Quantifying Weather-Related Risk in the Development of Offshore Oil & Gas Fields (Part I)

Arnold Doray

Cost management is a key factor to the successful exploitation of offshore assets. The ideal is to minimize project costs while also lowering the associated risk. Unfortunately, in most realistic scenarios it is challenging to do both; instead, the goal is to strike the right balance between risk and cost.

In this article, we focus on one aspect of cost management, that of quantifying weather related risks for drilling campaigns. Our assumption is that in the exploration of offshore fields, the selected solution for a drilling campaign will be one involving a weather-sensitive, mobile drilling facility, whether a jack up, semi-submersible or drill ship.

The Goal of Risk Analysis

Risk analysis seeks to determine the probability-cost curves ('S-curves') associated with a given location and vessel heading. Thus, given the type of drilling rig, mooring configuration, critical limits and costs for (de)mobilzation and other relevant costs (eg, cost of hire), we wish to calculate the probability that the campaign cost will exceed a given amount. This data is often displayed graphically as a plot of probability against cost, and necessarily takes a sigmoid ('S') shape.

This approach, called Quantitive Risk Analysis (QRA) is certainly not novel and is used in many other industries. Our goal is to customize the QRA approach specifically for the deployment of mobile drilling facilities in adverse-weather locations. Our hope is that the 'domain knowledge' is captured in a well-laid out process, with input that are clearly defined.

Motivation

The rationale for using QRA is to quantify the risk-cost tradeoff as accurately as possible. To understand why this is necessary, we consider two alternatives currently used in the industry:

  1. Heuristics ('rules of thumb'): These are rules that are often adhered to as 'industry standards', often imported from situations with different weather, and from times when costs were different from the present. Heuristics are useful as a reality check, but might often lead to either unnecessarily costly solutions or fail to adequately address weather related risk.
  2. Compiled Statistics: Compiled statistics (eg, from metocean studies) are supposed to help the operator understand local weather conditions and adapt the drilling solution & timing appropriately. This is certainly better than using a heuristic, but unfortunately, the time dimension is lost in the compilation of statistics. For example, let's say an exceedence table tells us the waves exceed the operating threshold 10% of the time over the proposed drilling period. Does this necessarily translate into a 10% chance of downtime? A little thought experiment shows it doesn't. Take two cases: If the 10% exceedence waves were caused by a 100-year storm, then the probability of downtime is only 1% (once in a 100 years). On the other hand, if the waves exceeded the threshold every year, but only 10% cumulatively, then there is a 100% chance of downtime over the proposed drilling period! The point of this illustration is that compiled statistics lose the time dimension, which is critical in assessing risk.

Given these obvious drawbacks why are heuristics and compiled statistics so widely used in the industry? There are a number of reasons:

  • Accurate weather models and cheap computing power have only been available recently.
  • Previously, ship observations made at widely-spaced time/space intervals were the only available data, which are naturally amenable to statistical compilations.
  • Inertia on the part of major meteorological & oceanographic service providers to expand their suite of services outside their traditional domains of expertise.
  • Decision makers in oil companies are unaware of the pitfalls of applying existing techniques to cost management or are unaware of better alternatives.

Assessing Weather-Related Risk

The basic idea is to feed time series weather data into a cost model (outlined below), to get the probability-cost curves for a proposed deployment scenario.

Simple Cost Model of a Drilling Campaign

From the point of view of cost management, a drilling campaign may be in one of five states at any given time, each with an associated cost:

StateMeaning
PreparingThe campaign is just underway. There may be a wait due to bad weather.
MobilizingThe facility is moving to the drilling location. This could be at the start of the campaign or going back to a location after a severe weather event.
WorkingThe facility is working.
Stop WorkingWork has stopped due to adverse weather.
DemobilizingThe facility is moving away from the drilling location. This could be either at the end of the campaign or in response to a weather forecast.
The rules for transiting from one cost state to another are supplied by the operator. The basic decision tool for making transitions is the weather forecast.
Table 1: Cost States for a Drilling Campaign

It is certainly true that this cost model is simplistic. For example, each drilling activity are likely to have different critical thresholds.

