Parametric Study

Parametric Study

This page will present the paramters that can be changed and how it affects the results

 

Analysis

In order to convey how responsive the results of the interactive tool are, a sensitivity analysis has been performed. The purpose of this section is to:

  • Determine the robustness of the model by changing the values of the parameters used

  • Observe how sensitive the model outputs are in these variations

  • Investigate how the results are affected with different inputs

Firstly, the most significant factors of the O&M model , i.e. Number of turbines and Distance, were given different values, as presented in Table 10. The numbers were carefully chosen as to cover as much range of the different characteristics of wind farms as possible, from ‘near shore’ (20 km) to far offshore (150 km), and from smaller to larger scale. At the same time the step selected for each of those two variables, had to be large enough in order to see the difference between the outputs. For those variables, using the input parameters as indicated in Table 11, the output of the model was as the one presented in Proposed Solution row of Table 10, with detailed explanation of the solutions in Table 12. For instance, the solution of No Offshore Accommodation and Spot-market charter of the CTV will be defined as N-S, while the Jack-up purchase and Spot-market charter of the CTV as JP-S, etc.

Table 10: Potential scenarios representing different Number of turbines and Distance parameters

Number of Turbines 40 100 200
Distance (km) 20 50 100 150 20 50 100 150 20 50 100 150
Proposed Solution N-S N-S N-S MS-S N-S N-S MS-S MS-S N-P JP-S JP-S JP-S

Table 11: Initial values selected for sensitivity analysis

Input Value
Capacity of turbines (MW) 6
Work shift hours/day 12
Speed (knots) 20
Cost of personnel (pounds/hour) 40
Lifetime (years) 25
Speed (km/h) 37
Operators (SHARING) % 100%
Discount rate 2%

Table 12: Abbreviations for potential solutions for O&M model

O&M Vessel Solution Financial Scenario
No Offshore Accommodation (N) Spot-market (S)
Mothership (M) Long-term charter (L)
Jack-up (J) Purchase (P)

Next, several values have been modified, and for each case of Number of turbines and Distance the proposed solution is indicated. The parameters chosen were the ones for which the uncertainty was considered higher (waiting time due to bad weather, logistic times) and the ones that were more ambiguous (cost of vessels). The values selected are presented in Table 13. It is important to remember that when each of these parameters was modified, the rest of them remained unchanged, i.e. the values stayed the same as in Table 11.

Table 13: Parameters for sensitivity analysis and their variations

Logistic Times (h)
50% (Less than initial values)
300% (Greater than initial values)
Weather delay Times (h)
CTV & Jack-up = 50 (Initial Value = 70)
Mothership = 50 (Initial Value = 30)
Lifetime period (yr)
5
10
15
20
25
30
Turbine Capacity (MW)
3
8
Vessel Costs (£)
CTV
70% of initial value
130% of initial value
Mothership
70% of initial value
150% of initial value
Jack-up
70% of initial value
200% of initial value
Preventive maintenance days limit per year(days)
90
120
150
180
Failure rates( annual failures per turbine)
Simple Danish Concept
Direct Drive

The range that was chosen (meaning the lower and higher values compared to the ones that have been calculated, taken from literature or assumed for the implementation of the model) was that so the changes in the output to be apparent, but at the same time the value to remain in a sensible area. The outputs of this part of the sensitivity analysis have a form similar as the one shown in Table 5 .

Table 5 : Sensitivity analysis and results formation.

Number of Turbines 40
Distance (km) 20 ...
Proposed Solution N-S ...
Logistic Times (h) N-S ...
50% (Less than initial values) N-S
300% (Greater than initial values) N-S
Weather delay Times (h) N-S
CTV & Jack-up = 50 ( Initial Value = 70) N-S
Mothership = 50 ( Initial Value = 30) N-S
Lifetime period (yr) N-S
5 N-S
10 N-S
15 N-S
20 N-S
25 N-S
30 N-S
Turbine Capacity (MW) N-S
3 N-S
8 N-S
Vessel Costs (£) N-S
CTV N-S
70% of initial value N-S
130% of initial value N-S
Mothership N-S
70% of initial value N-S
150% of initial value N-S
Jack-up N-S
70% of initial value N-S
200% of initial value N-S
Preventive maintenance days limit per year(days) N-S
90 N-S
120 N-S
150 N-S
180 N-S
Failure rates( annual failures per turbine) N-S
Simple Danish Concept N-S
Direct Drive N-S

The second part of the analysis examines the possible outcomes of sharing the O&M vessels for different scales and distances of wind farms.  Since the main disadvantage for both solutions is the high cost, there is a possibility of reducing it by sharing the vessel between wind farms and operators.  In this analysis the input parameters have again the values as presented in Table 11, and the representation of the results is shown in Table 15

Table 15: Parametric investigation for sharing O&M vessels among operators.

Number of Turbines 50 ... 600
Distance (km) 20 50 100 150 20 50 100 150 20 50 100 150
Sharing percentage of operator (%) Solution
25 N-P JP-P JP-P JP-P ... JP-P JP-P JP-P JP-P
50 N-P N-P N-P MS-S JP-P JP-P JP-P JP-P
75 N-S N-P N-P MS-S JP-P JP-P JP-P JP-P

The detailed calculations with all different scenarios for both the analyses was implemented into the Excel tool. By clicking one of interactive command buttons, new analysis is run for every time any of the inputs changes. That has been included and applied using Visual Basic code.

 

Discussion of Results

Logistics Time

 As before, what is meant by logistic times is the time spent to organise the crew, procure the materials or equipment, and charter the appropriate vessel. Here, it has different values in each of the categories of failures, the highest one (500h) being when a jack-up vessel is necessary to be deployed to perform the corrective tasks. Considering that the statistical data used for our analysis for these times are not always representing the uncertainty of real-life scenarios, the values of logistic times were given between 50% and 300% of its initial value.

