Research Of The Recognition And Extraction Methods Of Component Waves In Array Acoustic Logging

Array acoustic logging is a specific logging method among the well logging methods.Array acoustic logging tool produces an acoustic signal using a transmitter and receives the signal by an array of equally spaced receivers.In this way,the multiple channels of duplicate signals can make a compensate to the information loss caused by factors such as the variation of borehole diameter.Thus array acoustic logging has some advantages over conventional acoustic logging.There is abundant information contained in array acoustic logging data.What's more,the information of the extracted component waves can be used for the calculation of elastic parameters of the rocks,the analysis of borehole stability,the study of formation anisotropy,the analysis of in-situ stress,the identification of gas reservoir,the evaluation of reservoir permeability and so on.As a result,in the exploration and exploitation of geological resources,the application of array acoustic logging becomes more and more widespread.The processing of array acoustic data is a significant segment in the process of logging information.Researchers are paying continuous attentionto the efficiency and precision of the processing.With the development of petroleum industry and the increase of difficulties in exploration and exploitation,conventional logging methods cannot fully meet the demands of production.In conclusion,new methods and new techniques are desperately needed in array acoustic logging processing and interpretation.In this paper,the processing and interpretation methods of array acoustic logging data is mainly discussed including the extraction methods of component waves in array acoustic logging data considering efficiency and precision,respectively.In order to enhance efficiency,this paper introduces anew slowness-time coherence method based on dichotomy method.The new method combines energy ratio of short time window versus long time window method and slowness-time coherence method.First of all,the arrivals of component waves are calculatedby energy ratio of short time window versus longtime window method.Then the calculated arrivals are input into the slowness-time coherence method,so that the two-dimensional slowness-time searching are simplified to one-dimensional slowness searching.To make further improvement,the dichotomy method is utilized so that the ergodic searching pattern in conventional slowness-time coherence method is replaced.Hence the searching times is less and the amount of calculation is largely reduced.In conclusion,the aim of enhancing the calculating efficiency is successfully achieved.In order to enhance the precision in the extraction of component waves,a new slowness-time coherence method is prompted based on simulated annealing algorithm.The core of the algorithm is to transform the correlation coefficient of slowness-time coherence method into the energy function of simulated annealing algorithm so that the ability of jumping out of local maximum can be taken advantage of in the slowness searching procedure of component waves.For compressional wave,the energy ratio of short time window versus longtime window method can be applied to the computation of arrival before annealing,so that the two-dimensional search can be simplifiedto one-dimensional;for other component waves,the algorithm is used respectively to extract the slowness by two-dimensional search of slowness and arrival.In the annealing(global optimization)process,current temperature is reduced gradually;under every temperature,the slowness of the component wave is updated step by step.At the same time,the Metropolis principle is utilizedto judge if the new solution is accepted.After the temperature is lowered for sufficient times,the slowness of a certain component wave converges to a result that is closer to that of conventional logging data than conventional slowness-time coherence method.As a result,the precision is improved.