| Popis: |
As full waveform inversion (FWI) has become a topic of great interest in the oil and gas industry, its field application becomes a focus. With factors being studied like huge computational cost, strict requirement of input data, and so on, an approach of FWI on GPU clusters is proposed. Testing on synthetic datasets proves that the method is able to describe and build theoretically high-precision velocity model, where an acceleration ratio of more than forty times is achieved. Trial processing of land field datasets proves that the velocity model inverted from FWI in the Shengli Oilfield leads to promising imaging result than that from other velocity model building methods. Through further analysis, this paper points out the strategy of applying FWI on land datasets. |
