Identification of 4D-BIM Barriers in Offshore Construction Projects Using Fuzzy Structural Equation Modeling

Abstract

Planners face many obstacles during the planning phase of any new project, especially project scheduling due to the numerous details and complexity of each uniquely designed project; previous experience in similar projects and traditional scheduling methods are not sufficient. The 4D-BIM schedule is the best solution, as it can be integrated with other modern technologies such as UAS, which improves project scheduling by allowing the project team to access project plans, details, and time-related information to simulate construction sequences. Despite the benefits of using 4D-BIM, planners prefer traditional planning and scheduling methods because of the barriers to using modern technologies. This research proposes a structural equation model focusing on estimating the weights of BIM barriers in offshore construction projects, including: the investigation of barriers keeping the construction industry standing at 3D-BIM; the relationships between them; and the evaluation of the probability and impact of each, using fuzzy theory. To validate the proposed model, a case study of an offshore project was conducted. The most important latent variables were awareness, risk, demand, and management, while the most prominent observed variables were the uncertainty of the return on investment and the lack of experienced users. This research guides project managers on actions that can be taken for each key variable to enhance and develop the use of 4D-BIM in project scheduling.

1. Introduction

Many countries focus on the construction industry as a source of progress since it accounts for the majority of the GDP and is critical to the country’s economic prosperity [1] through constructing various types of projects, including ports, tunnels, and skyscrapers, and employing different project management approaches and technical breakthroughs. When scheduling such projects, planners tend to rely on their previous experience in similar projects and the use of traditional scheduling approaches; this frequently results in numerous changes to the project schedule by adding new activities, changing their durations and relationships, and causing an increase in time and costs [2]. The planners require a complete visualization of new projects to create and adjust a precise schedule with the necessary activities and appropriate work breakdown structure; therefore, not just relying on prior experiences alone, as each project has a unique design and does not contain the same details. Traditional plans are often produced utilizing computer technologies such as 2D diagrams or drawings, without the spatial characteristics of actual construction such as the activity-based critical path approach (CPM). This approach is frequently used in construction planning [3] and does not provide detailed information about the spatial configuration and the project components’ complexities [4]. These recurring issues highlight the importance of implementing new planning and scheduling methods that can improve communication among stakeholders and aid in an accurate understanding of the project, design, and planning well in advance of implementation. The preferred and cost-effective option for the construction industry is building information modeling (BIM). Depending on BIM as a project management tool has the most beneficial influence on project quality, cost, and timeliness [5]; as it enhances stakeholder collaboration, minimizes misunderstanding, and increases project clarity by utilizing a single model that has all the information and data.

Several strategies and approaches have been developed throughout the years by governments in various nations to begin implementing BIM in projects. As a result of this transition in the construction sector, the majority of industrialized countries have modified their project management techniques and started striving to boost BIM reliance by adopting it into their construction markets, abandoning old project management methodologies [6], and incorporating BIM with emerging technologies such as unmanned aerial systems (UASs) or drones to manage data and coordinate all elements of construction, including planning and design activities [7]. After creating the 3D-BIM model (level 1), the first step in BIM level 2 is 4D-BIM, which is also called 4D planning. It is an intelligent link between 3D CAD elements and their corresponding time or schedule information through the entering of schedule dates into the model elements or components creating the as-planned model [8]. The created BIM model can be integrated with UAS or drones that are supplied with different types of sensing systems such as infrared cameras, light detection and ranging (LIDAR), and other ranging systems to enrich the BIM model with accurate real-time data [7,9]. This integration provides planners and project managers with a strong, effective tool to monitor the moving sequence of the construction process; follow the status of each component; obtain a complete visualization of the project details; control the project by comparing as-planned model with as-built; and enhance the performance of the users [10,11]. Despite the benefits, project managers and planners prefer using traditional methods in project management and scheduling in most of their work, depending on pure 2D planning rather than on BIM and UAS for various reasons that prevent them from adopting such a new way of project management [12].

Many studies have discussed the implementation barriers of BIM and 4D-BIM as independent factors through surveys and interviews in terms of effect only, without taking into account the probability of occurrence of each factor, in addition to not taking into account the relationship between the measured variables and the latent variables. Hence, this study proposes a structure equation model to measure the overall obstacles to the construction of offshore works, taking into account the latent variables, probability, the impact of each barrier, and the relationship between them. The model also detects the main latent causes of barriers that need to be solved to enhance BIM and 4D-BIM adoption.

