Research And Implementation Of Residual Static Correction Based On Hybrid Parallel Computing

With an increasingly important role the energy including oil and gas plays in today’s society, the importance of seismic exploration technology has reached an unprecedented height. Seismic exploration technology is the key to improve the accuracy of hydrocarbon exploration. As an important part of seismic exploration technology, residual static correction plays a vital role for the analysis of the location of hydrocarbon. However, the actual amount of seismic data, which is used for residual static correction, is large-scale. It will take huge time for the traditional ways which used serial computation to do residual static correction in general, but parallel computing not. Parallel computing has obvious advantages in solving a problem with large amount of data and high time complexity. What the paper major studies is the implementation of residual static correction based on Hybrid Parallel Computing. The main researches are as follow:(1) Research on the basic theory of residual statics correction, mastered the nonlinear and big data of residual static correction. In order to pick residual statics correction, we focus on learning the fundamental of the methods which are used to solve residual statics correction and the process how the methods solve. The methods include stack-power maximitation algorithm, simulated annealing algorithm, genetic algorithm and hybrid optimization algorithm.(2) Study on the basic theory of parallel computing. Especially, an in-depth research on MPI and OpenMP we take, we mainly focus on the design patterns and programming basics of the two parallel programming models: MPI and OpenMP. At last, we unflod experiment simulations to test the two parallel programming models and efficiency analysis.(3) Detail analysis of how to achieve residual static correction, including the choice of objective function and the method of model trace building. Research in-depth on the parameter selection and solution step of all methods refered in(1) in solving residual statics correction. Simultaneously, we complete the achievement of residual static correction in conventional serial computation model with all methods refered in(1). We test them with both theoretic and real seismic data, and analyze the effect and efficiency of each method above all.(4) An implementation model of residual static correction based on MPI-OpenMP Hybrid Parallel Computing is presented. We also complete the achievement of residual static correction in this model with all methods refered in(1). And we test them with the same data as serial computation model do. In order to reflect the advantage of hybrid parallel computing model, we compare the results with serial computation model’s in effectiveness and efficiency. At last, we implement the software modules of residual static correction with C on Linux operating system.