spiral.imperial.ac.uk · web view2020. 10. 13. · instead of using fault tree as a tool for...

Understanding and Visualizing Schedule Deviations in Construction Projects using Fault Tree Analysis

Abstract

Purpose

Delays in construction projects are both disruptive and expensive. Thus, potential causes of schedule deviation need to be identified and mitigated. In previous research, delay factors were predominantly identified through surveys administered to stakeholders in construction projects. Such delay factors are typically considered individually and presented at the same level without explicitly examining their sequence of occurrence and interrelationships. In reality, owing to the complex structure of construction projects and long execution time, non-conformance to schedule occurs by a chain of cascading events. An understanding of these linkages is important not only for minimising the delays, but also for revealing the liability of stakeholders. To explicitly illustrate the cause-effect and logical relationship between delay factors and further identify the primary factors which possess the highest significance toward the overall project schedule delay, the fault tree analysis (FTA) method, a widely implemented approach to root cause problems in safety-critical systems, has been systematically and rigorously executed.

Design/methodology/approach

Using a case study, the in-depth analysis for identifying the most fundamental delay factors has been fulfilled through FTA’s tree structure. The logical deduction for mapping and visualising the chronological and cause-effect relationships between various delay factors have been conducted through the logical gate functions of FTA based on the data collected from the site event log, prefabricated structural component manufacturing log and face-to-face interview with project stakeholders.

Findings

The analysis identified multiple delay factors and showed how they are linked logically and chronologically from the primary causes to the ultimate undesired event in a rigorous manner. A comparison was performed between the proposed FTA model and the conventional investigation method for revealing the responsibility employed in the construction industry, consisting of event logs and problem reports. The results indicate that the FTA model provides richer information and a clearer picture of the network of delay factors. Importantly, the ability of FTA in revealing the causal connection between the events leading to the undesired delays and in comprehending their prominence in the real-world construction project have been clearly displayed.

Originality/ value

This study demonstrates a new application of FTA in the construction sector allowing the delay factors to be understood and visualised from a new perspective. The new approach has practical use in finding and removing root causes of the delay, as well as clarifying the attribution of responsibility that causes the delay.

1. Introduction

A construction project is considered to be successful when it is completed on time, within budget and all the stakeholders are satisfied with its quality (Gündüz et al. 2013). However, unique features of construction projects, such as their long execution period, complicated processes, high sensitivity to environmental influences, diverse interests of the different stakeholders and the dynamic structure of project teams, often make the delivery of projects on time very challenging (Zou et al. 2007). Indeed, delays have become considered as inevitable in construction (Aibinu and Odeyinka, 2006).

To control and eliminate the obstacles to complete a project on time, the causes of construction delays have to be understood and mitigated. As it currently stands, many studies have been carried out to disclose the causes of lateness. The methods typically applied are surveys conducted using interviews or questionnaires (e.g. Assaf et al. 1995; Gündüz et al. 2013), and also through reviews of existing literature (e.g. Aibinu and Odeyinka, 2006; Pethkar and Birajdar, 2015).

Such extant studies disclose the most common delay factors in construction projects. However, the results of these studies have generally regarded the causes of non-conformance in construction schedules to be at the same level. The cause-effect and the chronological relationships between delay causes are typically missing. In other words, it is unclear if a certain identified delay factor is the primary root, or the intermediate event that is triggered by a factor prior to it and subsequently elicits other delay events or is the ultimate undesired symptoms for the entire project.

Yet, the lateness in construction schedules results from a chain or sequence of events. The result of delay is triggered by intermediate events and these, in turn, are caused by primary events. Hence, delay factors in a construction project are connected not only chronologically but also logically. In practice, it is only by eradicating the primary delay factors in a project, that the likelihood of the subsequent undesired events to happen can be eliminated. In addition, knowing the causal relationship between delay factors can help clarify the responsibility of stakeholders and resolve disputes.

To acquire this knowledge, Fault Tree Analysis (FTA), an established tool to root cause problems in safety-critical systems is introduced and employed rigorously and systematically in this research. FTA uses a deductive approach to conduct a top-down analysis, in which an undesired event is investigated through a chain of lower-level events using Boolean logic to identify its primary causes.

There are only a few studies have adopted FTA in construction management. Yet, in these studies (Al-Humaidi and Hadipriono Tan 2010; Swarna and Venkatakrishnaiah 2014; Shahhosseini et al. 2018), FTA was used in a simplified way, and the fault tree’s ability to resolve the root causes of undesired events (e.g. cost overrun or schedule delay) has not been fully exploited. For example, the many intermediate events within the fault trees constructed in previous research were regarded as the root causes of the top tier event. Intermediate events such as rework owing to quality deficiency and owner’s changes to the design are taken as root cause directly, but they, in fact, can be further traced to more fundamental factors. Such circumstance might be due to not enough data being collected to support a more in-depth deduction and the logical mapping for identifying the primary reasons triggering the undesired events.

Another limitation of these studies is that the data to construct the fault trees is based on literature review and questionnaires used to collect the potential reasons for causing the undesired events. Therefore, in these studies the fault tree was used primarily as a tool to visualise and classify delay factors identified from previous work and by industrial stakeholders rather than deduce them by investigation. In other words, FTA was not utilised as a tool for identifying the cause-effect and logical relationships between the delay factors in a construction project as it is designed for.

This study aims to investigate the factors behind schedule deviations in construction projects and their causal and time relationships by analysing the fault tree that has been fully constructed according to a thorough interpretation of the available records of construction projects. The research presents a new application of the FTA method, and is conducted based on a case study which involves a large infrastructure construction project composed of various types of prefabricated concrete components.

2. Identifying delays in the construction industry

2.1. Surveys and literature reviews

Many studies have been carried out for obtaining a general understanding of the reasons for contributing to schedule deviation in construction projects. These investigations were carried out in different geographical areas, but the research methods applied were practically alike.

Alkass et al. (1996), Akintoye and MacLeod (1997), Bing (2005), Meng (2011), Braimah (2013), McCord (2015) and Parry (2015) all sent out questionnaires and conducted interviews with both contractors and consultancy companies in the UK to investigate the reasons for construction delays. Choong Kog (2018) applied the same investigation methods in the US and UK for discovering the main reasons for project prolongations. Love et al. (2010, 2013) and Zidane and Andersen (2018) also identified a list of potential reasons for causing the construction delay in Europe and Australia by reviewing the existing literatures and then further sent out questionnaires to various stakeholders in the construction sector for ranking these factors.

