| With the increasing emphas is on thin interbed ded reservoir in oil exp loratio n,how to use existing techno logy to identify, it is becoming the ma in researchproble m in the geolo gical workers. Many o f our reservo irs are thin la yers, it has animportant guid ing significance to identify thin layer of e xploration a nd mining.Now it becomes a focus on how to improve the resolution of the thin laye r, also itappears many methods, technology.The seis mic signals of vario us frequenc y co mponents of the statistica lproperties of unstab le changes over time, t hese changes and abnorma lities record alot of wealth information that reflects the characteristics of subs urface reflectormed ia. In conventio na l seis mic data, spectra l ana lys is, the d iffere nt frequenc yco mponents of seis mic s igna ls is underutilized, failed to fully exp lore many of thepotentia l informatio n contained in the seis mic s igna l. In partic ular, appea red in thininterbedded strata, due to the thin-layer tuning effect, the mutation was verydiffic ult to reflect in the origina l s igna l.In order to observe the s ubtle cha nges ofthe signal better, this paper uses the wavelet tra nsform to ana lyzes the seis mic data.Based on the freq uency ana lys is method of the wa velet tra ns form, this paperuse sits time window resolution capability, high-freq uency s ide of the hightemporal resolutio n and low end of the high-freque ncy resolutio n signa ldecompositio n for multi-la yer signa l, extract the feature vectors toreconstructachieved very good results. In add ition to the reflection coeffic ient o fthin alternating layers of ana lys is, thus to reflect the spectrum characteristics of thethin layer of reflective thin layer. The main contents are:(1)The basic concept of the thin la yer.Through the research on the d iffere ntsizes of the reflection coeffic ie nt, differe nt po larity, and change s in the timeinterva l model. This paper gets the factors that affect the characteristics ofthin-layer spectrum. (2) Through the ana lys is of the wave let signa l process, this thes is finds out asuitable wa ve let functio n to ana lys is the thin interbedded, And by co mbining withthe time-frequenc y information, this thesis did a decomposition reconstruction oneach s igna l of the synthetic seis mic data, Through this method,we increas ed theresolutio n of the thin la yer. And verified the accuracy the method by co mbining thedb3wave let and time-freq uency informatio n of the thin layer to deal with theseismic data.(3) This thes is uses the selected wavelet to analyze the wedge model and thethin layer model that are in different thickness, and finds the freque ncy signa l thatconta ins the signa l mutation, by reconstruct ing of the signa l compone nts conta iningthe mutation point, and co mparing with the forward model, thin layer that is morethan10m can be identified. This paper also do an experiment on the thin-la yermodel that of the doub le5m mutual re flection, by this method we can find thebottom of the first layer of the bilayer.(4) Thro ugh wave let time-freq uency ana lys is and multisca le analys is of the rea lseis mic data getting in the target zo ne ShuSha nHe gro up,we can identify a thinla yer. That is about12m in he ight. By this ana lys is, this thes is proved theeffectiveness of the method by comb ining the wave let transform time-frequenc yanalysis and the multiresolution analysis in the practical seismic data processing. |
PDF Full Text Request