文档库 最新最全的文档下载
当前位置:文档库 › 地震属性约束地质建模001

地震属性约束地质建模001

地震属性约束地质建模001
地震属性约束地质建模001

?.Journal of Petroleum Science and Engineering 28200065–79

www.elsevier.nl r locate r jpetscieng

Integrating seismic attribute maps and well logs for porosity modeling in a west Texas carbonate reservoir:addressing the

scale and precision problem

Tingting Yao a,),Andre G.Journel b,1

a

Department of Geological and En ?ironmental Sciences,Stanford Uni ?ersity,Stanford,CA 94305,USA

b

Department of Petroleum Engineering,Stanford Uni ?ersity,Stanford,CA 94305,USA

Received 7February 2000;accepted 14July 2000

Abstract

A 3-D porosity field in a west Texas carbonate reservoir is modeled conditional to both A hard

B porosity data sampled by wells and 2-D seismic attribute map with less vertical resolution than the well log data.The difference-of-scales between the two sources of data is resolved by a prior 2-D estimation of vertically averaged porosity using well and seismic data.These 2-D estimates are then used to condition the 3-D stochastic simulation of porosity.The algorithm used to merge 2-D average values and 3-D data values,i.e.,to solve the difference-of-scale problem,is a form of block kriging,which ensures that vertical averages of the 3-D estimates reproduce exactly the 2-D average data values.The precision of the 2-D conditioning data is also addressed.Several new geostatistical algorithms,such as automatic covariance modeling and direct sequential simulation algorithms,are weaved into the application.These new algorithms facilitate the process of integration of the soft data in petrophysical modeling.q 2000Elsevier Science B.V.All rights reserved.

Keywords:reservoir;carbonate;geostatistics;porosity;modeling;seismic

1.Introduction

Geostatistics is used in the petroleum industry to model the spatial distribution of petrophysical prop-erties,such as porosity.These 3-D numerical models of petrophysical properties provide the input into the

)Corresponding author.Current address:ExxonMobil Up-stream Research Company,ST 3209,PO Box 2189,Houston,TX 77252-2189,USA.Tel.:q 1-713-431-7174;fax:q 1-713-431-6336.

?.E-mail addresses:tingting@https://www.wendangku.net/doc/e210078914.html, T.Yao ,?.journel@https://www.wendangku.net/doc/e210078914.html, A.G.Journel .1

Tel.:q 1-650-723-1594.

flow simulator for reservoir performance prediction.

Because sampling with well log data is sparse,it is critical to integrate information from seismic data,?which delivers better areal coverage Abrahamsen et .al.,1996;Fournier,1995;.However,the utilization of seismic data for modeling 3-D petrophysical prop-erties faces some severe problems,possibly the most ?important being that of scale difference Haas and .Dubrule,1994;Gorell,1995;Doyen et al.,1997.?.The well data usually considered as the A hard B data are defined on a much smaller volume support than the seismic data.Well data are distributed in the 3-D ?space providing high vertical resolution assuming

0920-4105r 00r $-see front matter q 2000Elsevier Science B.V.All rights reserved.?.PII:S 0920-41050000068-1

()

T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering28200065–79 66

.

vertical wells,while seismic data often do not pro-vide the same vertical resolution and represent only some vertically averaged information over the layer being modeled.Integration of data defined on such different scales is a difficult challenge.Behrens et al.?.?.

1996and Deutsch et al.1996offer reviews of algorithms presently used to address this challenge. Most algorithms address the scale difference by, implicitly or explicitly,repeating the2-D seismic attribute map along each vertical slice,thus,creating artificially dense3-D seismic data and generating vertical banding artifacts in the resulting petrophysi-cal model.Simulated annealing algorithms can by-pass this problem,but they require delicate tuning of the annealing schedule parameters to reach conver-gence.In addition,they may be CPU intensive.

?.

Following an original lead by Xu et al.1992,

?.

Behrens et al.1996suggested integrating the seis-mic data by first performing a2-D cokriging of the vertically averaged porosity values using vertical average well data and the2-D seismic attribute map as a covariate.The resulting estimated vertically averaged porosity map is then used to condition a 3-D stochastic simulation of the3-D porosity field. In this second step,the authors solve the difference-of-scale problem with a A block kriging B procedure, whereby the previously A estimated B vertically aver-aged porosity values are assimilated to A true B actual vertically averaged data values.Thus,the error of estimation of the2-D vertically averaged values is ignored.In addition,the block kriging is performed

?

on the normal score transforms of the data both well

.

data and seismic data.The vertical linear averaging is not preserved by such non-linear transform.What is needed is a simulation algorithm which can oper-ate directly on the original data without any prior non-linear transform.Nevertheless,the results shown

?.

by Behrens et al.1996applied to two different Nigerian fields are remarkable,and their algorithm is worth revisiting and correcting.

In this paper,the direct sequential simulation algorithm is used to avoid the non-linear normal score transform as in sequential Gaussian simulation, hence,the vertical linear averaging is preserved in the final model.The precision problem of the previ-ously estimated2-D conditioning data is addressed by associating those2-D conditioning data with an error item,which is represented by their estimation variance.Some other new geostatistical algorithms, such as automatic covariance modeling approach,are weaved into this application.

2.Methodology

?.

In a recent paper,Journel1999recalled a little known property of cokriging estimates,which allows matching,not only the hard data values at their locations,but also,any volume averaged data values, as long as the averaging process is linear.

In the following application,we used2-D cokrig-ing to generate an A estimated B2-D field of vertically averaged porosity values conditioned to both seismic and well data.Instead of an analytical model for the ?.

cross covariances,we used the newly developed

?.?technique of modeling cross spectrum tables Yao

.

and Journel,1998.

Next,we proceeded to a3-D cokriging of poros-ity values conditioned to the3-D well data and to the previously estimated2-D vertically averaged poros-ity values.This3-D kriging accounts for the error variance associated with the previous estimation of the2-D vertically averaged porosity values:esti-mated vertical averages are not reproduced exactly, whereas actually measured vertical average values at well locations are matched exactly.The3-D porosity field honors the sample cross-correlation between the 2-D averaged porosity and the2-D seismic attribute map.

Last,direct sequential simulation of3-D porosity values is performed by drawing values from a posi-

?.

tive distribution lognormal,whose mean is the estimated porosity value and the variance,the corre-

?.

sponding kriging error variance.This results in several alternative,equiprobable,3-D realizations of the porosity field which:

?match the hard3-D porosity data along the wells;?reproduce the sample scattergram between the vertically averaged porosity data and the collo-cated seismic data;and

?reproduce the sample3-D porosity histogram and variogram.

The realizations do not show any artifact of smoothing or vertical banding.The implementation

()

T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering28200065–7967

of the method is documented in detail in the follow-ing case study.

3.A west Texas carbonate reservoir

The case study developed here utilizes well log and seismic data from a carbonate reservoir in a Permian basin in west Texas.The production interval is a shallowing upward,prograding,carbonate shelf sequence composed primarily of alternating

?. dolomites and siltstones Chambers et al.,1994.The dolomites have slightly poorer reservoir quality than the siltstones because of hydritic plugging of the pore space.Within the study area,there appears to be a trend of decreasing siltstone proportion to the west with a corresponding decrease in porosity.The local complexity of the dolomite r siltstone geome-tries did not allow separating these two facies from the data available.In the following presentation,all data and coordinate values were scaled to preserve confidentiality.

The volume of study here is limited to a well defined producing layer that has an area of10400=?.?. 10400ft3170=3170m and is50ft15m thick. This layer is seen in a seismic profile from a3-D survey as a single reflection event.Therefore,a2-D seismic attribute map represents the vertically aver-aged porosity rather than the individual3-D porosity values.The corresponding estimation r simulation grid is discretized into65=65horizontal grid nodes

?. having50vertical layers,each1ft0.305m thick.

?. The horizontal cell size is160=160ft50=50m, approximately that of the seismic data available.

The study focuses on the integration of the2-D seismic attribute map for the estimation r simulation of a3-D volume of porosity values,each defined on a very small support volume assimilated to a quasi A point B support.The hard porosity data are log-de-rived from62wells within the study area A.The location map and three of the well–log curves from the62wells are shown in Fig.1:the greyscale represents the vertically averaged porosity values expressed in percentage.The corresponding his-tograms of the3-D log-derived porosity values and 2-D vertically averaged porosity values are given in Fig. 2.The porosity sample histogram shows an 8.40%mean and positive

skewness.

?.

Fig.1.A Location map of the62wells.The grey scale repre-sents the value of the vertical averaged porosity at each well ?.

location.B Porosity curves from the three wells.