When using lower values than the initial ones, no significant changes in the results were noticed.  The only time that the results were different from the base case solutions (Table 10) was when a case with 200 turbines & 20km were considered, and where No Offshore Accommodation solution is proposed instead of Mother-ship. Due to the fact that the logistic times becomes lower, its negative impact on the cost stops to be so important since the distance is small.

On the other hand, a considerably higher value has been selected (300%). This did not seem to cause the offshore accommodation to become more viable option, even with a closer distance or smaller number of turbines. That means that this parameter, which actually increases the downtimes, does not have such a big impact on the results.

 

Weather Delay Time

Weather conditions will play an important part when planning repair tasks or travelling from the shore to the wind farm. Predicting weather conditions require needs sophisticated methodology in order to capture its uncertainty (due to wave and wind changeable states). Therefore a time series analysis needs to be considered. However, in this project results from different literature sources have been considered to see if they are these ones that make the one solution more beneficial to the others. During our sensitivity analysis the weather delay parameter was changed to observe which vessel would be most beneficial for different weather conditions. The results showed that even though the mother-ship solution has the advantage of lower weather delay time, the result is not affected at all by changing that variable.

 

Lifetime Period

Different values of lifetime periods of the investment have been investigated in order to observe how this would affect the solution chosen. The short lifetime of 5 years examines the scenario of deploying the O&M solution at an already existing operating wind farm. The results show that a longer lifetime make the high-cost solution of Jack-up more beneficial, whereas in shorter lifetime period it is not possible to pay off the cost of a Jack-up vessel. Hence in the shorter lifetime period, No Offshore Accommodation becomes the cheaper option.  In some cases, the lifetime parameter will also influence which financial option should be chosen for a chosen vessel (Table 12).

 

Turbine Capacity

The turbine capacity defines the energy output of the wind farm. Hence as the capacity of the wind turbines increases, in the event of failure, the wind farm will experience bigger losses. Therefore, the larger cost is caused for the same period of time that turbine is off. Thus, as failure rates increase, it is important to reduce the repair time and hence minimize the downtime. When comparing a 3 MW turbine with a 6 MW one, the results seem to be the same, except from the case of [100 turbines, 100 km] where the O&M is not beneficial, as in base case. However, if 8 MW turbines are installed, there are cases that the O&M becomes viable, for instance for 200 turbines 20 km from shore.

 

Vessel Cost

Vessel cost is considered to be the most defining factor of the modelling approach and it would affect the results significantly. The vessel costs used in our model have been based on the information found in the literature. In some cases, if a cheaper mother-ship (less than £ 10,000,000)  is considered, this solution becomes viable even for relatively close to shore wind farms [100 turbines, 50 km].

Even though, it is a very important parameter, the difference in costs between the 3 solutions is quite large. Hence even if a solution is cheaper than assumed, it is a bit difficult to still overcome the necessary value to become feasible for nearer locations of the site. When considering higher prices (than the ones in the base case scenario), although a jack-up vessel could be as expensive as £200,000 per day. However it is still beneficial in large distances and with large number of turbines, due to the bigger savings obtained over the expenditures.

 

Preventative Maintenance Days Limit

As previously described, the preventive maintenance tasks have to be performed in an allocated time frame. A large number of turbines and distance with a specific number of CTV would cause the days required to be as high as 300 days, making it impractical since the access to the site is not accessible that many days per year. Therefore, larger number of CTV has to be used in order for them to run in parallel and perform the preventive tasks in a specific time frame. Knowing that these tasks are usually carried out during the summer, 2 months have been considered as the time limit. This low value means that a larger number of CTV has to be used, increasing the cost of No Offshore Accommodation significantly. For this reason the different values of 3,4,5, and 6 months have been considered to determine the impact in the decision-making process. The results showed that in the case of [100 turbines, 100 km] the offshore accommodation is not the most beneficial solution as calculated before. However, in most examples the results do not differ significantly from the base case scenario. A difference observed is that as the time limit increases, the days required increase as well, because the tasks are performed with less CTV. That sometimes leads to the option of purchasing CTV being more beneficial than chartering it on the Spot-market, which was found in the base case scenario.

 

Failure Rates

The input failure rates used to determine the corrective maintenance tasks have been described in previous sections. In [24] two different sets of failure rates were mentioned: first one corresponds to Simple Danish Concept technology, which is considered to be more mature technology, and second one to  Direct Drive concept. These rates indicate how frequently each component of the wind turbine fails. Those two technologies have been considered due to the differences in the results (compare to base case scenario), especially in the case of direct drive technology that seems to have reduced reliability comparing to the conventional one.

The increase in failure rates would make the need for corrective tasks higher, therefore jack-up would be required more times per year. The downtimes would also raise, hence the cost from losses in energy production would be higher. So with failure rates more pessimistic than the base case scenario, the offshore accommodation solution should be viable in more cases and vice-versa. It was observed that when using the considerably high rates mentioned in [24] for direct-drive technology, the Jack-up solution is shown to be beneficial in all cases. This occurs because the failures in this scenario are almost 4 times higher than the base case, which is a value that is not typically  presented in the literature, but even this extreme case shows the severe impact of reliability in O&M costs.

 

Sharing of Operators

The different scenarios reveal interesting results that show that the potential of the proposed O&M can be further expanded. For instance, if 4 wind farms of 100 turbines each would decide to share a vessel, a jack-up would be the most beneficial solution even in the near distance of 20 km.