DE102014109280B4 - Method and apparatus for determining seismic damping based on a microseismic event - Google Patents

Abstract

A method for determining a seismic attenuation (Qi, Qs, Qt) based on a microseismic event (SQ) by applying an inversion using a forward calculation to microseismic recorded data of such a microseismic event (SQ) (S4).

Description

Method and apparatus for determining seismic damping based on a microseismic event The invention relates to a method and apparatus for determining seismic damping based on a microseismic event.

US 2012/0116680 A1 describes a localization of sources of microseismic events by calculating the Green function. US 2005/0190649 A1 describes a microseismic process in which waveform inversion is performed to determine parameters representing a source characteristic of the seismic event.

More particularly, the invention relates to a method of determining seismic damping using microseismic data as collected for monitoring hydraulic stimulation measures, e.g. This includes, for example, the development of geothermal reservoirs, underground gas and CO ₂ storage and the extraction of oil, gas and coal. Seismic damping is generally described by the so-called Q-factor, which is defined as the ratio of stored to dissipated energy. For total attenuation, also referred to as apparent attenuation Q _t , two effects contribute decisively: Q -1 / t = Q -1 / i + Q -1 / s (1)

In this case, a strength of an intrinsic damping Q _{t describes} a conversion of elastic energy into heat. By contrast, a scattering attenuation Q _s quantifies a redistribution of energy, in particular at later times, due to an interaction of a seismic wave with heterogeneities in the propagation medium, that is to say in particular an influence of scattering (English: scattering). The different types of damping depend on properties such as lithology, water saturation, permeability and / or porosity. Thus, damping is useful to derive specific reservoir characteristics or to develop more complex subsurface models for use in seismic inversion.

wobei v eine mittlere Wellengeschwindigkeit, D(r) eine geometrische Dämpfung, M(f) ein Quellspektrum und der Exponentialterm eine Amplitudenabnahme aufgrund der scheinbaren Dämpfung Q_t beschreibt. Der Quellterm wird vollständig beschrieben durch ein seismisches Moment M₀, eine Eckfrequenz f_c, eine Abstrahlcharakteristik ψ und einen Abfall γ des Amplitudenspektrums A(f, r) bei hohen Frequenzen.Use of damping in microseismic monitoring:
In the context of microseismic monitoring, the attenuation is very rarely determined. For modeling purposes, a mean, in particular constant value for the total attenuation or apparent attenuation Q _{t is} used, which was previously taken from external literature references. To determine the attenuation from real, microseismic data, an inversion method is used which is based on an adaptation of a theoretical curve to an observed amplitude spectrum A (f, r) as a function of a frequency f and a hypocentral distance r. The amplitude spectrum A (f, r) is calculated as a Fourier transform of a short data window of P-phase or S-phase, where the P-phase is a phase of a primary wave or pressure wave and the S-phase is a phase of a secondary or transversal wave, in each case in particular the first missions are considered. Typical window lengths are less than 1 second, whereby the duration of the data in the data window is understood as meaning the duration of the window. The spectrum of the seismic signal is parameterized as follows (Masuda and Suzuki, 1982):

where v denotes a mean wave velocity, D (r) a geometric attenuation, M (f) a source spectrum, and the exponential term an amplitude decrease due to the apparent attenuation Q _t . The source term is completely described by a seismic moment M ₀ , a corner frequency f _c , a radiation characteristic ψ and a drop γ of the amplitude spectrum A (f, r) at high frequencies.

In this context, a theoretical curve can be adapted either to the entire spectrum, in particular in the sense of a total inversion or to spectral ratios, in particular in the sense of spectral division using the least-squares method. The total inversion of the spectrum requires a priori knowledge of some parameters, while in spectral division the number of free model parameters is significantly reduced. Here, either the ratio of spectra at different stations with distance r and frequencies f (Jia, 1996) or seismic events (Ide et al., 2003) can be considered. Jia (1996) describes two inversion methods based on an inversion of the amplitude spectrum A (f, r), which allow the user to determine the apparent attenuation Q _{t in a} frequency-dependent manner Q _t (f) = Q ₀ · f ⁿ . The inversion results show especially in the Determination of a frequency-dependent Q-model strong dependencies between the different model parameters. The inversion is therefore usually limited to the determination of a frequency-independent apparent attenuation Q _t . In addition, most assumptions about the nature of the source spectrum, such. B. a so-called ω ^-2 model (Aki and Richards, 1980), integrated directly into the inversion algorithm.

