component failure, exter- for the risk control measures, become the main target nal environment changes and steeplejack error of dynamic risk management of architecture con- factors). struction engineering. On the basis of above research Using FTA it is possible to determine the probability achievements, from the point of view of dynamic of any undesired events, also known as top event, change, this paper modularizing analyzes the pre- given the probability of the basic events, which are sented model; transforms the dynamic fault tree mod- the lowest-level event causes. Therefore, it is necessary ule into the Bayesian network model to quantitatively to quantify the failure probability of each basic event analyze the impact of occupational safety risk events to be able to measure the probability of the top event. of the whole construction system, and presents The criticality of the basic events can also be deter- a novel occupational safety risk analysis method of mined by calculating their relative contributions to the high-altitude fall accident in construction project occurrence of the top event (Yazdi, Kabir, and Walkerb management for decision-making based on inte- 2019). In the DFTA (dynamic fault tree analysis) grated Dynamic Fault Tree Analysis (DFTA) and method, some dynamic logic gates (such as sequence Bayesian network method. enforcer gates, functional dependency gates, priority and gates, cold spare gates, warm spare gates and so on) are added, which can help to analyze the dynamic 3. The dynamic fault tree model of and sequential functions of the system. occupational safety risk analysis of Figure 1 takes the occupational safety risk factors of high-altitude fall accident high-altitude fall from the operating platform as an example. This paper establishes a dynamic fault tree. Among the different probabilistic risk assessment This system contains two dynamic logic gates, the prior- (PRA) methods, FTA is the most-widely used ity and the gates of T3 and the warm spare gates of T6. approach for system safety and reliability evalua- Some contents of Figure 1 can be explained as follows: tion. Fault tree analysis (FTA) is a quantitative risk (1) Sequence Enforcer (SEQ) gates of architecture analysis method and can be applied to major construction engineering unwanted events (Yazdi, Kabir, and Walkerb 2019). Multiple bottom events (or intermediate events) FTA method is an effective tool for the occupational that may cause the system to fail must occur and in safety analysis of complex system, it looks the fault a certain sequence. If disrupting the order of some Figure 1. Dynamic fault tree of high altitude falls from operation platform. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 317 Figure 3. The functional dependency gates. Figure 2. Sequence enforcer gates. bottom events (or intermediate events), the top event will not occur. This gate is shown in Figure 2. (2) Functional dependencies (FDEP) of architecture construction engineering The function dependency gate is composed of more than one basic event and a trigger event. The occurrence of the trigger event will certainly cause the occurrence of all other basic events and will lead to the occurrence of the top event, which causes the whole system’s failure. However, if all the relevant basic events occur, and the trigger event does not occur, the system will not fail, as shown in Figure 3. (3) Priority and (PAND) gates Priority and gate refers to two bottom events or intermediate events that may result in the failure of the system in both occurring and one of the events preceding the other one event (or coincidence), will cause the fault occurrence, reversing the order or any one event does not occur can not lead to the