A number of2-D seismic attribute maps were available to crossplot with the vertically averaged

()T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering 28200065–79

68Fig.2.Histograms of log-derived 3-D porosity data and vertically averaged 2-D porosity data.

porosity data.The seismic attribute,which had the highest collocated correlation with the vertically av-eraged porosity,was the low-frequency reflection energy.There is no specific rock physics considera-tion relating this seismic attribute to vertically aver-aged porosity.The low-frequency reflection energy was retained because it has a higher correlation with the vertically averaged porosity based on pure statis-tical grounds.Fig.3shows the greyscale map of the seismic data.Fig.4shows the scattergram of the 62collocated pairs of seismic vs.vertically averaged porosity values.The higher seismic values in the NE corner are consistent with the higher sample porosity values of Fig.1.The 2-D correlation coefficient is significant at 0.60.

3.1.Cokriging ?ertically a ?eraged porosity The first step of the study consists of generating a 2-D field of vertically averaged porosity values that

will be used to condition the final 3-D simulation of point-support porosity.Cokriging was used whereby,at each location x of the 65=65seismic grid,the a vertically averaged porosity is estimated from a lin-ear combination of up to 12neighboring vertically ?.averaged porosity data f from wells and 12seis-mic data:

12

)f x y m s

l f x y m ?.?.Y

f a a f

11a s 1

112

q

h s x y m 1?.?.

Y

a a s

22a s 1

2?.where x is the set of horizontal coordinates,f x a 1is the vertically averaged porosity data at the well ?.location x ,and s x is the seismic data at a a 12location x .The m and m are the corresponding a f s 2porosity and seismic mean values,l

and h are a a 12the cokriging weights.

Such cokriging requires three covariance models,the porosity covariance,the seismic covariance and

Fig.3.Map and histogram of the seismic data.

()T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering 28200065–7969

Fig. 4.Calibration cross-plot between the vertically averaged porosity and the collocated seismic data.

the cross-covariance.In this study,instead of follow-ing the traditional route of calculating experimental covariances and modeling them with a linear model ?.of coregionalization Goovaerts,1997,the direct covariance table approach proposed by Yao and ?.Journel 1998was retained.More precisely,1.the three experimental correlogram tables are cal-culated and the results are presented in Fig.5?.Deutsch and Journel,1998,

2.these correlogram tables are smoothed and filled-in for missing entries by a preliminary smoothing,

3.the completed correlogram tables are transformed into quasi-spectrum tables through fast Fourier ?.transform FFT ,

4.the quasi-spectrum tables are corrected into per-?missible spectrum tables positive definiteness .condition ,and

5.they are inverse Fourier transformed into permis-sible,jointly positive definite correlogram tables ?.Fig.

6.All correlogram and cross-correlogram values re-quired by the cokriging process are read directly from the tables of Fig.6.This approach,in addition to being fast and easy to implement,allows a more accurate modeling of the original experimental cor-relogram values,and it is not constrained by any closed-form analytical expression.

The greyscale map of the cokriging estimates of the vertically averaged porosity and their histogram is given in Fig.7.Note that the estimated values are

Fig.5.Experimental auto-and cross-correlograms of the vertically averaged porosity and seismic attribute.

()T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering 28200065–79

70Fig.6.Final smoothed auto-and cross-correlograms of the verti-cally averaged porosity and seismic

attribute.

Fig.7.Grey-scale map and histogram of estimated average poros-ity using cokriging.

high in the NE corner due to the influence of the seismic data.Also,the influence of the conditioning well data is reflected in the central lower half of the study area.The overall estimated mean is f s 8.25,a value slightly less than the sample mean of 8.40.This smaller value can be explained by the decluster-ing effect of kriging,which underweights the cluster of wells with high porosity values in the NE corner,see location map of Fig.1.

The scattergram of estimated,vertically averaged porosity values vs.the collocated seismic value is given in Fig.8:the reproduction of the sample scattergram of Fig.4is good,although the correla-?.tion coefficient is too high 0.71instead of 0.60.This is typical of cokriging.

An alternative to the full cokriging approach of ?.?Eq.1could have been collocated cokriging Xu et .al.,1992,retaining the single seismic datum value ?.s x at the location x being estimated.We chose to

()T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering 28200065–79

71

Fig.8.Scattergram of estimated vertically averaged porosity vs.the seismic low-frequency attribute.

?.present the more general formulation as in Eq.1,which would be required if the simulation grid is much finer than the seismic grid.

As expected from any regression-type estimation algorithm,cokriging yields a smoothed image.The ?.2variance of the estimated values in Fig.7is 1.62,a value smaller than the corresponding variance of ?.2?.the well data,i.e. 1.91Fig.2.3.2.3-D cokriging of porosity

A 3-D estimated map of quasi point-support porosity values is generated conditional to the 3-D ?.porosity data from wells and the previously ob-tained 2-D vertically averaged porosity values which carry the seismic information.Consider the collo-cated cokriging estimate:

n

)f u y m s

l f u y m ?.?.Y

f a a f

a s 1

)q l f x y m 2?.?.

0f

?.where u s x ,z is the set of 3-D coordinates:x is the vector of horizontal coordinates and z is the )?.depth coordinate,f u is the estimated 3-D poros-)?.?.ity,f u are the 3-D porosity data,x is the a previously estimated vertically averaged porosity at horizontal location x ,l and l are the correspond-a 0ing kriging weights,and m is the global mean f value of porosity.

The corresponding cokriging system is given in Appendix A.

If this cokriging process is applied,the resulting )?.3-D estimated values f u are fully exact in the sense that:

?.1.they match the well data values f u at their n a 3-D locations u :

a f )u s f u ,;a s 1,...,n ,and 3?.?.?.

a a 2.their vertical averages match the 2-D conditioning data:

N z

1))f x ,z s f x ,

;x 4?.?.?.

Yk N z

k s 1

where k is the index of the vertical level z and N k z ?is the number of vertical layers N s 50in this case z .study .

This 3-D cokriging process requires inference of only the 3-D porosity variogram model.

The analytical semivariogram model retained is ?.sill standardized to 1:

g h ?.2

2

2

h h h x y

z

s 0.6Sph

q

q

(?

/?/?/?/

3000

1000

12

q 0.4Sph =

2

2

2

h h h x

y z

q

q

(?

/?

/?/?/

6000

30000

50

5?.

where Sph designates a spherical model of unit-range ?.and sill,h s h ,h ,h are the coordinates of a x y z separation vector h in the EW,NS and vertical directions,respectively.

This anisotropic model is displayed in Fig.9,together with the corresponding sample semivari-ograms.

)?.In order for the 3-D cokriging estimates f u to match the vertically averaged data,the same data

()

72

T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering28200065–79

?.

Fig.9.Experimental variograms of3-D porosity and the fitted model in three directions sill standardized to1.

()T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering 28200065–7973

?.configuration must be used for all locations x ,z k along the same vertical at x .In this study,the 50?.locations x ,z ,k s 1,...,50are estimated from k the same 33data,more precisely:

?four data points taken at regular intervals z X ,k k X s 10,20,30,40from the eight wells closest to location x ;and

)?.?the collocated,vertically averaged value x ,as obtained from the previous 2-D cokriging us-ing the seismic data.Fig.10gives two vertical sections,NS at x s 30and EW at y s 30.The vertical bands seen in Fig.10are the consequences of conditioning the esti-mates of porosity along a vertical column to the same vertically averaged datum value.This banding is actually a by-product of the smoothing effect of kriging and will be corrected by the process of ?.simulation;see later Fig.16.Behrens et al.1996noted the same banding artifact with a 3-D cokriging approach,which was also corrected in their simula-tion results.

Fig.11is a crossplot to verify that the vertical averages of the 3-D estimated porosity values match exactly the conditioning 2-D vertical averages.

The characteristic smoothing of kriging is re-vealed by the histogram and q –q plot of Fig.12.The estimation process is exact and unbiased:it reproduces exactly the mean 8.25of the 2-D condi-tioning data of Fig.7,but the variance of the 3-D ?.2estimated values is only 2.0,much lower than

the

Fig.10.Two vertical sections of the estimated 3-D porosity.Artifact banding will be corrected by

simulation.Fig.11.Scatterplot of vertical averages of the 3-D estimated porosity vs.the previously estimated 2-D average porosity.

?.2sample variance 3.37given in Fig.2.Again,the process of simulation will correct that

smoothing.

Fig.12.Histogram of the estimated 3-D porosity values and their quantile plot with the well-porosity data.

()T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering 28200065–79

74Fig.13.Scatterplot of vertical averages of the 3-D estimated porosity vs.the previously estimated 2-D average porosity,ac-counting for precision of the 2-D conditioning data.

3.3.Addressing the precision problem

In the A block B kriging system,providing the esti-)?.?.mate f u defined in Eq.2,if the covariances )?.associated with the estimates f x are made equal to the covariances associated with the true vertical ?.averages f x ,the exactitude property of kriging ?would apply,resulting in the exactitude relation Eq.)?..?.4.The data f x are,however,estimated val-ues,hence,they should be reproduced only up to

?.2?.their estimation kriging variances s x which

SK are generated as by-products of the 2-D kriging process.If the covariances associated with the esti-)?.mates,x ,account for the corresponding estima-2?.?tion variance s x ,the exactitude relation Eq.