Similar research approaches have also been applied in the middle-east area, Assaf et al. (1995), Al-Momani (2000), Odeh and Battaineh (2002), Assaf and Al-Hejji (2006), El-Sayegh (2008), Gündüz et al. (2013) and Rashid (2020) identified construction delay factors by performing literature reviews to collect common reasons for causing schedule deviations and sending questionnaires to the industry sector in Saudi Arabia, Jordan, Turkey and United Arab Emirates, respectively.

The identical investigation methods of adopting literature review and questionnaire survey have also been implemented in the Asia area by Chan and Kumaraswamy (1997) and Lo et al. (2006) in Hong Kong, Zou et al. (2007) and Chen et al. (2019) in China, and Khoiry et al. (2018) in Malaysia, Doloi et al. (2012) in India and Islam et al. (2015) in Bangladesh.

Overall it appears that the methods for studying delay factors are common and widespread. A list of such studies is compiled and shown in Table 1.

(Table 1. A list of studies regarding the delay in the conventional projects continue)

2.2. Site event logs and daily reports

Construction companies usually keep track of events and activities at sites on a daily basis using daily reports and event logs. A daily report includes the numbers of labour, equipment and materials employed on a site and also states the project performance indices, such as actual cost of work performed (ACWP) and budgeted cost of work performed (BCWP). The report also records the progress of all ongoing construction activities and the occurrence of any adverse events (Goedert and Meadati 2008). For small-scale projects with short construction durations, event logs are employed which only use bullet points to list out the significant events on-site. These records serve as proof of finished works during the settlement of payment. In addition, when the project cannot be completed within the contract date, the contents of these report can be used to clarify which stakeholder should be considered responsible for the delay (Hamzah et al. 2011).

Although the daily reports and event logs provide a clear view of the facts that happen on the site, identifying the reasons for schedule deviation requires further investigation and logical deductions. In this study, these systematically recorded data are taken as the basis for performing the root cause analysis to determine the primary delay factors.

2.3. Root cause analysis

Various methods have been proposed to identify the causes of undesired events in a process or product for preventing them from reoccurring. Root-cause analysis (RCA) is one of the approaches developed for fulfilling this goal (Wilson et al. 1993).

The most commonly implemented technique for executing RCA is FTA tool, which carries out deductive investigations of the facts embedded in the event tree (Sklet, 2004). The concept of FTA was first conceived in 1961 by H. A. Watson of Bell Laboratories (Ericson, 1999). In 1963, Boeing was the first commercial company to develop the FTA tools for investigating failures in aircraft mechanical components.

A fault tree is capable of displaying the details of all the factors involved in an undesired incident, and in the meantime eliciting their cause-effect, time-dependent and logical relationships (Goodman, 1988; Joshua and Garber, 1992; Rausand and Hoyland, 2004; Hsu et al. 2017). It is generally recognised that FTA can provide a more in-depth understanding of the connections between events than other RCA approaches such as Fish-bone diagram, why-because analysis and Failure mode and effects analysis (Al-Humaidi and Hadipriono Tan 2010; Burhan, 2010; Swarna and Venkatakrishnaiah, 2014 and Shahhosseini et al., 2018). This is because the fault tree can efficiently direct the analysts to consider only the leading causes that contribute to the undesired events through the application of logical gates and the event-tree data structure (Lee et al. 1985).

FTA has been extensively implemented in various industry sectors for safety assessment and risk analysis, including the nuclear power industry (NRC, 1981), space shuttles (NASA, 2002), marine engineering (Miri-Lavasani et al., 2011), transportation (Xia et al., 2012) and mining (Iverson et al. 2001).

In the construction sector, however, there have been only a few studies on the application of the RCA approach. Rosenfeld (2013) used a technique called “expand and focus” to understand the primary causes of construction cost overrun in Israel. As a result, 15 leading causes were determined and a fishbone diagram was subsequently used to present the research outcome. Though, the diagram specified the cause-effect relationship between the cost overrun factors, the logical and the historical relationships among them are implicit.

Swarna and Venkatakrishnaiah (2014) and Shahhosseini et al. (2018) both used FTA to investigate the most prominent reasons for causing construction cost overrun. Their research outcomes have proven that FTA not only can disclose the primary reasons through logical deduction, but also can visualise the chronological order of major adverse events that happened during the project execution. However, their results also showed that the quantity and quality of the input information for constructing the fault tree has a profound influence on generating the correct primary cost overrun factors.

In Swarna and Venkatakrishnaiah (2014), the subject and the conditions of the research were not well delineated, i.e. the boundary of the fault tree is blurred, thus no tangible scope or coverage is specified for what data should be used to construct the fault tree. Under this circumstance, the importance of factors is hardly differentiated. The fault tree of Shahhosseini et al. (2018) was oversimplified (contains only three tiers), and all the factors related to the same attribute, regardless of their roles as either the cause or the effect are gathered in the same branch and placed at the same tier in the tree Thus, the outcome of the fault tree can only be regarded as a classification of the reasons contributing to construction risks. Consequently, to yield results in high quality and reliability, the procedures and rules for constructing fault trees must be strictly followed, and the data involved in undesired incidents must be included as detailed as possible.

RCA has also been employed to investigate the safety of building structures. Battikha (2008) used the FTA to identify the causes of building collapse. Photos of a collapsed building were scrutinized to observe the cracks on the building structures and further compared with the original structural design. Then the causal relationship between the cracks and the reasons for causing them was illustrated in a tree-like diagram and the primary reasons for causing the building collapse were shown in the bottom of the tree. Burhan (2010) also built a fault tree by examining problematic site photos and consulting experts. The purpose of their research was to identify the reasons for structural failure in a commercial building.

2.4 Closing remarks

The above review clearly shows that, compiling a list of delay factors, categorising and ranking them based on their impact and frequencies can develop an understanding toward the schedule deviation. However, in practice, to completely prevent delay from occurring, additional information is needed, i.e. the primary root causes of delay must be identified. Such knowledge can be obtained through analysis of the cause-effect and chronological relationships between various delay factors.