While this objection is certainly true, it is not entirely relevant. Our goal is to obtain a better estimate of the true probability-cost profile, not necessarily a perfect one.

The Role of Weather Forecasts

Weather forecasts provide the necessary information for the campaign to move from one state to another. The hindcast weather time series data that is the primary input of our cost model may be used to provide perfect 'forecasts' (3, 4 or 5 day) for this purpose. The forecast length may affect the analysis since operators may have different protocols for demobilization in the case of bad weather.

The Simulation

Our basic approach is now clear:

  1. Fix the rig choice, heading, mooring, and other parameters that are decided upon at the planning stage before the campaign.
  2. Assess the operability for each state in Table 1.
  3. Fix the transition rules from one state to another. This implies fixing the operating limits and operational procedures that have to be adhered to based on weather reports (eg, storms within x miles, squalls, etc).
  4. Simulate the state transitions (and therefore, costs) based on the weather input over a number of years.

In this way, we may obtain the cost of the entire campaign under a variety of different hindcast weather scenarios.

Even more scenarios may be generated by taking into account the approximate error of any one weather datum (wind, wave, currents, sea ice, water temperature) and generating additional, 'surrogate' weather data by adding small errors to these

The Risk Report

The final 'deliverable' of the QRA is a Risk Report. Since this does not address design and other weather-related aspects of the drilling campaign, it complements (and does not supplant) the traditional metocean study. The Risk Report details the:

  1. Rig type and other relevant parameters that are being assessed in the study. We call this the 'deployment configuration'.
  2. Operational procedures used in response to forecast bad weather, (eg, where to seek safe harbor, when to stop work, etc.).
  3. Operating thresholds used,
  4. Cost assumptions,
  5. Details on the weather data sources/models used for the analysis.

The ultimate result of the analysis is one probability-cost curve for each 'deployment configuration', with a discussion to evaluate each configuration from a risk perspective.

Extensions

So far, we have only touched on Drilling Operations, but field development encompasses much more:

  1. Seismic Survey operations
  2. Site Survey of Drilling and other site locations
  3. Drilling Operations
  4. Field Development - Installation & commissioning
  5. Field Development - production operations

In all these stages of field development, weather is an important risk factor. At the moment the risk assessment of these activities are based more on subjective perception than on a holistic probabilistic approach.

The weather factors (such as sea state, wind speed, etc) can be easily provided for the first three operations. The complexity enters the equation in the Field Development stages, but even here the main bottleneck is the weather. Future extensions would include refining the risk models involved to make them more realistic and flexible.

Conclusion

QRA provides a formal means for identifying and managing the project risk from day one, giving the operator an overall assessment of their offshore operations, which can be refined and enhanced as the project progresses. Some specific merits:

  • A holistic overview of what the operator is proposing to undertake in producing their assets.
  • Enables the operator to have an improved selection process when identifying the costs and the schedule impact of each of the five operations.
  • Reduces the complexity of the decision making process, since even complex deployment configurations are represented by just one S-curve.
  • Provides a cross check of the traditional advice function, highlighting any discrepancies and allowing a better risk profile to be adopted.

Obviously, while we have focused on Oil and Gas field development, this approach is also useful to almost any weather-sensitive offshore activity, for example, tow route studies, port development, etc. - the operator gets to see the risk profile immediately, instead of having to guess at it based on compiled statistics.

Acknowledgements

My sincere thanks to David Fellowes (Just Developments) and Mike Stefanov (BP Exploration) for making valuable suggestions during the development of this article. My thanks also to Jon Dunstan (London Marine Consultants) for taking the time to explain some technical aspects of vessel response.

Any errors are of course the sole responsibility of the author.

Subscribe to KnowledgeVault

If you'd like to get the latest updates of articles, enter your email below and click "subscribe". We'll send the latest articles to you!