PROCEDE DE LOCALISATION ET D'IDENTIFICATION DES ANOMALIES D'UN MILIEU

METHOD FOR LOCATING AND IDENTIFYING SITE ANOMALIES

Abrégé français

Procédé de localisation et d'identification des anomalies d'un milieu. Il est caractérisé en ce qu'il consiste à: utiliser un bloc sismique donné (1) composé de traces sismiques repérées par leurs coordonnées spatiales; délimiter dans ledit bloc sismique au moins un intervalle temporel entre un niveau supérieur (2) et un niveau inférieur (3); sélectionner un modèle temporel (5) d'une anomalie; corréler ledit modèle (5) avec chacune desdites traces (4) comprises dans ledit intervalle temporel; déterminer, pour chaque trace, la corrélation maximum ( GAMMA M) et le temps (ti) correspondant de ladite corrélation maximum; réaliser une carte (7) des corrélations maximales égale aux dimensions spatiales du bloc sismique (1) et une carte (8) desdits temps correspondant aux maxima de corrélation; ladite carte des temps étant de mêmes dimensions et repérée dans le même système d'axes (x, y) que ladite carte des maxima de corrélation.

Abrégé anglais

This invention features a method for locating and identifying
site anomalies. It is characterized by the steps of:
- Using a given seismic block (1) consisting of seismic traces
located by their space coordinates;
- Demarcating in the said seismic block at least one time
interval between a higher level (2) and a lower level (3);
- Selecting a time-model (5) of one anomaly;
- Correlating the said model (5) with each of the said traces (4)
within the said time interval;
- Calculating for each trace the maximum correlation (.GAMMA.M) and
the corresponding time (t i) of the said maximum correlation;
- Drawing a maxima correlation chart (7) equal to the spatial
dimensions of the seismic block (1) and a chart (8) of the said times
corresponding to the correlation maxima;
the said time charg being of the same dimensions and located in the same
system of coordinates (x, y) as the said maxima correlation chart.

Revendications

7
CLAIMS
1. Method for locating and identifying the anomalies of a medium and
consisting in :
- Using a given seismic block (1) composed of seismic traces (4)
located from their spatial coordinates;
- Delimiting in said seismic block at least one temporal interval
between an upper level (2) and a lower level (3);
- Selecting a temporal model (5) of an anomaly ;
- Correlating said model (5) with each of said traces (4) within said
temporal interval;
- Determining, for each trace, the maximum correlation (.GAMMA. M) and time
(t i) corresponding to the said maximum correlation;
- Carrying out a map (7) of the maximum correlations equal to the
spatial dimensions of the seismic block (1) and a map (8) of said times
corresponding to the maxima correlation; said time map having the same
dimensions and referenced in the axial system (x, y) as the said map of the
maxima correlation.
2. Method according to claim 1, wherein the time map is segmented in several
of connex and homogeneous zones (C1 to C4), each connex and homogeneous
zone is such that a point in said zone comprises at least a neighbouring point
not containing discontinuity in time higher than a given threshold (.about.t).
3. Method according to claim 2, wherein are selected all the homogeneous
zones where there is at least one point of maximum correlation above a
predetermined threshold of correlation (S).
4. Method according to claim 2, wherein each homogeneous zone is of a
surface greater than a given value.

8
5. Method according to claim 2 to 4, wherein each homogeneous zone is dealt
individually by a propagator to spread the above mentioned zone in all
directions while controlling the correlation with the neighbouring seismic
traces.
6. Method according to claim 1 to 5, wherein the homogeneous zones and
their extension make anomalous zones which are organized in a number of
layers (10 to 13) in such a way that in each layer, two anomalous zones,
whatsoever, do not cover one another.
7. Method according to claim 6, wherein the layers are in decreasing order
relative to their maxima correlation.
8. Method according to one of claims 1 to 7, wherein each anomaly is
validated by controlling the timecorrelation relationship so that only the
anomalies having a maximum correlation for a minimum time are kept.