A disadvantage of such methods is the great effort to determine a damping, in particular a total inversion of the spectrum, but also a spectral division using different frequencies, stations and events such. B. a need for the occurrence of similar earthquakes.

Use of damping in seismological monitoring:
In the field of seismology in general it is known, with the aid of inverted damping parameters, a scattering coefficient g * and an absorption parameter b, to derive values for the damping, in particular for the Q factors Q _s and Q _i . Previous applications of this method (Sens-Schönfelder & Wegler, 2006; Padhy et al., 2007; Przybilla et al., 2009) are limited to an analysis of tectonic events with magnitudes M _L = 3.5-6.0 in order to increase the attenuation properties at the regional level to study.

In particular, a seismogram duration or window lengths of at least 100 s up to more than 500 s are analyzed in order to evaluate not only a direct wave but also code waves, ie waves scattered along an obstacle along the path and for the seismological analysis, in particular for a separation of the different types of attenuation (Q _s and Q _i ) are required.

Disadvantageously, however, microseismic data have other characteristics, such as. B. a higher frequency content and lower microseismic magnitudes M _L ≤ 1.0, in particular less energy released, in particular a poorer signal-to-noise ratio (SNR).

From a microseismic point of view, there are no or only a few scattered waves due to short running distances, which, in addition to a direct route, also propagate via indirect pathways between the microseismic source and a geophone or seismometer or hydrophone. This results in a particularly shorter signal duration, which is why the application of the known seismological procedure is not transferable to microseismic data.

The object of the invention is to improve a method and an apparatus for determining a seismic damping based on a micro-seismic event, in particular to enable as extensive automation as possible.

This object is achieved by the method having the features of patent claim 1 and the device having the features of patent claim 10. Advantageous embodiments are the subject of dependent claims.

Accordingly, preferred is a method of determining seismic damping based on a microseismic event by applying an inversion using forward calculation to microseismic acquired data of such a microseismic event.

Thus, to determine the attenuation, a procedure is used in which a seismological method known per se can be used to analyze microseismic data. Amongst the recorded data, seismogram data recorded or registered or registered with a possibly different arrangement is understood in particular, which are read out from a storage medium for carrying out the method.

One embodiment is that the inversion is applied to the recorded data via an inversion time window, wherein the duration of the inversion time window is shorter than 150 seconds, in particular shorter than 100 seconds, in particular shorter than 75 seconds, preferably shorter than 50 seconds and in particular a duration of Inversion time window longer than 8 seconds, especially longer than 9 seconds, preferably longer than 10 seconds.

Thus, the duration is so long that sufficient data of the particular entire S-wave train, which contains in particular the direct S-wave and in particular the S-coda of the wave train, or an envelope of the recorded data in the inversion. An envelope is in particular determinable as a squared seismogram track and in particular the sum of the envelopes of the recorded components. The use of the envelope means in particular that a poor amplitude Signal-to-noise ratio of a single seismogram component does not necessarily result in a poor signal-to-noise ratio of the envelope. For example, the S-wave is primarily recorded on the horizontal components and the P-wave on the vertical component, but not exclusively, which is why the summation is advantageous. In particular, this procedure can advantageously be used for the seismological procedure, since the envelope designated here is linearly proportional to the energy density distribution of a seismic event, in particular of a microseismic event SQ. A start time of the inversion time window is determined in particular by the automatically determined operating time of the S phase.

One embodiment is that from the recorded data, a data area is used in which a signal-to-noise ratio of 12, in particular 14, preferably 16, is exceeded for at least one point of the envelope.

As a result, only seismic data receiving stations are selected for inversion whose inversion timeslot or their recorded S-wave train contains a distinct direct S-wave. In particular, this also ensures that the envelope of the direct S wave and the envelope of the coda of the wave train or of the recorded data are received in sufficiently high degree in the inversion.

One embodiment is that from the recorded data a data area is used in which a signal-to-noise ratio of the envelope of 4, in particular of 5, preferably of 6 is not fallen below.

As a result, a data range for the inversion time window is selected or determined in which sufficiently significant, in particular exclusively significant, portions of the particular entire S-wave train are included in the inversion. An end time of the data area or the inversion time window is determined according to a development in particular also by the fact that over a predetermined period of z. B. a few seconds, in particular 1 second or less or over a predetermined number of sampling times this required signal-to-noise ratio is exceeded.

