By Katie Mazerov, contributing editor
The bounty of rich hydrocarbons that lies deep beneath the ocean floor holds great promise for an industry that is continually searching for new production frontiers. But tapping that prize brings a host of challenges and obstacles, not the least of which lies in an element that can trigger a multimillion dollar failure if it is not handled properly: sand.
“Deepwater wells are like a space shot,” says Luis Rivas, completion engineer, deepwater exploration and production for Chevron North America. “You have to get it right the first time. If you start producing sand, it can cut out, just through erosion, your tree, flowlines, jumpers, chokes and other well components, and cause a well to fail. And when these wells fail, they fail big.”
Nowhere is this more evident than the Gulf of Mexico, with its high-pressure conditions, long intervals, unprecedented well depths and unconsolidated sands. Controlling sand production is a key completion challenge that Mr Rivas and Chevron have faced in the GOM’s Tahiti Field, where wells are producing at 26,000 ft and 20,000-psi bottomhole pressure.
“Whether to engage sand-control measures has been probably the single biggest completion design decision,” he said.
Sand control is currently considered the default solution for deepwater production wells. Making that decision requires balancing the cost of implementing sand-control measures with the overall economic viability of the well. “Completions must be cost-effective, reliable for the long term and optimize the production stream from the well,” Mr Rivas said. The process is difficult because operators don’t have a lot of history in deepwater, so not much reliable well data are available.
Several factors determine whether sand-control measures should be engaged, he explained. Characteristics of the rock are generally the first order of business and involve cutting, then testing, core samples from the well itself.
“Sometimes we’ll do hollow cylinder tests to try and mimic what is happening downhole while producing a well, and see whether the sand is going to crumble, come apart, fill the wellbore and cause a failure of the completion,” he said.
Other factors include the length of time the well will be producing, the degree of drawdown, effects of production on the formation pressure and whether the formation has strong aquifer support such that the pressure won’t decline.
“If I exceed the safe drawdown pressure (SDP) on the well, if I pull the well harder than, for example, 1,500 psi, I will probably need sand control,” Mr Rivas noted. “If I stay below 1,500 psi drawdown, or within the limits of the calculated SDP, I don’t need to install sand control. Or, if reservoir pressure drops below a certain value, which affects SDP, I need sand control at that time. I’m probably not going to pull the well apart and install sand control at that point. I’ll install sand control upfront during the initial completion, which is a much more cost-effective approach.
“To a certain degree, if the pressure in the reservoir rock goes below a critical value from producing the well, or if the aquifer support was not there or was weak and the reservoir becomes depleted, the overburden tends to crush the rock at producing intervals, and the well fails. Then we start producing sand.”
In many cases, deepwater projects present multiple layers of sand that are stacked up three or four high, Mr Rivas continued. “We sand-control each layer, but if one fails, especially one of the lower ones, the cost of remediating that whole completion is even higher because now we have to pull out completions that were in good shape to go down and get the offending one that has come apart, and reinstall it. That is costly and high-risk, and we may lose the wellbore as we are trying to clean everything out. And the fix is generally never as good as the initial completion.”
Further complicating the sand-control dilemma is the notion that sand can be managed rather than controlled, an idea that appeals to some in the industry who believe that managing versus controlling sand can reduce the completion’s complexity and cost.
“Managing sand means we won’t install sand-control measures, and we will operate the well by reducing the drawdown and the choke,” Mr Rivas explained. “We’ll have sand detectors in the tree, at the platform, in the flowline, and we will dial that well’s operating condition in to where we will limit the amount of sand that well produces to a pound per barrel or a pound per five barrels, or whatever the surface facilities can handle.”
That approach has been used in the North Sea, but formations there behave differently from those in the GOM. “The North Sea wells in which sand management has been successful can be operated in such a manner that when they start producing too much sand, the operators can choke them back and reduce the amount of sand produced and still have a competent well that is producing oil, water, gas and some sand,” Mr Rivas explained.
“But in the GOM, if we exceed a certain safe drawdown, the sand doesn’t come apart like snowfall in reverse, lifting off one grain at a time. It tends to fail catastrophically to when the well fails, everything – the tubing, the flowlines, even the separators – are full of sand.
“The problem in the Gulf of Mexico is that no one has really been able to accurately predict how we can manage sand production,” he continued. “We don’t know how to determine how much sand will be produced and the particle size distribution. We are pretty good, as an industry, in forecasting the reservoir pressures and drawdown at which the sand will come apart, but we don’t know the other critical parameters, like the rate of sand production and whether the sand will come apart in a nice controlled manner or all at once.
“All we can do is guess how much sand will be produced if we go below a certain pressure in the well, and that’s the critical question. And in the Gulf of Mexico, nobody wants to be the first to jump off a cliff with a deepwater well that costs $150 million, and investigate sand management by opening up the choke until a well fails.”
The big downside to sand management is that it can restrict the well production rate and thus jeopardize profitability, Mr Rivas said.
