Multiple Wave AVA Inversion Based On Chaotic Particle Swarm

Lithologic parameter is very important to oil or gas detection and lithologyidentification. we pick lithologic parameters from the relationship of changes in theamplitude characteristic in seismic exploration. The information of amplitudecharacteristic in pre-stack seismic data is far more than post-stack seismic data. Theprecision, based on pre-stack AVA inversion of seismic data, to extract the formationof longitudinal wave velocity, shear wave velocity and density is reliable. Due to theinversion of the nonlinear characteristics, a lot of nonlinear algorithm is more andmore widely used in geophysical field. In this paper, the particle swarm optimization(PSO) algorithm in the application of multiple wavesAVAinversion is studied.The forward theory research is the basis of the inversion study. Zoeppritzequation is the basis of the forward research, in homogeneous isotropic medium,according to the transmission of the wave equation of plane wave and its kinematicsand dynamics on the interface, the reflection and transmission coefficient of planewave on the boundary is deduced. In this paper, gotten P-P and P-SV wave AVAcurve by the theoretical model through Zoeppritz formula and its approximateformulas, at the same time, analyzed the deviation between approximation formulaswith precise formula respectively. Based on convolution model, use the minimumphase wavelet to get these two types of wave angle gathers seismic records. ParticleSwarm Optimization algorithm (PSO), is based on Swarm intelligence Optimizationalgorithm, searching optimization through the information sharing mechanismbetween groups. This algorithm, firstly, initialize a group of random solutions, and inthe process of iteration algorithm, by the interaction between individual and group,individual and individual, search the global optimal solution. Because the algorithmis easy to fall into local extremum and search precision is not high, so this paper, thechaotic will be introduced to particle swarm optimization algorithm for optimizationproblems. Chaotic systems, as one of nonlinear phenomenon, it is seemingly randomand complex, but has the greatly strengthened internal regularity. The randomness,ergodicity and sensitivity of initial value of chaotic, compared with the randomsearch，has more advantages to avoid the algorithm into local optimum, thus improvethe accuracy of the algorithm. In this paper, the state of chaos based on logisticmapping are analyzed, and combines Logistic mapping and particle swarmoptimization (PSO) into the chaos particle swarm optimization (CPSO), through thetest function test, the results show that the superiority of the proposed algorithm. First, P-P and P-SV and the two types of joint inversion of simple model whichis two layers, the results showed that the accuracy of joint inversion is higher than theaccuracy of the single wave inversion. Followed, by using joint inversion ofmultilayer model, the results showed that as the layer increasing, the parameters of theinversion increases accordingly, chaotic particle swarm optimization algorithm alsohas its advantage. In the parameter given under different initial scope, the result ofinversion shows that the model precision of algorithm will be effect by initial searchscope. Analyzing sensitivity of parameters, through disturbance caused by thedisturbance of calculation parameters, it can be seen that various parameters playdifferent role on the inversion result, and demonstrates the feasibility of the selectedobjective function. Inversion of a sample point model, at last, the algorithm wasapplied to the actual seismic data processing.