Research Of Logging Curve Similarity Algorithm

Depth correction is an important part before casing perforation, and its requirement of accuracy is high. Inaccurate depth correction will make the subsequent perforation job to appear deviation, reduce the production of oil as well as affect the evaluation of the block directly, and further affect its future development prospect. The core of depth correction is to align the similarity points before and after case, so the accuracy of similarity algorithm is the key problem of the research.This paper mainly studies similarity algorithm suitable for logging curve and how to integrate it into depth correction efficiently and accurately. In order to make the similarity algorithm have higher accuracy and less time-consuming, preprocessing of curve has become an indispensable part. Curve burr noise affects the accuracy, huge data volume affects the computational complexity, namely affects the calculation time. By studying the subsistent methods of eliminating curve burr, we found that the threshold method in the field of wavelet is very suitable for the logging curve processing. On the basis of analyzing the existing threshold functions and threshold, an improved adaptive wavelet threshold method is educed and applied to the well logging curve. Through simulation experiment, we found that this method can not only eliminate noise but also keep the important information of the curve. After comparison and discussion with several linear stepwise methods, an important characteristic value linear stepwise method is selected to process the curve after the process of the signal denoise, and the amount of data compression can be achieved in this way. The curve composed by the points extracted by important characteristic value is the piecewise linear curve. Through simulation experiment observation on processed curve shape and the amount of data, we found that the method can save the local important information and compress the data volume at the same time. There are a lot of similarity algorithms combining with the characteristics of logging curves. The dynamic time warping distance is selected as the similarity core algorithm of curve. The relaxation method for dynamic time warping has some limitations, but it’s very suitable for logging curve correction. To improve the original relaxation point algorithm, the calculation time can be further shortened. The ideas of artificial depth correction and the programming method are completely different, and two kinds of schemes are put forward, integrating the original improved time warping method and the improved relaxation point method into the solution suitable for their programs respectively, through comparing the results obtained by simulation experiment with manual correction, it is clear that both of these schemes can be used for depth correction. After discussion and comparison, with higher accuracy and less time consuming, "No. 2 scheme" can be used for depth correction in well logging curve, and it has significant value to reduce productivity and save cost.