Research On Swarm Intelligent Inversion Algorithm Of Electromagnetic Logging While Drilling

This article mainly aims at the current mainstream while drilling electromagnetic wave logging inversion correction methods are mostly used is highly dependent on the initial model of linear iterative inversion,such as gauss Newton iteration,the principle of which is by using a nonlinear inverse problem is converted to linear inversion problem for iteration,but in order to cater to diverse solution,compulsive,highly nonlinear characteristics of while drilling electromagnetic logging inversion,selection of the swarm intelligence optimization algorithm of particle swarm optimization algorithm for related research.For slow convergence speed and low convergence accuracy question of fractional order Darwinian particle swarm optimization(FDPSO)algorithm,improved the algorithm of fractional order velocity update strategy,and introduce the improved algorithm into CHPSO algorithm that have algorithm running time advantage,greatly improve the convergence speed and precision of the algorithm,but found that FCHPSO algorithm in dealing with a multimodal function than FDPSO algorithm have slightly insufficient.On the basis the FDPSO algorithm of improved fractional order velocity update strategy,introduce the Logistic model mixed fractional order adaptive dynamic adjustment strategy,and get an improved adaptive particle swarm optimization(LFDPSO)fractional order Darwinian algorithm,and through the corresponding theoretical analysis,prove the convergence of the algorithm under given conditions.The corresponding numerical experiments are carried out by six classical functions,and it show that different from the FDPSO algorithm,NFDPSO algorithm is improved on precision and convergence speed and it is also proved the accuracy of the velocity update strategy in this paper from the other side,secondly on the convergence accuracy and convergence speed,introduced Logistic type hybrid adaptive fractional order Darwinian particle swarm(LFDPSO)algorithm has been effectively improved and the enhancement,the escape ability of particles in local optimum,global optimization and intelligent search ability have achieved effective improvement.Finally this paper gives the corresponding inversion correction experiment of inhomogeneous medium and two layer geological model,by comparing the proposed LFDPSO on the inversion results is feasible,and the corresponding results are analyzed.