Downhole Microseismic Geophone Array Tools and Length

With the geophone tool arrays being deployed on fiber optic or 7-conductor, there became some inherent limitations to the sampling rate and the number of geophones tools that could be deployed as well as the spacing. While some vendors would deploy 8 tools with 100' spacing others were deploying 12 tools at ~30' spacing. With time, some vendors started adding more tools or even stacking some of the geophone tools. DAS systems offer the opportunity to employ 1000 individual sensors along the length of the array in which to locate events.

So what does this mean from a processor's perspective and from a client's perspective? Is the value really there?

Case Study in Extremes: 8 Geophone Tools vs. 40 Geophone tools Several years ago, two companies were brought in to monitor a multi-stage multi-well project at the same time. Company A deployed 8 tools via a 7-conductor wireline in one well where as the domestic Company B, deployed 40 tools on their fiber optic array. Company A was able to locate about 10x the number of events during realtime processing as the Company B and was able to define their velocity model and orientations within a few minutes after the treatment started. Company B, missed the first stage completely and had difficulty picking and locating events during realtime. After their calibration was complete, Company B was still unable to locate as many events as company A. Amusingly, after the first days results were in, Company B muted all but 8 geophones. Both companies used the same processing software.

So what happened? Why did Company B, which had more deployment options, more tools, and a faster wireleine, ultimately lose to their competitor?

Orienting a 40 level geophone array simply takes more time than orienting 8 levels. Locating events within a larger grid search at a high resolution (<3 ft) also takes significantly more time. Simply doubling the resolution (or array length) results in an 8-fold increase in processing power needed.

In post processing, Company A out performed Company B with superior equipment and more tools as well. Company A had significantly more events, better clustering and was able to locate the data with a single velocity model whereas the Company B had more scatter, more artifacts, lower location resolution, and more velocity models to describe the same data.

So what is the ideal Geophone Array Length?

The length of the array may be limited by the equipment and the sampling rate desired. (TANGENT ALERT) Since ~2010, the industry has moved to recording most data at 0.25 ms sampling rate which helps in locating the events by providing more samples to describe the waveform, more accurately pick the breaks/orient the array. The effect on the actual event location is minimal between 0.5ms and 0.25ms which for a single receiver is <5 ft (for a typical Marcellus Shale velocty). Since most P-wave arrivals are high frequency (~125-250 Hz) this allows for 16 or 8 samples at the higher end. A practice in orientation is to use as many samples as possible to orient the tools with the highest linearity and this is usually observed on the first 1/4 of the wavelet which is approximately 4 samples at .25ms sampling or 2 at .5ms sampling assuming the sampling interval starts perfectly at the zero crossing. As more of the wavelet is included in the orientation process the linearity often suffers as additional noise comes into play. That being said, I have processed data at both sampling rates successfully, but noticed more consistency in orientations with the higher sampling rate. I have also recorded the calibration shots at a higher sampling and then lowered the sampling rate for data transmission purposes. (/TANGENT ALERT)

The answer is ultimately "it depends on the project". If the geophone array is going to straddle the formation or locate slightly above the lateral then running an 600 ft array with 12 tools will adequately locate the events. If there is fear of faulting or upward migration, additional tools can be added. In the very common case where the geophone array is located above the formation, then a very long array may not be very useful as the signal may attenuate and the top tools will offer little to no contribution. Additionally, the geophone array can often be used with mixed interconnect lengths which will allow a higher density of geophones closest to the formation and while having a larger overall aperture than could be had using a smaller interconnect distance. Recently the trend in the past few years has been to record microseismic data with 20-40 geophone array tools, with little discussion on the actual benefits to the event location error.

What is the take home message?

Geophysical equipment choice and design in hydraulic fracture monitoring is a system of trade-offs. While it is intrinsically understood more sensors are better for the overall picture, there are real difficulties within locating within a large location area, modeling ray paths through additional layers and calibrating velocity models over a larger area.

As more companies are shifting to migration based processing the discussion changes somewhat. Additional tools can be helpful as the user no longer needs to explore the cement bond log to avoid placing equipment. Some are likely to roll back the sampling in order to run more equipment to transmit the same file size to their processing centers. Contrary to what a leading service provider claims, realtime processing can be performed in the field or remotely by hand picking and locating events.