However, only a few studies on undesired events (e.g. cost overrun or schedule delay) in the construction sector have applied FTA (Al-Humaidi and Hadipriono Tan 2010; Swarna & Venkatakrishnaiah 2014 and Shahhosseini et al. 2018). Although they have put forward a new point of view, they have limitations.

Instead of using fault tree as a tool for identifying the root causes in a single problematic construction project, these previous studies were carried out aiming at gaining a general understanding of the root causes that contributed to project delay or cost overrun in the overall construction field, so the data for constructing fault trees were collected by sending questionnaires to consultant companies, contractors and project owners asking what are the most prominent reasons for causing the undesired events. The fault trees were then merely employed as an auxiliary tool for visualising the result of the questionnaire. The most frequently mentioned delay or cost overrun factors were regarded as the primary reasons in the constructed fault tree, which had only two tiers, consisting of an undesired event and primary reasons. Under this circumstance, the fault tree actually only acted as a diagram for structuring and classifying the delay factors and disclosing the main reasons causing them and which stakeholder should be blamed for, but the function of logical deduction between different tires of the tree is not explored.

The second limitation of these studies is that the fault tree’s capability of resolving the root causes of undesired events was not revealed. In the fault trees constructed in the aforementioned studies, many primary causes identified were actually intermediate events which can be triggered by more fundamental causes. For instance, rework was mentioned as one of the root causes for making project cost overrun (see Swarna & Venkatakrishnaiah (2014) and Shahhosseini et al. (2018)), but the reason for causing rework has not been explained. The causes of these intermediate events should have been further explored if more information and evidence was collected. Unfortunately, in these studies, the questionnaire respondents did not provide further information about the possible reasons for trigging the most frequently mentioned delay factors.

In this study, a brand-new application of the FTA method in the construction sector is deployed. The FTA method is applied based on the daily reports and event logs from a manufacturing factory in a systematic and rigorous way. This study’s application demonstrates how these conventionally written records can be transferred into tangible proofs showing logical and causal relationships between the construction schedule delay and the primary reasons for trigging it.

3. Case study and FTA development process

The following sections introduce the case study upon which the research is based, the data collection method and the modelling technique used to analyse delays.

3.1 Background to the case study

A large infrastructure project was studied to understand the causes of the delays that occurred during its execution. The infrastructure is mainly composed of precast concrete structural components such as beams and walkways. Each component is about 10-meter long and weighs more than 8 tons. The scheduled construction duration was 53 weeks. During the execution of the project, different delays were experienced. The most severe delay, totalling 17 weeks, occurred at the stage of manufacturing the structural components.

The project involved collaboration between five leading companies, see Figure 1. The design and build of the infrastructure were awarded to a large construction company (Company A), which subcontracted various design work and operations to other partners. In particular, the design of the main structural components was subcontracted to a design and consultancy company (Company B). The detailed design of the structural components was assigned to a company specialised in CAD (computer-aided design) (Company C). The design and manufacture of the moulds were undertaken by a small manufacturing company (Company D). Finally, some operations were outsourced to a manufacturing company with expertise in precast concrete (Company E). The relationship between the five companies is presented in Figure 1, and the numbers from 1 to 8 represent the time sequence of activities occurred in the project.

(Figure 1. The relationship between the stakeholders in the case study.)

3.2 Data collection methods

To understand the root causes of the delays in the case study, project data was collected from the collaborating company. The data include a face to face interview with the project manager and two project documents. The interview was semi-structured and lasted approximately 60 min. During the interview, notes were taken by the first author. The documents include a problem report consisting of two pages of A4 in Word format and an events log in Excel format. The problem report is mainly composed of short bullet points presenting the events that caused the project to deviate from its original timeline. A representative example of the problem report is shown in Table 2.

(Table 2. Fragment of the problem report obtained from the construction company.)

Schedule deviations that occurred in the manufacturing factory and on-site were logged by the construction company as events in spreadsheets. Fragments of the spreadsheets are shown in Table 3 and 4, respectively. The timelines of the events in Tables 2, 3 and 4 are given in Figure 2.

(Table 3. Fragment of the building site events log report.)

(Table 4. Fragment of the factory events log report.)

(Figure 2. Timelines of the design, site and factory events.)

3.3 Modelling method: Fault Tree Analysis

The FTA technique was adopted to model the chain of events that occurred during the project execution for deducing the primary causes of the delays. A fault tree is generally composed of multiple tiers. As shown in Figure 3, the undesired event is at the top. The intermediate events that trigger the undesired event are located in the middle, and at the bottom of the tree, there are the primary events. In addition, the logical relationships between events are indicated by specific logical gate symbols. The AND-logical gate indicates that the event in the tier above would be triggered when all the events in the tier below happen. Differently, the logical OR-gates means that if any one of the events in the tire below happens, the event in the tier above will be triggered.

(Figure 3. The structure of a fault tree. The logical symbols including AND and OR-gates are defined on the right-hand side of the figure.)

According to Wilson (1993) and Andersen and Fagerhaug (2006), the first step in constructing the FTA is to define the undesired event. In this study, the undesired event is “the construction project cannot be completed by the contract calendar date”. The boundary of the FTA was defined as all the stakeholders and places involved in the construction project under study. The second step is to determine the immediate necessary and sufficient events that result in the top event and draw the logical symbols, which best describes the relationship between them. Once the undesired event is resolved into its immediate causes, the third step is to treat each intermediate event as an intermediate level top event, and determine its immediate, necessary and sufficient causes, while resolving the logical relationships between causes. This will be repeated until the primary causes are identified.

3.4 Data analysis method: construction of the Fault Tree

The construction of the fault tree in this study was carried out in three steps:

(1) Identification of the delay factors.

(2) Organisation of the delay factors based on the date and their cause-effect relationship.

(3) Linking of the delay factors by drawing logic gates.

The delay factors were identified from the events log and the problem report, as well as the written record of the face-to-face interview. For instance, items mentioned in the problem report (see Table 1) such as “Late design release from Designer”, “The opportunity to produce detailed drawing at the factory was missed”, and “Design change for all walkways a week before planned first cast” are all captured as causes of the delay in the overall construction schedule. Similarly, items listed in the events log (see Tables 2 and 3) such as “A beam was not delivered today”, “Lorry driver collected the wrong load from the factory” and “Winded off from 10.30 am, installed 4 out of 12” are also recognised as delay factors.