2. Literature Review

One of the most important tasks in projects that advance and expand the construction industry is planning. As effective pre-construction planning cuts down on building time by minimizing delays, increasing production rates, lowering costs, and making the most use of available space. The planners have a difficult job since the finished result will be viewed and scrutinized upon completion [2]. Additionally, projects are usually characterized by difficulties and ambiguity as a consequence of potential causes and situations during execution, which makes it challenging for the many project participants to arrive at an accurate and comprehensive plan [13]. Planners tend to use their prior experiences from comparable projects to assess and analyze the activities while researching the their implementation effects [2]. However, as a result of the technological progress in designing projects, the huge amount of information and the complexities associated with the project details, combined with the accelerating pace of construction, have made these methods of management and planning insufficient. Therefore, there was an urgent need to adopt new technology that can keep pace with this technological development in construction management.

BIM has evolved as an innovative concept in the past few years, and it is widely seen as the future transformation of the engineering and construction business [14]. Many studies conducted using BIM in project scheduling have demonstrated that it is a more effective technique for planning and scheduling than conventional methods that do not have a direct link to the design or construction model, leading to exceeding the planned schedule. In addition, the traditional methods do not include the spatial or resource requirements of activities [15]. The barriers to BIM and 4D-BIM scheduling in building construction projects have been discussed in many studies conducted in different industrialized nations. They indicated that the obstacles are different from one nation to another according to various factors affecting each government’s construction industry. The most common barrier to implementing 4D planning in the UK has been the lack of tangible benefits for all parties involved; the lack of understanding of the business value of BIM; the lack of experience within the workforce; the lack of global use; resistance to change; or a contract type and project delivery method which prevents technology adoption due to time and cost [14].

The adoption of barriers also have been delayed in China due to the insufficient government leadership to provide a direction or motivation to start relying on BIM as a management tool, along with some regulatory and legal issues, the high cost of the application, and resistance to change [16]. In Australia, despite the efforts the Building Smart Organization is taking, there is a low adoption rate of using 4D-BIM due to the lack of investment cost with the lack of evident rate of return [17]. Furthermore, despite the research progress in Sweden in the field of BIM, and its use in the construction industry, 4D-BIM faces many barriers preventing its use as in in many other countries. For example, there are owners’ refusals to adopt 4D-BIM due to a lack of understanding of its importance in projects; the lack of 4D-BIM experts in the market; a lack of certified standards to guide the spread and understand 4D-BIM; and cultural resistance believing that the existing scheduling programs and techniques are adequate. A survey of the German AEC market revealed that planners are utilizing their advanced applications for conventional 2D planning in much more than 60% of their entire usage, while for construction companies it is much more than 70% [12]. Unfortunately, most of the previous research did not examine the overall barriers. Previous studies were also based on the construction industry as a whole, and not on specific businesses such as offshore works. The BIM/4D-BIM adoption barriers are shown in

Table 1. BIM/4D-BIM adoption barriers in construction projects.

Reference	BIM/4D-BIM Barriers
	Awareness	Experience	Time	Cost	Culture	Contract	Training	Standards	Demand	Risk	Management
[15]		√	√	√			√			√
[14]	√	√	√	√	√	√	√	√	√		√
[17]		√		√	√			√	√	√	√
[18]				√	√	√		√	√	√
[19]	√	√	√	√			√
[20]	√	√	√	√		√			√		√
[21]	√	√									√
[22]				√	√		√			√	√
[23]	√	√		√	√	√	√	√	√	√	√
[24]		√		√			√			√	√
[25]	√	√		√	√		√	√		√	√
[26]		√		√			√	√			√
[27]	√		√		√				√		√
[28]	√	√			√					√
[29]		√		√	√		√	√
[30]		√	√	√	√			√			√
[31]	√	√		√				√			√
[32]	√	√	√	√	√		√	√	√	√	√
[33]	√	√		√	√		√		√	√	√
[34]	√	√	√	√	√		√	√	√		√
[35]	√	√	√	√	√	√	√	√	√	√	√
[36]				√	√		√	√		√
[37]	√	√	√	√	√		√	√	√		√
[38]			√	√	√	√	√	√	√	√	√
[39]	√			√	√	√		√		√	√
[40]	√		√	√		√	√	√	√	√	√
[41]	√	√	√	√	√			√	√	√	√
[42]			√	√	√				√	√	√
[43]	√	√	√	√	√	√		√	√	√	√
[44]	√	√		√			√		√		√
[45]				√	√	√	√				√
[46]	√	√		√	√	√			√	√	√
[47]	√	√	√	√	√	√		√	√	√	√
[48]		√		√	√			√	√	√	√
[49]	√	√			√		√		√		√

√ indicates the identification of the barrier in the corresponding reference.