An embodiment is that at least one of a start time or a start and an end time or an end of the inversion time window or a coda of the recorded data relative to a noise level is determined, particularly relative to a noise level before a first primary wave is used recorded data of the microseismic event is determined.

One embodiment is that the inversion is carried out individually over mutually different frequency bands of the recorded data, wherein the frequency bands are defined overlapping each other, in particular at 10 to 30 percent, in particular 30 to 50 percent, in particular 40 to 60 percentage, in particular 60 to 80 percent overlapping.

An appropriate selection of the frequency bands and filtering of the recorded data in the provision of the recorded data allows the individual analysis of different frequency bands, which in turn makes it possible to more accurately map the frequency dependence of the attenuation parameters. Thus, a determination of the attenuation parameters for the individual frequency bands is preferably first carried out before a determination of the frequency dependence, in particular an adaptation of a regression line, is carried out. By taking into account the time and frequency dependence (time-frequency approach) in the analysis of the energy density distribution, both damping mechanisms can be quantified independently of each other.

One embodiment is that only data from stations in which a station noise level is less than twice an average noise level is used from the acquired data.

An embodiment is that prior to the inversion a modeling of the recorded data is carried out by means of an analytical determination of a Green's function, in particular the calculation of such Green's functions is carried out for a predeterminable set of scattering coefficients.

One embodiment is that the inversion is applied to an S-wave train of the recorded data, in particular to an entire S-wave train including a direct S-wave and its S-coda of the recorded data.

Thus, only one data area or one inversion time window is used in the inversion, in which in particular all information relevant for the inversion and necessary for determining the attenuation is contained.

Accordingly, there is also preferably provided a device for determining seismic damping based on a microseismic event with a receptacle for providing microseismic data, and a controller adapted and / or programmed to apply such method to provided microseismic data for determining seismic attenuation.

In particular, the use of the described seismological procedure on data from microseismics is made possible by such methods of data selection for inversion and by the use of such signal processing of the recorded data. In particular, a simple adaptation of an inversion methodology to the analysis of microseismic data to determine the attenuation parameters is made possible. In particular, specially developed inversion criteria, adapted to the particular conditions of microseismic data, and secondly an inversion of the particular entire S-wave train using a single inversion time window and a fast forward calculation are used.

The method has the notable advantage of not having to make assumptions about fracture, near-surface site enhancement, or the mode of attenuation.

By designing a corresponding filter database, the frequency dependence of the attenuation parameters in the inversion is directly determined.

An embodiment will be explained in more detail with reference to the drawing. Show it:

1 a measurement situation in a microseismic event;

2 exemplary time histories of a registered microseismic event and of its envelope over time;

3 a flowchart of an exemplary method for determining a seismic damping based on a microseismic event;

4 Data diagrams illustrating elements of a seismological procedure with modeling a set of Green's functions, fitting a regression line, and minimizing errors; and

5 Data diagrams to illustrate the final results of an exemplary process implementation with a comparison of the measured and modeled energy density and with a frequency-dependent attenuation.

1 shows a background in which a microseismic event SQ occurs. This microseismic event SQ is intended by means of a recording device 1 detected and by means of a control device 2 be evaluated. The recording device 1 is in particular a geophone, which, depending on the design, can also be implemented as a seismometer or hydrophone. The control device 2 is in particular a device with a processor and software, which of the receiving device 1 evaluates received data and processes. The software or the control device 2 In particular, it is designed to determine a seismic damping from an in particular 3-component seismogram, the seismograms representing such a microseismic event SQ.

Optionally, as a recording device and a data input of the control device 2 be viewed via the corresponding Seismogrammdaten in the controller 2 or a device having this are transmitted. In particular, such a device can have further components, such as a memory device for storing such seismogram data, a computer and / or software system for automated standard evaluation, in particular for determining the cooking time, localization of the hypocenter and determination of the operating times of distinctive phases and an evaluation software for determining Attenuation parameters from such seismogram data of a microseismic event SQ.

