Hydraulic fracturing is ubiquitous throughout the domestic oil and gas industry: more than 1 million wells have been fracture-stimulated in the U.S. since the technology was introduced more than five decades ago. In 1999 alone, some 20,000 wells were fractured in North America at a total cost of $850 million. Going forward, the Society of Petroleum Engineers estimates that in excess of $1 billion may be spent each year on hydraulic fracturing. The wells that are fractured each year represent a tremendous outlay of capital, horsepower and materials, but not all of these stimulations are physically or economically successful. Historically, engineers have relied on indirect methods-the calculations of surface pressures and pump volumes, the response of the well in production-to determine the effectiveness of the treatments. Direct methods such as tracer surveys revealed some clues, but these were stringently limited to near-wellbore effects. A new technology is seeking to change that. Microseismic diagnostic technology focuses on the acoustic signals that are associated with fracturing in the subsurface. These subtle signals, which can be detected with today's ultrasensitive receivers, can provide real-time, direct looks at how and where fractures propagate during well stimulations. The technology works because hydraulic fracture treatments induce microearthquakes. The high-pressure pumping of fluid disturbs the reservoir, and it slips along planes of weakness, bedding planes, natural fractures or faults. These slippages cause very small earthquakes-on the Richter scale, they would be negative numbers. Comparing the differences in arrival times of shear and compressional waves reveals the distance of the microseismic events from the sensors. Because the sensors are triaxial, the direction is also known. (The inherent assumption, of course, is that the microseismic events are directly related to the induced fractures.) The resulting microseismic map is quite detailed, and provides a three-dimensional view of the subsurface. Engineers can resolve the length, height and strike of induced and natural subsurface fractures. The results of microseismic surveys often surprise people, much as the images from 3-D seismic volumes changed the conventional understanding of the subsurface. Prior to the widespread use of 3-D, pictures of the subsurface were often quite simple; it took the superior imaging power of the 3-D volumes to reveal the intensely complex habitats of many petroleum accumulations. Likewise, the various indirect methods of estimating the results of a hydraulic stimulation are proving to be wide of the mark. "Microseismic monitoring has revolutionized the view of fracture patterns," says Dick Zinno, director of hydraulic fracture monitoring and permanent borehole seismics for Schlumberger. "In naturally fractured reservoirs, the fracture patterns start to look like TV antennas. We're finding that preexisting structures very much control where the stimulation goes, what the drainage is like and how the reservoir should be developed." With a good data set, the microseismic technique can map out the fracture system that is created during a stimulation, as well as the types and orientations of the fracture planes, he says. "We're starting to learn a lot more about the fracturing process." Reality is indeed more complicated than the previous models indicated. "Probably 60% of the time, we find that the fracture dimensions we measure are different than what the model had predicted," says Kevin Fisher, vice president of business development for Pinnacle Technologies. Pinnacle, a privately held company based in San Francisco, focuses its business on fracture optimization. It has long experience with tiltmeters, another diagnostic technique for fracture monitoring. In January 2001, Pinnacle negotiated an exclusive license with the Gas Technology Institute for technology it had developed for microseismic mapping. That license covers software and certain equipment. "The microseismic information can be used in several ways," says Steve Wolhart, former GTI manager and now project manager with Pinnacle. "We can make sure all the pay intervals are covered by the stimulation, and that the stimulation is effective. In cases where we are trying to treat several separate zones, we can tell if we are treating all of them successfully." Another crucial use is to determine proper well spacing for field development, as the direction and length of the induced fractures affects well placement. Microseismic techniques were initially developed decades ago in the mining industry, where it is a mature and routine technology, says Dr. Shawn Maxwell, manager of petroleum services for Engineering Seismology Group, based in Kingston, Ontario. Mine safety is its primary purpose, as in a number of deep heavy-metal mines in northern Ontario that are plagued by rock bursts, caused by induced seismicity. Microseismic analysis is used to assess the sites for potential problems, and to determine in real time exactly where rock bursts have occurred. From its original involvement in mining, ESG has since diversified into microseismic applications for nuclear waste storage and civil engineering uses. Since 1997, the company has been developing its microseismic business for oil and gas. In the oil industry, efforts to use microseismic monitoring to map induced fractures began in earnest about 15 years ago. GTI, several industry operators, and Sandia and Los Alamos national laboratories ran a series of experiments to prove the technology. Another group, Teledyne Geotech, was instrumental in the early development of microseismic techniques. "A common problem at that time was having good enough tools to record the high-frequency signatures of the microseismic events," says Zinno. To solve this, people tried cementing permanent geophones behind casing in the observation wells. That solution, however, was very expensive and research efforts stagnated. Stymied by the lack of data sets and case histories, many researchers shelved their microseismic efforts. One study that finally caused oil and service companies to take new notice of the technology's potential was done in 1997 in East Texas, however. The Cotton Valley Consortium, undertaken by an