.

3. Methodology

A combined method was used in the research methodology, consisting of six stages, as shown in Figure 1. The first stage is related to identifying the barriers affecting the adoption of the use of BIM and 4D-BIM in projects, by selecting the correct database to collect sufficient information about BIM, traditional scheduling problems, barriers to the use of 3D and 4D-BIM in construction projects and how 4D-BIM can enhance the scheduling process. Then Delphi technology was used with BIM specialists to review barriers to the adoption of BIM and 4D-BIM in the construction industry, identify new barriers if any, and examine relationships between barriers. The first stage yielded a list of 23 obstacles in six categories.

In the second stage, a two-section questionnaire was designed and sent online to evaluate the probability and impact of each barrier. The questionnaire data were collected from different perspectives of clients, consultants, and contractors related to the offshore construction industry in the third stage. The collected data were analyzed in the fourth stage using fuzzy set theory (FST) to turn the linguistic variables (high, moderate, and low) into crisp values that can be used in the analysis. According to the list and categories, the fifth stage started with an SEM creation using AMOS 23 software to represent the barriers and their relationships, then the model modifications and goodness-of-fit (GOF) standards were created. The FST output results were entered first into IBM SPSS software and then exported to the proposed model on AMOS.

To validate the created SEM, in the sixth stage a case study of an offshore project was conducted. The model was accepted according to the results that were tested using AMOS fit measurements. After SEM validation, the latent and observed dimensions were calculated and then ranked.

4. Barriers’ Identification

The first stage started with the selection of databases such as Scopus, Science Direct, and Google Scholar that are commonly used to create adequate literature reviews. Through careful previous studies with 30 publications in Scopus, 30 in Science Direct, and 70 in Google Scholar, 35 BIM barriers were initially collected. These barriers were listed, and then the Delphi technique was adopted by seven experts with at least 15 years of experience in BIM utilization to identify the related main barriers to 4D-BIM adoption in construction projects. The expert’s information is represented in

Table 2. Experts’ characteristics.

Expert Number	Job Occupation
Expert 1	Project Manager of an offshore project in a consultant engineering company
Expert 2	Senior Infrastructure Engineer
Expert 3	Senior Offshore Engineer with MSc Degree in offshore structures
Expert 4	Project Coordinator Engineer
Expert 5	Senior Planning Engineer with MSc Degree in project management
Expert 6	Project Manager of Infrastructure projects with MSc in BIM development
Expert 7	Senior Quality Control Engineer

.

For the first round, the initial 35 barriers were reviewed by the experts to see if there were new barriers that needed to be added or barriers that needed to be modified or excluded from the list. The results showed that 14 barriers were removed and 2 new barriers were added. Then for the second round, the experts were asked to develop proper relationships between the 14 barriers by categorizing them into groups. Responses were collected, summarized, and re-sent to the experts for further comments and review. A consensus among the experts was achieved in the third round. Finally, a list of 23 barriers was reached, including two newly identified barriers (observed dimensions). The barriers were categorized into six groups (six latent dimensions) as follows: awareness (A), cost (C), demand (D), management (M), time (T), and risk (R). The barriers with their categories are shown in

Table 3. The identified barriers.

ID	Barrier	Category	BIM Dimension
A1	Lack of academic support	Awareness	3D, 4D
A2	Lack of experienced users familiar with 4D-BIM	Awareness	4D
A3	Lack of awareness and long-term benefits	Awareness	4D
A4	Lack of information in the 3D model that disables the development of the 4D model	Awareness	4D
D1	Insufficient government lead or instructions towards implementing 4D-BIM	Demand	4D
D2	The client does not require the use of 4D-BIM	Demand	4D
D3	Lack of marketing for adopting 4D-BIM technology	Demand	4D
D4	The belief that existing techniques and software are adequate	Demand	4D
D5	The strong resistance to change and industry culture	Demand	4D
D6	No clear methodology from companies’ management	Demand	3D
M1	Contract types or project delivery methods	Management	3D
M2	Organizational structure does not support workflow change	Management	3D
M3	Lack of standards and guidelines	Management	4D
T1	A longer process to create the 4D-BIM model with detailed data	Time	4D
T2	Time-consuming to apply the ongoing projects, with a negative impact on current productivity	Time	4D
T3	Long time to learn and understand/learning curve	Time	4D
C1	High cost of training	Cost	4D
C2	High cost of implementation in projects and companies	Cost	3D
C3	High cost of software and updates	Cost	3D, 4D
C4	High cost of the proper hardware upgrade	Cost	3D, 4D
R1	Uncertainty over design liability	Risk	3D, 4D
R2	Uncertainty about the return on investment (ROI)	Risk	4D
R3	Losing data between software/exchanging data problems between software	Risk	4D

.

