| The demand for oil and gas resources in the national economy continues to grow.At the same time,many oil fields in China are in the middle and late stage of development.In order to further tap the potential of oil fields,it is necessary to supplement corresponding data for some problematic old wells and carry out dynamic monitoring of oil reservoirs,so as to improve the output of oil fields.Through casing density logging is one of the important means to tap the potential of old wells,but it is also facing some challenges as follows:first,casing damage,corrosion and perforation are easy to occur,which seriously affects the service life of oil wells and oil and gas production.At present,there is a lack of a set of means and methods to accurately evaluate casing damage in through casing density logging;Second,how to use the through casing density logging to measure the cement sheath thickness and density more accurately is a problem that has not been completely solved at present;Third,with the depletion of oil and gas rich reservoirs,the accurate measurement of thin-layer reservoir properties in old oilfields is very important for determining the location of gas producing reservoirs and correctly evaluating oil and gas potential.At present,there is still a lack of thin-layer evaluation means based on through casing density logging technology.In view of the above challenges,in order to better evaluate the formation density and the integrity of casing and cement sheath under the condition of cased well,based on the four detector density logging instrument,we studied the influence of well fluid,casing thickness,cement sheath thickness,formation density,cement sheath density and other factors on the gamma count of four detectors,determined the main factors affecting the count rate,and obtained the response relationship between the detector count rate and the parameters of cased well.Based on the experimental data and the numerical simulation results of gamma transport,the forward modeling model of four detector density logging tool for cased well parameters is established through conventional theoretical formula.The casing well parameters are inversely calculated by using the established forward model,and compared with the real value.The inversion results show that the casing thickness is in good agreement with the real value,but the other three parameters affect each other,and the inversion accuracy is not ideal.A new technical route is needed to study the multi parameter correlation.Therefore,we have established an artificial intelligence neural network prediction model for logging parameters of cased wells.In this model,the counts of different energy windows of four detectors are taken as the input variables of the model,and the parameters such as formation and cement sheath density,cement sheath and casing thickness are taken as the output.The research results show that the neural network prediction model can be better applied to the evaluation of formation density and integrity of casing and cement sheath,including formation density,cement sheath thickness,casing thickness.The average absolute errors of formation density,cement sheath thickness,casing thickness and cement sheath density are 0.06g/cm~3,0.52cm,0.04cm and 0.05g/cm~3 respectively,and the accuracy is much higher than the forward model of conventional theoretical formula.At the same time,in order to correctly evaluate the thin layer,the detection characteristics of the four detector density instrument are studied,mainly including the detection depth and vertical resolution of each probe.Finally,based on the characteristics of the detection characteristics of each probe in the four detector density instrument,the density curve is processed by resolution matching and wavelet transform.This method has achieved good results in the identification of thin layer. |
PDF Full Text Request