The chronological links between the identified delay factors were obtained from the events log, which have recorded the date of each event. More so, the cause-effect relationship, as well as the logical connection between events at different tiers of the fault tree, were derived from the interview with the project manager.

In many instances, the relationships between delay factors could also be extracted from the linguistic expressions documented in the problem report and events log. For example, in the statement “Forced to seek external detailers due to the programmed delay”, the wording “due to” is a clear indication of the cause-effect relationship. Similarly, in the statement “Additional resource consumed to cope with changes in the programme”, the wording “to cope with” is also a hint of the connection between the time and monetary resources devoted to deal with this unplanned circumstance and the alternation in the assembly sequence.

3.5 Evaluation method

A comparative evaluation between the results given by the problem report, events log and the FTA model was performed. The numbers of nodes and the logical links identified in the fault tree that are absent in the problem report and events log were calculated to measure the extent of the outperformance that FTA can deliver. With this approach, the practitioners can comprehend the fault tree’s capability of capturing and visualising implicit information and relationships between various delay factors in different locations and at different times and understand how they contribute to the ultimate schedule delay.

4. Root cause analysis: delays in the construction of a public infrastructure

The fault tree constructed for the case study is shown in Figure 4. It is composed of two main branches. The left branch explains the delays arising within the construction site and the right branch illustrates the delays happening within the manufacturing factories.

The primary causes are found between the fourth and seventh tier below the top undesired event. Every event shown in the fault tree represents a factor for causing schedule deviation in this case study, and their chronological relationship is externalised from bottom-up in all branches.

The left main branch breaks down into two sub-branches. The one on the left-hand side describes why the original construction schedule was inaccurate, while the one on the right-hand side presents the causes of not fulfilling the original construction schedule.

The right main branch is composed of three sub-branches, which depict, from left to right, why the moulds for the prefabricated components had to be sent back to the manufacturer, why the production capacity in the factory was insufficient, and why the production quantity was below target.

(Figure 4. The fault tree analysis for the delays in the construction of the public infrastructure under study. The primary delay factors are circled by oval-shaped frames and numbered.)

For the two delay events presented in the grey-shaded boxes in Figure 4, remedy measures were taken by the construction company, which, however, resulted in further delays. These are captured by two additional FTA branches presented in Figure 5. The primary causes for the delays traced through the FTA, as shown in the bottom tiers of the branches in Figures 4 and 5, are numbered 1 through 12, and listed in Table 5.

(Figure 5. The fault tree analyses for the delays caused by the remedies to address the events presented in the grey-shaded boxes in Figure 4. The primary delay factors identified are circled by oval-shaped frames and numbered.)

(Table 5. The primary delay factors found in the case study)

As presented in the FTA, the ultimate undesired event was stated as “Components were not delivered as expected”. The primary causes of such event were found to be mainly associated with lack of information and inexperience.

As an example, lack of information and experience were found to result in the inaccurate scheduling for the assembly process, which then forced the assembly sequence to be changed twice (primary delay factor 1 and 2). Lack of manufacturing experience and clear understanding of the client’s requirements by the designers in Company B resulted first in the ill design of the walkways and beams, and then in the incorrect design of the moulds by Company D (primary delay factor 8 and 9). As a result, the moulds had to be sent back to the manufacturer (Company D) for revision, which severely slowed down the production of modular components.

Furthermore, the design of the walkways and beams was not delivered on time, which further delayed the mould design and the start of the component production, leaving insufficient time to make the number of components requested by the site.

In addition, the manufacturing operators in Company A lacked the experience in making structural components in a factory environment (primary delay factor 6). For instance, defects were found on the surfaces of many finished components, reinforcements were not properly installed in some components, and other components were not produced in accordance with the detail design drawings. When these faults were identified on the site, the non-compliant components had to be sent back to recast. The inexperience of the manufacturing operators in Company A also led to fluctuations in productivity (primary delay factor 11), resulting in production below the target quantity.

The delay was also caused by the lack of a systematic management plan in Company A. Without a strict quality check mechanism for finished components, components with defects were sent to the site and got rejected (primary delay factor 7). Furthermore, without an effective tagging system, wrong components were sent to the site on the wrong date (primary delay factor 5).

Other factors that contributed to the lateness of the project are strong winds which forced the crane to stop operating (primary delay factor 3) and the traffic congestion between the factory and the site that made the delivery on time rather challenging (primary delay factor 4).

Interestingly, the attempts to correct the two delays (shown in the grey-shaded boxes in Figure 4) were also affected by inexperience and insufficient communication between the parties. As shown in Figure 5, insufficient information was given to the outsourcing manufacturing company (Company E), and much time was needed to provide explanations, slowing down the production of the structural components (primary delay factor 12, the left path in Figure 5). Similarly, not enough information for producing the drawings of the moulds was provided to Company C, causing the delay in receiving the detail drawings (primary delay factor 12, the right path in Figure 5).

5. Evaluation

In the fault trees shown in Figures 5 and 6 there are a total of 43 event nodes, and 40 logical links (33 nodes and 32 links in Figure 4, and 10 nodes and 8 links in Figure 5) representing all the schedule delay incidents occurred in the case study project.

A detailed analysis of the information sources behind the FTA nodes and links showed that only 23 event nodes and 14 logical links are explicitly stated in the written records accounting for 53% and 35% of the entire information space respectively (see Table 6). The remaining events and the logical relationships were retrieved from the face-to-face interview with the project manager, suggesting that it sometimes requires human memory to complement incomplete written records.

(Table 6. A comparison of the quantity of information captured by

FTA and conventional methods)

It is essential to recognise what events were not captured in the events log and problem reports. The problem report acknowledged that the schedule delay was caused only by: (1) the late release of the structural component design, which was subcontracted to the design and consultancy company (Company B); and (2) a misunderstanding of the design by the company responsible of producing the detail design of the structural components (Company C).

Compared to the FTA presented in section 4, the conclusions of the problem report did not capture the delay owing to adverse events on-site, nor did the delay happen in the manufacturing and transportation process. These delays were, however, captured in the events log, which indicates that not all the schedule deviation incidents recorded in the events log were discussed in the problem report.