SK ?..?4prevails only at well locations x see Appendix a .)?.A .More precisely,the 3-D estimated porosity f u values are now such that:

?.1.they match the 3-D well data f u at their n a 3-D locations u :

a f )u s f u ,;a s 1,...,n ,and 6?.?.?.

a a 2.their vertical averages are equal to the 2-D mea-sured values,only at the well locations x :

a N z

1)f x ,t s f x 7?.?.

?.

Ya z a N z

k s 1

The scatterplot between the vertical averages of the estimated 3-D porosity values and the condition-ing 2-D vertical averages shown in Fig.13indicates that the estimated vertical averages are not repro-duced exactly.This is due to the consideration of the precision of the 2-D conditioning data.For partial-?.penetrating wells,the f x could be approximated a by partial vertical average,associated with some error to take account for the precision problem.3.4.3-D simulation of porosity

To preserve the linear relation between the 2-D vertically averaged porosity and the 3-D quasi-point porosity,direct sequential simulation is performed without any normal score transform.The theory of direct sequential simulation calls for drawing simu-?l .?.lated values ff

u of the 3-D porosity from a

Fig.14.Histogram of simulated 3-D porosity and quantile plot vs.well data.

()T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering 28200065–79

75

Fig.15.Two different horizontal sections of the simulated 3-D porosity

values.

Fig.16.Two vertical sections of the simulated 3-D porosity

values.Fig.17.Histogram of vertical averages calculated from the 3-D simulated realization and scatterplot with the conditioning 2-D average porosity of Fig.

7.

Fig.18.Scattergram of simulated vertically averaged porosity vs.collocated seismic data.

()

76

T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering28200065–79

?.?.?.

Fig.19.Sample variogram dots,input variogram model continuous line and the variogram of simulated values dash line in three directions.

()T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering 28200065–7977

sequence of conditional distributions whose means and variances are identified by the kriging means

)?.2?.f u and variances s u .For this kriging,we

K retain the algorithm that accounts for the precision of the estimated 2-D vertical average data.The condi-tional distribution type can be anything.For this application,we retain a lognormal distribution type,which yields only positive simulated values.The lognormal parameters a and b 2of each lognormal distribution at location u are given by the relation:

a s ln f )u y

b 2r 2

?.b 2s ln 1q s 2

u r f )2u 8?.?.?.

?.

K A random number g is then drawn from a stan-dard normal distribution,the simulated porosity value at u is:

?l .w x ff u s exp a q b g 9?.?.

?.The superscript l indicates that the value relates ??l .?.4to the l th simulated realization ff u ,u g A .?l .?.That simulated value ff u is immediately stored as a datum value to be used for cokriging of )?X .X f u at any subsequent and neighboring node u ;recall the sequential simulation paradigm, e.g.in ?.)?.Goovaerts 1997.The cokriging of f u considers for data the previously obtained,2-D estimated,ver-tically averaged porosity of Fig.7in addition to the 3-D original f data and the neighboring,previously simulated ff ?l .values.

The last step of the simulation algorithm consists ??l .?.4of transforming each realization ff u ,u g A to approximate the sample porosity histogram of Fig.2,or any other target distribution deemed appropri-ate.The rank order-preserving transform algorithm and program trans have been used for this purpose ?.Deutsch and Journel,1998.This algorithm pre-?l .?.serves the exactitude of the simulation ff u vs.the hard 3-D well data,i.e.,the final simulated ?l .?.values f u are such that:

f ?l .u s ff ?l .u s f u ,;a s 1,...,n

?.?.?.a a a 10?.Fig.14shows that the target sample histogram ?.well data has been well approximated by the simu-??l .?.4lated values f u ,u g A of the first realization ?.l s 1produced;compare to the sample statistics of the well data in Fig.2.

Figs.15and 16show several horizontal and vertical sections from the first simulated realization.When compared to estimation as depicted in Fig.10,it appears that the process of simulation has cor-rected the artifact smoothing and vertical banding of estimation.

To check the impact of conditioning the simula-tion to vertical averages,the 3-D simulated porosity values were vertically averaged.The histogram of these simulated vertical averages and their scatter-gram with the corresponding 2-D vertical averages of Fig.7are given in Fig.17.Although the simula-tion does not reproduce exactly at each location x ?.the conditioning estimated vertical average value,the statistics are closely reproduced and the collo-cated 2-D correlation is 0.90.Note that since the )?.vertically averaged values f x are only esti-?.mated,they need not should not be reproduced exactly.

Reproduction of the sample correlation of verti-cally averaged porosity vs.seismic data is checked through Fig.18and compared to Fig.4.Both the ?.non-linear shape and the correlation value 0.60of the sample cloud are well reproduced by the simula-tion.The exact match of the correlation value is a coincidence.

Fig.19provides a final check,plotting the three-directional variograms calculated from the simulated realizations vs.the input variogram model and the sample variogram calculated from the original well data.

The variograms from the simulated realization are reasonably close to the sample variograms,actually closer to them than to the input model.In the presence of dense conditioning data,as is the case ?.here 62wells ,the sample statistics prevail over whatever model is given as input:this is a positive consequence of exact conditioning to the well data values.

4.Conclusions

The block kriging procedure initially proposed by ?.Behrens et al.1996is improved to provide an algorithm for simulating 3-D fields of a petrophysi-cal property conditioned to 3-D hard data and 2-D ?seismic attribute maps with lower vertical resolu-

()

T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering28200065–79 78

.

tion.The difference-of-scale problem is solved by generating first a2-D estimated field of vertically averaged values conditioned to the2-D seismic at-tribute map,followed by a3-D A block B kriging-type estimation to generate alternative simulated3-D real-izations by direct-sequential simulation.

This paper weaves together several recently de-veloped geostatistical algorithms for data processing and integration,more precisely:

1.A fast modeling of covariance and cross-covari-

ance tables with FFT,which does not require choosing a parametric analytical model and cir-cumvents completely the tedious fitting of a linear model of coregionalization.

2.The concept of A block kriging B,which allows

exact reproduction of large-scale data values ex-pressed as linear averages of small-scale attribute values.This algorithm is critical for integrating data defined on different volume supports.

3.The utilization of an estimation variance to ac-

count for the precision of the A data B,which are not actually measured,but are the result of a prior

?.

estimation process see Appendix A.

4.The direct sequential simulation algorithm,which

generalizes the well established sequential Gauss-ian simulation algorithm.No prior normal score transform is required;simulation is performed on the original data values,which allow preserving the linear average characteristic of some data and

??.

the related exactitude property see point2just .

above.

Any of the four previous algorithms may not be familiar to the reader.The original contribution of this paper,however,is the combination of these four algorithms into a general methodology for the inte-gration of data of different sources and resolution accounting for their reliability.

Application to seismic data integration for3-D porosity mapping within a west Texas carbonate field has yielded excellent results with none of the artifact smoothing associated to using data with widely different resolutions.

Nomenclature

a mean of the lognormal distribution

b2variance of the lognormal distribution g variogram function

s2kriging variance

K

m mean porosity value

f

m mean seismic value

s

s seismic value

z the k th vertical level in3-D model

k

?.

u x,z,3-D coordinates vector

k

A interested study area

f porosity value

f)estimated porosity value

ff?ll.simulated porosity value

f?l.transformed porosity value to identify tar-get histogram

(.x vertically averaged porosity value at2-D location x

;for all

g within

Appendix A.Block cokriging accounting for scale and precision

?.

Recall Eq.2of the3-D cokriging estimate:

n

)

f u y m s l f u y m

?.?.

Y

f a a f

a s1

)

q l f x y m A-1

?.?.

0f

)?.

where f x is the estimated2-D vertical average porosity at horizontal location x using both3-D well data and2-D estimates of vertically averaged poros-ity values.These estimates carry the seismic infor-

?. mation.The corresponding true vertical average f x can be written as:

)

f x s f x q e x A-2?.?.?.?.

?.

where the error e x is,by definition of kriging ?.

Journel and Huijbregts,1978,orthogonal to the )?.

estimator f x,in which case:

)2

Var f x s Var f x y s x G0A-3?.?.?.?.

?4?4

SK

2?.??.4?. where s x s Var e x is the kriging error SK

variance at x.

The previous relation expresses the smoothing

)?.

effect of kriging:the estimated value f x displays a variance deficiency equal to the kriging variance.

()T.Yao,A.G.Journel r Journal of Petroleum Science and Engineering 28200065–7979

?.Assuming independence of the error e x at x ?X .with any variable f u at a location with horizontal coordinates x X /x ,the following covariance is writ-ten:

X X )Cov f u ,f x s Cov f u ,f x q 0

?.?.?.?.?4?4A-4?.with for the point-to-vertical average covariance:X Cov f u ,x ?.?.?4N z

1X X s Cov f u ,f x ,z s C u ,x ?4?.?.?.

Y

k N z

k s 1

A-5?.