failure, as shown in Figure 4. (4) Warm spare (WSP) gates Warm spare gates have one basic event and one or more optional basic events. Basic events occur in Figure 4. Priority and gates. entering the work state from the beginning of the system, and the basic optional events are looked as gates is cold spare gates, that is, it will not fail before an alternative to the basic event, and when all the entering the working state, as shown in Figure 5. basic events occur, the warm spare gates start to To solve complex system fault tree, our paper focuses work. Before entering the working condition, there is on the construction of a Bayesian network model for a certain failure rate, and the failure probability is dynamic fault tree and applies it to the occupational greater than zero. In contrast to the warm spare 318 X.-P. BAI AND Y.-H. ZHAO The Bayesian network (simply referred to as BN), also known as Belief (probability distribution) network, is a directed acyclic graph (DAG), which is formed by the nodes representing variables and directed edges connecting these nodes. Figure 6 shows a simple Bayesian network example of six nodes (not include conditional probability distribution). The Bayesian network is composed of nodes V and the arrow line connecting the nodes. Node V represents variables in the system, which is an abstract representation of a subsystem, component or personnel problems of a dynamic fault tree. The arrow lines between the nodes indicate the inci- dence relation between the variables. Among them, the node that pointed out by the line is known as the parent node. The node that pointed by the line is called a sub node. A node that has no parent node is called the root node (such as node V1 in Figure 1), and the nodes without child nodes are called leaf nodes (such as node V4, V6 in Figure 1), and these nodes and the arrow lines are composed of a Bayesian network. In the directed acyclic graph, there is a conditional independence assumption, namely for arbitrary node V , in the given parent node set pa (V ), V and non-descendant node set i i A (V ) is conditionally independent: Figure 5. The warm spare gates. ðPðVjpaðV Þ; AðVÞÞ ¼ PðVjpaðVÞÞ (1) i i i i i safety risk fault tree analysis of high-altitude fall accident The conditional probability distribution that related to in architecture construction engineering. each node is indicated by P. From the conditional independence assumption of Bayesian networks can be known, the conditional probability distribution can 4. The presented integrated dynamic fault use P (V | Pa (V )) to describe, it is an expression of the i i tree analysis (DFTA) and Bayesian network quantitative relationship between node and its parent method node. Through the given prior probability distribution of the root node and the conditional probability dis- 4.1. Dynamic fault tree analysis (DFTA) based on tribution of the non-root node, then can calculate the Bayesian network transformation probability of the top event. Dynamic fault tree transforming to the Bayesian net- According to the logical relationship of the dynamic work model can effectively avoid some disadvantages, fault tree, the two-state Bayesian network mapping mod- such as state space explosion and transformation pro- els are modeled, that is, the nodes in the Bayesian net- gram error, with respect to the traditional fault tree work can only have two states: 1 indicates that the event transformation to the Markov chain method. It also occurring and 0 indicates that the event does not occur. benefits from the inherent advantages of Bayesian net- Each basic event in the dynamic fault tree subjects to works, such as probability updating, etc. a node in a Bayesian network (repeated occurrence of the Figure 6. Bayesian network. