Seismic prediction of porosity in tight reservoirs based on transformer

Porosity is a crucial index in reservoir evaluation. In tight reservoirs, the porosity low, resulting weak seismic responses to changes porosity. Moreover, relationship between and response complex, making accurate inversion prediction challenging. This paper proposes Transformer-based multi-attribute method for address this issue. The proposed takes multiple attributes as input data output data. Transformer mapping transformation network consists of an encoder, multi-head attention layer, decoder optimized training with gating mechanism variable selection module. Applying actual from sandstone gas exploration area Sichuan Basin yielded coincidence rate 95% well