industry group led by Union Pacific Resources, was a groundbreaking test. Zinno, then at UPR, designed the arrays in association with research and development groups at Texaco and Arco. Significant efforts were aimed at developing state-of-the-art tools and processes to effectively collect and interpret microseismic data. The study focused on the Cotton Valley sands in Carthage Field, Panola County. It proved that microseismic data could be valid at depths approaching 10,000 feet, and that the arrays could be successfully deployed in open holes. The results were eye-opening, for unlike the expectations of the completion engineers, the induced fractures did not radiate out equally on each side of the wellbore, and the maximum fracture lengths were reached very early in the fracture treatments. The success of the project renewed enthusiasm in the technology. Microseismic monitoring was finally pushed into the commercial realm about two years ago by a number of advances, both in instrumentation and in computing power. Stafford, Texas-based Oyo Geospace developed reliable and highly sensitive geophones for downhole sensors, and another thrust was provided by the advent of fiber-optic wirelines for the transmission of the tremendous data volumes. Too, computers can now easily manipulate gigabytes of data. In the years since the first microseismic efforts, there have been about 125 jobs performed worldwide, and 75 of those jobs were performed last year, says Fisher. "Clearly, the technology is taking off." Real-world applications Today, operators consider fracture diagnostics for several reasons. Often, it's because their well completions are not performing as they expect. There are many problems with stimulations, ranging from incomplete coverage of the zone, poor fluid diversion, out-of-zone growth, and twisting or horizontal fractures, says Chris Wright, Pinnacle's president. "It's very important for companies to calibrate their frac models with data from direct fracture-detection methods." The technology also makes sense when a company plans a fracturing program of more than 20 jobs, says Fisher. "Microseismic monitoring gives the company information on how many stages are needed to stimulate the whole pay zone, and on how to properly size the frac job. Correctly designing a large frac program can save a company a great deal of money and result in better wells." A typical microseismic survey might cost $60,000; 10% to 15% per job can be saved on multiple surveys in the same area, says Wright. At an average cost of $50,000 per single frac stage, the advantage of an accurate understanding of a stimulation's effectiveness is apparent. To run a microseismic survey, downhole sensors are installed in a monitoring well-one monitoring well is more than adequate. Pinnacle suggests that the observation well should either be a new, unperforated well or a producing well in which the zone of interest can be isolated. A string of eight to 12 sensors is run above, below and across the target zones, covering a distance of 200 to 800 feet. The operator must know the formation velocities, which can be obtained from sonic logs or cross-well tomograms. Schlumberger, which uses a different set of algorithms, specifies that it have access to an observation well that is within 1,500 to 2,000 feet of the treatment well. It places the geophone array within a vertical separation of about 1,000 feet of the target zone; the zone of interest does not need to be isolated; and velocities need only be known to an initial estimate. "The dimensions are reservoir dependent, and in some applications the requirements can be looser or tighter," says Zinno. Generally, the tools are run on wireline into the observation hole, and the reservoir is monitored while the frac job is being pumped. The raw data doesn't look much like traditional seismic data, notes Wolhart. "Microseismic has more in common with locating earthquakes than with regular oilfield reflection seismic." Field processing is done on PCs; final processing is completed shortly thereafter in the office. "We can do the analysis in real time, depending on the quality of the data. We can give the customer a map in the field." Zinno notes that Schlumberger and ESG commonly produce real-time results as the frac job is pumped, and have been offering this capability since January 2000. In the years ahead Excitement is now growing over using microseismic as a reservoir-monitoring tool. The technique can detect acoustic signals that are generated by events such as production from a reservoir, or the movement of water- or steam fronts through a reservoir. "The frac-monitoring abilities of microseismics are well proven," says Maxwell, "but the reservoir characterization applications are still in their infancies." So far, these applications have only been attempted on a limited basis. A classic study was conducted at Ekofisk Field, operated by Phillips Petroleum in the Norwegian sector of the North Sea. The operator, faced with alarming seafloor subsidence and compaction of the field caused by production, instrumented one of the wellbores under the platform. It listened for 18 days to the acoustic emissions from the field and was able to map a pattern of seismicity in a 4,000-foot radius around the platform. The microseismic emissions showed the location of reactivated fault planes that were involved in the compaction, and detailed the compartmentalization of the field. A further finding was that microseismic activity was dampened around the wells that were injecting seawater into the reservoir to counteract the subsidence, and was heightened around the production wells. Today, ESG has a number of permanent systems around the world, placed to image fracture networks, waterflood fronts and steam injection fronts. "A lot of our reservoir characterization work is part of time-lapse VSP and 4-D seismic programs," says Maxwell. "Passive imaging is used to fill in the gaps between the cycle times." Clearly, the development process for microseismic monitoring will be long, but already it has demonstrated tremendous economic value. "Microseismic analysis is giving us a new view of completion and reservoir-drainage issues," says Zinno. "There's a big, bright future out there for this technology."