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 319 basic event is merged into one node), and the probability of occurrence of the fault tree’s bottom event corre- sponds to the prior probability of the node in a Bayesian network, the logic gate of the fault tree corre- sponds to the directed arc of a Bayesian network, to represent the logical relationship of the event (connec- tion strength). Through the prior probability of the root node, we can calculate the conditional probability of Figure 7. Bayesian network transformation of sequence each leaf node, so as to get the probability of the top enforcer. event from the bottom up. For the root node, the occurrence probability of the bottom event in the fault tree, the failure rate λ of the P ¼ PðC ¼ ½ðz 1ÞΔ; z ΔÞjB ¼ ½ðz 1ÞΔ; z ΔÞÞ ¼ z ;z c c b b root node can be obtained by combining the expert 0ð0hz hz � nÞ b c scoring method and the data analysis, and then can < ðλ λ ÞΔ 1 λ ðe 1Þ ð0hz ¼ z � nÞ b c λ Δ calculate the prior probability distribution of the root b ðλ λ Þðe 1Þ b c ðλ λcÞΔ λ Δ ðλ λcÞðzc z ÞΔ node x(Assuming that the bottom events are exponen- > λ ðe b 1Þðe b 1Þe b b ð0hz hz � nÞ λ Δ c b ðλ λ Þðe 1Þ b c tially distributed). (4) λ z Δðeλ Δ 1Þ x x x Pðx ¼ ½ðz 1ÞΔ; z ΔÞÞ ¼ e ;ð0hz � nÞ (2) x x P ¼ PðC ¼ ½T ;1ÞjB ¼ ½ðz 1ÞΔ; z ΔÞÞ z ;1 0 b b ð ¼ 1 P (5) z ;z 1 b c λ t λ T x x 0 0hz � n Pðx ¼ ½T ;1ÞÞ ¼ λ e dt ¼ e (3) 0 x P ¼ PðC ¼ ½T ;1ÞjA ¼½T ;1ÞÞ ¼ 1 (6) 1;1 0 0 Among them, x ¼ ½ðz 1ÞΔ; z ΔÞ indicates that x x x fail in the interval ½ðz 1ÞΔ; z ΔÞ, Δ indicates the time x x The conditional probability of root node T is: interval; x ¼ ½T ;1Þindicates that x does not fail in the P ¼ PðT ¼ ½ðz 1ÞΔ; z ΔÞjC ¼ ½ðz 1ÞΔ; z ΔÞÞ z ;z c c c c c c task time T ¼ 1 Here is the transformation of the dynamic logic gate (7) into Bayesian networks introduced above, the transfor- mation of static logic gate (And gates, Or gates, etc.) P ¼ PðT ¼ ½T ;1ÞjC ¼ ½T ;1ÞÞ ¼ 1 (8) 1;1 0 0 has been described in detail in the literature, here do not repeat them. It also can be calculated according to the rule of bucket elimination, so as to get the probability of top event T: (1) Sequence Enforcer (SEQ) gates X X X PðTÞ ¼ PðTjCÞ PðCjBÞ PðAÞPðBjAÞ (9) C B A The Bayesian network transformation of Sequence Enforcer is shown as Figure 7. Functional dependencies gates: According to the temporal logic relationship of the The Bayesian network transformation of Functional sequence enforcer gates, we can get the correspond- dependencies is shown as Figure 8. The conditional ing Bayesian network, and the conditional probability probability of node A and node B are the same, takes of node B and node C is the same, takes node C as an node A as an example, its conditional probability dis- example, its conditional probability distribution is: tribution is: Figure 8. Bayesian network transformation of functional dependencies. 320 X.-P. BAI AND Y.-H. ZHAO The conditional probability of top event T is: P ¼ PðA ¼ ½ðz 1ÞΔ; z ΔÞjC ¼½ðz 1ÞΔ; z ΔÞÞ ¼ z ;z a a c c c a λz Δ λΔ e ðe 1Þ ð0hz hz � nÞ > a c < P P ¼ PðT ¼ ½ðz 1ÞΔ; z ΔÞjA ¼ ½ðz 1ÞΔ; z ΔÞ; z ;z ;zt t t a a a b 1 P ð0hz ¼ z � nÞ z ;z a c c a B ¼ ½ðz 1ÞΔ; z ΔÞÞ > z hz b b a c 0 ð0hz hz � nÞ 1ð0hz � z ¼ z � nÞ c a a b t (10) 0ð0hz hz � nÞ b a (15) P ¼ PðA ¼ ½ðz 1ÞΔ; z ΔÞjC ¼½T ;1ÞÞ 1;z a a 0 λz Δ λΔ ¼ e ðe 1Þ (11) P ¼ ðT z ;1;1 ¼ ½T ;1ÞjA ¼ ½ðz ÞΔ; z ΔÞ; B ¼ ½T ;1ÞÞ ¼ 1 0 a 1 a 0 P ¼ PðA ¼ ½T ;1ÞjC ¼½T ;1ÞÞ ¼ 1 P 1;1 0 0 1;z (16) 0hz hn (12) P ¼ ðT ¼ ½T ;1ÞjA ¼½T ;1Þ; B ¼ ½T ;1ÞÞ ¼ 1 The Bayesian network transformation of Priority and 1;1;1 0 0 0 Gate is shown as Figure 9. The conditional probability (17) of root node T is: (3) Warm Spare Gates P ¼ PðT ¼ ½ðz 1ÞΔ; z ΔÞjC ¼½ðz 1ÞΔ; z ΔÞÞ ¼ 1 z ;z c c c c c c The Bayesian network transformation of Warm (13) Spare Gates is shown as Figure 10. The conditional probability of node S is: (β indicates P ¼ PðT ¼ ½T ;1ÞjC ¼½T ;1ÞÞ ¼ 1 (14) 1;1 0 0 the sleep factor of the warm spare gates S) Figure 9. Bayesian network transformation of priority and gate. Figure 10. Bayesian network transformation of warm spare gates. