A precise knowledge of the inherent risks of a project leads to better drilling results.

Operators are coming closer to that perfect combination of the lowest cost and the best construction of oil wells - the goal they've tried to reach since the industry began dropping cable-tool bits into holes.
The search has led to more efficient rigs, tools and bits, better muds, directional and horizontal drilling techniques, multilaterals, measurement while drilling and logging while drilling; but in many ways it is the human side of the drilling operation that could use a little help.

One management process is the system called P1, developed by The Peak Group in Aberdeen, United Kingdom.

"P1 software allows a probabilistic approach to be taken by an operator, enabling a step change in well construction efficiency," said Andrew Paterson, managing director.

He said this management and planning technique was a key component in Petro-Canada UK's (formerly Veba UK) position as best-in-class driller in the North Sea, in the semisubmersible development well category, in the latest Rushmore study. The Peak Group acts as Petro-Canada's contract drilling arm in the North Sea.

The process' structured interface allows an operator to create a map or model of the well. This map charts all operational phases required to drill and complete the well, with each phase broken down into the relevant operational activities with associated risks. Each operational activity has risks, and no matter how small, these can be assigned a likelihood (probability of occurrence) and a distribution of effect (how long it would take to deal with that risk, how much it could cost).

For example, if a loss zone is encountered, the time to treat the mud system, the value of the lost mud and the cost of the remedial mud work can be allocated in the model. A Monte Carlo simulation then can be run on the model.

"We can model anything we can envisage happening in the well. Actually, it's not so much a drilling tool as a performance management tool that is focused on aiding people's planning and conducting the well operation. It aids the understanding of risk both within the asset management team and the well operations team. It allows quantification of both risk and opportunities within the project and highlights any opportunities for improvement," he said.

When an operator drills a well, it makes several assumptions about operational activities. These assumptions include rate of penetration, hole stability and a host of other events. However, the level of certainty in these assumptions is highly subjective when that operator is drilling a well with 3 miles (5 km) of departure and is dependent on inferred data gathered from measurement-while-drilling data, mud chemistry and drillstring behavior.
The modeling process allows offset well data to be fed back into the loop of future well planning. The offset and historical well data enables better data distributions to be entered and real probabilities to be used in the model, increasing the ability of the software to model the "real" well. For example, if five out of 10 offset or historical wells encounter borehole stability problems linked to mud chemistry, a new model of a well would include this as a risk. However the opportunity to use a different drilling fluid would be highlighted as a way to minimize this risk, and a further model using a new drilling fluid could be built and used to compare the overall time and cost implication.

"Risk is not common sense, and it's not intuitive," Paterson said, quoting experts at BP. Wells are too complex these days for intuition to perceive everything affecting the operation.

A good risk-management program should accomplish several objectives, including:

* development of more realistic business and project planning;
* increased certainty of achieving business and project objectives;
* reduced cost and time;
* assumption of rational risk;
* improved project understanding;
* improved confidence in the project team;
* promotion of a blame-free culture; and
* improved learning.

In a classic, often repeated case, an asset manager will ask the drilling manager for the cost to drill a development well. The drilling manager quotes a price of US $13.5 million, based on the best experience available. The asset manager shakes his head, tells the drilling manager he promised partners to do the job for $12.5 million and asks the drilling manager for his very best number. The drilling manager massages the figures and comes back with a final total of $11.75 million.

This situation comes from a sequence of subjective interpretations, and every time the figure is reduced, the chance of an overrun increases, because someone has cut corners somewhere. That's just common sense, since there is no reason for the drilling manager to give the asset manager anything but the most appropriate number when first asked.

However, consider another sequence of events. When a company begins planning a well, it should look at the range of possible risks associated with the activities to be performed and assign probabilities to those risks. The company plugs those risks into the model. Once the model is built the team can come up with mitigation plans for all the risks considered in the model. A Monte Carlo simulation then can be run on the model, with the output presented in a Microsoft Excel spreadsheet as a series of distributions between no probability (P0) and 100% probability (P100).

The activity with the largest range between the P10 and P90 probability is the operation that represents the highest risk and is therefore the activity that the design and operations team should most wish to get right. This ability to identify risk ranges is a key attribute in managing performance.
This well planning and modeling process has focused team-engineering skills.

"In one well, we proved the 16-in. hole section had more risk that the 8-in. high-temperature, high-pressure section, thus allowing the engineering team to refocus their attention to where the greatest well risk was," Paterson said.

This example makes a couple of points. First, proper planning can save money and reduce risk. Second, the greatest gains in well-drilling efficiency can be made in the planning stage by preparing for events (should they occur) in the operating stage. It also shows that even the best engineer cannot intuitively and consistently identify risk without the aid of a structured program.
This kind of statistical analysis isn't new. Shell uses it in its Deliver the Limit (DTL) system. Shell's DTL manual states that good project definition and execution delivers the highest value. Reduce the quality of either definition or execution and value falls. Take away both and value plunges.