X ?.Thus,the covariance value u ,x is merely ?X averaged from the point 3-D covariance model C u .y u .

The A block kriging B system corresponding to the 3-D kriging estimate can nowbe developed as ?.Journel and Huijbregts,1978:

n

l C u y u q l C u ,x s C u y u ,

?.?.?.Y

b b a 0a a b s 1a s 1,...,n A-6?.

n

)l C u ,x q l Var f x s C u ,x ?.?.?.

?4Y

b b 0b s 1

)2??.4??.4?.where Var x s Var x y s x .

SK 2?.If s x s 0,that is,if the estimated data value SK )?.?.x is assimilated to the true value x ,then the exactitude property of kriging would apply,en-?.2?.tailing relation 3.If s x /0,then the data

SK )?.value f x is not any more reproduced exactly.

)?.The weight l given to the data value f x in the 0cokriging expression is controlled by its estimation

2?.variance s x .

SK References

Abrahamsen,P.,Hektoen,A.L.,Holden,L.,Munthe,K.L.,1996.Seismic impedance and porosity:support effect.In:Baffi,?.Schofield Eds.,Geostat.Wollongong 96vol.1Kluwer Aca-demic Publishing,pp.489–500.

Behrens,R.A.,Macleod,M.K.,Tran,T.T.,1996.Incorporating seismic attribute maps in 3-D reservoir models.SPE 36499,31–36.

Chambers,R.L.,Zinger,M.A.,Kelly,M.C.,1994.Constraining geostatistical reservoir descriptions with 3-D seismic data to reduce uncertainty.Stochastic Modeling and Geostatistics;Principles,Methods,and Case Studies.In:Yarus,J.M.,Cham-?.bers,R.L.Eds.,Comput.Appl.Geol.3pp.143–157.

Deutsch,C.V.,Srinivasan,S.,Mo,Y.,1996.Geostatistical reser-voir modeling accounting for precision and scale of seismic data.SPE 36497,9–15.

Deutsch,C.V.,Journel,A.G.,1998.GSLIB:Geostatistical Soft-ware Library and User’s Guide.Oxford Univ.Press.

Doyen,P.M.,Psaila,D.E.,den Boer,L.D.,1997.Reconciling data at seismic and well log scales in 3D Earth modeling.SPE Paper 38698.1997SPE Annual Technical Conference,San Antonio,TX.pp.465–474.

Fournier,F.,1995.Integration of 3D seismic data in reservoir stochastic simulations:a case study.SPE Paper 30564.1995SPE Annual Technical Conference,Dallas,TX.pp.343–356.Goovaerts,P.,1997.Geostatistics for Natural Resources Evalua-tion.Oxford Univ.Press.

Gorell,S.B.,1995.Creating 3D reservoir models using areal geostatistical techniques combined with vertical well data.SPE Paper 29670.Western Regional Meeting,Bakersfield,CA.pp.967–974.

Haas,A.,Dubrule,O.,1994.Geostatistical inversion:a sequential method of stochastic reservoir modeling constrained by seis-?.mic data.First Break 1211,561–569.

Journel,A.G.,Huijbregts,Ch.J.,1978.Mining Geostatistics.Aca-demic Press.

Journel, A.G.,1999.Conditioning geostatistical operations to ?.non-linear volume averages.Math.Geol.318,931–953.Xu,W.,Tran,T.T.,Srivastava,R.M.,Journel,A.G.,1992.Inte-grating seismic data in reservoir modeling:the collocated cokriging alternative.SPE 24742,833–842.

?.Yao,T.,Journel, A.G.,1998.Automatic modeling of cross covariance tables using fast Fourier transform.Math.Geol.30?.6,569–615.

河南省地质矿产勘查开发局局直属单位变更名称

变更后名称 变更前名称 1 河南省地质调查院 河南省地质调查院 2 河南省地质矿产勘查开发局第一地质矿产调查院河南省地质矿产勘查开发局第一地质调查队 3 河南省地质矿产勘查开发局第二地质矿产调查院河南省地质矿产勘查开发局第二地质队 4 河南省地质矿产勘查开发局第三地质矿产调查院河南省地质矿产勘查开发局第三地质调查队 5 河南省地质矿产勘查开发局第四地质矿产调查院河南省地质矿产勘查开发局第十一地质队 6 河南省地质矿产勘查开发局第一地质勘查院 河南省地质矿产勘查开发局第一地质勘查院 7 河南省地质矿产勘查开发局第二地质勘查院 河南省地质矿产勘查开发局第二地质勘查院 8 河南省地质矿产勘查开发局第三地质勘查院 河南省地质矿产勘查开发局第三地质探矿队 9 河南省地质矿产勘查开发局第四地质勘查院 河南省地质矿产勘查开发局第四地质探矿队 10 河南省地质矿产勘查开发局第五地质勘查院 河南省地质矿产勘查开发局第一地质工程院

11 河南省地质矿产勘查开发局第一地质环境调查院 河南省地质矿产勘查开发局第一水文地质工程地质队 省地矿物资供应中心 12 河南省地质矿产勘查开发局第二地质环境调查院 河南省地质矿产勘查开发局第二水文地质工程地质队 13 河南省地质科学研究所 河南省地质矿产勘查开发局区域地质调查队 14 河南省航空物探遥感中心 河南省地质矿产勘查开发局地球物理勘查队 省探矿机械院 15 河南省地质矿产勘查开发局测绘地理信息院 河南省地质矿产勘查开发局测绘队 16 河南省岩石矿物测试中心 河南省岩石矿物测试中心 17 河南省地质工程技术学校 河南省经贸工程技术学校 18 河南省地质职工学校 河南省地质职工学校

地震资料解释基础 复习题

地震解释基础 复习题 1.为什么并非每一个地质界面都对应一个反射同相轴? 子波有一定的延续长度,若地层很薄,相邻分界面的信号可能会重叠到一起形成复合波,导致无法分辨界面。所以一个反射同相轴可能包含多个地质界面。 2.影响地震资料纵向分辨率的因素有哪些?提高分辨率的实质是什么? 1)激发条件——激发宽频带子波——井深、药量、激发岩性、虚反射、激发组合 2)接收条件——检波器类型、地表岩性、检波器耦合、组合方式、仪器响应 3)近地表低降速带的影响 4)大地滤波作用、地层速度 实质:提高主频,拓宽频带 3.提高横向分辨率的方法是什么?为什么它能提高横向分辨率? 偏移是提高地震勘探横向分辨率的根本方法 提高横向分辨率的核心是减小菲涅尔带的大小, 菲涅尔带的极限 : 要想减小菲涅尔带的大小就要减小h ,偏移将地表向下延拓到地下界面,使h=0,所以 菲涅尔带减小到极限L=λ/4,所以偏移能提高横向分辨率。 4.地震剖面的对比方法 1)掌握地质规律、统观全局 在对比之前,要收集和分析勘探区的各种资料。研究规律性的地质构造特征,用地质规律指导对比解释。了解地震资料采集和处理的方法及相关因素,以便准确识别和判断出剖面假象。 2)从主测线开始对比 在一个工区有多条地震剖面,应先从主测线开始对比工作,然后从主测线的反射层延伸到其他测线上去。(主测线:指垂直构造走向、横穿主要构造,并且信噪比高、反射同相轴连续性好的测线。它还应有一定的延伸长度,最好能经过钻探井位。) 3)重点对比标准层 对某条测线而言,可能有几个反射层,应重点对比目标层(或称为标准层,标准层:具有较强振幅、连续性较好、可在整个工区内追踪的目标反射层。它往往是主要的地层或岩性的分界面,与生油层或储集层有一定的关系,或本身就为生油层、储油层)。 4)相位对比 反射波的初至难以辨认,采用相位对比。若选振幅最强、连续性最好的某同相轴进行追()222042164h L O C h h h λλλλ??'==+-=+== ???