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 321 PðEi ¼ 1jEj ¼ 1Þ P ¼ PðS ¼ ½ðz 1ÞΔ; z ΔÞjA ¼ ½ðz 1ÞΔ; z ΔÞÞ ¼ z ;z s s a a a s PðEk ¼ ek; Ei ¼ 1; Ej ¼ 1; 1 � k � M; k� i; k� jÞ βλz Δ βλΔ s E1;��� ; Ei 1; Eiþ 1;��� ; e ðe 1Þð0hz hz � nÞ s a Ej 1; Ejþ 1;��� ; EM βλz Δ 1þβ e ð1þβÞλΔ βλΔ 1 ðe e þ Þð0hz ¼ z � nÞð18Þ λΔ a s PðEj ¼ 1Þ e 1 β β βλΔ (27) : e 1 λðz z βÞΔ a s e ð0hz hz � nÞ a s P ¼ PðS ¼ ½T ;1ÞjA ¼ ½ðz 1ÞΔ; z ΔÞÞ z ;1 0 a a 4.3. Importance degree ¼ 1 P (19) z ;z a s 0hz � n The important degree is the influence degree of a basic bottom event on the occurrence of the top event; it P ¼ PðS ¼ ½ðz 1ÞΔ; z ΔÞjA ¼ ½T ;1ÞÞ 1;z s s 0 generally can be divided into the probability impor- βλz Δ βλΔ ¼ e ðe 1Þ (20) tance degree and the structure important degree. The occurrence probability of the bottom events can affect the occurrence probability of the top event. The impor- P ¼ PðS ¼ ½T ;1ÞjA ¼ ½T ;1ÞÞ ¼ 1 P 1;1 0 0 1;z 0hz � n tant degree is the occupational safety and reliability parameters of the system structure and the functional (21) components, and controlling the important degree of The conditional probability of root node T is: basic components can change the failure probability of P ¼ ½ðZ 1ÞΔ; z ΔÞA ¼ ½ðz 1ÞΔ; z ΔÞ; the top event, and can assess the magnitude of the za;zs;zt t t a a event and its change. By ranking the important degree S ¼ ½ðz 1ÞΔ; of the components, the important parts are prior to be 1 ðZ ¼ maxðZ ; Z ÞÞ t a s analyzed, and the reliability of the system can be z ΔÞ ¼ 0 ðZ � maxðZ ; Z ÞÞ s t a s greatly improved. In dynamic Bayesian networks, it is easy to get the (22) probability important degree of bottom events at any time t by using its inference algorithm. P ¼ PðT ¼ ½T ;1ÞjA ¼ ½T ;1Þ; S 1;z ;1 0 0 Pr Ii ðtÞ ¼ PðT ¼ 1jE ¼ 1Þ PðT ¼ 1jE ¼ 0Þ (28) ¼ ½ðz 1ÞΔ; z ΔÞÞ ¼ 1 (23) ti ti s s E indicates the state of the root node E at the time t. If ti i P ¼ PðT ¼ ½T ;1ÞjA ¼ ½ðz 1ÞΔ; z ΔÞ; S z ;1;1 0 a a the structural importance degree of a basic event is ¼ ½T ;1ÞÞ ¼ 1 (24) obtained, assuming that the failure probability of all the other bottom events is 0.5 and the probability P ¼ PðT ¼ ½T ;1ÞjA ¼ ½T ;1Þ; S ¼ ½T ;1ÞÞ ¼ 1 1;1;1 0 0 0 important degree of the bottom event is its structural (25) importance. 4.2. The probability of the top event T 5. Case analysis In the Bayesian network transformed by dynamic fault Taking the concrete occupational safety problems of tree, it is no need to solve the cut set, and the joint architecture construction engineering site as an exam- probability distribution can be used to calculate the ple, it is used to prove that the above demonstration. occurrence probability of the top event T: In architecture construction engineering, there are many occupational safety risks, and many occupational PðT ¼ 1Þ ¼ PðE1 ¼ e1; . . . ; Ei ¼ ei; . . . ; safety incidents often occur. Traditional occupational E1; EM 1 ��� (26) safety risk management methods, such as occupa- EM 1 ¼ eM 1; T ¼ 1Þ tional safety inspection table, occupational safety There the node Eið1 � i � M 1Þ corresponds to responsibility system and other static methods, gener- the intermediate events and the bottom events in ally have the disadvantages of lagging. Therefore, the dynamic fault tree, e i �f0; 1g is used to show Dynamic Fault Tree Analysis (DFTA) method can use whether the event occurs or not, M is the number of time or time period