After many months of harnessing mind-bending ideas, a Houston-based team of engineers, geologists and geophysicists has developed technology to monitor hydraulic fractures from the surface and wellbore simultaneously.
Adding multiple dimensions to the monitoring process, they say, allows operators to more clearly understand drainage patterns of hydrocarbons in shale reservoirs and, more importantly, know with greater certainty where to drill the next well, how to optimize completions and maximize asset value.
Though the two monitoring techniques have been used effectively as individual applications for years, the team at MicroSeismic Inc. recently saw its hard work pay off with their first commercial success in the field using the new combined method.
What’s more, their technology has received formal recognition by the industry.
AAPG member Peter Duncan, co-chairman and founder of the company, accepted the Virgil Kauffman Gold Medal at the Society of Exploration Geophysicists’ conference last October in Denver. Duncan, a former SEG president, has been an AAPG-SEG Distinguished Lecturer.
The award – which recognized the surface monitoring technology Duncan developed that led to his tandem microseismic monitoring achievements – recognizes scientists who have made notable contributions to the advancement of geophysical exploration in the past five years.
Not bad for a geophysicist and his team of four that began working out of a converted apartment leasing office in 2003. They were determined to be the first to monitor hydraulic stimulation from the surface when the rest of the industry was monitoring – and often running into problems – using downhole acquisition.
The plan for monitoring hydraulic fractures from the surface soon followed. Once Duncan mastered that, the notion of going downhole seemed simple enough: Take the company’s proven surface monitoring technology, turn it on its side and insert it into a borehole.
“It seemed to me, and to others, that we should get the same results from the surface and from downhole,” Duncan said. “We knew we would be able to acquire more precise information from downhole, yet see the entire fracture network from the surface.”
However, the signals traveling from the mini, underground earthquakes to the surface and wellbore proved difficult to synchronize. After months of staring at countless seismic wavelets and working to align them, things started to take shape. A detailed image of a complete fracture network came into view.
“There is no single individual who made this great ‘ah ha’ moment possible,” Duncan said. Instead, he quickly credits his team for learning how to successfully synchronize data from five different dimensions: three spatial dimensions, time, and the dimension of the wave fields from the surface and downhole.
But there is still room for improvement. Data from the surface and wellbore are currently being processed separately. Duncan and his chief technology officer, Mike Thornton, who played a large role in developing the technology, are currently working to combine the two.
Evolution of Ideas
Initially out to conquer the field of reservoir monitoring, Duncan took a list of potential services he could provide clients and drew a red “x” through “hydraulic fracture monitoring,” as the market was already covered by the big players monitoring vertical wells, he said.
But as the shale boom began to sweep the industry, the number of horizontal wells surged.
Operators who relied on downhole monitoring of fractures ran into issues as their lateral wells stretched farther and farther from the monitor well. Lacking the additional monitor wells needed to effectively observe the fractures, many approached Duncan for an alternative.
“The original practice of downhole microseismic allowed for drilling a monitor well a few hundred or a thousand feet from the frac well,” Duncan said. “The trouble is, as you get more than a couple thousand feet away from the monitor well, you lose the ability to detect the arrival time of waves to the millisecond,” thereby losing accuracy in the data acquired.
Because it was too costly for companies to drill multiple monitor wells, seismologists often relied on observing well pressures, injecting fluid and modeling the rock response to determine the location of fractures. Because this is in large part a guessing game, incorrect models were inadvertently developed, Thornton explained.
Duncan and Thornton began to explore how to monitor fractures in lateral wells from the surface. They reasoned that they could lay thousands of geophones over many square miles and monitor an entire fracture network.
“The challenge is recognizing the fracture’s pattern, the signature of the rock snap, and where it took place,” Thornton said. “The industry doesn’t care about the snap, crackle and pop of the rock. They care about what the fracture network is like so they can ultimately understand the type of drainage they will get.”
Cornering the Market
Duncan recalled techniques used by Russian geoscientists looking for geothermal deposits in Iceland beginning in the 1980s.
They used geophones on the surface to locate the pipes and vents in subsurface rocks. They stacked their data and aligned entire wavelets to detect hot water percolating through porous rocks.
Duncan opted to use the entire wavelet, as the Russians did, rather than rely solely on the arrival times of the compressional (P) waves and the shear (S) waves – which can be difficult to pick out because of surface noise and subsurface structures that often distort seismic waves. He called his technique Passive Seismic Emission Technology (PSET) and patented it.