Description

, . _.._ __ . .
CA 02225511 1997-12-11
1
METHOD FOR LOCATING AND IDENTIFYING SITE ANOMALIES
This invention is about a procedure to localize and identify
anomalies of a medium like the ones found during seismic campaigns.
The image of a seismic area is generally shown as one or many
two-dimensional seismic section, referred to as seismic 2D, defined by
axis x and t, or by three-dimensional seismic sections, referred to as
seismic 3D, defined by axis x, y and t or z, where t is time and z is depth.
In a seismic block, a seismic event is found partly by one or
many shot points and from the receivers associated with the shot point,
defined by their coordinates at axis x and y, and partly by the time t it
takes to go from the shot point to the corresponding receiver or the depth
z where it is located.
An anomaly is viewed by interpreters as a seismic event. The
study of anomalies in a medium allows for better understanding of the
medium, as some anomalies can be clues on the presence or absence of
hydrocarbons (water and oil) in the given medium.
The detection of anomalies on a seismic section comprised of a
large number of seismic traces previously assembled within given criteria
as for example in common middle points (CMP), at a common receiver,
etc... is done manually by the interpreter. Subsequently, the selection or
rejection of an anomaly depends solely on the judgment of the interpreter
and his ability to interpret the seismic section correctly. Figure 1
represents a seismic section (x, t) on which anomalies, Al and B 1 for
example, are platted by the interpreter. The plotted anomalies can be
considered as either different, identical or of the same nature. As can be
seen in Figure 1, to the left and in the second lower half, many anomalies
Al that overlap can be found, making it difficult to determine their nature.
In the presence of a fault in a medium, it is often faced two anomalies
which are unrelated. These two anomalies are on different levels and can
be viewed as the same from one end to the other of the fault while the
interpreter views each as divided and delimited by the fault without being
able to clear the ambiguity.

CA 02225511 1997-12-11
2
In any case, the manual detection of an anomaly cannot take
into account all anomalies especially those that are barely visible or that
are hidden by other anomalies.
This invention presents a method to localize and identify every
anomaly which exist between two predetermined levels that are either real
or fictitious horizons of the medium to be explored.
This invention consists of a method which consists of:
- Using a given seismic block composed of seismic traces
located from their spatial locations;
- Delimiting in the said seismic block by at least one temporal
interval between an upper level and a lower level;
- Selecting the temporal model of an anomaly;
- Correlating the said model with each of the said traces within
the said temporal interval;
- Determining for each trace, the maximum correlation and
time corresponding to the said maximum correlation;
- Developing a map of the maximum correlations equal to the
spatial dimensions of the seismic block and a map of the said times
corresponding to the maximum correlations, said time map having the
same dimensions and located in the axial system than said map of the
maxima correlation.
An advantage of this invention is that it automatically sweeps a
whole seismic section by marking successive intervals, adjacent or not, to
the above mentioned seismic section.
According to another characteristic of this invention, the map
of times is segmented in an assembly of zones connected and
homogeneous, every one of which being like a point in the above
mentioned zone consists of at least one adjacent point not containing
discontinuity in time superior to a given threshold (At).
Another advantage is to detect every anomaly by their connex
components.
According to another characteristic, every homogeneous zone
where at least one point represents a maximum correlation higher than a
predetermined correlation threshold (S) is selected, every zone
representing advantageously a surface higher than a given value.

CA 02225511 1997-12-11
3
An advantage lies in the fact that every anomaly is taken into
consideration and in that the relevant anomalies of predetermined criteria
can be selected. In this case, every anomaly that is of non or little
relevance are rejected and considered insignificant.
According to another characteristic, every homogeneous zone is
considered individually by a propagator in a way to cover the above-
mentioned zone in every direction while controlling the correlation with
neighboring seismic traces.
Thus, it is possible, thanks to this invention, to detect
anomalies that are covered and difficult to detect manually while limiting
their outline.
According to another characteristic, homogeneous zones and
their extensions constitute anomalous zones that are organized in a number
of layers such that within every layer, two anomalous zones are not
covered by one another, the layers being, for example, in an order
decreasing from the maxima correlation.
An advantage that lies in this characteristic comes from the fact
that anomalies can be categorized in relation to each other, for example
following a decreasing maxima correlation, and the different levels of
anomalies places in memory.
Finally, according to another characteristic, every anomaly is
validated by controlling the time-correlation relations in a way that only
the anomalies with a maximum correlation over a minimum time are kept
which allows, amongst other things, to validate every anomaly and make
marker cards for each anomaly. Every card can include amplitude, origin,
spatial coordinates, surface, etc.
Other characteristics and advantages will become more clear
during the reading of a preferred embodiment of the invention, as well as
drawings in the appendix in which :
- figure 1 is a 2D seismic section;
- figure 2 is a schematic representation of a seismic block
(x, y, t);
- figure 3 is a schematic representation of a seismic trace in a
given interval and the result of the correlation with an
anomalous model;