Furthermore, the delays incurred by the remedy actions were also not explicitly covered in the problem report. In conclusion, the report was found to lack a systematic and structured approach to capturing the full picture of the delay problem.

Keeping an event log is a common practice in every construction site and factory. In the adopted case study, the events that happened on the site were recorded but not classified into different categories. Also, the events log did not indicate how events happened and who should be responsible, nor identify which event had a higher priority to be resolved.

The same situation occurred in the events log for the factory. There was no comprehensive and holistic record of the production process and the undesired issues that happened. At present construction companies tend to file a problem report directly based on event logs without conducting further detailed analysis (Hamzah et al. 2011). Instead, they should attempt to make use of tools such as FTA to deduce the primary causes.

Table 7 presents a comparison between the FTA tool and the practice employed by the construction company with respect to the investigation of delay factors. The items listed in Table 7 were obtained from Andersen and Fagerhaug (2006), who have generalised the most common features of the methods used for finding reasons causing undesired events. In conclusion, the fault tree is established based on written records/interview, but the later can become more meaningful and useful through further FTA.

(Table 7. The comparison between the FTA and the common practice)

6. Discussion

The conventional approach for studying delays in construction by using survey or interview indeed can find a list of generalised delay factors that are of high impact and/or high frequency of occurrence. Nevertheless, these methods are of limited use when the goal is to identify the root cause of delays in a specific project. This study demonstrates that FTA is a powerful tool for analysing a complex network of delay events in a construction project and identifying the roots of the delays. If the FTA method is embraced by the construction sector, the reasons for causing the lateness of projects can be identified more efficiently and precisely. Thus, once the root causes have been eradicated, the chance of running into the same type of schedule deviations can be minimised.

The most important benefit of employing the FTA is that it can reveal the cause-effect relationship between the delay factors in a tangible structure with chronological order. It also discloses that some delay factors are more influential than others by implementing a rigorous logical deduction.

Furthermore, the delay factors, which happened in different locations (e.g. building site and factory) can be captured in distinct branches of a single fault tree, thus clearly and concisely showing the whole picture of how they contribute to the top undesired event. The remedial actions for the adverse event can be illustrated by the fault tree as well.

In contrast, the practice adopted in the industry is such that delay events are typically scattered on different pages of an event log and even in different spreadsheets or books, making it extremely difficult to identify the real primary cause of delays unbiasedly.

It is worth mentioning that in the construction industry a simple investigation based on the facts recorded in site event logs and problem reports is the most prevalent way of identifying the reason for project delays. The purpose of the investigation is merely to find out which stakeholder has not reached the project milestone within the contract time or has consumed the float time belonging to another stakeholder. The outcome of this process is to clarify the responsibility when there is a dispute between the stakeholders, or a stakeholder is being sued by the project owner for late project delivery. In reality, most of the construction companies do not systematically review the reason why a project is delayed, because they usually regard construction projects to be a one-off thing. Thus, the true primary reasons for causing schedule deviation are hard to eradicate, and the probability of a project being delayed remains high.

The FTA can be embraced by the construction sector simply by consulting the facts recorded in the exiting site event log and the preliminary problem report and carrying out further analysis and information compilation. The procedures for constructing a fault tree for tracing the root cause of a construction project that cannot be finished on time has been introduced in section 3 of this research. Once, the FTA has been executed systemically on a project, the revealed primary reasons for causing the delay can serve as the basis of lateness prevention for the similar type of project in the future.

Furthermore, the FTA method can be applied to the investigation of the most prominent reason for causing project cost overrun. Therefore, introducing the FTA method to the construction sector can provide motivation for stakeholders to keep detailed records for all events, because FTA can impose the value of cost mitigation and lateness prevention in the future. In this context, the stereotype that a construction project is a one-off thing may be altered.

Using the FTA tool, this research mapped multiple delay factors, which were traced all the way to their primary causes as shown in Figure 4 and 5. This approach to modelling shows that some delay factors possess greater significance than others as shown in Table 5.

However, it is worthy to note that FTA often is project-specific, i.e. each fault tree is tailored for a particular construction project.

7. Conclusions

In construction sectors, delay factors are commonly investigated employing surveys based on interviews or questionnaires. With these methods, the logical and chronological relationships between delay factors are usually hidden. In reality, construction projects have unique traits such as multiple and dynamic stakeholders, environmental sensitivity and long execution periods, which contribute to delays being triggered by a sequence of events.

Using a modular construction project to develop a large public infrastructure as a case study, this research has proposed the FTA tool as a new paradigm to illustrate the cause-effect link between the primary delay factors and the main undesired event.

Different from the previous works, this study focused on demonstrating a new application of the FTA method in the construction sector, which provides insightful information about the primary reasons for trigging adverse events in a construction project. This was achieved by systemically analysing the causal and logical relationships between all the events recorded in the site event logs and problem reports of a delayed construction project. Therefore, the new application adds new value to these conventional event logs and problem reports.

By implementing the FTA tool to investigate lateness in a construction project, the authors found that approximately half of the delay events and two-thirds of the logical relationships between these events were not uncovered by the conventional investigation method used in construction.

8. Limitations and future works

The logical deduction of the FTA is heavily depending on the quality and completeness of the collected data for illustrating the interrelationship between all the cascading events which contributes to the ultimate undesired event. In this context, a prerequisite for adopting the FTA by the construction sector is that the site event log must record all the events occurred on the site not only in detail but also symmetrically. If the project involves the method of prefabrication/ modular construction, the event logs recording the manufacturing and transport processes also become essential. Thus, a limitation of the FTA application is that the data collection may be very time consuming and expensive.

The future work can be extended to involve the probabilistic fault tree which indicates the probability of occurrence for the intermediate and the primary events linked by the “OR” logical gate. Under this circumstance, the importance of each of the multiple root causes at the bottom of the tree can be assigned by observing its probability of occurrence through the fault tree. This can be regarded as a method for the quantification of stochasticity.

Acknowledgements

The authors acknowledge the support of the Taiwan Top University Strategic Alliance-Imperial PhD scholarship.