After classifying the barriers into categories, a two-section questionnaire was created and sent online. The first section focused on the individual respondents’ data, including their occupations, and years of experience. The second section evaluated the likelihood and potential effects of each barrier to BIM adoption in offshore construction projects through a three-point Likert scale structured questionnaire. The second section of the questionnaire included two parts, the first part to evaluate the probability and the second to evaluate the impact of each barrier (high, moderate, and low).

5. Data Collection

The data collection stage started with obtaining the sample size of the questionnaire using Equation (1).

SS = z² × P × (1 − P)/c²

(1)

where ‘SS’ stands for sample size; ‘Z’ for the degree of confidence, which is 1.96 (95%); ‘P’ for picking percentage, which is 0.33 as there are three choices (high, moderate, and low); and ‘C’ for the margin of error. The questionnaire was sent online to engineers representing various parties including owners, consultants, and contractors, and 200 valid responses were received. The a minimum sample size of 150 is recommended for a simple structural path model [50]; hence, the margin of error was 6.5%. The respondents were divided up according to their experiences and field of profession, as shown in

Table 4. Distribution of the respondents.

Profession	No.	%	Years of Experience	No.	%	Affiliation	No.	%
Offshore Engineers	70	35	3–5	26	13	Contractor	98	49
Quality Engineers	34	17	5–10	32	16	Sub-contractor	40	20
Planners	26	13	10–20	77	39	Consultant	33	17
Infrastructure Engineers	22	11	>20	65	32	Client	29	14
Project Managers	12	6
Civil Engineers	36	18

.

6. Fuzzy Set Theory

Fuzzy sets have been extensively used in many disciplines to address a range of fuzziness-related practical issues. Fuzzy set theory has the benefit of combining qualitative and quantitative analyses into a single algorithm. It provides scientists a highly potent new mathematical tool that is split 50/50 between verbal, conceptual, and analytical thinking. To increase measurement accuracy, the fuzzy Likert scale (FLS) specifically includes the ideas of partial membership and de-cluttering from fuzzy set theory in its design.

Previous studies have demonstrated that the fuzzy scale is more accurate than the conventional scale. Low, moderate, and high are the language phrases used in questionnaire replies to describe the data. These language concepts are converted into fuzzy values using the FLS so that fuzzy interference rules can be applied to them. After that, these fuzzy values are defuzzied to produce concrete values which are more precise and can be used in analysis. The adopted methodology was similar to the work in [51], where the trapezoidal distribution (a, b, c, d) can describe a fuzzy number (M) of the universe of discourse (U) as shown in Figure 2. The trapezoidal value is defuzzied into (e). The crisp value (e) can be estimated using Equation (2) [52].

(e − b) * (1) + ½ (b − a) * (1) = (c − a) * (1) + ½ (d − c) * (1)

(e − b) − (c − e) = ½ (d − c) − ½ (b − a)

2e = (d − c − b + a)/2 + (2b + 2c)/2

2e = (a + b + c + b)/2

e = (a + b + c + b)/4

(2)

The FLS model has two main inputs which are the probability of the barrier and its impact, collected from the questionnaire data with 200 responses, and then the Mamdani rules.

Table 5. Probability and impact matrix.

		Impact
		Low	Moderate	High
Probability	Low	Very Low	Low	Moderate
	Moderate	Low	Moderate	High
	High	Moderate	High	Very High

shows the Mamdani rules which connect these input variables with the output values based on the fuzzy state description of the linguistic variables.

Many forms of membership functions can be used to associate the points in a fuzzy set with a concrete number between the intervals of 0 and 1 [53]. The trapezoidal and triangular form functions were utilized here, since it is one of the simplest functions that is frequently used to de-fuzzy input dimensions and convert them to fuzzy values. The membership functions for the linguistic variables were illustrated in Figure 3.

Defuzzification can have many strategies to define the crisp value [54].

Table 6. Linguistic terms with the output corresponding defuzzied value.

Linguistic Term	Very Low	Low	Medium	High	Very High
Fuzzy numbers	(0, 0, 0, 0.3)	(0, 0.3, 0.3, 0.5)	(0.2, 0.5, 0.5, 0.8)	(0.5, 0.7, 0.7, 1)	(0.7, 1, 1, 1)
Crisp value	0.075	0.275	0.5	0.725	0.925

shows the output crisp values calculated using Equation (2), that are used in the validation and development of the multi-dimensional relationships 4D-BIM barriers’ model [51].