as the lead wire, calculate the [4] nodes in the Bayesian network . The Bayesian net- reliability of the system in a certain period of time or work model also can be used to diagnose the main time period, according to the occupational safety risk cause of the accident after the failure occurring, and to fault tree and the probability of the occurrence of verify the posterior probability of each node in the occupational safety risk events, so as to finding the dynamic fault tree, in addition to its reasoning ability. main occupational safety risk sources and its dynamic For example, after the event E , the posterior probabil- change at different stages of the system, through qua- ity of the occurrence of other events is: litative and quantitative analysis, to provide a more 322 X.-P. BAI AND Y.-H. ZHAO effective reference for the development of occupa- From the results of the analysis, the main event influ - tional safety risk response strategies. encing the occurrence of high-altitude fall accident in In the following section, taking the dynamic fault construction site is barbaric construction, absenting of tree of occupational safety risk factors that high- protection does not meet the requirements and not altitude fall from an operating platform of wearing a safety belt, and so on, which shows that the a construction project in the above section as an exam- system once the failed (a high-altitude fall accident), it ple analyzes the construction and application of is likely that these aspects of the failure, so that the Dynamic Fault Tree Analysis (DFTA) and Bayesian net- avoid measures should focus on these types of occu- work model. The corresponding Bayesian network is pational safety risk factors. Along with the progress of shown in Figure 11. architecture construction engineering, these occupa- The probability of the bottom events is shown in tional safety risk factors are constantly changing, and Table 1: the structure of the dynamic fault tree should be con- stantly updated, so as to analyze the main occupa- tional safety risk factors in different stages of the 6. Discussion construction project life cycle. According to the transformation formula (15) ~ (17) and (18) ~ (25) of the dynamic logic gate to the 7. Conclusions Bayesian network in the upper section, the conditional probability of each leaf node is obtained. In which There are already many types of research to study the assuming the sleep factor of the warm spare gates occupational safety risk of architecture construction T8, β = 0.2, the time intervalΔ=1. A Bayesian network engineering, and these references are mainly focused model of high-altitude falling accident is used to cal- on the static case, with no special concern about the culate the probability of a high fall accident is 22.1% by complexity and variability of occupational safety risk Bayesian simulation software Genie2.0, as shown in of high-altitude fall accident in architecture construc- Figure 12. According to the probability of a known tion engineering. In this paper, a novel integrated height fall accident is 22.1%, uses the backward rea- Dynamic Fault Tree Analysis (DFTA) and Bayesian soning ability of the Bayesian network, can calculate network method is presented, including the defini - the posterior probability of each bottom event, as tion of dynamic logic gates, the transformation of shown in Table 2 (where State 0 shows the event dynamic logic gates to Bayesian networks and the does not occur, State1 shows the incident occur). probability calculation, and apply it to the Figure 11. The Bayesian network of high-altitude fall accident in architecture construction engineering. Table 1. The probability of the bottom events. Occurrence Code Name probability