| Monte Carlo method is a numerical simulation method based on probabilistic model.In neutron logging,Monte Carlo method establishes a random sampling model for the target problem,tracks the reaction process and motion trajectory of particles,and completes an accurate simulation of particle transport.However,the Monte Carlo method requires a large number of random event simulations,and its computation time is often long especially when high precision calculation results need to be obtained.Therefore,in order to solve the problems of its long computation time and low computational efficiency,the research on the acceleration of Monte Carlo methods is very crucial.In this thesis,an adaptive mesh generation method is proposed based on the global reduced variance FWCADIS method,which can achieve optimal mesh for simulation models with different environments and detector positions,thus improving the adaptability of the particle importance map to the current Monte Carlo model and increasing the computational accuracy and efficiency of FW-CADIS method.The analysis of the influence of neutron logging environment and the validation of the adaptive mesh generation method are completed using the traditional correction plate method.And the high computational efficiency of the method is applied to the study of neutron logging inversion problems.First,the thesis conducts an in-depth investigation and analysis of advanced Monte Carlo acceleration methods at home and abroad,and the FW-CADIS method is selected to carry out the accelerated study,which has good acceleration effect in the multi-detector problem and depth penetration problem.Based on the neutron diffusion theory and contribution theory,the mesh generation indicator with adaptivity to different environmental parameters and detector positions is derived.And the optimal power window mesh is obtained utilizing a combination of coarsening and refining of the grid according to the depth of investigation of the neutron logging tool,which is used to accurately guide the resampling process of particles and optimize the calculation accuracy and calculation time of the Monte Carlo method.Secondly,the environmental impact model database of neutron porosity logging is constructed using FW-CADIS method with adaptive mesh,and the effects of environmental parameters such as well diameter,tool eccentricity distance,casing thickness,cement thickness and so on are analyzed,and the actual logging data are corrected using the traditional correction plat method to further verify the computational accuracy of the adaptive mesh generation method under different environments.Finally,the inversion method of well diameter and porosity of the array neutron tool is completed using the regularization method based on the environmental impact database acquired by the adaptive mesh generation method.The adaptive mesh generation method and the regularized inversion method are also modularized,and integrated into a visual operation platform to reduce the technical threshold of users and complete the algorithm invocation and data statistics in a more intuitive and convenient way. |