7. Structural Equation Modeling

To evaluate the overall barriers facing the application of 4D-BIM in offshore construction projects and investigate the rank of these barriers considering the latent variables, a model based on the barrier categories was developed using structural equation modeling (SEM). The SEM is a broad statistical modeling method that is commonly employed in the behavioral sciences, in academics, and in industries where professionals preach theories [55]. It is a combination between path analysis and factor analysis in four logical steps: identifying the model; estimating model parameters; measuring model fit by evaluating the suitability of the proposed model with the entered data; and developing the model [56]. It provides a flexible framework for constructing and analyzing complicated interactions across various variables, allowing researchers to use empirical models to evaluate the validity of a theory [57]. The proposed SEM was constructed using AMOS 23 software, which produces more accurate models than traditional multivariate statistical approaches and is simple to use because of its graphical user interface [58]. Even without the accomplishment of standard indicators, the first structural equation model based on theoretical forecasts and earlier empirical data was initially acceptable. The final model had to meet the recommended goodness-of-fit (GOF) standards. The model specification stage is first, where ellipses are used to represent the latent dimensions and rectangles to define the observable (measured) dimensions in the graphical path diagram [59]. As seen in Figure 4, single-headed arrows indicate that the variable at the tail of the arrow affects the variable at the point. Since it is expected that the spatial component would not completely predict the model’s observable dimensions, all of the observed dimensions have circles on them to reflect the error in measurements [57].

The GOF includes a chi-square “X²” that can be approximated using Equation (3), in which “N” represents the sample size. Consequently, the chi-square test result is dependent on the sample size, which makes it significant in large samples; hence, it should be rejected. The normed chi-square “NX²” presented in Equation (4) is an alternative that can be regarded as a good indicator because the sample size is not taken into account, and “DOF” stands for the degree of freedom.

X² = (N − 1) × DOF_ML

(3)

NX² = X²/DOF

(4)

The suggested normed chi-square value should be between one and two but not greater than three [60]. The root mean square error of approximation (RMSEA), presented in Equation (5), indicates the suitability of a model that has been ideally determined by unknown variables, with the population of a variance matrix.

$$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{E}\mathrm{A}=\sqrt{\frac{\mathrm{Max}\left(0,{{\mathrm{X}}^2}_{\mathrm{M}\mathrm{L}}-{\mathrm{D}\mathrm{O}\mathrm{F}}_{\mathrm{M}\mathrm{L}}\right)}{{\mathrm{D}\mathrm{O}\mathrm{F}}_{\mathrm{M}\mathrm{L}}\times \left(\mathrm{N}-1\right)}}$$

(5)

The ‘$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{E}\mathrm{A}$’ value is recommended to be between 0.05 and 0.08, which indicates that the model is a good fit [61]. The normed fit index (NFI) evaluates the model by comparing the default model’s chi-square to the null model’s chi-square, which reflects the worst-case scenario in which all observed factors are unrelated. The NFI has a value between zero and one. The comparative fit index (CFI) is the second generation of the NFI in which the size of the sample is taken into account. The CFI is calculated using Equation (6); its values vary from zero to one, and the closer the results are to one, the better the model. The CFI is thought to perform better than the NFI when the sample size is small [56,57]. The ‘${\mathrm{X}}_{\mathrm{M}}^2$’ represents chi-square of the model, ‘${\mathrm{D}\mathrm{O}\mathrm{F}}_{\mathrm{M}}$’ the model of the degree of freedom, ‘${\mathrm{X}}_{\mathrm{B}}^2$’ the null model chi-square, and ‘${\mathrm{D}\mathrm{O}\mathrm{F}}_{\mathrm{B}}$’ the null model degree of freedom.

$$\mathrm{C}\mathrm{F}\mathrm{I}=\frac{\mathrm{M}\mathrm{a}\mathrm{x}(0,{\mathrm{X}}_{\mathrm{M}}^2-{\mathrm{D}\mathrm{O}\mathrm{F}}_{\mathrm{M}})}{\max \left(0,{\mathrm{X}}_{\mathrm{B}}^2-{\mathrm{D}\mathrm{O}\mathrm{F}}_{\mathrm{B}}\right)}$$

(6)

The goodness-of-fit index (GFI), expected cross-validation index (ECVI), and modified expected cross-validation index (MECVI) were also picked to evaluate the model’s validity [62]. The GFI represents the variance proportion for the estimated population covariance that was picked to evaluate the model’s validity [63]; while the ECVI and MECVI express the single-sample indicator that shows whether the model produced from one sample is likely to match another sample from the same population of the same size. The ECVI calculates the difference between the covariance matrix in the studied sample and the predicted covariance matrix from another sample [58].