References

Abdelgawad, M., and Fayek, A. R. (2011). “Comprehensive hybrid framework for risk analysis in the construction industry using combined failure mode and effect analysis, fault trees, event trees, and fuzzy logic.” J. Constr. Eng. Manag., 138(5), 642-651.

Abd El-Razek, M. E., Bassioni, H. A., and Mobarak, A. M. (2008). “Causes of delay in building construction projects in Egypt.” J. Constr. Eng. Manag., 134(11), 831-841.

Aibinu, A. A., and Odeyinka, H. A. (2006). “Construction delays and their causative factors in Nigeria.” J. Constr. Eng. Manag., 132(7), 667-677.

Alazzaz, F., and Whyte, A. (2015). “A review of current barriers (real and perceived) to off-site construction.” 21st IIER International Conference, Australia, 27-31.

Al-Humaidi, H. M., & Hadipriono Tan, F. (2010). A fuzzy logic approach to model delays in construction projects using rotational fuzzy fault tree models. Civil Engineering and Environmental Systems, 27(4), 329-351.

Al-Momani, A. H. (2000). “Construction delay: a quantitative analysis.” management.” Int. J. Proj. Manag., 18(1), 51-59.

Andersen, B., and Fagerhaug, T. (2006). Root cause analysis, in: Hytinen, A., O’Mara, P. (Eds.), Root cause analysis: Simplified tools and techniques, 2nd ed., ASQ Press, Milwaukee.

Assaf, S. A., Al-Khalil, M., and Al-Hazmi, M. (1995). “Causes of delay in large building construction projects.” J. Manag. Eng., 11(2), 45-50.

Assaf, S. A., and Al-Hejji, S. (2006). “Causes of delay in large construction projects.” Int. J. Proj. Manag., 24(4), 349-357.

Battikha, M. G., (2008). “Reasoning mechanism for construction non-conformance root-cause analysis.” J. Constr. Eng. Manag., 134(4), 280-288.

Blismas, N. G., Pendlebury, M., Gibb, A., and Pasquire, C. (2005). “Constraints to the use of off-site production on construction projects.” Archit. Eng. Design. Manag., 1(3), 153-162.

Burhan, A. M. (2010). “Fault tree analysis as a modern technique for investigating causes of some construction project problems.” J. Eng., 2(16), 5214-5224.

Chan, D. W., and Kumaraswamy, M. M. (1997). “A comparative study of causes of time overruns in Hong Kong construction projects.” Int. J. Proj. Manag., 15(1), 55-63.

Choong Kog, Y. (2018). Major Construction Delay Factors in Portugal, the UK, and the US. Practice Periodical on Structural Design and Construction, 23(4), 04018024.

Chang, S. Y., Lin, C. R., and Chang, C. T. (2002). “A fuzzy diagnosis approach using dynamic fault trees.” Chem. Eng. Sci., 57(15), 2971-2985.

Chen, G. X., Shan, M., Chan, A. P., Liu, X., & Zhao, Y. Q. (2019). Investigating the causes of delay in grain bin construction projects: the case of China. Int. J. Cons. Manag., 19(1), 1-14.

Doloi, H., Sawhney, A., Iyer, K. C., and Rentala, S. (2012). “Analysing factors affecting delays in Indian construction projects.” Int. J. Proj. Manag., 30(4), 479-489.

El-Sayegh, S. M. (2008). “Risk assessment and allocation in the UAE construction industry.” Int. J. Proj. Manag., 26(4), 431-438.

Ericson, C. (1999). “Fault tree analysis: A history.” Proc. 17th International System Safety Conference, International System Safety Society, Orlando., 1-9.

Goedert, J. D., and Meadati, P. (2008). “Integrating construction process documentation into building information modelling.” J. Constr. Eng. Manag., 134(7), 509-516.

Goodman, G. V. (1988). “An assessment of coal mine escape way reliability using fault tree analysis.” Min. Sci. Tech., 7(2), 205-215.

Gündüz, M., Nielsen, Y., and Özdemir, M. (2013). “Quantification of delay factors using the relative importance index method for construction projects in Turkey.” J. Manag. Eng., 29(2), 133-139.

Hamzah, N., Khoiry, M. A., Arshad, I., Tawil, N. M., and Ani, A. C. (2011). “Cause of construction delay-Theoretical framework.” Procedia. Eng., 20, 490-495.

Hsu, P. Y., Aurisicchio, M., and Angeloudis, P. (2017) “Investigating Schedule Deviation in Construction Projects through Root Cause Analysis.” Procedia Comput. Sci., 121, 732-739.

Islam, M. S., Trigunarsyah, B., Hassanain, M., and Assaf, S. (2015). “Causes of delay in construction projects in Bangladesh.” 6th International Conference on Construction Engineering and Project Management, Busan, Korea., 82-86.

Iverson, S., Kerkering, J.C., and Coleman, P. (2001). “Using fault tree analysis to focus mine safety research.” 108th annual meeting of the society for mining, metallurgy and exploration, 1-10.

Joshua, S. C., and Garber, N. J. (1992). “A causal analysis of large vehicle accidents through fault‐tree analysis.” Risk Anal., 12(2), 173-188.

Katsakiori, P., Sakellaropoulos, G., and Manatakis, E. (2009). “Towards an evaluation of accident investigation methods in terms of their alignment with accident causation models.” Saf. Sci., 47(7), 1007-1015.

Khoiry, M. A., Kalaisilven, S., & Abdullah, A. (2018). A Review of Minimizing Delay in Construction Industries. In E3S Web of Conferences, 65, 03004.

Lee, W. S., Grosh, D. L., Tillman, F. A., and Lie, C. H. (1985). “Fault Tree Analysis, Methods, and Applications A Review.” IEEE T. Reliab., 34(3), 194-203.

Lo, T. Y., Fung, I. W., and Tung, K. C. (2006). “Construction delays in Hong Kong civil engineering projects.” J. Constr. Eng. Manag., 132(6), 636-649.

Love, P. E., Li, H., Irani, Z., Treloar, G. J., and Faniran, O. O. (2000, November). MiDiCON: A model for mitigating delays in construction. In 1st International Conference on Systems Thinking in Management, ICSTM2000. Geelong, Australia (pp. 8-10).

Love, P. E., Sing, C. P., Wang, X., Edwards, D. J., and Odeyinka, H. (2013). Probability distribution fitting of schedule overruns in construction projects. J. Oper. Res. Soci, 64(8), 1231-1247.