Code Name Occurrence probability E1 Not wearing a seat belt (0.067, 0.933) E6 Absence of protective fence (0.0008, 0.9992) E2 Safety belt damaging (0.0097, 0.9903) E7 The protective does not meet the requirements (0.00151, 0.99849) E3 Scaffolding not fixed (0.0084, 0.9916) E8 Absence of safety net (0.0003, 0.9997) E4 Scaffolding not covered (0.00398, 0.99602) E9 The safety net does not meet the requirements (0.00262, 0.99738) E5 The workers step on the outside (0.00194, 0.99806) E10 Barbaric construction (0.186, 0.814) JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 323 Figure 12. The probability of a steeplejack fall accident. management of high-altitude fall accident in architec- Table 2. The posterior probability of the bottom events. ture construction engineering. Posterior Posterior Number probability Sort Number probability Sort E1 (94.1, 5.9) 3 E6 (96.9, 3.1) 4 E2 (98.9, 1.1) 6 E7 (93.7, 6.3) 2 Disclosure statement E3 (99.4, 0.6) 7 E8 (99.6, 0.4) 8 E4 (97.6, 2.4) 5 E9 (99.7, 0.3) 9 No potential conflict of interest was reported by the author. E5 (99.9, 0.1) 10 E10 (81.7, 18.3) 1 Funding occupational safety analysis of high-altitude fall acci- dent of architecture construction engineering. The This work was supported in part by the National Natural Science Foundation of China (NSFC) [51774228]; Shanxi main factors are obtained by using the Bayesian net- Province Education Department Research Project work simulation software GeNIe2.0, then can quanti- [12JK0803] and Shanxi Province Natural Science Basic tative analysis the impact degree of various factors Foundation [2016JM5088]. on the occurrence of the accident, can conclude that barbaric construction, absenting of protection does not meet the requirements and not wearing a safety Notes on contributors belt are the major occupational safety risk of acci- Xiao-Ping Bai is currently an associate professor in Xi’an dent, provided an effective basis for the prevention University of Architecture & Technology, Xi’an; PRC . His and control of the accident. research interests include project management, system engi- The flowcharts and veracity of presented methods neering, and etc. His articles have appeared in Frontiers of Structural and Civil Engineering (SCIE), Kybernetes(SCIE), are mainly limited by the accuracy of the changed Discrete dynamics in nature and society(SCIE), Scientific data, these changed data and information may contain World Journal(SCIE), Applied Mathematics & Information substantial inaccuracies due to the imperfection of Sciences(SCIE), Tsinghua Science and Technology, etc. data acquisition ways in the high-altitude fall accident. Yu-hong Zhao is master in Xi’an University of Architecture & As a result, the study has certain limitations. For this Technology, Xi’an; PRC. Her research interests include project reason, the future direction of research is to study management, system engineering, and etc. better data acquisition ways. The presented analysis method can also be applied References to other aspects of the construction project risk man- agement, such as schedule risk, cost risk, quality risk, Abdo, H., and J.-M. Flaus. 2016. “Monte Carlo Simulation to etc. The presented novel method and detailed realiza- Solve Fuzzy Dynamic Fault Tree.” IFAC Papers 49 (12): tion process can offer a useful reference for the risk 1886–1891. doi:10.1016/j.ifacol.2016.07.905. 324 X.-P. BAI AND Y.-H. ZHAO Boudali, H., and J. Bechta Dugan. 2006. “A Continuous-time Network.” Expert Systems with Applications 36 (1): Bayesian Network Reliability Modeling, and Analysis 5880–5887. doi:10.1016/j.eswa.2008.07.057. Framework.” IEEE Transactions on Reliability 55 (1): 86–97. Leu, S.-S., and C.-M. Chang. 