System for intelligently recovering water flooded layer of oil reservoir based on spiking neural network

Abstract

A system for intelligently recovering water-flooded layer of oil reservoir based on spiking neural network is provided. The system includes: a data acquiring module, a data preprocessing module, a model training module, a resistivity recovering module, and a water-flooded layer interpreting module. The data acquiring module is configured to acquire multi-modal data such as conventional and electrical well logging curves. The data preprocessing module is configured to perform range normalization and Z-Score standardization on the data. The model training module is configured to construct a resistivity recovering model based on a spiking neural network, and the resistivity recovering model is improved and verified by various manners. The resistivity recovering module is configured to recover an original resistivity. The water-flooded layer interpreting module is configured to calculate correlative parameters based on the original resistivity to determine a water-flooded layer and classify and interpret it.

What is claimed is: 
     
         1 . A system for intelligently recovering a water-flooded layer of an oil reservoir based on a spiking neural network, comprising: a data acquiring module, a data preprocessing module, a model training module, a resistivity recovering module, and a water-flooded layer interpreting module;
 wherein the data acquiring module is configured to acquire data of multi-modal well logging curves, and wherein the multi-modal well logging curves comprise conventional well logging curves and electrical well logging curves;   wherein the data preprocessing module is configured to perform range normalization and Z-Score standardization on the acquired data;   wherein the model training module is configured to construct a resistivity recovering model based on a spiking neural network, wherein the resistivity recovering model adopts a Leaky Integrate-and-Fire neuron model as a basic unit and receives acquired data via a feedforward-loop hybrid structure to generate a prediction value of resistivity, and wherein the model training module is configured to combine plasticity of synapses and global error backpropagation to adjust weights of the synapses to optimize performance of the resistivity recovering model;   wherein the resistivity recovering module is configured to recover an original resistivity of an oil layer according to the optimized resistivity recovering model; and,   wherein the water-flooded layer interpreting module is configured to calculate oil saturation and water production rate based on the original resistivity to determine a water-flooded layer and quantitatively classify and interpret the water-flooded layer.   
     
     
         2 . The system for intelligently recovering a water-flooded layer of an oil reservoir based on a spiking neural network according to  claim 1 , wherein the conventional well logging curves comprise natural gamma-ray, acoustic interval transit time, compensated neutron log, density, and self potential; and the electrical well logging curves comprise deep laterolog resistivity and shallow laterolog resistivity. 
     
     
         3 . The system for intelligently recovering a water-flooded layer of an oil reservoir based on a spiking neural network according to  claim 1 , wherein the range normalization is expressed as: 
       
         
           
             
               
                 
                   x 
                   norm 
                 
                 = 
                 
                   
                     x 
                     - 
                     
                       x 
                       min 
                     
                   
                   
                     
                       x 
                       max 
                     
                     - 
                     
                       x 
                       min 
                     
                   
                 
               
               ; 
             
           
         
         wherein X norm  is a normalized data, x min  is a minimum value in original data, x max  is a maximum value in the original data, and x max −x min  is a range of data; 
         and the Z-Score standardization is expressed as: 
       
       
         
           
             
               
                 
                   x 
                   std 
                 
                 = 
                 
                   
                     x 
                     - 
                     μ 
                   
                   σ 
                 
               
               ; 
             
           
         
         wherein x std  is a standardized data, is a mean of the original data, and a is a standard deviation of the original data. 
       
     
     
         4 . The system for intelligently recovering a water-flooded layer of an oil reservoir based on a spiking neural network according to  claim 1 , wherein a dynamic expression of the Leaky Integrate-and-Fire neuron model is: 
       
         
           
             
               
                 
                   
                     τ 
                     m 
                   
                   ⁢ 
                   
                     dV 
                     dt 
                   
                 
                 = 
                 
                   
                     - 
                     
                       ( 
                       
                         V 
                         - 
                         
                           V 
                           rest 
                         
                       
                       ) 
                     
                   
                   + 
                   
                     
                       I 
                       syn 
                     
                     ( 
                     t 
                     ) 
                   
                 
               
               ; 
             
           
         
         wherein V is a membrane potential, τ m  is a membrane time constant, V rest  is a resting potential, and I syn (t) is an input current of the synapses; 
         wherein prediction of a resistivity is performed as follows: when the membrane potential exceeds a threshold V th , neurons emit a pulse and reset to V reset , the feedforward-loop hybrid structure is engaged, an input layer of the resistivity recovering model receives the acquired data, a hidden layer of the resistivity recovering model contains pulse neuron nodes, and an output layer of the resistivity recovering model generates the prediction value of resistivity via pulse frequency encoding. 
       
