Multi-wave Seismic Oil And Gas Reservoir Prediction Based On Deep Neural Network

Multi wave seismic data has rich seismic reservoir information.It is helpful to improve the accuracy of reservoir prediction by effectively using the difference of sensitivity between P-wave and S-wave.Deep neural network is widely used in reservoir feature extraction,classification,recognition and prediction because of its strong robustness,generalization ability and high stability.Based on it,this study combines the distribution law of oil and gas reservoir and its characteristics in seismic response,based on multi wave seismic data and artificial intelligence technology.A method of multiwave seismic oil and gas identification and prediction based on deep neural network is designed.The unsupervised and supervised learning algorithm is used to extract oil and gas characteristics,so as to realize the accurate identification and prediction of seismic oil and gas reservoir.The deep neural network learning model for multiwave seismic reservoir distribution prediction mainly includes:First,we use clustering unsupervised learning to optimize the seismic attributes of P-and S-waves,and obtain the seismic attributes of P-and S-waves that are sensitive to the response of seismic oil and gas reservoirs.Then,multi wave attribute combination is carried out based on the difference of P-wave and S-wave response to oil-gas reservoir to enhance the characteristics of seismic oil-gas response.And the multi wave composite seismic attributes obtained through the above processing are processed by principal component technology to reduce data redundancy and highlight the characteristics of multi wave seismic oil-gas reservoir.Finally,the seismic attributes at well points are used as training samples,three kinds of principal component information are used as input for deep neural network learning and prediction under supervised learning,so as to realize the prediction of seismic oil and gas reservoir from known to unknown.Compared with the traditional neural network,the deep neural network has a stronger ability of data mining,and the oil-gas reservoir boundary is more clearly,and it has achieved the expected effect.Based on the geological data of the study area and the previous research results,the above scheme have been applied to the prediction of seismic reservoirs in HG area,and a better prediction result was obtained.The scope of oil and gas reservoirs delineated by this method was clearer.Combined with the geological data of the study area and previous research results,the oil and gas reservoir described by this method is more clearly and it has a good coincidence with the actual situation and makes a prediction of the favorable exploration area.It shows that the scheme is feasible and effective and provides a new way for the accurate identification and prediction of oil and gas reservoirs.