2015. “Bayesian-network-based doi:10.1109/tr.2005.859228. Fall Risk Evaluation of Steel Construction Projects by Fault Boudali, H., P. Crouzen, and M. Stoelinga. 2010. “A Rigorous, Tree Transformation.” Journal of Civil Engineering and Compositional, and Extensible Framework for Dynamic Fault Management 21 (3): 334–342. doi:10.3846/ Tree Analysis.” IEEE Transactions on Dependable & Secure 13923730.2014.890643. Computing 7 (2): 128–143. doi:10.1109/TDSC.2009.45. Manno, G., F. Chiacchio, L. Compagno, D. D’Urso, and Chebila, M. 2018. “Bayesian Networks for Frequency Analysis N. Trapani. 2012. “An Integrated FT and Monte Carlo in Dependability.” Journal of Failure Analysis & Prevention Simulink Tool for the Reliability Assessment of Dynamic 18 (3): 538–544. doi:10.1007/s11668-018-0433-5. Fault Tree.” Expert Systems with Applications 39 (12): Chena, T.-T., and S.-S. Leub. 2014. “Fall Risk Assessment of 10334–10342. doi:10.1016/j.eswa.2011.12.020. Cantilever Bridge Projects Using Bayesian Network.” Safety Merle, G., J.-M. Roussel, and -J.-J. Lesage. 2011. “Algebraic Science 70: 161–171. doi:10.1016/j.ssci.2014.05.011. Determination of the Structure Function of Dynamic Fault Chiacchio, F., M. Cacioppo, D. Durso, G. Manno, N. Trapani, Trees.” Reliability Engineering and System Safety 96 (2): and L. A. Compagno. 2013. “Weibull-based Compositional 267–277. doi:10.1016/j.ress.2010.10.001. Approach for Hierarchical Dynamic Fault Trees.” Reliability Merle, G., J. M. Roussel, J. J. Lesage, and A. Bobbio. 2010. Engineering and System Safety 109: 45–52. doi:10.1016/j. “Probabilistic Algebraic Analysis of Fault Trees with ress.2012.07.005. Priority Dynamic Gates and Repeated Events.” IEEE Codetta-Raiteri, D., and L. Portinale. 2017. “Generalized Transactions on Reliability 59 (1): 250–261. doi:10.1109/ Continuous Time Bayesian Networks as a Modelling and tr.2009.2035793. Analysis Formalism for Dependable Systems.” Reliability Merle, G., J.-M. Roussel, -J.-J. Lesage, V. Perchet, and Engineering & System Safety 167: 639–651. doi:10.1016/j. N. Vayatis. 2016. “Quantitative Analysis of Dynamic Fault ress.2017.04.014. Trees Based on the Coupling of Structure Functions and Dąbrowski, M., and J. Górski. 2019. “Influence of the Milling Monte Carlo Simulation.” Quality and Reliability Tool Setup on Occupational Safety in Furniture Making.” Engineering International 32 (1): 7–18. doi:10.1002/ International Journal of Occupational Safety and qre.1728. Ergonomics 25 (2): 1–30. doi:10.1080/ Miyauchi, Y., and H. Nishimura. 2016. “Construction and 10803548.2018.1514134. Evaluation of Bayesian Networks Related to the Specific Dugan, J. B., S. J. Bavuso, and M. A. Boyd. 1993. “Fault Trees Health Checkup and Guidance on Metabolic Syndrome.” and Markov Models for Reliability Analysis of Smart Innovation, Systems and Technologies 45: 183–193. Fault-tolerant Digital Systems.” Reliability Engineering & doi:10.1007/978-3-319-23024-5_17. System Safety 39 (3): 291–307. doi:10.1016/0951-8320(93) Mohammadfam, I., F. Ghasemi, O. Kalatpour, and 90005-J. A. Moghimbeigi. 2017. “Constructing a Bayesian Network Dugan, J. B., S. J. Bavuso, and M. A. Boyd. 1992. “Dynamic Model for Improving Safety Behavior of Employees at Fault-tree Models for Fault-tolerant Computer Systems.” Workplaces.” Applied Ergonomics 58: 35–47. doi:10.1016/j. IEEE Transactions on Reliability 41 (3): 363–377. apergo.2016.05.006. doi:10.1109/24.159800. Rahme, J., and H. Xu. 2015. “A Software Reliability Model for Ge, D., M. Lin, Y. Yang, R. Zhang, and Q. Chou. 2015. Cloud-based Software Rejuvenation Using Dynamic Fault “Quantitative Analysis of Dynamic Fault Trees Using Trees.” International Journal of Software Engineering and Improved Sequential Binary Decision Diagrams.” Knowledge Engineering 25 (9–10): 1491–1513. doi:10.1142/ Reliability Engineering & System Safety 142: 289–299. S021819401540029X. doi:10.1016/j.ress.2015.06.001. Rauzy, A., and C. Blériot-Fabre. 