甘肃省地质矿产勘查开发局第三地质矿产勘察院_中标190923

招标投标企业报告 甘肃省地质矿产勘查开发局第三地质矿产勘察 院

本报告于 2019年9月23日 生成 您所看到的报告内容为截至该时间点该公司的数据快照 目录 1. 基本信息:工商信息 2. 招投标情况:中标/投标数量、中标/投标情况、中标/投标行业分布、参与投标 的甲方排名、合作甲方排名 3. 股东及出资信息 4. 风险信息:经营异常、股权出资、动产抵押、税务信息、行政处罚 5. 企业信息:工程人员、企业资质 * 敬启者:本报告内容是中国比地招标网接收您的委托,查询公开信息所得结果。中国比地招标网不对该查询结果的全面、准确、真实性负责。本报告应仅为您的决策提供参考。

一、基本信息 1. 工商信息 企业名称:甘肃省地质矿产勘查开发局第三地质矿产勘察院统一社会信用代码:/工商注册号:/组织机构代码:/法定代表人:/成立日期:/企业类型:/经营状态:/注册资本:/ 注册地址:/ 营业期限:/ 至 / 营业范围:/ 联系电话:*********** 二、招投标分析 2.1 中标/投标数量 9 企业中标/投标数: 个 (数据统计时间:2017年至报告生成时间)

2.2 中标/投标情况(近一年) 企业近十二个月中,中标/投标最多的月份为,该月份共有个投标项目。 2019年06月1 序号地区日期标题中标情况1甘肃2019-08-22甘肃省地质矿产勘查开发局第三地质矿产勘察院中标 2兰州2019-07-09西固区范广片区兰州石化公司工业固体废物填埋场场地初步调查 及风险评估报告编制服务项目 中标 3甘肃2019-06-20甘肃省地质矿产勘查开发局第三地质矿产勘察院中标2.3 中标/投标行业分布(近一年)

GeoFrame_地震属性分析和应用

SIS 软件软件技术应用技术应用技术应用之一之一 斯伦贝谢伦贝谢科技服务科技服务科技服务((北京北京))有限公司 2007年3月 GeoFrame 地震属性分析和应用地震属性分析和应用

1 地震属性分析和应用 应用地震属性开展储层横向预测是地震资料综合解释的重要研究内容。随着地球物理理论、数学理论的不断发展,通过各种计算方法能够提取和分析的地震属性越来越多,如何从众多的地震属性中选择能够反映客观地质现象的属性对目的层储层开展分析,这是地球物理人员在实际工作中面对的一个主要问题。 GeoFrame 综合地学平台为地球物理人员开展储层横向预测研究提供了一套完善的工具。SATK 、SeisClass 、LPM 以及GeoViz 的组合应用,可以帮助研究人员完成从属性提取、属性优化、定性分析到定量计算的储层预测全过程。本文重点阐述GeoFrame 储层预测的基本思路及地震属性的地质应用。 1、地震属性储层预测的基本思路 地震地层学原理假定,地震剖面上的反射波同相轴具有年代分界面的意义,要研究地层岩性和沉积相主要依据的是地震反射特征及其横向变化,也就是地震属性的变化,这是应用地震属性进行储层预测的基本理论依据。 应用地震属性进行储层横向预测要解决的主要问题是多解性问题,即:一种地震属性参数的变化受多种地质因素的影响,而一种地质现象的改变,也会造成多种地震属性的异常。 因此,在对地震属性分析预测过程中,如何从众多的地球物理参数中选取能反映地质特征变化的参数,是地震属性预测的主要问题。实际工作表明,必须做好以下两项工作: ① 正确认识地震属性 正确认识地震属性是做好属性预测的基础,不同的地震属性参数,它的地球物理含义、数学含义不一样,反映的地质规律也不一样。如:半时能量和总能量,尽管都是振幅类参数,但具体的展布规律却不一样(图1)。 图1 1 相同地区相同地区相同地区半时能量半时能量半时能量和和总能量总能量对比图对比图对比图 半时能量半时能量((Energy half-time ) 总能量总能量((Total Energy )

地震资料解释课程教学大纲

地震资料解释课程教学大纲 课程代码:74190110 课程中文名称:地震资料解释 课程英文名称:Seismic Interpretation 学分:2.0 周学时:1.5-1.0 面向对象: 预修要求:地层学、构造地质学、海洋沉积学、地球海洋物理学 一、课程介绍 (一)中文简介 《地震资料解释》是海洋科学专业的一门专业必修课,其总目标是结合地震资料解释实习课,使学生能够理解地震资料解释的基本原理和概念、掌握复杂地质条件下的层序地层、构造和地震相分析等地震资料解释的基本方法。 (二)英文简介 “Seismic Interpretation” is a compulsory course for the students majored in Marine Science. In combination with associated practice course, the students who attend this course would: (1) understand the fundamentals and basic concepts in interpreting the seismic data; (2) master basic skills and methodology to analyze the sequence stratigraphy, structure and seismic facies in the subsurface with complex geological conditions. 二、教学目标 (一)学习目标 通过本课程系统学习,要求学生全面掌握地震地质解释的地球物理基础和地震地质解释方法;学会应用地震资料进行地质解释的技能;了解地震资料地质解释的现状及发展方向。 (二)可测量结果 (1)掌握沉积层序的概念、沉积层序的边界类型、层序划分的原则和方法,能在地震剖面

GeoFrame地震属性列表

GeoFrame地震属性列表 传统的CSA计算的地震属性: RMS Amplitude RMS 振幅 Energy half-time 半幅能量 Average Magnitude 平均能量 Maximum Magnitude 最大能量 Computed Inst. Frequency 瞬时频率算术平均值 Computed Inst. Phase 瞬时相位算术平均值 Max. Amplitude 最大振幅 Min. Amplitude 最小振幅 Mean Amplitude 中值振幅 Average Peak Value 平均波峰值 Ave. Peak Value(zero X) 过零最大平均波峰值 Ave. Trough Value 平均波谷值 Ave. Trough Value(zero X) 过零最大平均波谷值 Arc Length 弧形长度 Threshold Value 门槛值 Average Energy 平均能量 Number of Zero Crossings 过零个数 Ratio of Pos to Neg samples(RPN) 正/负样点比 Dominant Frequency 主频 Bandwidth 带宽 Bandwidth Rating(Bias) Bandwidth Rating(Debias) 带宽比(偏差)校偏频宽比(去斜) Sum of Amplitudes 总振幅 Sum of Magnitudes 总能量 Window Length 时窗 Blip Horizon 假想标志层 Local Attributes Lower Loop Duration 下半周时间 Upper Loop Duration 上半周时间 Lower Loop Area 上半周面积 Upper Loop Area 下半周环面积 Upper Loop Skewness 上半周偏移 Upper Loop Kurtosis 上半周尖峰 Upper Loop Asymmetry 上半周环不对称 Duration Attributes Average Duration of Negative Loops 负周时间平均值 Average Duration of Positive Loops 正周时间平均值 Average Duration 平均周时间 Minimum Loop Duration 最小周时间 Maximum Loop Duration 最大周时间 Standard Deviation of Loop Duration 周时间的标志偏差

地震资料地质解释总复习

地震层序分析 不整合面的分类: 1)按地层产状特征分类: 可分为平行不整合和角度不整合两大类 A 平行不整合:地质标志为冲刷面,底砾岩,古土壤层,赤铁矿,钙质结核等。 B 角度不整合:受地壳运动的影响使岩层发生倾斜或褶皱。 2)按成因分类: 1.动力作用不整合,因构造等动力活动是地层产状发生变化造成时间缺失的不整合。 包括:a .褶皱不整合:由于褶皱作用而地层弯曲遭受剥蚀 b.掀斜不整合:由于掀斜作)用而使抬升一侧的地层遭受剥蚀 c.块断不整合:因差异升降而使断凸遭受剥蚀形成的不整合 d.抬升不整合:因整体抬升而形成,一般为平行不整合 e.岩浆侵入不整合:因岩浆岩后期侵入形成时间反转(相当于逆断层),形成的不整合。 f.塑性岩侵入不整合:因塑性岩层侵入造成界面间出现时间间断,形成的不整合。 2.外动力作用不整合:在没有构造变动的情况下,主要由于沉积、侵蚀等外动力地质作用造成地层中的时间缺失而形成的不整合。包括: a.河谷下切不整合 b.海底峡谷下切不整合 c.淹没不整合:因海平面快速上升从而使碳酸盐台地停止发育而形成的不整合。 d.沉积过路:海平面相对静止时期,形成沉积物的进积作用,在沉积基准面附近,沉积作用与侵蚀作用达到动态平衡,即形成沉积过路。 e.沉积间歇:沉积间歇是规模较小,持续时间相对较短的沉积间断。无明显地层侵蚀造成沉积间歇的原因可以是水平面的高频相对变化界面。围小到中等。 (3)按分布围分类 1、区际不整合:多个相邻盆地同时发育 2、区域不整合:在盆地大部分地区发育 3、4、局部不整合:在盆地局部发育 不整一界面(5种): 削截,视削截,顶超,上超,下超

地震属性的含义

*说明:谱属性(Spectral Attribute)谱分解(Spectral Decompose)轨迹属性类(Local Attribute)

*

瞬时频率(Inst Frequency ):定义为瞬时相位对时间的导数,用Hz 表示。经常用来估计地震振幅的衰减,往往油气的存在引起高频成分的衰减,可用这一属性检测油气。 瞬时相位(Inst Phase ): 表示在所选样点上各道的相位值,以度或弧度表示。主要用于增强油藏内弱同相轴,对噪音也有放大作用,最终成图的彩色色标应考虑到 反射强度(Reflection Magnitudes ):反映了岩性差异、地层连续、地层空间、孔隙度的变化。 反(负)二阶微商变换(Negative of Second Derivative ) :显著地提升了连续性,有助于更快、更准确的层位解释。 道积分(Integrated Seismic Trace ):能起到伪波阻抗剖面的作用. 并不是说用它替代反演, 它可以起到快速指示孔隙度变化的作用. 谱分解技术(Spectral Decomposition )—— 分频:用于揭示薄层岩性横向的变化,指示可能的含烃地层圈闭。最后分频属性和井砂岩厚度结合作出目标层段的砂岩厚度图。由于不同频率段所看到的东西是有区别的,所以分频还可以观察到河道的形状更清晰,河道内的岩性细节变化。 砂岩厚度图流程图: Find the Power Spectrum using SYNTHETICS Extract Tuning Frequency SATK Run Spectral Decomposition SATK Net Thickness Determination Correlate using LPM