By using full waveform imaging, the entire fracture network could be monitored from the surface in the absence of a monitor well.
The technique, Thornton said, has several advantages – because the entire wavelet is being analyzed, a computer can perform much of the monitoring and analysis on its own.
“This automated process looks at the data in a way that allows us to detect when the P wave arrives, position the wavelets, and locate the breaks in the rock in a more objective way,” he explained. “You don't have to worry about the analyst picking arrival times and eventually getting tired or having two different analysts working on it and worry about consistency.”
Completing the Picture
While their technique worked well for surface microseismic monitoring, it lacked certain aspects that downhole monitoring provided, namely the ability to detect really small events, as the surface geophones were located farther away from the fracture.
“We found that the geophones sometimes picked up too much noise from the surface and could only pick waves up to 10 milliseconds,” Duncan said.
Furthermore, Duncan and his team discovered that not all reservoirs were easy to monitor from the surface. For example, some areas in the Permian Basin have large, loose gravel composition that absorbs seismic signals, and salts in the subsurface layers distort them.
The problem: Monitoring fractures strictly from the surface allows an entire fracture network to be seen, yet only includes the large, more detectable events. Monitoring fractures strictly from a wellbore allows the detection of smaller events, yet without the ability to view an entire fracture network.
The solution: Combine the two technologies for a more complete picture.
Not wanting to replicate their competitors, Duncan and his team worked diligently to apply its technique downhole.
The limited space in a wellbore allowed only a couple dozen geophones to be used rather than the thousands that could be spread out over the surface, Thornton said. Geophones in a wellbore had to be placed in a line and function as truncated antennas. To compensate for the limited number of sensors and data, both P waves and S waves needed to be captured.
Typically, seismologists monitor fractures in a wellbore by handpicking the arrival of P and S waves. They calculate the difference in time between the wave arrivals, multiply that amount by the velocity of the P wave, and determine the distance from the sensor to the actual fracture.
However, “when you throw the wavelet away and just rely on the P and S waves’ arrival times, you have nothing to ground it on,” Duncan said. “If you have different people picking P and S waves, there can be a subjective difference in the interpretations.”
Using the full waveform, Duncan and Thornton were able to detect the P and S waves downhole and estimate event locations by aligning the entire signal – not just the picked arrival times. This enabled them to see more clearly the smaller events and drive the detection threshold even lower.
By combining data from surface and downhole, they achieved an all-encompassing and detailed view of the fracture network. With that, they applied the technology in the Permian Basin with a client several months ago and achieved great success.
“It was clear from the beginning that a methodology that captured the best aspects of the downhole and the surface technique would yield superior results,” Duncan said. “By adapting our PSET algorithm to a downhole receiver, we have gone even one step further.”
Easier Said Than Done
The next step is to combine the two datasets for a single reading, and it’s not so easy. Subsurface waves, which travel horizontally to the wellbore, bounce between multiple layers of salt, shale and sandstone, which constantly change the waves’ velocity.
When looking at both surface and downhole signals, they essentially look like scribbles.
“We are using two cameras with different angles now instead of one, so to speak,” Thornton said.
He likened the problem to using both Google Earth’s aerial views and Google Maps’ street views to see one particular location.
“If you have the right source location, and the right velocity, and you correct for travel time, all the wavelets should be in alignment,” Duncan said.
While surface signals use a simple velocity calculation, downhole signals are much more complex, as wave propagation parallel to subsurface layers can be more difficult to model.
The key is building a “reasonable” velocity model that can account for two sets of signals that travel at different velocities, detect the time and location of a fracture, and not trigger on false positives or miss actual events, Duncan said.
“Nobody in the market has been able to merge those two datasets yet,” he added. “They can’t yet be processed in a single algorithm. So far, all the methods we have for creating alignment of the waves oversimplify assumptions about the velocity at which the sound is traveling. We are working on that now.”
Others in the industry are doing the same.
Throughout the process, Duncan said he has learned that the models currently used by the industry to locate fractures provide a satisfactory evaluation of a fracture network, but bypass many reserves.
“I believe every frac should be monitored,” he said. “Today, only about 4 or 5 percent of fracs pumped are monitored. Monitoring should be as important to frac’ing as logging is to drilling.”
The lack of microseismic monitoring is preventing the industry from fully exploiting the shale potential, Duncan added. “We have huge amounts of data, and we are just beginning to integrate it and interpret it in a way that will revolutionize our industry,” he said. “We need to continue pushing that boundary to penetrate the market further.”