As exemplified, the exemplary microseismic event SQ occurs with a range distance r to the receiving device 1 , Accordingly, waves that are triggered by the microseismic event SQ, in particular as direct waves, travel directly to the receiving device via a direct path dw from the location of the microseismic event SQ 1 , In particular, a primary wave P and a secondary wave S are the first more prominent events by the receiving device 1 recorded. In addition, further waves travel via indirect paths iw from the source of the microseismic event SQ to the receiving device 1 , Scattered waves received via such indirect paths form the so-called coda C of the seismogram.

2 shows over the time specified in seconds s exemplary courses of a microseismic event SQ, by means of such a recording device 1 was recorded. It is in the upper portion of the figure through the receiving device 1 exemplary z recorded component Z of the seismogram outlined. This has an initial section with only noise, especially background noise. This is followed by the arrival of the primary wave P of the microseismic event SQ as the first striking event. A time later, the largest striking event is the arrival of the secondary wave S.

In the further course of time, the amplitude of the recorded microseismic event decays, with further, in particular, also primary and secondary waves arriving via non-direct routes from the source to the receiving device 1 have arrived. These form as the coda C in particular a proportion of the amplitudes or excursions of the seismogram. The representation with a concrete time extension over a little more than 20 s is to be regarded as preferred and merely as an example.

In the lower part of the figure, an envelope E calculated using this seismogram component is plotted over time t. Marked sections form a noise window NW, which can be determined before the first use of the primary or P-wave P, a time window MTW moving for the data processing and an inversion time window ITW. The inversion time window ITW has an initial time value as start time t _min and a time end value as end time t _max . The start time t _min preferably begins with the first use of the secondary or S wave S, and the end time t _max is determined in such a way that a portion of the coda C sufficient for determining the attenuation is processed. In particular, the total duration of the inversion time window ITW is less than 100 s, ie significantly smaller than in an evaluation of seismological events with magnitudes M _w greater than 3 or 4.

3 outlines preferred method steps of an exemplary method for determining seismic damping based on recorded seismogram data of a microseismic event.

In a first step S1, data is recorded. Among them, a collection of data in the form of a seismogram is understandable. In particular, including a recording of such previously recorded data of a seismogram is understood, the data are already stored on a storage medium and retrieved from the storage medium for processing and in particular in a main memory of a processor of the control device 2 be read. In particular, a collection of data from an automated standard evaluation, which includes in particular a localization of the hypocenter, a determination of the stovetop time and the operating times of the P and S phases, is carried out.

In a second step, a data selection is made from this data or from its envelope. The data selection means, in particular, that a time segment of the data to be used for the inversion method is selected from the envelope of the seismogram data of each individual station. This is feasible fully automated by z. B. the appropriately configured control device 2 , by manual selection by an operator by means of an interface of the control device 2 or combined by operator and controller 2 ,

A criterion of the data selection is that a signal-to-noise ratio (SNR) of 12, in particular of 14, preferably 16, is exceeded, in particular for at least one point of the envelope. A criterion of the data selection is that, in particular, a minimum signal-to-noise ratio of 4, in particular of 5, preferably of 6, is not undershot.

A criterion of the data selection is that in particular a duration of the inversion time window ITW is longer than 8 seconds, in particular longer than 9 seconds, preferably longer than 10 seconds. Furthermore is a duration of the inversion time window ITW in particular shorter than 100 seconds, in particular shorter than 75 seconds, preferably shorter than 50 seconds.

A criterion of data selection is that, in particular, a station noise level is less than twice a mean noise level.

As part of the data selection or preset, the inversion determines a frequency range to be used and the frequency bands to be used. The frequency bands are defined overlapping each other.

In a third step S3, modeling based on the determined station parameters such as hypocentral distance r and deployment time of the S-phase is performed. In particular, an analytical calculation of the Green's function is carried out for a predeterminable family of scattering coefficients.

In a fourth step S4, an inversion of the selected data is performed, wherein in particular a determination of the inversion parameters and in particular a determination of the attenuation parameters for the individual, overlapping frequency bands and then a determination of the frequency dependence of these parameters is performed by means of regression lines.