The estimating procedure began with evaluating the model’s fit, performance, and general suitability with the entered data. The SEM employs sophisticated algorithms to maximize model fit while considering all model constraints and a variety of estimation techniques. To ensure the convergent validity of the model, the first-order factor loadings for each observable variable from the SEM model and the second-order standardized factor loadings calculated from AMOS for each latent dimension were examined to see if they were significant. From the first estimation, any measured variable with factor loading less than 0.30 from the 23 barriers was excluded from the analysis to strengthen the proposed model.

The model validation was tested using AMOS software Version (24). The model parameters and overall model goodness assessment were accepted, according to the resultant output data from the AMOS model and the recommended values as represented in

Table 7. Proposed SEM-AMOS model fit assessment.

Fit Index	Amos Model Fit Output	Recommended Values
RMSEA	0.075	≤0.08 [61]
X2/DOF	2.197	≤3.0 [60]
GFI	0.925	≥0.9 [63]
ECVI	0.69	Lower limit = 0.573 and Upper limit = 0.840 [64]
MECVI	0.70	Lower limit = 0.573 and Upper limit = 0.840 [64]

. The CFI value was 0.82, indicating that the model is 82% better than the baseline model, which assumes no relationships between the variables. Moreover, the Tucker–Lewis (TLS) score was 0.76, and the incremental fit index (IFI) score was 0.83, demonstrating that the second-order confirmatory factor analysis (CFA) model is valid and can be used in offshore construction projects.

8. Results

To evaluate the effect of the latent dimensions presented in the SEM model, confirmatory factor analysis was performed using similar methods to those used by [51] and [65]. The relative weights are calculated using Equation (7): where “WLi” represents the relative weight of the latent dimension; “SPCL_i“ represents the standardized path coefficients of the latent dimensions; and “∑SPCL_i“ is the summation of the latent dimensions’ standardized path coefficients. The latent dimensions relative weights’ results are presented in

Table 8. Latent dimensions’ relative weights.

Latent Dimension	Standardized Path Coefficient	Latent Dimension Relative Weight
Awareness	1.05	0.325
Risk	0.90	0.279
Management	0.82	0.254
Demand	0.46	0.142

.

WLi = SPCL_i/∑SPCL_i

(7)

Also, the observable dimensions’ relative weights (WOj), are calculated using the ratio between the factor loading of each observable variable (FLj) and the summation of the factor loadings of observable dimensions (∑FLj), in the corresponding latent dimension, as in Equation (8). The values of the observable variable factor loadings and relative weights are presented in

Table 9. Observable dimensions’ relative weights.

ID	Observable Dimensions	Factor Loading	WOj
A1	Lack of academic support	0.34	0.265
A2	Lack of experienced users familiar with 4D-BIM software in the workforce	0.57	0.445
A3	Lack of awareness and long-term benefits	0.37	0.289
M1	Contract types or project delivery methods	0.70	0.391
M2	Organizational structure does not support workflow change by adopting such modern technology	0.58	0.324
M3	Lack of standards and guidelines	0.51	0.285
D1	Insufficient Government lead or instructions toward implementing 4D-BIM	0.51	0.327
D2	The client does not request the use of 4D-BIM	0.56	0.359
D3	Lack of marketing for adopting 4D-BIM technology	0.49	0.314
R1	Uncertainty over design liability	0.35	0.438
R2	Uncertainty about the return on investment (ROI)	0.45	0.563

.

WLi = FLj/∑FLj

(8)

The relative effect (RE) of each observable variable is first calculated using Equation (9) and then ranked as shown in

Table 10. Barriers’ effect ranking on the adoption of 4D-BIM scheduling.

ID	Observable Dimensions (Adoption Barriers)	Latent Dimension	Relative Effect	Rank
R2	Uncertainty about the return on investment (ROI)	Risk	0.157	1
A2	Lack of experienced users familiar with 4D-BIM in the workforce	Awareness	0.145	2
R1	Uncertainty over design liability	Risk	0.122	3
M1	Contract types or project delivery methods	Management	0.099	4
A3	Lack of awareness and long-term benefits	Awareness	0.094	5
A1	Lack of academic support	Awareness	0.086	6
M2	Organizational structure does not support workflow change by adopting such modern technology	Management	0.082	7
M3	Lack of standards and guidelines	Management	0.072	8
D2	The client does not request the use of 4D-BIM	Demand	0.051	9
D1	Insufficient Government lead or instructions towards implementing 4D-BIM	Demand	0.046	10
D3	Lack of marketing for adopting 4D-BIM technology	Demand	0.045	11

.