地震属性分析技术综述

【全文】地震属性分析技术综述 [摘要] 地震属性是从地震资料中提取的隐藏有用信息,因而地震属性分析技术近几年在油气勘探开发中得到了广泛的应用与研究。本文对地震属性分析技术的发展状况进行了归纳、总结,简单阐述了地震属性分析技术的在不同时期所用到的基本原理和方法。特别对新地震属性进行了具体介绍。最后对该技术进一步的研究工作进行了总结和展望。 摘要:在勘探和开发周期的各个阶段,地震资料在复杂油藏系统的解释过程中,扮演着至关重要的角色。然而,缺少一种有效地将地质知识应用于地震解释中的上具。随着一系列属性新技术的出现,对地震属性进行充分研究,就给地质家提供了快速地从三维地震数据中获得地质信息的能力。尤其在用常规解释手段难以识别日的储层的情况下,属性分析技术更是给地质上作人员指出了新的方向。 [关键词] 地震属性储层预测叠前数据叠后数据 关键词:储层;波形分析;地震属性 1.引言 地震属性是指叠前或叠后的地震数据经过数学变换而导出的有关地震波的几何形态、运动学特征、动力学特征和统计学特征的特殊度量值。地震属性的发展大致从20世纪60年代的直接烃类检测和亮点、暗点、平点技术开始,经历了70年代的瞬时属性(主要是振幅属性)和复数道分析,90年代的多维属性(特别是相干体属性)分析,21世纪的地震相分析等阶段[1一SJ。随着地震属性分析技术的发展与研究,该技术已广泛应用于储层预测、油气藏动态监测、油气藏特征描述等领域,并取得了很好的效果。总之,地震属性分析技术可以从地震资料中提取隐藏其中的多种有用信息,这为油气勘探与开发提供了丰富宝贵的资料,也为解决复杂地质体评价提供了实用的分析手段。因此,对该技术进行深人调查研究具有很强的现实意义。 地震属性是指从地震数据中导出的关于儿何学、运动学、动力学及统计特性的特殊度量值。它可包括时问属性、振幅属性、频率属性和吸收衰减属性,不同的属性可指示不同的地质现象。地震属性分析则是从地震资料中提取其中的有用信息,并结合钻井资料,从不同角度分析各种地震信息在纵向和横向上的变化,以揭示出原始地震剖面中不易被发现的地质异常现象及含油气情况。 地震属性分析技术的研究已由线、面信息扩展到三维体信息,从分类提取扰化发展为一项系统的应用技术。随着地震技术的日趋成熟,地震属性技术近儿年也发展迅速,其中有多属性联合解释技术、波形分析技术、吸收滤波技术等。应用地震属性分析技术去完善勘探生产中的油藏描述工作,已经成为油藏地球物理的核心内容。利用地震属性分析技术预测岩性和有利储集体,描述油藏特征及孔隙度变化,寻找难以发现的隐蔽油区,以至于监测流体运动和进行其它综合研究,一直是石油工作人员追求的目标。 1波形分析技术的研究与应用 通常的层段属性只是表示了某儿个地震信号的物理参数(振幅、相位、频率等),但它们没有一个能够单独描述地震信号的异常,而地震信号的任何物理参数的变化总是对应着反映地震道形状的变化,所以,研究和分析地震资料中代表各种属性总体特征的地震道形状(波形),应该能有非常不错的效果[,]。 1. 1波形分析技术的原理及处理过程

地震地质综合解释基础知识试卷

地震地质综合解释基础知识试卷 一、填空题(每题2分) 1.地震反射同相轴的基本属性振幅、频率、相位。 2.影响地震速度的主要因素岩性、流体、埋深、温度、压力、密度等。3.AVO是指地震反射波振幅随炮检距的关系, AVA是指地震反射波振幅随方位角的关系。 4.振幅类地震属性主要有均方根振幅、平均振幅、最大峰值振幅、最大谷值振幅、平均能量、瞬时真振幅、反射强度、视极性平均振动能量、波峰振幅极大值、波谷振幅极大值总能量等(答对三种即可); 地震解释中振幅类地震属性主要用于识别油气流体的聚集、岩性概况、孔隙度情况、三角洲与河道砂的展布、礁体异常、不整合、调谐效应、层序变迁等(答对三种即可) 5.地震解释中相干属性主要用于识别断裂构造、岩性变化、地层物性、流体变化等 (答对两种即可)。 6.圈闭的要素有储集层、盖层、遮挡条件。 7.含油气盆地由基底、周边和沉积地层三个基本部分组成。 8.孔隙类型视其划分依据不同而异,主要流行三种方案:按孔隙的成因,可将孔隙类型分为原生孔隙、次生孔隙和混合孔隙;据孔隙的大小,分为超毛细管孔隙、毛细管孔隙和微毛细管孔隙;将成因和大小结合的分类,如E.D.Pittman对碎屑岩储集体孔隙分为粒间孔、溶蚀孔、微孔和裂缝。 9.成藏期的构造应力场必将有利于正确揭示油气藏的形成条件、分布规律和高产富集控制因素,同时对指导油气田的勘探和开发以及油气田施工设计都具有重要意义。 10.测井曲线的形态是岩性、物性和所含流体的综合反映,因此测井曲线的对比实质上就是岩性对比。 二、名词解释(每题5分) 1.圈闭:圈闭是具有储集层,盖层和遮挡条件,使油气能够在其中聚集并形成油气的场所。

地震属性含义及其应用

地震属性含义及其应用 一、 瞬时属性 19 假定复数道表示为:)t (iy )t (x )t (u +=,则 1. 瞬时实振幅 IReAmp ( Instantaneous Amplitude ) 是在选定的采样点上地震道时域振动振幅。是振幅属性的基本参数。 广泛用于构造和地层学解释。用来圈定高或低振幅异常,即亮点、暗点。反映不同储集层、含气、油、水情况及厚度预测。 2. 瞬时虚振幅 IQuadAmp (Inst. Quadrature Amplitude) 是复数地震道的虚部,与复数地震道的相位为90o时的时域振动振幅。即正交道,为虚振幅。 因它只能在特定的相位观测到,多用来识别与薄储层中的AVO 异常。 3. 瞬时相位IPhase ( Instantaneous Phase) ))t (x )t (y tan(A )t (=γ, 定义为正切,输出相位已转换为角度,数值范围是 [-180o ,180o ]。为q(t)/f(t)的一个角,是采样点处地震道的相位。 有助于加强储层内部的弱反射同相轴,但同时也加强了噪声,可用于指示横向连续性;显示与波传播有关的相位部分;用于计算相速度;因为没有振幅信息因此能够显示所有同相轴;用于显示不连续;断层、显示层序边界。由于烃类聚集常引起局部相位变化,也可以做烃类直接指示之一。 4. 瞬时相位余弦 CIP ( Cosine of Inst. Phase ) 是瞬时相位导出的属性。其计算式为))t ((Cos γ 常用来改进瞬时相位的变异显示。并用于相位追踪和检查地震剖面对比、解释的质量。多与瞬时相位联用。 5. 瞬时频率 IFreq (Inst. Frequeney) 定义为瞬时相位对时间的函数 dt )t (d γ(以度/毫秒或弧度/毫秒表示),其量纲为频率的量纲(Hz),是地震道在频率方面的瞬时属性。 用来计算、估算地震波的衰减。油气储层常引起高频成分衰减及杂乱反射显示,所以横向上可用于碳氢指示。高频成份多显示为尖锐的界面或薄层,亦可反映岩相的粗、细变化及地层旋回。

地震地质综合解释实习报告

成都理工大学地球物理学院2014年地震地质综合解释实习报告 姓名: 学号: 专业:勘查技术与工程 时间:2014.11.20

目录 一、实习目的和任务 (1) 二、软件介绍 (1) 三、操作步骤 (1) 1.建立工区 (1) 2.加载数据 (2) 3.工区建立完成 (4) 4.显示层位 (5) 5.加载井位 (6) 6.加载井数据 (6) 四.实习总结 (6)