     
     
         5 . The system for intelligently recovering a water-flooded layer of an oil reservoir based on a spiking neural network according to  claim 1 , wherein an expression for adjusting weights of the synapses is: 
       
         
           
             
               
                 
                   Δ 
                   ⁢ 
                   
                     w 
                     ij 
                   
                 
                 = 
                 
                   η 
                   · 
                   
                     
                       ∑ 
                       t 
                     
                     
                       
                         
                           δ 
                           j 
                         
                         ( 
                         t 
                         ) 
                       
                       · 
                       
                         
                           S 
                           i 
                         
                         ( 
                         
                           t 
                           - 
                           
                             Δ 
                             ⁢ 
                             t 
                           
                         
                         ) 
                       
                     
                   
                 
               
               ; 
             
           
         
         wherein η is a learning rate, δ j (t) is an error gradient, Δw ij  is a change in weights of the synapses, and S i (t−Δt) is a pulse state before time Δt; 
         and wherein expressions for evaluating the resistivity recovering model are: 
       
       
         
           
             
               
                 
                   
                     
                       RMSE 
                       = 
                       
                         
                           
                             1 
                             n 
                           
                           ⁢ 
                           
                             
                               ∑ 
                               
                                 i 
                                 = 
                                 1 
                               
                               n 
                             
                             
                               
                                 ( 
                                 
                                   
                                     R 
                                     pred 
                                     
                                       ( 
                                       i 
                                       ) 
                                     
                                   
                                   - 
                                   
                                     R 
                                     true 
                                     
                                       ( 
                                       i 
                                       ) 
                                     
                                   
                                 
                                 ) 
                               
                               2 
                             
                           
                         
                       
                     
                     ; 
                   
                 
               
               
                 
                   
                     
                       MAE 
                       = 
                       
                         
                           
                             1 
                             n 
                           
                           
                             ∑ 
                             
                               i 
                               = 
                               1 
                             
                             n 
                           
                         
                         ❘ 
                         
                           
                             R 
                             pred 
                             
                               ( 
                               i 
                               ) 
                             
                           
                           - 
                           
                             R 
                             true 
                             
                               ( 
                               i 
                               ) 
                             
                           
                         
                         ❘ 
                       
                     
                     ; 
                   
                 
               
               
                 
                   
                     
                       
                         R 
                         2 
                       
                       = 
                       
                         1 
                         - 
                         
                           
                             
                               ∑ 
                               
                                 i 
                                 = 
                                 1 
                               
                               n 
                             
                             
                               
                                 ( 
                                 
                                   
                                     R 
                                     pred 
                                     
                                       ( 
                                       i 
                                       ) 
                                     
                                   
                                   - 
                                   
                                     R 
                                     true 
                                     
                                       ( 
                                       i 
                                       ) 
                                     
                                   
                                 
                                 ) 
                               
                               2 
                             
                           
                           
                             
                               ∑ 
                               
                                 i 
                                 = 
                                 1 
                               
                               n 
                             
                             
                               
                                 ( 
                                 
                                   
                                     R 
                                     pred 
                                     
                                       ( 
                                       i 
                                       ) 
                                     
                                   
                                   - 
                                   
                                     
                                       R 
                                       true 
                                     
                                     _ 
                                   
                                 
                                 ) 
                               
                               2 
                             
                           
                         
                       
                     
                     ; 
                   
                 
               
             
           
         
         wherein RMSE is a root mean square error, MAE is a mean absolute error, R 2  is a coefficient of determination, 
       
       
         
           
             
               R 
               pred 
               
                 ( 
                 i 
                 ) 
               
             
           
         
       
       is a predicted value of an i-th resistivity, and 
       
         
           
             
               R 
               true 
               
                 ( 
                 i 
                 ) 
               
             
           
         
       
       is a true value of the i-th resistivity. 
     
     
         6 . The system for intelligently recovering a water-flooded layer of an oil reservoir based on a spiking neural network according to  claim 1 , wherein an expression of the oil saturation is: 
       
         
           
             
               
                 
                   S 
                   o 
                 
                 = 
                 
                   1 
                   - 
                   
                     
                       ( 
                       
                         
                           
                             RT 
                             original 
                           
                           
                             RT 
                             current 
                           
                         
                         · 
                         
                           
                             a 
                             · 
                             
                               ϕ 
                               
                                 - 
                                 m 
                               
                             
                           
                           
                             σ 
                             w 
                           
                         
                       
                       ) 
                     
                     
                       1 
                       n 
                     
                   
                 
               
               ; 
             
           
         
         wherein S o  is a current oil saturation of a reservoir body, RT original  is the original resistivity, RT current  is a current resistivity, a, m, n are coefficients related to rock porosity structure and saturation, ϕ is an effective porosity of stratum, and σ w  is a viscosity of water; 
         and an expression for the water production rate is: 
       
       
         
           
             
               
                 
                   F 
                   w 
                 
                 = 
                 
                   
                     
                       K 
                       rw 
                     
                     
                       μ 
                       w 
                     
                   
                   
                     
                       
                         K 
                         rw 
                       
                       
                         μ 
                         w 
                       
                     
                     + 
                     
                       
                         K 
                         ro 
                       
                       
                         μ 
                         o 
                       
                     
                   
                 
               
               ; 
             
           
         
         wherein K rw  is a relative permeability of an aqueous phase, K ro  is a relative permeability of an oil phase, μ w  is a viscosity of the aqueous phase, and μ o  is a viscosity of the oil phase. 
       
     
     
         7 . The system for intelligently recovering a water-flooded layer of an oil reservoir based on a spiking neural network according to  claim 1 , wherein quantitatively classifying is performed as follows: the water-flooded layer is quantitatively classified and interpreted based on an oil saturation-change value ΔS o  and a resistivity-change value ΔRT, and the water-flooded layer is quantitatively classified to be four levels: unflooded, weakly flooded, moderately-strongly flooded, and strongly flooded.