REj = WLi × WOj

(9)

9. Sensitivity Analysis

A sensitivity analysis was used to determine the impact of the input variable on the model output. It offers a rating of the significance of input variables with an objective to minimize uncertainty [66]. This is accomplished by examining the effect of small modifications in the independent inputs on the output variables. Highly sensitive variables have a greater impact on the final results. Determining the most critical variables improves the proposed model’s performance. The analysis was performed using IBM SPSS V23, after selecting the highest ranked effective variable which had ROI as a target. The analysis output is illustrated in Figure 5.

10. Discussion

The study proposed a model that can be used to estimate the overall BIM and 4D-BIM barriers in offshore projects in Egypt. This model was created using a hybrid methodology that starts with a systematic literature review to collect the barriers facing BIM in general and especially 4D-BIM, then the Delphi approach to filter the collected data with BIM experts until reaching the first barriers list including 35 barriers. Then a questionnaire survey was created and sent online. The responses were collected and analyzed using FST to change the linguistic variables into crisp values, and finally the data were sent to the proposed SEM. After the estimation process using AMOS, the final model reached four latent dimensions with 11 observable variables (barriers). The goodness-of-fit of the SEM was tested using AMOS fit tables, with a normed chi-squared value of 2.197, a root mean square error value of 0.004, and a comparative fit index of 0.82. As a result, the developed model fulfilled all the recommendation indices. Furthermore, the root mean square error of approximation was 7.5%, which is acceptable.

According to the model’s calculated estimates, awareness came first among latent dimensions as the highest effective barrier facing 4D-BIM scheduling in offshore projects, followed by demand, risk, and finally management. While the uncertainty about return on investment (ROI) ranks first among the observable dimensions of 4D-BIM scheduling adoption barriers, most companies in the construction industry prefer to remain working with the traditional methods. The uncertainty about the ROI barrier was found and featured in prior research in several different countries [32]. The lack of experienced users familiar with 4D-BIM in the workforce ranked second among the adoption barriers [14], followed by uncertainty over design liability [31]. As it is considered a little-used project scheduling method, the unstable nature of projects makes planners hesitant to take risks in using and learning a new way of scheduling. Moreover, the absence of contracts that define the framework for BIM usage with its different dimensions poses a risk and a barrier to its adoption in projects. The lack of awareness of the value of BIM, as Kassem and Dawood mentioned in their research [14], may be the main reason for believing that it is not helpful. Therefore, it is not considered as a valuable subject for teaching and discussion in academic settings (educational barriers), just as in the German AEC market [12]. The lack of knowledge of BIM and its long-term benefits affects the surrounding environment, which makes some institutions and companies view it as changing the workflow and its usual pattern in projects. The integration of UAS with 4D-BIM can help in overcoming some barriers, such as the lack of information in the 3D model that disables the development of the 4D model (A4), as it offers highly precise point cloud scanning and substitutes manpower for project inspections [67], can ease the data collection process in projects with accessibility challenges, such as bridges and offshore projects [68]. It also helps in producing the progress update report that expresses the differences between the actual state and the scheduled target of the project, so the project managers can take the proper actions to increase the productivity and efficiency of the project [11].

11. Case Study

The case study was conducted on a maritime project involving the establishment of a multi-purpose terminal station, Tahya Masr Terminal (TMT), on the Mediterranean Sea in Alexandria port, Egypt. It consists of sea berths with a length of 2530 m, as well as infrastructure networks and buildings for the multi-purpose terminal yard with a total area of 545,179 square meters. The infrastructure includes water pipelines, electricity, sewage, firefighting and storm networks, and backyards. The main project stakeholders are the Group of Multi-Purpose Terminals (EGMT) which represents the client; the Dar Al-Handasah Company, which represents the consultant who designed, and supervises the project; and two contractors, El-Dawliyah for Engineering and Contracting (EDECS), and the alliance of the Offshore Industries and Services Authority with the Al-Gharabli Company for Integrated Engineering Works (GIECO).

The project had several delays because of poor planning and a vague view of the project’s details at the beginning. It was started on the first of March 2020 and was planned to be completed, working, and receiving the first shipment ship on the first of January 2023; however, because of the use of traditional 2D methods in planning and scheduling, the actual schedule for the completion date has been postponed to July 2023.

