A report commissioned from Arup by the Department of Business, Energy and Industrial Strategy (BEIS) in the UK suggests the cost of decommissioning offshore windfarms at the end of their useful lives is unlikely to significantly affect the cost of the energy they produce.
Assuming no scrap value and considering that the costs will be incurred 26 years or more after the start of construction, Arup’s report for BEIS, Cost Estimation and Liabilities in Decommissioning Offshore Wind Installations, found the cost of decommissioning offshore windfarms increases the levelised cost of energy (LCoE) for offshore wind by £0.86/MWh (in 2014 prices), which is less than 1%.
This is in contrast to decommissioning in the offshore oil and gas industry, for instance, where costs are expected to be much higher.
As highlighted here and here the offshore wind industry won’t have to deal with some of the high-cost aspects of decommissioning an oil field – such as plugging and abandoning infrastructure in which hydrocarbons are involved – but the industry needsto focus on decommissioning now if costs aren’t to escalate later.
To support development of a cost model for decommissioning windfarms, Arup consulted with the offshore wind industry, gathering stakeholder views and considering future scenarios that may influence the decommissioning costs.
The cost model was used to estimate a range of decommissioning costs for 37 offshore windfarms at various stages of development (either operating, in construction or pre-construction).
Arup put the total estimated decommissioning cost at £1.28Bn (US$1.69Bn) to £3.64Bn, of which the liability to BEIS is estimated to be of £1.03Bn to £2.94Bn. The Crown Estate and The Scottish Government are potentially liable for the balance.
The range of costs considered factors which may impact cost outturn including potential for change in regulation which may affect whether some infrastructure can be left in situ; uncertainty in the decommissioning methodology; and uncertainty in a number of key cost drivers, such as future vessel charter rates.
Arup noted that as a result of the current early nature of the industry and these uncertainties, the estimated decommissioning cost range is wide. However, as more information becomes available this uncertainty will reduce.
Arup said its model is sensitive to vessel day rates and to certain activity durations. It is also subject to uncertainties in decommissioning methodology and the decommissioning philosophy. These uncertainties result in a multitude of scenarios, each resulting in a different estimate for the total decommissioning cost.
The key inputs in the model which drive the total decommissioning cost are:
- Vessel day rates.
- The duration required to complete various activities.
- The choice of vessel used to complete various tasks.
- The assumptions for overhead costs and wait on weather.
The total decommissioning cost is heavily dependent on vessel day rates and vessel mobilisation cost (which is itself a multiple of day rates).
The duration required to complete activities dictates how long the vessel is required on site and therefore drives the total cost of the vessel activities. Activities completed by the jack-up vessels have the highest cost. Therefore, the time taken to complete these activities – such as removing the generator, cutting and removing the foundation – has a large impact on the overall decommissioning cost.
Within the model vessels were selected for each activity. There are three decision points:
Wind turbine generator removal vessel – a low-, medium- or high-end jack-up vessel is selected based on hub height and nacelle weight.
Wind turbine generator transport vessel – either the same vessel as for WTG removal is selected or a separate cargo barge is chosen. The baseline estimate assumes the same vessel is used for generator removal and transport. This is an assumption in many developers’ decommissioning plans, but in many cases, it is possibly more cost-effective to use a separate cargo barge.
Foundation removal vessel – either the same vessel as for the generator is selected or a separate, additional, jack-up vessel is chosen. The baseline estimate assumes that different vessels are used to allow foundation removal and generator removal in parallel. This is an assumption in many developers’ decommissioning plans.
An overhead cost for engineering, project management and port fees was added as a percentage of the marine operations cost. A cost for waiting on weather was also added in a similar way. The percentage applied for these costs has a direct impact on the total cost.
Arup noted that until more offshore wind decommissioning projects have been executed, and more activity-based estimating of costs becomes possible, it is reasonable to assume that cost estimates could vary somewhat.
The conservative upper estimate recognises the high degree of uncertainty involved in estimating decommissioning cost, given that no large-scale projects have yet been executed and the majority of work is planned for 10-30 years in the future. This uncertainty can be attributed to uncertainty in vessel rates and decommissioning methodologies, particularly the timing of activities and the decommissioning philosophy.
A lower estimate acknowledged there may be some conservatism in the central estimate and there are opportunities for cost reduction. These opportunities include the development of new technologies, techniques and efficiencies over the period of the decommissioning works.
The cost of decommissioning the offshore transmission (OFTO) assets such as the offshore substation and the offshore export cable, was calculated separately for each windfarm. It was calculated based on the use of a heavy lift vessel used to remove the substation topside and foundation; and the export cable being decommissioned and left in situ.
The decommissioning cost model estimate for the total cost of decommissioning OFTO assets for the 37 windfarms is £158M. Only two OFTO decommissioning cost estimates were available from OFTO operators for comparison. Little information as to how these costs were calculated was available given the very small scope of this dataset.