2015. “Towards a Sound Haddawy, P., A.H.M.I. Hasan, R. Kasantikul. 2018. Semantics for Dynamic Fault Trees.” Reliability “Spatiotemporal Bayesian Networks for Malaria Engineering and System Safety 142: 184–191. doi:10.1016/ Prediction.” Artificial Intelligence in Medicine 84: 127–138. j.ress.2015.04.017. doi:10.1016/j.artmed.2017.12.002. Sutrisnowati, R. A., H. Bae, and M. Song. 2015. “Bayesian Kabir, S. 2017. “An Overview of Fault Tree Analysis and Its Network Construction from Event Log for Lateness Application in Model Based Dependability Analysis.” Analysis in Port Logistics.” Computers and Industrial Expert Systems with Applications 77: 114–135. Engineering 89: 53–66. doi:10.1016/j.cie.2014.11.003. doi:10.1016/j.eswa.2017.01.058. Tabar, V. R., and F. Elahi. 2018. “A Novel Method for Kabir, S., M. Yazdi, J. I. Aizpurua, and Y. Papadopoulos. 2018. Aggregation of Bayesian Networks without considering “Uncertainty Aware Dynamic Reliability Analysis an Ancestral Ordering.” Applied Artificial Intelligence 32 Framework for Complex Systems.” IEEE Access 6: (2): 1–14. doi:10.1080/08839514.2018.1451134. 29499–29515. doi:10.1109/ACCESS.2018.2843166. Vesely, W., J. Dugan, J. Fragola, J. Minarick, and J. Railsback. kabir, S., T. K. geok, M. kumar, M. yazdi, and F. hossain. 2020. 2002. “Fault Tree Handbook with Aerospace Applications.” “A Method for Temporal Fault Tree Analysis Using Technical Report Version 1.1. Washington, DC, USA: NASA Intuitionistic Fuzzy Set and Expert Elicitation.” Ieee Access Office Saf. Mission Assurance. 8: 980–996. doi:10.1109/ACCESS.2019.2961953. Walker, M. 2009. “Pandora: A Logic for the Qualitative Laitila, P., and K. Virtanen. 2016. “Improving Construction of Analysis of Temporal Fault Trees.” Ph.D. dissertation, Conditional Probability Tables for Ranked Nodes in Dept. Comput. Sci., Univ. Hull, Hull, U.K. Bayesian Networks.” IEEE Transactions on Knowledge and Xin Ge, W., L. Tang, and Y. Zhe. 2009. “Comprehensive Data Engineering 28 (7): 1691–1705. doi:10.1109/ Analysis Methods of Dynamic Fault Tree.” Equipment tkde.2016.2535229. Manufacturing Technology 8: 22–24. Lee, E., Y. Park, and J. G. Shin. 2009. “Large Engineering Xing, L., O. Tannous, and J. B. Dugan. 2012. “Reliability Analysis Project Risk Management Using a Bayesian Belief of Nonrepairable Cold-standby Systems Using Sequential JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 325 Binary Decision Diagrams.” IEEE Transactions on Systems, Reliability Engineering International 32 (4): 1509–1520. Man and Cybernetics, Part A (Systems and Humans) 42 (3): doi:10.1002/qre.1861. 715–726. doi:10.1109/tsmca.2011.2170415. Zhu, P., Y. Guo, S. Si, and J. Han. 2017. “A Stochastic Analysis Yazdi, M., S. Kabir, and M. Walkerb. 2019. “Uncertainty of Competing Failures with Propagation Effects in Handling in Fault Tree Based Risk Assessment: State of Functional Dependency Gates.” IISE Transactions 49 (11): the Art and Future Perspectives.” Process Safety and 1050–1064. doi:10.1080/24725854.2017.1342056. Environmental Protection 131: 89–104. doi:10.1016/j. Zhu, P., J. Han, L. Liu, and F. Lombardi. 2016. “Reliability psep.2019.09.003. Evaluation of Phased Mission Systems Using Stochastic Yevkin, O. 2016. “An Efficient Approximate Markov Chain Computation.” IEEE Transactions on Reliability 65 (3): Method in Dynamic Fault Tree Analysis.” Quality and 1612–1623. doi:10.1109/TR.2016.2570565.
Journal
Journal of Asian Architecture and Building Engineering
– Taylor & Francis
Published: May 4, 2021
Keywords: Architecture engineering; construction engineering; safety; accident; risk