Hereinafter, particularly preferred steps of such a method for attenuation determination by inversion of envelopes, in particular Seismogrammeinhüllenden (English envelope inversion) are described in more detail. This method follows a stochastic, time- and frequency-dependent approach, which is based on the so-called "energy transfer theory" (English: Radiative Transfer Theory). To determine the attenuation, the heterogeneous subsoil is regarded as a statistical random medium, and the energy propagation of multiply scattered seismic waves in this random medium is considered. In particular, this process is fully described by the energy transfer equation. The inversion is based on the adaptation of theoretical curves to an observed or acquired energy density distribution E _obs (f, t, r) of the recorded data of a recorded microseismic event. A measured energy density distribution E _obs (f, t, r) is determined from real seismogram data by the calculation of the envelopes of the recorded velocity seismograms. In particular, the envelope is determinable as a smoothed trace of the squared seismograms and, in particular, the sum of the envelopes of the recorded seismogram components. The envelope is then in particular linearly proportional to the measured energy density at time t and at the distance r from the source of the microseismic event SQ. A theoretical energy density E _syn (f, t, r) at a time t and a hearth distance r is parameterized as follows: E _syn (f, t, r _j ) = S (f) * R _j (f) * G (t, r, f) * I (t, f), (3) where S (f) and R _j (f) describe a source term as well as a near-surface location-dependent gain factor. G (t, r, f) corresponds to Green's function of the energy transfer equation and describes a complex redistribution of energy due to multiple scattering. I (t, f) contains the independent portion of the intrinsic damping Q _i , which is described by the exponential term e ^-bt and quantified by an absorption ^parameter b as a damping ^parameter . A scattering coefficient g * is another attenuation parameter that controls the intensity of the scattering. Corresponding values for the Q factors, ie the intrinsic damping Q _i and the scattering attenuation Q _{s can be} derived from the inverted damping parameters g * and b: Q -1 / s = g * v / 2πf Q -1 / i = b / 2πf (4)

The total attenuation is then calculated according to equation (1). It should be emphasized that the appropriate method of data selection for inversion as well as an application of a combination of seismology and signal processing techniques known per se are used, which are used for the adaptation of the inversion method to the analysis of microseismic data. In particular, the components of the intrinsic damping Q _i and the scattering damping Q _s are taken into account independently of one another in the energy transfer theory.

Microseismic data have different characteristics from seismic techniques for analyzing tectonic events. Accordingly, in the determination of attenuation microseismic data z. As a higher frequency content, a low signal-to-noise ratio and a shorter signal duration considered. The following describes a corresponding procedure for application to microseismic data.

wobei ρ₀ eine Dichte des Ausbreitungsmediums ist, durch welches die mikroseismische Welle gelaufen ist. Die sogenannte Hilbert-Transformation H wird dabei bevorzugt verwendet, um glatte Kurven zu erzeugen, die wiederum besser zum Vergleich mit theoretischen Kurven geeignet sind.In method step S2, the data selection, the recorded data, ie the microseismic data set, passes through a filter database in order to subdivide the data into corresponding frequency bands. In order to more accurately map the frequency dependence of the attenuation parameters, the filter bank is constructed such that the frequency bands are in particular 10 to 30 percent, in particular 30 to 50 percent, in particular 40 to 60 percent, in particular 60 to 80 percent, in particular have a 50 percent overlap. The observed or determinable energy density distribution E _obs (f, t, r) of the microseismic event at time t, range distance r and frequency f, in particular center frequency of the bandpass used, results from the particular three components speed seismogram u → (t, r) according to :

where ρ _{0 is} a density of the propagation medium through which the microseismic wave has passed. The so-called Hilbert transform H is preferably used to generate smooth curves, which in turn are better suited for comparison with theoretical curves.

For the inversion, the appropriate selection of the inversion time window ITW uses the data of the particular entire S-wave seismogram, including in particular the direct S-wave and in particular the code waves. The end of the coda C is thereby determined with the aid of the noise level, measured in the noise window NW before the P-wave insert. In order to take account of the special features of microseismic data, the inversion time windows of the envelopes are selected in particular according to the criteria according to method step S2.

By way of example, a maximum SNR in at least one point greater than 16, a minimum SNR, in particular for determining the S-coda end greater than 6, the window length of the inversion time window ITW greater than 10 seconds and the station noise level (English individual noise level) smaller as twice the mean noise level. Such a noise level is determined, in particular, as an average over all stations with which the recorded data of the microseismic event were recorded, wherein the calculation can be carried out individually, in particular in each frequency band.

