Study On The Effect Of Mudcake And Its Correction Method In Sweep-Frequency Dielectric Logging

Conventional logging methods are now very commonly used,but in some complex geological environments,such as in the case of high resistance water and low resistance oil formations,dielectric logging is more efficient in identifying reservoir fluids compared to conventional resistivity logging by using the difference in pore fluid dielectric constants.Due to the limitation of measurement environment,conventional dielectric logging has not been widely used.With the in-depth research on the mechanism of rock dielectric dispersion,the practical application of swept frequency dielectric logging has become possible.In this paper,we discuss the rapid solution method of the analytical solution of swept frequency dielectric logging under the hybrid wave model,simulate the response of the instrument,and study the influence of mudcake and its correction method on this basis to provide theoretical support for the development of domestic instruments.The rock dielectric dispersion mechanism and instrumentation measurement principle are introduced,the analytical solution of swept frequency dielectric logging under the hybrid wave model is derived,the fast Hankel transform method and the discrete complex mirror method are verified,and the feasibility of quickly solving the analytical solution of the instrumentation response containing Sommerfeld’s equation is explored.By calculating the instrument response,the influence law of the swept frequency dielectric logging response is analyzed from several aspects,including mudcake thickness,mudcake propagation coefficient,formation propagation coefficient,and transmitting and receiving distance,and the influence of mudcake on the instrument response is summarized and analyzed: the correction amount of mudcake is proportional to the thickness of mudcake,relative permittivity and conductivity of mudcake,and inversely proportional to the relative permittivity and conductivity of formation and the distance from transmitting antenna to receiving antenna.According to the influence law,the correction method of mudcake is investigated.Firstly,the correction plates of mudcake are drawn from several dimensions,and the mudcake correction can be performed in the range of formation relative permittivity10～80 and formation conductivity 0.2S/m～1S/m.Then the prediction of the mudcake correction amount is performed using machine learning algorithms,including Bagging,Boosting,random subspace classification,K-nearest neighbor,decision tree and random forest algorithms,among which the random forest algorithm has the highest prediction accuracy and is reliable for the mudcake correction amount.