CA 02225511 1997-12-11
4
- figure 4 is an enlarged view of part of a seismic section (x, t)
containing anomalies;
- figures 5a and 5b are schematic and partial representations of
a map of maxima correlation and a isochronal map,
respectively;
- figure 6 is a schematic representation of related components;
- figure 7 is a schematic representation of a ranking or sorting
of anomalies;
- figure 8 is a schematic representation of a validated anomaly;
- figures 9 and 10 are representations of anomalies sorted on
different levels and originating from the sorting of the
schematic in figure 7.
According to the invention, a 3D seismic block is carried out
that represents the medium (figure 2) and that contains a large number of
seismic traces as a result, for example, of a collection of traces in
common mid point (CMP). In the block 1, we define a delimited temporal
interval by an upper level 2 and a lower level 3 indefined, the upper levels
2 and lower levels 3 corresponding to real or fictitious horizons but that
for all practical purposes, correspond for the considered interval to a
given minimum time t,,,;n and a maximum time t,,,aX.
On figure 3, it is represented a portion of the seismic trace 4
within levels 2 and 3 that correspond to the times tmin and t,,,,,x
respectively.
A model 5, representing an anomaly is represented as a signal.
In a first step, the model 5 is correlated with a portion of trace
4 in a way to get a correlated signal 6 where the maximum correlation is
I,M. The time or index to of the maximum correlation I-'M is taken. Then
this step is carried out for every portion of the seismic traces within levels
2 and 3 in a way to obtain values of maximum correlation r'M and indices
tl. This allows to create two maps 7 and 8 where one 7 corresponds to the
maximum correlation I' (figure 5a) and the other 8 to the index t; (figure
5b), the later called isochronal map. The two maps 7 and 8 are of equal
dimension to the spatial dimensions of block 1 and located in the same
axial system, x, y for example.
In a second step, the anomalies are sorted. A threshold S of
correlation is determined and only those anomalies with a maximum

CA 02225511 1997-12-11
correlation I'M over the threshold S are kept, then each anomaly is
extented to find the connex components of the anomaly, the above
mentioned extensions being performed on the anomalies where the
maximum correlation is greater than S. The search for connex components
5 is done on an isochronal map (figure 6) on which for example four points
P1 to P4 of coordinates (t,, x,), (t2, x2), (t3, x3), and (t4, x4) are
reported
and corresponding to four consecutive seismic traces. Two points, P, Q
belong to a connex component if there is a path formed by the points of
the connex component linking P to Q. Two neighboring points P; (x;, t;)
and Pj (xj, Q belong to the same connex component if I I < At where
At is a value of the predetermined threshold. Thus, on figure 6 points P1
to P3 belong to the same connex component because I t2-tl I < At and
I t3-t2 1 < At. On the other hand, point P4(x4, Q does not belong to this
connex component since I t'-t3 I = Ot' > At.
Another criteria selection or sorting could consist of rejecting
all the anomalies with a maximum correlation higher than S, but where
the size is smaller than a given size.
The connex components C1, C2, C3 and C,t... are then
referenced or numbered so that the isochronal map has numbered connex
components.
In a third step, every anomaly is extended with a propagator to
solve the problem of hidden anomalies.
On figure 4, it can be seen that the anomaly A2 is unique and
has not been overlapped by another nearby anomaly. Anomalies A3 and A4
partially overlap each other but, due to the extension created by the
propagator, they are distinguished from one another with distinct
boundaries.
The extension of every anomaly is carried out in every
direction and is controlled closely by their correlation with the seismic
traces of the boundaries or outline of the above mentioned anomaly with
neighbouring seismic traces. The extension of the anomaly is stopped
when the above mentioned controlled correlation becomes a maximum
correlation below the threshold S. Thus, the propagator finds the
complement of the anomaly, partially hidden by one of several other
anomalies.

CA 02225511 1997-12-11
6
In a fourth step, the anomalies which are stretched by the
propagator are classified in the form of multiple maps whereby each
contains anomalies which do not cover each other. Preferably, the map of
these anomalies (figure 7) are classified by the decreasing order of their
maximum correlation. The upper map 10 in figure 7 corresponds to the
highest maximum correlation while the lower map 11 corresponds to the
weakest maximum correlation, the two other maps 12 and 13
corresponding to intermediate maxima correlation.
In a fifth step, the anomalies are validated by pointing out the
maxima correlation corresponding to minimum times (figure 8).
It is possible to establish marker cards for each anomaly, every
marker card containing information relating to the above mentioned
anomaly like for example, the amplitude, the size, the surface, the origin,
etc...

Dessin représentatif