一、实习目的和任务 地震资料综合解释是物探的重点课程之一,也是当前油气勘探领域最重要的一门学科,本次实习是一次综合性的地震解释训练,利用所学地震地质学来解释地震剖面的各种地质现象,通过实习,旨在提高我们的解释技巧,学会合理判断和分析各种地震信息,并初步学会SMT软件的使用方法,并完成实习报告。 二、软件介绍 SMT 解释系统是由美国Seismic Micro-Technology, Inc.公司研制开发的基于Windows 操作系统平台的地震资料解释系统。该系统包括了基础地震解释所要求的所有功能。一旦加载了地震数据、井数据和人文信息以后,便可以完成层位解释、计算网格和等值线、产生深度图,以及绘制高品质图件。包括在系统中的主要功能包有:2d/3dPAK、VuPAK、SynPAK、TracePAK 和ModPAK。 该模块支持深度域和时间域的解释,或是两者间的联合解释,并且支持多用户环境。使用该模块可以创建和管理层位、断层以及井信息。此外,还可以完成时深转换、生成等值线、网格化、地层属性计算、体计算,以及闭合差分析等工作。 三、操作步骤 1.建立工区 建立工区名字,用户名字,选择数据库:

选择工区高程为4800m,不使用已经存在的坐标系,选择“否” 2.加载数据 选择加载SEG-Y格式数据,并选择3D数据 选择振幅数据

地震资料解释

第五章:地震资料解释 用地震资料解释地下的地质问题,是地震勘探的最终目的。 §5.1地震反射信息的构造解释 序:构造油气藏的类型见P183,构造解释就是用地震资料识别解释构造油气藏。 一、地震时间剖面与地质剖面的对应关系 1.地震反射界面与地质界面的对应关系 (1)二者往往一致 只要有波阻抗,就有反射,所以地质上的层面,断层面,侵入接触面,不整合面,流体分界面,都有地震反射同相轴与之对应。 (2)二者不完全一致 例1:古老的地层,长期的构造运动和地层压力作用,使相邻的两套地层可能有相近的波阻扰,这种地质层面没有反射。 例2:有些地层界面,虽然两侧物性差异很大,但界面太短或太粗糙,地震上没有明显的反射,如珊瑚礁只有零星反射。 例3:同一岩性的地层,无层面又无岩性面,但由于含流体成份不同而形成反射界面产生反射波,但却不是地质界面。 例4:声波、面波、干扰波没有对应的地质界面。 2.地震的反射同相轴有地质年代意义 地质上主要以不整合面划分地质年代,这样的不整合面、层面有对应的反射同相轴,所以同相轴也就有了年代意义:上新下老。(北海模型)3.地震反射同相轴形状与地质构造的关系(熊粉书P52) 一般情况下反射同相轴的形状反映构造的形态,但有速度陷阱。 R 012

下拉 021021 4.地震反射与岩性有关 介质的岩性、反射系数、速度、密度、吸收等对地震波的波形有影响,对振幅、频率影响较大。反过来说不同的波形、不同的振幅、不同的频率反映不同的岩性。 总之: 现代的地震剖面与地质剖面极相似,因为地震剖面是地质剖面对地震波的响应。地下的构造特点,岩性特点决定了地震时间剖面的特点,二者有联系但又不完全一一对应,必须去伪存真,找出地质上有用的东西,这就要进行解释。 二、时间剖面的对比 (一)反射波的识别标志(北海模型) 1.波的对比 在时间剖面上,反射层是以同相轴的形式出现的,追踪反射层就变成 了对同相轴的追踪,只有同一个界面的反射波才能反映构造的形态,追踪 .. 同一个界面的反射波的同相轴叫做波的对比 ...................。 2.波形相似 同一界面的反射波,其上覆层的性质、深度、岩性、产状等,在一定 的范围内变化不大,在相邻的记录道上有相似性,因而导致同一个界面的 ...... 反射波相似 .....。主要有三个相似特点,也叫反射波对比的三大标志。 3.反射波对比的三大标志 (1)强振幅标志 反射波振幅比干扰波振幅明显强 ..............,因为各种野外方法、处理方法都在加强一次波。

常用地震属性的意义

常用地震属性的意义 地震反射波来自地下地层,地下地层特征的横向变化,将导致地震反射波特征的横向变化,进而影响地震属性的变化,因此,地震属性中携带有地下地层信息,这是利用地震属性预测油气储层参数的物理基础。随着地震属性处理及提取技术的大量涌现,属性种类多达几百种,实际应用人员应用起来遇到了很大困难,迫切需要按实用的角度,总结各地震属性参数与储层特征参数间的内在联系,为进一步研究建立地震信息与储层参数之间的关系提供可靠的前提条件,做到信息提取有方向、有目标。为了达到这一目的,首先按类别较全面总结了目前常用地震属性,从算法开始,分析了各属性所表达的在地震波波形上的意义,从正向上分析地震属性变化与油气储层特征变化的关系,进而探讨总结了它的潜在地质应用。 1、属性体、属性剖面 这类属性是按剖面(或体)处理的,是一个体文件(或剖面文件),属性值对应 、属性值),可以用于常规地震剖面的方式显示与使用,常空间位置,即(x、y、t 用的属性有:相干体(方差体、相似体等)、波阻抗、道积分数据体,经希尔伯特变换得到的瞬时属性体、倾角、倾向数据体等,这些属性体可以直接应用于解释,也可以用解释层位提取出来转变为属性层,下表为常用属性体属性意义及潜在地质应用一览表。

2、沿层地震属性 这种属性是用解释层位在地震数据体(剖面)中提取出来的属性,它的数值对应一个层位或一套地层,每个属性值对应一个x、y坐标。提取方式有两类:沿一个解释层开一个常数时窗,在此时窗内提取地震属性,提取方式有4种(图2-1a)。用两个解释层提取某一段地层对应的地震属性,提取方式也有4种(图2-1b)。 常用地震属性的计算方法总结如下: (1)、均方根振幅(RMS Amplitude) 均方根振幅是将振幅平方的平均值开平方。由于振幅值在平均前平方了,因此,它对特别大的振幅非常敏感。

Landmark主要地震属性及其地质意义

Landmark主要地震属性及其地质意义利用地震进行储层预测时主要从振幅属性及其延伸属性出发,分析属性的变化特征,然后与钻井和地质进行标定,赋予属性地质意义。 为了将已知井上的岩性信息,在整个工区进行有效的外推,需要优选出在该区对岩性参数和含油气性反映敏感的属性,我们通过两个层次来完成这一个工作。第一个层次是选择对岩性变化相对敏感的地震属性,这部分工作在属性提取时已完成,其最基本的理论基础是:时间派生的属性有利于对构造的细节进行解释;振幅和频率派生的属性用于解决地层和储层特征; 一般认为振幅是最稳健和有价值的属性;频率属性更有利于揭示地层的细节; 混合属性包含振幅和频率的因素,因此更有利于地震特征的测量;同时在对所提取的地震属性的物理意义的理解也有助于对地震属性的提取第二个层次是使用数学和信息学的方法优选属性。“地震属性和井数据采样伪相关在独立的井数据较少或者参加考虑的独立的地震属性过多时产生的概率较大”(CYNTHIA T. KALKOMEY),由于对于该区已知的独立井信息多数情况下较少,勉强满足统计分析的样本要求,单纯使用相关分析方法产生伪相关的概率较大,因此我们在经过第一个层次的筛选之后,采用数据相关和信息优化组合方法进行属性优选。 目前属性种类很多,属性软件也非常多,这里转列landmark软件中的PAL 属性,供大家参考选择使用:Average Reflection Strength 平均反射强度:识别振幅异常,追踪三角洲、河道、含气砂岩等引起的地震振幅异常;指示主要的岩性变化、不整合、天然气或流体的聚集;该属性为预测砂岩厚度的常用属性; Slope Half Time 能量半衰时的斜率:突出砂岩/泥岩分布的突变点;预测砂岩厚度的常用属性; Number of Thoughs 波谷数:可以有效的识别薄层,为预测砂岩厚度的常用属性;Average Trough Amplitude 平均波谷振幅:用于识别岩性变化、含气砂岩或地层。可以有效的区分整合沉积物、丘状沉积物、杂乱的沉积物等;预测含油气性的常用属性; Average Instantaneous Phase 平均瞬时相位:由于相位的横向变化可能与地

地震资料综合解释资料

名词解释: 1.褶积模型:地震记录的褶积模型是当今地震勘探中三大环节的主要理论基础之一,其应用十分广泛,主要表现在三大方面:正演、反演和子波处理。层状介质的一次反射波通常用线性褶积模型表示,即:式中:w(t)为系统子波;r(t)为反射系数函数,符号“*”表示褶积运算。 2.分辨率:分辨能力是指区分两个靠近物体的能力。度量分辨能力强弱的两种表示:一是距离表示,分辨的垂向距离或横向范围越小,则分辨能力越强;二是时间表示,在地震时间剖面上,相邻地层时间间隔dt 越小,则分辨能力越强。时间间隔dt 的倒数为分辨率。垂向分辨率是指沿地层垂直方向所能分辨的最薄地层厚度。横向分辨率是指横向上所能分辨的最小地质体宽度。 3.薄层解释原理:Dt

地震地质综合解释2.0 (1)