The start time t _{min of} the inversion time window ITW is determined in particular on the time of the first use of the direct S wave S or before, in particular a few milliseconds before. The end time t _{max of} the inversion time window ITW is correspondingly calculated as a function of the predefinable, minimum signal-to-noise ratio in relation to the determined value in the region of the noise window NW.

wobei δ(t) und H(x) die Dirac'sche Deltafunktion sowie die Heaviside-Funktion bezeichnen. Entsprechende Kurven der Green'schen Funktion sind beispielhaft für eine Schar von 3 beliebig ausgewählten Streukoeffizienten in 4 links oben skizziert. Dieser analytische Ansatz ist gültig für ein insbesondere verwendetes einfaches Halbraummodell unter der Annahme einer isotropen Quelle und isotroper Streuung von S-Wellen und weist nur im Zeitbereich nach der direkten S-Welle minimale Abweichungen, insbesondere unter 5% Abweichung von der exakten Lösung nach Zeng et al. (1991), auf. Mehrfachstreuung basiert auf der Annahme, dass durch Konversion S-Wellen zur dominanten Phase im späteren Teil des Seismogramms werden. Aufgrund dieser Tatsache ist mit dem hier dargestellten Verfahren insbesondere die Bestimmung der seismischen Dämpfung von S-Wellen möglich.The modeling (forward modeling) carried out according to method step S3 consists in particular of a determination of a Green's function G (t, r) of the energy density distribution, which is in particular a solution of the energy transfer equation, and is determined by different methods, e.g. B. Monte Carlo or finite difference simulation, depending on the required complexity of the resulting subsurface model feasible. In order to ensure a rapid application, in particular, an analytical approximation method for the Green's function G (t, r) is implemented in the modeling algorithm:

where δ (t) and H (x) denote the Dirac delta function and the Heaviside function. Corresponding curves of Green's function are exemplary of a family of 3 arbitrarily selected scattering coefficients in FIG 4 sketched above left. This analytical approach is valid for a particularly used simple half-space model assuming an isotropic source and isotropic scattering of S-waves and exhibits minimal deviations only in the time domain after the direct S-wave, in particular below 5% deviation from the exact solution according to Zeng et al. (1991). Multiple scattering is based on the assumption that by conversion S-waves become the dominant phase in the later part of the seismogram. Due to this fact, the determination of the seismic damping of S waves is possible in particular with the method presented here.

The inversion performed according to method step S3 in particular uses an inversion problem which is over-determined with the simultaneous determination of the source term S, the scattering coefficient g *, the absorption parameter b and the location factors R _j for all stations j with which the microseismic event SQ was recorded. Therefore, in the first part of inversion, R _j (f) = 1 is assumed and then the remaining quantities S, g * and b are determined individually for each frequency band. In particular, the core of inversion consists of linear regression and least squares error minimization. To linearize the inversion problem, the natural logarithm is applied to equation (3):

mit Hilfe einer 1D Gittersuche berechnet. Die Bestimmung des Fehlerminimums ist beispielhaft in 4 unten skizziert.By a particularly automated adaptation of a compensation or regression line ld, which exemplifies in 4 right above, the source term S and the absorption parameter b are determined as a function of frequency for a fixed value of the scattering coefficient g * and a Green's function G calculated therefrom. The problem is solved individually for each scattering coefficient g *, but preferably all stations are used simultaneously. The scattering coefficient g * finally becomes the minimum of the error function

calculated using a 1D grid search. The determination of the error minimum is exemplary in 4 outlined below.

For the application of a stochastic method in seismology, the formation of an ensemble average is required, which is defined as an average over various realizations of the propagation medium of the microseismic wave. Since there are only real data for a single realization for the inversion, in the inversion algorithm the share is replaced by a time average, in particular according to the so-called ergodic principle.

Finally, in an additional step, the location gain R _j (f) can be determined as mean deviation between measured and inverted energy density at station j.

A feasibility of this method can be demonstrated on the basis of the evaluation of a microseismic data set with a maximum magnitude M _w ≦ 1. The data used as an example were registered during the hydraulic stimulation on the KTB (Continental Deep Drilling Program) in 2000. Using the described inversion method, a very good agreement between theoretical and observed energy density distribution can be achieved, which can be demonstrated in 5 outlined on the left. By inversion of several microseismic events, mean values for the two damping factors Q _i and Q _{s can be} determined, which coincide within the scope of the measurement accuracy with regional attenuation models for southern Germany. The total attenuation, shown in 5 right, is best described in the range between 5 Hz and 95 Hz by a frequency-dependent Q model with Q (f) = Q ₀ f ⁿ , where Q ₀ and n describe the Q value at 1 Hz and the frequency dependence. The window length used for the determination of the noise level NW, the signal-to-noise ratio and the time average are T = 1 s.