Miri-Lavasani, M. R., Wang, J., Yang, Z., and Finlay, J. (2011). “Application of fuzzy fault tree analysis on oil and gas offshore pipelines.” Int. J. Mar. Sci. Eng., 1(1), 29-42.

National Aeronautics and Space Administration (NASA), (2002). Fault tree handbook with aerospace applications. Office of Safety and Mission Assurance, Washington, DC.

Odeh, A. M., and Battaineh, H. T. (2002). “Causes of construction delay: traditional contracts.” Int. J. Proj. Manag., 20(1), 67-73.

Olawale, Y. A., and Sun, M. (2010). “Cost and time control of construction projects: inhibiting factors and mitigating measures in practice.” Constr. Manag. Econ., 28(5), 509-526.

Pethkar, P. M., and Birajdar, B. V. (2015). “Literature review on causes of delay in building construction projects.” J. Info. Know. Res. Civ. Eng., 3(2), 217-220.

Rausand, M., and Hoyland, A. (2004). Qualitative system analysis, in: David, J. (Eds.), System reliability theory: models, statistical methods, and applications.2nd ed. John Wiley & Sons, Inc., Hoboken.

Ramkumar, A., and Gopalakrishnan, S. (2014). “Root cause analysis of issues in construction industry.” Int. J. Innov. Res. Sci. Eng. Tech., 3(1), 1-8.

Rashid, Y. (2020). Analysis of delay factors and their effects on construction projects. Manag. Sci. Letter, 10(6), 1197-1204.

Rosenfeld, Y. (2013). “Root-cause analysis of construction-cost overruns.” J. Constr. Eng. Manag., 140(1), 1-10.

Shahhosseini, V., Afshar, M. R., and Amiri, O. (2018). “The Root Causes of Construction Project Failure.” Scientia Iranica, 25(1), 93-108

Sklet, S. (2004). “Comparison of some selected methods for accident investigation.” J. Haz. Mtl., 111(1), 29-37.

Swarna, A. M., and Venkatakrishnaiah, R. (2014). “Fault tree analysis in construction industry for risk management.” Int. J. Adv. Res. Civ, Struct. Envir. Infrastruct. Eng. Dvlpg, 2(1), 15-21.

Sweis, G., Sweis, R., Hammad, A. A., and Shboul, A. (2008). “Delays in construction projects: The case of Jordan.” Int. J. Proj. Manag., 26(6), 665-674.

Tako, A. A., and Robinson, S. (2012). “The application of discrete event simulation and system dynamics in the logistics and supply chain context.” Deci. Sup. Sys., 52(4), 802-815.

Tiware, M. V., and Kulkarni, S. S. (2013). “Root Cause Analysis of Variations in Construction Tasks and Developing Effective Strategies to Reduce Variations.” Int. J. Sci. Eng. Res., 4(9), 51-58.

United States Nuclear Regulatory Commission (NRC) (1981). Fault tree handbook, NUREG-0492. Systems and Reliability Research Office of Nuclear Regulatory Research, Washington, D.C.

Wang, D., Zhang, P., and Chen, L. (2013). “Fuzzy fault tree analysis for fire and explosion of crude oil tanks.” J. Loss Prev. Proc. Ind., 26, 1390–1398.

Wilson, P., F., Dell, L. D., and Anderson, G. F. (1993). Introduction, in: Lau, J. L., Nilles, B. S. (Eds.), Root cause analysis: A tool for total quality management. ASQ Press, Milwaukee, 1-18.

Xia, M., Li, X., Jiang, F., and Wang, S. (2012). “Cause analysis and countermeasures of locomotive runaway accident based on fault tree analysis method.” Proc. Eng., 45, 38-42.

Yuhua, D., and Datao, Y. (2005). “Estimation of failure probability of oil and gas transmission pipelines by fuzzy fault tree analysis.” J. Loss Prev. Porc. Ind., 18(2), 83-88.

Zidane, Y. J. T., & Andersen, B. (2018). The top 10 universal delay factors in construction projects. Int. J. Proj. Manag. in Business. , 11(3), 650-672.

Zou, P. X., Zhang, G., and Wang, J. (2007). “Understanding the key risks in construction projects in China.” Int. J. Proj. Manag., 25(6), 601-614.

Table 1. A list of studies regarding the delay in the conventional projects continue

Authors

Country

Project type

Methods applied

Leading delay factors found

Assaf et al. (1995)

Saudi Arabia

Civil engineering

Questionnaire

Inexperience of contractor

Inappropriate site management

Inclement weather conditions

Alkass et al. (1996)

United Kingdom

Civil engineering

Interview

Questionnaire


Owner’s changes in design

Late payment to contractors

Akintoye & MacLeod (1997)

United Kingdom

Civil engineering

Interview

Questionnaire



Slow decision making

Chan & Kumaraswamy(1997)

Hong Kong

Civil engineering

Interview

Questionnaire

Poor site management


Frequent change in design

Al-Momani (2000)

Jordan

Civil engineering

Interview

Questionnaire




Aibinu & Jagboro (2002)

Nigeria

Building

Interview

Questionnaire

Financial problems



Odeh and Battaineh (2002)

Jordan

Civil engineering

Interview

Questionnaire

Owner over interference

Inexperience of contractors

Improper project planning

Bing (2005)

United Kingdom

Civil engineering

Interview

Questionnaire




Aibinu & Odeyinka (2006)

Nigeria

Building

Interview

Questionnaire

Financial problems



Assaf & Al-Hejji (2006)

Saudi Arabia

Building

Literature review

Questionnaire




Lo et al. (2006)

Hong Kong

Civil engineering

Interview

Questionnaire

Lack of capital

Unforeseen geological conditions

Inexperience of contractors

Sambasivan & Soon (2007)

Malaysia

Civil engineering

Interview

Questionnaire

Clients’ financial problems



Zou et al. (2007)

China

Civil engineering

Interview

Questionnaire




Abd El-Razek et al. (2008)

Egypt

Civil engineering

Literature review

Interview




El-Sayegh (2008)

United Arab Emirates

Civil engineering

Interview

Questionnaire




Sweis et al. (2008)