名词解释:6*5=30 简答:5*9=45 计算:1*10=10 论述:1*15=15 地震地质综合解释思考题参考答案 1、名词解释:(★) 1)地震子波:在震源附近,地震波以冲击波的形式传播,当传播到一定距离时,波形逐渐稳定,此时的地震波被称为地震子波。 2)反射系数:在垂直人射的情况下,纵波入射时将只考虑产生的反射纵波和透射纵波的情 况。这时界面的反射系数定义为: 3)弹性参数:地震波是在岩层中传播的弹性波,在弹性力学中,引入了一些物理量来描述弹性体的弹性特征,在研究地震波动力学时经常用到的弹性参数如下:杨氏模量E、泊松比σ、切变模量μ、体变模量K、λ系数、λ、μ合称拉梅常数 2)波阻抗:波在某介质中传播的速度与介质密度的乘积定义为该介质的波阻抗。 3)信噪比:所谓信噪比,通俗地讲就是有用的地震波与无用地震波的能量(振幅)之比。4)振幅:质点振动离开平衡位置的最大位移(幅度)称为振幅。 5)频谱:组成一个复杂振动的各个谐振动分量的特性与其频率的关系的总和,就称为这个振动的频谱 6)正花状构造:与压扭性走滑断裂相对水平运动相伴生的构造样式,在走滑断裂上部形成背形构造 7)负花状构造:与张扭性走滑断裂相对水平运动相伴生的构造样式,在走滑断裂上部形成向形构造 8)地震相:是一个可以在区域内固定的,由地震反射层组成的三维单元,其反射结构、振幅,连续性、频率和层速度等要素,与邻近相单元不同。 9)底超:是在一个沉积层序底界上的超覆尖灭现象,它又分为上超和下超两种基本类型。(1)上超:是一套当初是水平的地层对着一个原始倾斜面超覆尖灭,或者是一套原始倾斜的地层对着一个原始倾角更大的斜面逆倾向的超覆尖灭。

Petrel地震地质解释和建模使用技巧2013

Petrel 地震地质解释和建模 使用技巧 2013 斯伦贝谢科技服务(北京)有限公司

Copyright Notice ? 2009 Schlumberger. All rights reserved. No part of this manual may be reproduced, stored in a retrieval system, or translated in any form or by any means, electronic or mechanical, including photocopying and recording, without the prior written permission of Schlumberger Information Solutions, 5599 San Felipe, Suite 1700, Houston, TX 77056-2722. Disclaimer The License Agreement governs use of this product. Schlumberger makes no warranties, express, implied, or statutory, with respect to the product described herein and disclaims without limitation any warranties of merchantability or fitness for a particular purpose. Schlumberger reserves the right to revise the information in this manual at any time without notice. Trademark Information Software application names used in this publication are trademarks of Schlumberger. Certain other products and product names are trademarks or registered trademarks of their respective companies or organizations.

常用地震属性的意义

常用地震属性得意义 地震反射波来自地下地层,地下地层特征得横向变化,将导致地震反射波特征得横向变化,进而影响地震属性得变化,因此,地震属性中携带有地下地层信息,这就是利用地震属性预测油气储层参数得物理基础。随着地震属性处理及提取技术得大量涌现,属性种类多达几百种,实际应用人员应用起来遇到了很大困难,迫切需要按实用得角度,总结各地震属性参数与储层特征参数间得内在联系,为进一步研究建立地震信息与储层参数之间得关系提供可靠得前提条件,做到信息提取有方向、有目标。为了达到这一目得,首先按类别较全面总结了目前常用地震属性,从算法开始,分析了各属性所表达得在地震波波形上得意义,从正向上分析地震属性变化与油气储层特征变化得关系,进而探讨总结了它得潜在地质应用。 1、属性体、属性剖面 这类属性就是按剖面(或体)处理得,就是一个体文件(或剖面文件),属性值对应空间位置,即(x、y、t0、属性值),可以用于常规地震剖面得方式显示与使用,常用得属性有:相干体(方差体、相似体等)、波阻抗、道积分数据体,经希尔伯特变换得到得瞬时属性体、倾角、倾向数据体等,这些属性体可以直接应用于解释,也可以用解释层位提取出来转变为属性层,下表为常用属性体属性意义及潜在地质应用一览表。

相似体计算相邻地震道 得相似系数 同上 不但可以对三维体数据作 不连续分析,还可以对基于 层位得二维数据作相似性 预测,以及倾角、方位角,边 界检测与图象增强。还可以 沿层解释得层位作相似性 分析 波阻抗它将地震资料、测 井数据、地质解释 相结合,利用测井 资料具有较高得 垂向分辨率与地 震剖面有较好得 横向连续性得特 点,将地震剖面 “转换成”波阻抗 剖面 用于储集层得研究, 识别砂体得分布特征 与范围 将地震资料与测井资料连 接对比,能有效地对地层物 性参数得变化进行研究,对 储层特征进行描述 道积分对地震道进行积 分 识别砂体、岩性尖灭 点等 相对对数波阻抗 倾角倾向数据体计算同相轴得倾 角 识别尖灭点、不整合、 了解地层产状 2、沿层地震属性 这种属性就是用解释层位在地震数据体(剖面)中提取出来得属性,它得数值对应一个层位或一套地层,每个属性值对应一个x、y坐标。提取方式有两类:沿一个解释层开一个常数时窗,在此时窗内提取地震属性,提取方式有4种(图21a)。用两个解释层提取某一段地层对应得地震属性,提取方式也有4种(图21b)。 常用地震属性得计算方法总结如下: (1)、均方根振幅(RMS Amplitude) 均方根振幅就是将振幅平方得平均值开平方。由于振幅值在平均前平方了,因此,它对特别大得振幅非常敏感。 (2)、平均绝对值振幅(Average Absolute Amplitude) 平均绝对值振幅没有均方根振幅那样,对特别大得振幅敏感。 (3)、最大波峰振幅(Maximum Peak Amplitude) 最大波峰振幅得求取方法就是,对于每一道,PAL在分析时窗里做一抛物线,恰好通过最大正得振幅值与它两边得两个采样点,沿着这曲线内插可得到最大波峰值振幅值。

地震属性含义

1、属性名称:反射强度(Reflection Strength),振幅包络(Amplitude Envelope),瞬时振幅(Instaneous Amplitude)REFLSTAN (缩写) 定义: 在解释中的应用:用于振幅异常的品质分析;用于检测断层、河道、地下矿床、薄层调谐效应;从复合波中分辨出厚层反射。 属性特征:提供声阻抗差的信息。横向变化常与岩性及油气聚集有关。值总是正的。 2、属性名称:瞬时相位(Instaneous Phase)INSTPHAS(缩写) 定义:在解释中的应用:进行地震地层层序和特征的识别;加强同相轴的连续性,因此使得断层、尖灭、河道更易被发现。可对相位反转成图,有可能指示含气与否。 属性特征:描述了复相位图中实部和虚部之间的角度。它的值总在±180°之间。瞬时相位是不连续的,从+180°到-180°的反转可引起锯齿状波形 3、属性名称:瞬时频率(Instaneous Frequency)INSTFREQ(缩写) 定义:在解释中的应用:用于气体聚集带和低频带的识别;确定沉积厚度;显示尖灭、烃水界面边界等突变现象 属性特征:瞬时相位对时间的变化率。值域为(-fw, + fw)。然而,大多数瞬时相位都为正。可提供同相轴的有效频率吸收效应及裂缝影响和储层厚度的信息 4、属性名称:正交道(Quadrature Trace),希尔伯特变换(Hilbert Transform)QUADRATR(缩写) 定义:h(t)是f(t)的希尔伯特变换,也是f(t)的90°相移 在解释中的应用:用于复数道分析的品质控制 属性特征:当实地震道代表地震响应中质点位移的动能时,正交道相当于质点位移的势能 5、属性名称:视极性(Apparent Polarity)APPAPOLA(缩写) 定义:在振幅包络峰值处实地震道的极性 在解释中的应用:用于振幅异常的品质分析 属性特征:为实地震道的符号位,假设零相位子波、视极性与反射系数的极性相同 6、属性名称:响应相位(Response Phase)RESPPHAS(缩写) 定义:在振幅包络峰值处的瞬时相位值 在解释中的应用:地震地层层序的识别、检测。由于流体含量或岩性引起的横向变化,在具有相似的振幅响应时,用来区分有利和不利带 属性特征:强调反射界面的主相位特征。与瞬时相位的应用相同 7、属性名称:响应频率(Response Frequency)RESPFREQ(缩写) 定义:在振幅包络峰值处的瞬时频率值 在解释中的应用:识别与气藏聚集有关的可能区带 属性特征:相应频率在区域上更具可解释性。与瞬时频率的应用相同 8、属性名称:反射强度交流分量(Perigram)PERIGRAM(缩写) 定义:消除了反射强度中的均值(直流分量)部分后的偏差 在解释中的应用:用于振幅异常的品质分析。与反射强度的应用相同,但更适合于分析和处理,因为它有正负

相关文档