LIST OF REFERENCE NUMBERS

1

recording device

2

control device

A (f, r)

observed amplitude spectrum

b

absorption parameter

C

Coda

D (r)

geometric damping

dw

direct route

e

envelope

E _obs (f, t, r)

observed energy density distribution of a microseismic event

E _syn (f, t, r)

theoretical energy density

f

Frequency, in particular center frequency of the bandpass used

f _c

corner frequency

g ^*

scattering coefficient

G (t, r)

Green's function

H

Hilbert transform

H (x)

Heaviside function

I (t, f)

Proportion of intrinsic damping

ITW

Inversion time window

iw

indirect route

ld

Equalization or regression line

M (f)

source spectrum

M ₀

seismic moment

M _w

Moments of magnitude, in particular magnitude of a seismic event

MTW

moving time window

n

Frequency dependence of the Q model

northwest

noise window

P

primary wave

Q _i

intrinsic damping

Q _s

scattering attenuation

Q ₀

Q value at 1 Hz

Q _t

apparent damping or total damping

r

Stove removal, especially hypo-centric removal

R _j (f)

Location gain

t

Time

t _max

End time of the inversion time window

_min

Start time of the inversion time window

S

secondary wave

S (f)

Z (f) Source term of the theoretical energy density

SQ

microseismic event

u → (t, r)

3-component speed seismogram, in particular bandpass-filtered

u _i (t, r)

Single component of the velocity seismogram, e.g. B. Z component, in particular bandpass filtered

v

mean shaft speed

Z

Z component of a seismogram

γ

Drop in the amplitude spectrum

ρ ₀

Density of the propagation medium

δ (t)

Dirac delta function

ψ

radiation

Claims (10)

A method for determining a seismic attenuation (Q _i , Q _s , Q _t ) based on a microseismic event (SQ) by applying an inversion using forward calculation on microseismic recorded data of such microseismic event (SQ) (S4). The method of claim 1, wherein the inversion is applied to the recorded data via an inversion time window (ITW), wherein The duration of the inversion time window (ITW) is shorter than 150 seconds, in particular shorter than 100 seconds, in particular shorter than 75 seconds, preferably shorter than 50 seconds, and In particular, a duration of the inversion time window (ITW) is longer than 8 seconds, in particular longer than 9 seconds, preferably longer than 10 seconds. Method according to one of the preceding claims, in which a data range is used from the recorded data in which a signal-to-noise ratio (SNR) of 12, in particular 14, preferably 16, is exceeded for at least one point of the envelope. Method according to one of the preceding claims, wherein from the recorded data a data area is used in which a signal-to-noise ratio (SNR) of the envelope of 4, in particular of 5, preferably of 6 is not undershot. Method according to any preceding claim, wherein at least one of a start time or a start and an end time or an end of the inversion time window (ITW) or a coda (C) of the recorded data is determined relative to a noise level, in particular relative to one Noise level (NW) is determined before use of a first primary wave (P) of the recorded microseismic event (SQ) data. Method according to one of the preceding claims, in which the inversion (S4) is carried out individually over mutually different frequency bands of the recorded data, wherein the frequency bands are defined overlapping each other, in particular from 10 to 30 percent, in particular from 30 to 50 percent, in particular 40 to 60 percent, in particular to 60 to 80 percent overlapping. A method according to any preceding claim, wherein only data from stations in which a station noise level is less than twice an average noise level is used from the acquired data. Method according to one of the preceding claims, wherein before the inversion (S4) a modeling of the recorded data is carried out by means of an analytical determination of at least one Green function (S3), in particular the calculation of such Green's functions is carried out for a predeterminable family of scattering coefficients , Method according to any preceding claim, wherein the inversion (S4) is applied to an S-wave train of the recorded data, in particular to an entire S-wave train including the direct S-wave (S) and its S-coda (C) of the recorded data becomes. Device for determining a seismic damping (Q _i , Q _s , Q _t ) based on a microseismic event (SQ) with a recording device ( 1 ) for providing microseismic data (dw, C) and - a control device ( 2 ) adapted and / or programmed to apply a method according to any preceding claim to provided microseismic data (dw, C) for determining a seismic attenuation (Q _i , Q _s , Q _t ).

Images