“If we’re talking about curtailing production to manage sand, we also have to look at the revenue side of the equation to make that decision,” he said. “If we have 40 wells producing and we are at the limit of our facility, then we can maybe choke back a few wells and defer some production. But if we need the maximum production from each well, which requires the highest drawdown we can safely put on a well, we don’t want to curtail that production over time to save a few dollars upfront while installing the completion.”
Sand-control options range from frac packs to high-rate water or gravel packs and screens. In the GOM, frac packs are generally considered to be the best sand-control completion option. “Frac packs deliver a very reliable and efficient completion in the sense that there is very little pressure drop across the completion, and they have a low skin value, risking little damage to the wellbore,” Mr Rivas said.
“Frac packs generally deliver a skin in the zero to three range, which is very robust. In addition, the mean time to failure for a frac pack is much longer than it is for a high-rate water pack or gravel pack or for stand-alone screens.”
The downside of frac packs is that they are more expensive. “But that gets back to the issue of total cost of ownership,” he said. Because gravel packs cannot be subjected to the same amount of drawdown as safely as frac packs can, the completion won’t be as efficient. “A gravel pack puts sand or proppant in the perforation tunnels, which only extend six to eight inches at most,” Mr Rivas explained. “A frac pack can extend up to 120 feet out, giving us bigger proppant in a fracture that is more conductive, which allows us to draw the well down more and gain a longer mean time to failure and higher reliability in bringing production into the wellbore.”
Gravel packs can be viable in shallow waters or mature areas where reserves are smaller and production rates lower, and don’t justify the cost of bringing in expensive frac equipment. “These are completions that are short-lived, need minimal treatment to protect the facility from sand, and don’t have to optimize every single barrel because the economics don’t justify it,” Mr Rivas emphasized.
“Operational considerations drive economics,” he said. “In gas-producing areas, for example, stand-alone sand screens can be very efficient if it is a case of pure gas flowing from a rock that doesn’t create much pressure drop or friction. We have casings and perforations in a gas interval, and the screens are like downhole sand filters that keep the sand out of the tubing, assuming we have done everything properly.”
Gravel packs and screens also may be selected in parts of world where frac pack vessels are not available. “It is very tough to pump a frac pack, with all the chemical additives and crosslink gels, without a big, built-for-purpose frac vessel that can roll up to the platform and hook up to the well to do the stimulation,” Mr Rivas noted.
“Frac packs are not ideal in formations where there is an oil/water contact in the sand. In that case, with the oil floating on water and the formation more permeable to water than to oil, we would not want to frac that interval because that fracture will likely extend into the water and water out the well when we put it on production.”
As the industry continues to delve into deeper waters, it is continually faced with technology gaps that necessitate ongoing development of new technological alternatives and tools. Among those is a resin treatment, suited for low reserve and shelf areas, that involves cleaning the grains and gluing them so that they bond together. “This method, which is a replacement for a frac or gravel pack, is limited by the thickness of the interval, but it doesn’t require a frac vessel, is low cost and cuts permeability,” Mr Rivas explained.
“The resin must be pumped from top to bottom in the perforated intervals. If drawdown is put on any untreated intervals, those intervals will produce sand.”
Another emerging technology addresses what the industry calls fines mobility, a subset of sand control that occurs when bits of shale and silt, or fines, that form in highly compressed and pressurized sand formations become mobile as the drawdown is put on the well to allow production.
“If we are unfortunate to have a project that has nice permeability and potential for flow, but the fines are very moveable, we are going to have a real challenge,” Mr Rivas said. “These fines come right through the frac packs, gravel packs and screens and will be produced along with the gas or oil. Moveable fines have ruined some good deepwater projects.”
Frac packs can help in this situation because they control the flux, but there are some techniques, in the form of additives, that are being developed to make the fines stick to the proppant or sand, but out away from the wellbore so they don’t impede flow.
The need for new and better screen designs and better ways to predict the effects of erosion have also become apparent.
“A lot of work still needs to be done on downhole sand screens that can withstand the high rates going through them, as well as the differential pressures that they are asked to stand up to,” Mr Rivas said. “We don’t want to burst the screens, and we’re finding out that screen burst is not quite as well understood as we thought it was.”
In terms of erosion, the industry has learned that the old models for predicting the effects of erosion don’t work very well for sand flow. “Where we are now with erosion modeling is like where we were 20 years ago when we started trying to model fractures,” Mr Rivas said. “There were a lot of different ways to do it, but nobody really knew the best way to do it. This will be a lengthy process to investigate, with long lead times and difficult-to-run computational flow dynamics models for erosion. This is not an off-the-shelf application that a typical completion engineer can look at.”
While sand control and sand management has always been an integral part of completion in the GOM and other
producing areas with unconsolidated sands, it, along with other well design components, is being pushed to the limit as the industry is challenged by high costs, low permeability and marginal reserves in deeper waters.
“In order to deliver the plan economically and reliably, we need to understand the cost/benefit relationship of every aspect of well construction, of which sand control is one of the big hitters that has an impact throughout the life of the well,” Mr Rivas emphasized.
“We’ve got to look at the whole picture – drilling, completion and production facilities – when making the sand-control decision or we leave money on the table.”