# Microseismic processing

## Contents

## Overview

Microseismic is the energy coming from low magnitude Earthquakes and small-scale movements. They are caused by a man-made source (i.e. drilling, mining and hydrocarbon production). Micro-earthquakes are earthquakes with magnitude less than zero. Furthermore, they occur due to the volume change within the rock-mass or change in the shear-stress components. Knowing the microseismic event location is crucial for processing, interpretation and building a velocity model. Creating the velocity model is the first step in determining the microseismic event's locations. By that we can reduce uncertainty of position. Integrating the velocity model with sonic data (log data) and surface seismic data provides the best use of the microseismic data.

### Passive Seismicity

Microseismic events are detected through geophones planted on the survey surface area or down a borehole (see figure 1). Method is also known as Passive Seismicity^{[1]} which focuses on low frequency (between 0 to 10 Hz). Passive Seismicity is a real-time Microseismic record of the body and provides continuous footage of what is happening in the earth as industrial activities engage.

## Velocity Model

### Data modeling (Drilliginfo Transform— DI-transform)

A three stages horizontal well is color coded with the microseismic events that corresponds to the event timing (see figure 2). Another well is drilled to recover microseismic data (vertical black line to the right). We can rotate the images to see a better distribution within the reservoir, and how the reservoir responds to the three stages. Producing a time series data display enhances our understand of the microseismic data and reservoir state. Other factors to keep in mind as we process the data is surface pressure, flow rate, and concentration. Also, modeling microseismic data gives clear information about issues during injecting fluids into a pre-existing faults/fractures.

Real-time event graphs show the distances between the microseismic events and the source. Overlapping of microseismic events between stages is very common due to the disruption caused in previous event—original state. DI-transform provides a smooth jump to the statistical work.

### Location estimation uncertainty

Posterior covariance matrix approximates the estimation of covariance matrices[3] (see statistical equation formula, figure 3) on an unbiased basis^{[2]}. This statistical analysis of the level of confidence of the microseismic data location decreases uncertainty.
Upper portion of the ellipsoids are at 90% confidence interval(Figure 3 ^{[2]} ). However, uncertainty increases as we go deeper. The ratio of the actual confidence and the predicted confidence falls in the 10% error ellipsoids.
The true velocity model obtained from the microseismic estimated and true locations. We try to figure out the standard deviation of arrival times and compare it to the synthetic example. Eventually, we compare a contractor’s Vp/Vs ratio to the inverted Vp/Vs ratio estimate to the interferometry estimation (Figure 4 ^{[2]} ). Interferometry estimation of Vp/Vs can change depending on the number of microseismic events.

## Anisotropy

Some shale reservoirs have pre-existing fractures. Anisotropy usually occurs in such unconventional reservoirs. Anisotropy parameters recovery is possible through integrating both microseismic data and sonic log data. Such calibration gives a clear image about the reservoir’s dipping layer/multi-layer geometry. Calculating the velocity model as the fracturing occurs provides a more accurate representation of reservoir state. On-site computation reduces time error misfit. Hence, better determination of microseismic events location. ^{[3]}

Model calibration, including anisotropy computation can also be performed using the direct arrivals from perforation shots, taking advantage of the fact that the location and timing of the perforation shots are usually precisely recorded.

## Multiple data sets integration ^{[4]}

Advanced techniques are implemented to create such integration for modeling reservoirs. This requires the processor to have access to multiple datasets, and the geological background to build up a realistic model to represent the lithology of the reservoir. (see figure 6)

## Pitfalls

A standardized industrial assumption that misleads processors when dealing with microseismic data is that all microseismic data have a close ray paths. Also, assuming a simplified oil reservoir can result an overlook of significant characteristics, such characteristics are interpreted when looking at the microseismic. It is always better to keep an open mind during processing, and always process before you interpret. The less assumptions you have while processing, the fewer artifacts within interpretation.

## External Links

- Microseismic Signal Acquired from a Surface Array

During Hydraulic Fracturing in Pomerania Region in Poland (https://ac.els-cdn.com/S1877050917308931/1-s2.0-S1877050917308931-main.pdf?_tid=b35ea915-eab1-4c6d-afac-ef51d228b779&acdnat=1521602100_6e28e6d34f3b702f40ce9f818a6a50c0)

- Ghawar, Saudi Arabia. The king of giant fields (https://www.slb.com/~/media/Files/evaluation/industry_articles/201009_geoexpro_microseismic.pdf)

## References

- ↑ [1].
- ↑
^{2.0}^{2.1}^{2.2}[Zhang, Z., Rector, J. W., & Nava, M. J. (2015, December 17, https://library.seg.org/doi/abs/10.1190/segam2015-5919420.1). Improving Microseismic Event Location Accuracy with Head Wave Arrival Time: Case Study Using Marcellus Shale. Society of Exploration Geophysicists.]. - ↑ [2]
- ↑ [Auger, E., Aubin, F., Rajic, V., & Branan, A. (2011). Microseismic data: Understanding the uncertainty. Retrived November, 2011, https://www.cgg.com/technicaldocuments/cggv_0000013972.pdf]