**5. Application on a Natural Gas Compressor Station Construction Project**

The suggested framework was used in a construction project in Greece of the Kipi compressor station of the Trans Adriatic Pipeline, with all the buildings that are foreseen in the relative area. Initially, an administration building that accommodates offices of engineering and management staff that support the station function is built, and it consists of a guard house, offices of engineering and management staff, conference and break rooms, kitchen, sanitary rooms for men, women, and people with disabilities, HVAC, server, electrical room etc. Next, a stores and workshop building accommodates electrical and mechanical workshops, small parts storage, male and female lockers, showers, sanitary rooms, meeting room and break room with kitchen, electrical and server room, archive, workshop office, mechanical workshop, storage area. Also, a utility building provides space for station vehicles as well as space for heating system, instrument and plant air, firefighting material and equipment storage, and lube oil storage. Finally, an electrical and control building, which is the main building that provides space for electricity supply facilities and the operation and control of compressor station, such as transformers (connection to public medium voltage grid), switch gears (medium voltage, low voltage), uninterruptable power supply with battery room, cathodic protection facilities, station control system, and HVAC rooms. The following Table 3 summarizes the tasks of the project and the corresponding risks of each task.




**Table 3.** *Cont.*

The decision-maker, responsible for making the decisions needed by the multicriteria approaches, was the engineer serving as the risk manager of the project. This technique gives the risk managers a choice mechanism for effectively prioritizing hazards and subsequently leads to efficient allocation restricted budget for expenditures in accident prevention.

The present case study consists of a separate hierarchy for every single task, given that the AHP is applied for the risks of each project activity. For example, the hierarchy for the task "Working at height" consists of three risks (Fall of personnel, Fall of objects, Improper use of portable ladder), which will be assessed using the multicriteria method (Figure 3).

**Figure 3.** An example hierarchy for T11/Working at height task.

The expert risk manager has to apply evaluations and fill pairwise comparison matrices for the risks of every project activity.

Table 3 shows the risks of each task and the expert's choices. The influence of each risk on the overall level of safety in the workplace while carrying out each activity is established through the use of pairwise comparisons. The typical AHP technique generates the consistency ratios (CRs) in order to measure and assure that the judgments made by the decision-maker are consistent with one another. The appropriate local priorities that were computed using the typical AHP are outlined in Table 4 below. It is important to point out that every judgment turned out to be accurate, as evidenced by the fact that the CR for each matrix was less than 10%.

**Table 4.** The risk list, judgments, and results for each activity of the project.



**Table 4.** *Cont.*


**Table 4.** *Cont.*

Applying the AHP method allows the risk manager to extract more accurate information regarding the importance of each risk for every activity. More specifically, in the last two columns of Table 3, the score of each risk and the corresponding ranking are listed. Thus, the manager identifies the most influential risk factor and can allocate budget resources to reduce as much as possible the total project risk.

In addition, it is worth mentioning that three classes of activities are considered regarding the results. Firstly, there are some activities observed that the first ranked risk is more important than the others. In this analysis, we considered that risk is much more important than the others if it has a score of far more significance than 60% (namely, the sum of scores for the rest of the risks is by far smaller than 40%). In this class belong activities such as T2, T3, T4, T5, T6, T10, T11, T12, T15, T17, T18, T20, and T23. Next, we observed that for some tasks, there is a more critical risk (the one ranked as first), but its score is between 50% and 60%, namely, it is marginally responsible for the majority of the task's risk. In this group, we classified activities such as T1, T7, T13, T16, T21, and T22.

Finally, we considered a group of tasks for which there is a predominant risk factor, but the majority of the risk of the activity is not due to it. In this class belong activities such as T8, T9, T14, and T19. Figure 4 presents the weight of the first ranked risk for each task of the project.

**Figure 4.** The weights of the first ranked risk for every task.

These findings provide evidence that supports the hypothesis that was initially proposed for this research: that it is essential to make use of a multicriteria analysis method in order to determine the significance of risk factors for project activities. This is because there are activities for which the risk manager needs to revise the given judgments or allot a budget in order to ensure that other risk factors besides the predominant ones are taken into consideration.
