Acoustic Impendence Inversion Based On Particle Swarm Optimization Algorithm

With the development of exploration technology, the exploration of structuralhydrocarbon reservoir whose type is simple has been almost exhausted. So, how to findlithologic hydrocarbon reservoir will be the main research direction. Seismicexploration is an important method to find and determine lithologic hydrocarbonreservoir. Seismic impedance inversion is very useful for reservoir description andinterpretation, and it can accurately describe the rock formations and geologicalstructure, provides basis for drilling wells. In many practical engineering problems,their essence can be transformed into optimization. Seismic impedance inversion is amulti-parameter nonlinear optimization problem, its main purpose is to obtain thereflection coefficient of underground medium, and then get the density, velocity andother parameters. In recent years, various non-linear methods are developing rapidly,the heuristics inversion methods included genetic algorithm, simulated annealingalgorithm, ant colony algorithm and artificial neural network, these new emergingmethods bring hope and dawn for seismic inversion. Particle swarm optimization (PSO)algorithm originated from the movement behavior of birds, is a technique based onswarm intelligence evolutionary computation. Because of its principle is simple andeasy to implement, PSO currently has been applied to neural network training, functionoptimization, fuzzy system control, and many other fields and have achieved goodresults.This paper first introduces the topic basis and the general development of seismicwave impedance inversion and the particle swarm optimization algorithm, brieflyelaborates PSO theory and research status, and analysis the principle of PSO and itsimplementation flow. For particle swarm optimization algorithm, the particle is rapidconvergence at early, but at later it is easy to fall into local optimum. Hence, I putforward a nonlinear inertia weight particle swarm algorithm based on simulatedannealing to solve this situation. In the algorithm nonlinear inertia weight factor is toexpand the search space of the population in the early, to enhance the particles’convergence. The SA-ULWPSO Fusion Algorithm make in Swarm’s flight process, theSwam not only can accept good function value point, but also can accept worsefunction value point by certain probability. Through carried experiment, the result showSA-ULWPSO Fusion Algorithm increased the Swarm’s multiplicity, and strengthenedSwarm’s get rid of the partial optimal solution ability, can’t run into local minimumeasily, strengthen the overall situation detecting ability, and has higher convergence rateand precision. This paper is based on a large number of reading other articles and understandingthe theory of PSO algorithm fully, then applies PSO algorithm to one-dimensionalseismic impedance inversion. For the complicated and multi-dimensional inverseproblem of the conventional particle swarm optimization algorithm, thereare disadvantages of slow convergence and low accuracy. Therefore the paper appliesSA-ULWPSO Fusion Algorithm to one--dimensional and two-dimensional waveimpedance model, showing the feasibility and reliability the algorithm. Finally, theimproved algorithm is applied to the actual seismic data, and inversion results are alsobetter, showing the practicability and validity of the algorithm.