More simply, Paterson said, "Appropriate technology application can add value to your business, since it is one of the ways in which the opportunity gap between current performance and the technical limit can be closed."

Ideally, the asset team should get together to go over each phase of the project. Often it helps to have a facilitator at the meeting to direct and focus discussion when required. The facilitator can tone down the impact of the person who tends to push through decisions by volume and force of personality and encourage opinions from the shy but knowledgeable technician in the corner.

Some companies that keep extensive records will go to their databases and simply pull out the high, low and mean numbers recorded for an activity, for example, a 13 3/8-in. casing operation. Asset managers then will use that distribution in the model. That's a mistake for three reasons. First, it masks the risk of reaching a new low. Second, it cuts off the likelihood of achieving a new high. And third, if the operating team shoots for the mean figure, it is aspiring to mediocrity.

That's one of the reasons that the best, most competent experts available for this particular kind of well should prepare the drill-the-well-on-paper exercise. One of those experts probably will be the drilling engineer, but other experts might include bit salesmen or toolpushers. Any event envisioned in the drilling sequence deserves the attention of an expert.

Paterson readily admits this approach has weaknesses. For example, the quality and accuracy of the results depend on the available data and the project team's knowledge and skill. However these weaknesses also are present in and have greater detrimental effect in traditional deterministic engineering.

The object of the exercise is to cover the possibilities. For example, if a company gets a tool stuck in the hole, it may not recover the tool and face a $2.5 million bill from the supplier. Or it may recover the tool in a damaged condition and have to pay for repairs, or it may fish the tool with no damage at all. The cost of the risk in using that tool ranges from zero for a job with no problem to $2.5 million to pay for the tool plus an additional amount for time and expertise if the tool gets stuck irretrievably.

An unexpected lost-circulation zone might eat up hundreds or thousands of barrels of mud. That might be a risk worth taking, but a contingency plan should be in place to help the drilling operations team mitigate the loss and the impact on rig time and operational cost as well as any implications further down the well. Those implications might include the need for additional casing strings or the impact upon completion technique and style. If the operator accounts for the possibilities in the planning stage, early signs of trouble lead to the implementation of contingent plans and actions that can minimize overall project impact. These events often are overlooked when deterministic approaches are taken to well and project planning, but they are easy to model.

When a team models a well, it considers the risks involved in drilling the well and the costs entered to generate an authority for expenditure based on the likelihood of each risk and its overall impact on the project. The modeling process outputs a cost and time distribution (between P0 and P100) for each operational phase of the well.

A flow chart lists each activity and any associated risks, providing crews with an easily accessible forward plan should any risks occur.

Ideally, the drilling team should keep the model at the drillsite and enter the actual operational events as they occur. Subsequent simulations of the model with the actual data and planned activities allow the project team to monitor the drilling process and review how closely the well is following cost projections. The team also can define the residual business risk at any point in the operation.

This kind of drill-the-well-on-paper system also allows a company to compare options in a way that goes beyond the spreadsheet "what-if" analysis. By plugging in options, an operator can measure the probability of each option's contribution to risk or opportunity.

For example, "A slimhole well usually will be cheaper if everything goes right, but it could easily be more expensive is something goes wrong," Paterson said. The analysis might show that a slimhole well is better 45% of the time by as much as $2.2 million, but a standard well is better 55% of the time by as much as $2.8 million.

Enterprise Oil (now Shell) performed this type of comparative analysis on its Bijupira-Saleema field offshore Brazil to determine options from the available sand screen technology. It built three models with different products and costs and came up with an objective and compelling recommendation.
That ability also can help drilling managers obtain a cost-risked, objective answer to justify the cost of using a more expensive but more flexible or modern drilling rig.

Another job analyzed weather patterns to determine if a company was better off with three anchor-handling boats or five. The analysis showed five boats would be the most suitable choice, and post-operations analysis showed that if the company had used three boats, it would have run into weather delays that would have greatly increased costs.

A competent program also can show comparative risk and reward between old technology and new technology, what happens if something goes wrong, the probability of something going wrong and the costs associated with each option.

This kind of analysis is not a contest between the wildcatter who loves the risk element and the bean counter who hates risk; it's arithmetic that gives both a way to understand the consequences of the risks they take.

When the time comes to review the well, everything is in plain sight - operations, costs, times, risks avoided, risks encountered and mitigated and mistakes made. The model will highlight best practices and areas that need work. Those factors will be entered in the next well model to assure improved understanding of the operation and deliver enhanced performance.

The first question that comes to mind about using this kind of program concerns need. Benchmarking studies consistently show that companies in the top quartile of drilling consistently drill less expensive, higher producing wells than companies in the bottom quartile, often by large factors, and much of that extra costs stems from the difference in planning quality.

That doesn't mean an operator should use this kind of planning program on every well. If a company is drilling cookie-cutter coalbed methane wells to a well-known formation at a shallow depth in the Powder River Basin of Wyoming, it probably doesn't make economic sense.
This kind of management process can be a big help to companies that drill high-potential wells with large investment costs.



Comments