The New Methods Of Converted Wave Statics

Converted wave statics is one of difficulties in seimics processing. Converted wave statics is becoming hot spotting of domestic and international research.. Particularly the method of surface-consistent statics performs poorly in handling the long-wavelength satic components so far. The paper go research pointing to the problem. It goes research from short-wavelength and long-wavelength components and proposes the resolved method being test by the theoretical data. At last using high-order statistic method processes time-delay estimates and compares with crosscorrelation .Converted wave statics is composed of P-wave statics and SV-wave statics .The former is processed according to normal P-wave statics. Yet the latter it is divided into short-wavelength and long-wavelength . For the short- wavelength it is optimized common receiver stack traces . When structural dip is much heavy ,it is estimated by curve smoothing . The difference between of the two is receiver statics. The long-wavelength statics of converted wave computation: there are variant Vp/Vs values in the near-surface. Under this conditions it is computed PS-wave statics by correlation method, using processed P-wave stacksection to estimate structure term and to match with PS-wave by stretching . In addition ,using high-order cumulants estimates delay time.The paper goes to explain theoretically in time domain and frequency domain.Moveover using the mothod processes synthetic data and compares it with result of correlation method .By the result it is tested the high-order cumulants superiority.The paper is divided into 7 charter. In charter 1, at first several methods for calculating the coordinates of the convertion points are introduced ,and then pick up the CCP gather on the base of common convertion point.In charter 2,at first it is introduced the short-wavelength statics of converted wave difference with P-wave, and then introduced principle of convertion wave and the difference between statics pointing the difference and explained model trace forming ,getting moveout, that is getting total statics ,structural terms and receiver statics . At last for the convert-wave statics in 3-D showed some aspects. In charter3 ,at first it is showed some causes of long-wavelength statics , introduced theory of long-wavelength statics and introduced method of correction. In charter 4, it is the data computation part. It is processed short-wavelength data ,explained result , showed the method validity and processed practical data. For long-wavelength,have showed the method which can stretch P-wavelet to match PS-wave and computed receiver statics using practical data .At the same time, some noticing problems is proposed in the course of stretching and demonstrated the method feasibility. In charter 5,it is the high-order cumulants and time delay estimates. At first, it is defined the high-order cumulants and the high-order moments from the math , gave emphasis to introduce three-order cumulants rms parallel bispectrum estimates.And then it is introduced time delay estimation by high-order cumulants and demonstrated their estimates criterion from time domain and frequency domain.At last it is computed practical data using the high-order cumulants to estimate time delay . Through the paper ,it is arrived at following conclusions:It is very much important for convertion point assertion of converted wave to effectly stack . In fact, it is adopted to approximate formula computing according to Vp/Vs.2) Picking converted wave statics for the short- wavelength using trend analysis method can very much pick-up structure terms and separate the statics .The trend analysis can use 5 point 3 times smoothing method and the least –squares fit can see 5 point 3 times smoothing method to be betterly and more simple .Whether the static correction to perform well will depend on the results of analysis trend . The starts several traces and the last several traces need to interpolate the data to guarantee proceeding 5 points, the smoothing times chose by experiment...