January 2016 XMET – Four seasons in one day Chris Lovell The problemThe weather is infamously the UK’s favourite subject of conversation, and nowhere more so than in the offshore industries. Metocean conditions on the UKCS can be extreme as well as being ever-variable. Of these two characteristics, it is the impact of variability that is the hardest to accurately quantify on offshore intervention campaigns, whether greenfield development, brownfield modification or decommissioning projects. Engineering design ensures that installations can withstand the extremes that the most isobar-packed weather systems hurl at installations. Meanwhile, methods to make detailed short-term weather forecasts are well developed at the execution stages of projects. However, in Xodus’ experience, there is a dearth of methods and tools employed to evaluate the risk of downtime due to weather in the design and planning stages of project development. Furthermore, in the methods and tools that do exist, deficiencies can be found. These shortfalls and gaps can result in misleading and incorrect information on which to base risk decisions. In today’s cost-constrained climate, the need for continual improvement in quantification of risk at all stages of project development is evident. A solutionTo address the issue, Xodus has developed XMET – a rapid, efficient and cost-effective methodology for evaluating the risk of downtime due to weather that is appropriate for design and planning. The applications of XMET comprise of a range of risk decision processes from option screening/selection in concept design to risk mitigation cost and schedule planning. Below is a summary of its applications: Concept design – Option screening / selectionEngineering design – Cost optimisationTendering – Contracts negotiationPlanning – Contingency determination / risk mitigation XMET at a glanceIn a nutshell, XMET is an operational, time series simulation. Its objective is to replicate as closely as possible what would happen in reality – in a given set of conditions – tested over as many years of metocean data as is available. There are typically at least 50 years’ worth of hindcast metocean data available for locations on UKCS. This sample set enables a probabilistic output. Simulations are notoriously difficult to describe textually, and XMET is no exception, so the graphic below gives a brief overview of the inputs, decision logic and the outputs. The Cumulative Density output from XMET can be used directly to support risk decisions. Alternatively, if a project plan is in any stage of development, the Probability Density can be inputted seamlessly into industry standard Cost and Schedule Risk Analysis packages. XMET is just one of a suite of tools that Xodus’ Risk Decision Management Team uses to support project development and operations optimisation. The team is on hand to discuss any queries you may have. The examples below demonstrate ways in which XMET provides risk decision support. Crucially, they also show the benefits XMET offers when compared to other commonly used methods. Cost/benefit evaluationTwo heavy Lift Vessel (HLV) options were being considered for a partial topsides removal operation: HLV Option 1 required 24 hours with significant wave height (SWH) below 2mHLV Option 2 required 36 hours with SWH below 1.5m plus wave direction exclusion between 20 and 120 WbN of TN. The project team had decided to apply P90 level of contingency for downtime due to weather. The day rate for Option 2 was known to be £230,000 per day but the day rate for Option 1 had yet to be negotiated. The project team wanted to know what day rate for Option 1 would make it competitive with Option 2. The graph shows the output from XMET: It can be seen that the P90 downtimes due to weather for Options 1 and 2 are 10.5 and 15 days respectively. Solving for cost, the day rate for Option 1 must be £330,000 or lower in order to be competitive. The risk of downtime due to weather calculation methods that are not simulations can usually only give deterministic rather than probabilistic results, so percentile values which align to risk-based decisions with the project risk strategy are unavailable. Detailed contracts negotiationXMET can even differentiate between start date options less than a month apart as the following example shows: The operator of a construction support vessel (CSV) has stated that the vessel become available sometime between ‘early March and early April’ in the planned year of a flexible riser pull-in projectThe CSV day rate is non-negotiable during period in questionThe project team is to apply a P90 level of contingency for downtime due to weatherThe project team needs to know impact on cost and schedule of the range of potential start dates. The results show that the risk of downtime due to weather when starting on 5th April is approximately half that of starting on 8th March. This finding would be impossible by any other method. Contingency determinationThis example goes further in showing how XMET differs from other methods, even those that are other types of simulations. A diving campaign was being planned on an asset in the Northern North Sea with a ‘good weather duration’ estimated to be 58 days. The operability limits varied between 1.5m and 3m dependent on the particular operation. A diving support vessel (DSV) had already been contracted provisionally at a day rate of £250k per day. The project is to apply a P90 level of contingency for downtime due to weather but had yet to needed to determine how much contingency to allocate for downtime due to weather. Four methods had been tried before XMET was used. The results of the other four methods (all in grey) plus the XMET result (in red) are shown in the graph below. As can be seen, there is a large discrepancy between the results. To illustrate, if the most optimistic method had been relied upon, the project would have been exposed to approximately £1.75m of risk at P90. On the other hand, if the most pessimistic method had been used, £2.75m of contingency would have been unnecessarily allocated. Incidentally, it is important to note that is not a given that the XMET result would always fall approximately ‘in the middle’ of results from the range of other commonly used methods. Understanding the weather’s variabilitySo, can the discrepancies between the methods be explained? Yes, it is all down to the way in which the variability of the weather is being modelled (if at all). Statistical deficiencies can be identified in each the methods except XMET. Yet the other four methods are all-too-commonly employed in the offshore industries. Surprisingly, the result shown by the most pessimistic result is that obtained using the market-leading cost and schedule risk analysis package. It would be too convoluted here to describe each method mathematically. However, Xodus’ Risk Decision Management Team is happy to explain all the methods in the above example and the discrepancies between them.