Worldwide

Building

Literature review




Love. et al. (2010)

Australia

Civil engineering

Literature review

Interview

Change orders

Rework due to poor quality Inexperience of contractor

Hamzah et al. (2011)

Malaysia

Civil engineering

Literature review

Questionnaire

Inexperienced contractors

Ineffective project scheduling


Meng (2011)

United Kingdom

Civil engineering

Interview

Questionnaire



Late payment to contractors

Doloi et al. (2012)

India

Building

Interview

Questionnaire

Contractor’s financial problems



Braimah (2013)

United Kingdom

Civil engineering

Interview

Questionnaire




Gündüz et al. (2013)

Turkey

Civil engineering

Literature review

Interview


Ineffective project scheduling


Love. et al. (2013)

Australia

Civil engineering

Literature review

Interview

Change orders

Rework due to poor quality Unforeseen ground conditions

Islam et al. (2015)

Bangladesh

Civil engineering

Interview

Questionnaire

Low bidding price

Owner’s financial problem


McCord (2015)

United Kingdom

Civil engineering

Interview

Questionnaire




Parry (2015)

United Kingdom

Civil engineering

Interview

Questionnaire




Choong Kog (2018)

United State

United Kingdom

Civil engineering

Interview

Questionnaire


Slow government permission

Incomplete design drawing

Khoiry et al. (2018)

Malaysia

Civil engineering

Literature review



Frequent change in design

Zidane & Andersen (2018)

Norway

Civil engineering

Literature review

Questionnaire


Delays in payment

Poor planning and scheduling

Chen et al. (2019)

China

Building

Literature review

Questionnaire

Client frequently change order


Shortage of adequate equipment

Rashid (2020)

Saudi Arabia

Civil engineering

Literature review

Questionnaire


Client changes in design

Inefficient equipment

Table 2. Fragment of the problem report obtained from the construction company.

Design Intent & Programme

- T1* – planned design intent release

- Late design release from Designer – assumed release on T2

- T3 – final construction issue design received, 17 weeks late from T1

Detailing & Moulds

- Commitment from the project team to approve drawings within five days not actioned

- Forced to order the first mould in March based on preliminary drawings

- The opportunity to produce detailed drawing at the factory was missed

- Forced to seek external detailers due to the programmed delay

- External detailers draw in 2D – not compatible with the production plan

- Design change for all walkways a week before planned first cast

Manufacture & Install

- Change of install sequence on T4

- Lack of clarity on tolerances of reinforcement resulting in extra time spent fixing steel

- Factory outputs lower than planned

- Change of install sequence on T5

- Additional resource consumed to cope with changes in programme

*T# denote the date in which the designated event happened.

Table 3. Fragment of the building site events log report.

Date

Issue with

Summary

Description of Issue

S1

Factory

Wrong delivery sequence

The last delivery arrived first, and it arrived at 10 am

S2

Factory

Late delivery

11 am delivery late

S3

Factory

Panels have surface blemishes

Panels have surface blemishes

S4

Factory

Load missing beams

A beam was not delivered today

S5

Factory

Wrong delivery

Lorry driver collected the wrong load from factory

S6

Factory

Beam cast incorrectly

Cantilevered beams cast with no grout tubes

S7

Factory

Late delivery

The walkway arrived at 12 pm

S8

Factory

Late delivery

The beam arrived at 12 pm

S9

Factory

Beam cast incorrectly

The beam cast with insufficient cover at one end

S10

Site

Winded off cranes

Winded off from 10.30am, installed 4 out of 12

S11

Factory

Poor finish beams

Factory carried remedial works to repair cracks

S12

Factory

Poor surface finish on walkways

Walkway surface finish was poor

Table 4. Fragment of the factory events log report.

Date

Description

Required Control

F1

Walkway had no bullnose finish

Bullnose finish to be reformed

F2

Poor finish on walkway surface

Ensure brush finish has proper attention

F3

Site found cracking on beams

Factory confirms beams are safe and have no long-term effect

F4

Site found a walkway wrongly cast

The walkway was wrongly tagged but correctly cast

Table 5. The primary delay factors found in the case study

No.

Primary Delay factors

Responsible company

Group

1

Lack of experience by schedule engineers

A

Exp.

2

Lack of information at scheduling time

A

Exp.

3


External

FM.

4

Traffic congestion between factory and site

External

FM.

5

No tag system built for identifying each component

A

Mgmt.

6

Lack of experience in making components in the factory by manuf. operators

A

Exp.

7

No effective quality check mechanism for finished components

A

Mgmt.

8

Requirements were not understood by designers

B,C

Exp.

9

Lack of experience in design for manufacturing by the designers

B,C,D

Exp.

10

Construction company changed the design

A

Exp.

11

Fluctuations in workers’ productivity

A

Exp.

12

Inadequate information exchanged between parties

A,B,C,D,E

Mgmt.

Exp: experience related; Mgmt: management related; FM: force majeure related

Table 6. A comparison of the quantity of information captured by

FTA and conventional methods

Method

FTA

Conventional method

Conv. meth. /FTA (%)

Node

43

23

53

Link

40

14

35

Table 7. The comparison between the FTA and the common practice

Information revealed*

FTA

Common practice

The cause-effect relationship between delay factors

V

The logical relationship between delay factors

V

The chronological order of the delay factors

V

V

The significance (importance) of delay factors

V

The stakeholders responsible for the delay

V

V

The classification of delay factors

V

V

Alert that can elicit prompt responses

V

Display delay factors that happened in different locations and at a different time altogether

V

*The items listed in the left-most column are derived from Andersen and Fagerhaug (2006).

Figure 1. The relationship between the stakeholders in the case study.

Figure 2. Timelines of the design, site and factory events.

Figure 3. The structure of a fault tree. The logical symbols including AND and OR-gates are defined on the right-hand side of the figure.

Figure 4. The fault tree analysis for the delays in the construction of the public infrastructure under study. The primary delay factors are circled by oval-shaped frames and numbered.

Figure 5. The fault tree analyses for the delays caused by the remedies to address the events presented in the grey-shaded boxes in Figure 4. The primary delay factors identified are circled by oval-shaped frames and numbered.

21

spiral.imperial.ac.uk · web view2020. 10. 13. · instead of using fault tree as a tool for...

Documents