This delay occurred due to many factors, one of which was that an important activity was inadvertently dropped from the project schedule. This activity involved installing pipes and sleeves with a diameter of 33.5 cm in the sidewalk body before pouring the concrete, to be connected later to the yard storm network. This missed activity led to a stoppage of construction work for about three months to study how to implement this network in the berths after starting construction and finishing seven joints in the southern berths and eight joints in the northern berths. Moreover, delays occurred due to the use of conventional approaches and 2D drawings in infrastructure planning and solving the infrastructure network conflicts, which took a long time. The planning and preparations for infrastructure works’ execution started in August 2021, and the first partial plans for composite drawings were released in April 2022. The final plans for resolved clashes and engineering works were completed in November 2022, implying that without the use of BIM models and the lack of composite drawings, the study and coordination of infrastructure works took approximately a year and three months.

According to the meetings with the project’s clients and contractors, 4D-BIM was not used in scheduling the TMT project for several reasons that challenged the planning teams from both clients’ and contractors’ perspectives. The client believed that the return on investment in the use of BIM and 4D-BIM would not be beneficial and would lead to a high cost without a sufficient contribution to commitment and control of the project’s schedule, which would be a risk in such a fast-track project. In addition, there was a lack of a framework and standards that define how to use BIM and how it is applied in delay claims. The contractor thought that since the client did not specify a particular technique for project planning or requested using BIM for planning and scheduling, there was no need to adopt it. Furthermore, both the workforce and the organization were accustomed to traditional scheduling techniques. Among the general challenges it was seen that most of the senior engineers involved in the project did not have the essential expertise and understanding of 4D-BIM because it had not been used in their prior projects. Moreover, the project suffered various design and general layout modifications, which would have required a significant amount of time to continuously adjust in the BIM model. The case study demonstrates the similarity between the project’s challenges and the model’s output results; with the model including these project participants’ viewpoints, and more identified barriers.

12. Conclusions

This paper proposed an SEM multidimensional model to investigate the 4D-BIM scheduling in offshore projects, and the reasons why the industry is at a standstill at BIM level 1 (3D-modelling) by identifying the relationships between the barriers for its application in projects, and by estimating its weights through SEM. The model was validated using the FLS and the SEM in a case study of an offshore project in Egypt. First, from the extensive literature review, expert interviews, and the Delphi technique, a total of six groups (latent dimensions) and 23 barriers (observable dimensions) were collected. A two-section questionnaire, including a three-point Likert scale was created. After collecting the responses of 200 participants the data were defuzzied into crisp numbers using FLS and then entered into IBM SPSS software. Then a model was proposed using AMOS software illustrating the relationships between the 23 observable dimensions and six latent dimensions. In model estimations, only the significant barriers were chosen after exporting the data to AMOS, reaching 11 barriers with four categories. The AMOS fit tables were used to evaluate the SEM model’s goodness, which showed that the normed chi-squared value was 2.197, the root mean square error value was 0.004, and the comparative fit index was 0.82. As a result, the generated model met all the recommended indexes. Furthermore, the allowed root mean square error of approximation was 7.5%. From the SEM estimations for the input data, the largest influential latent barriers were awareness, then demand, risk, and management. The observed barrier-relative effects were calculated and ranked from the most effective to the least effective barriers. The most influential barriers affecting the adoption of 4D-BIM scheduling in offshore projects were: uncertainty about the return on investment (ROI); lack of experienced users familiar with 4D-BIM in the workforce; uncertainty over design liability; contract types or project delivery methods; lack of awareness of the long-term benefits; lack of academic support; the organizational structure does not support workflow change by adopting such modern technology; lack of standards and guidelines; the client does not request the use of 4D-BIM; insufficient government lead or instructions towards implementing 4D-BIM; and lack of marketing for adopting 4D-BIM technology.

13. Managerial Insights

The proposed SEM model helps to accurately identify the obstacles facing planners and project teams that prevent them from relying on BIM in project planning and scheduling, and make them prefer to use traditional methods in newly designed projects, despite the previously mentioned disadvantages. The reasons why BIM has not been used in planning offshore works before can also be reported. It will then be possible to think about the appropriate solution for each obstacle, by knowing the most influential latent barrier and the appropriate measures that can be taken by project managers, governments, and company stakeholders to increase the popularity of using BIM and move forward to the next BIM level.

14. Limitations of the Research

The conceptual model for evaluating 4D-BIM barriers in offshore construction projects applies to offshore projects in Egypt and other countries. Before utilizing this model in other countries or for other types of construction projects, the latent dimensions must be modified.

The five-point Likert scale is usually preferred for structural equation modeling analysis of 200 respondents.