Quantification Method of Physical Parameters Based on Neural Network Algorithm
A neural network algorithm and physical parameter technology, applied in the field of seismic data reservoir prediction, can solve problems such as weak convergence ability, incorrect relationship establishment, large deviation of prediction results, etc.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0029] A physical parameter quantification method based on a neural network algorithm, which extracts attributes related to physical parameters from seismic data and adds them to the input of the neural network; adds known physical parameters to the output of the neural network, and then according to the neural network The algorithm establishes the relationship between output and input, and applies this relationship to all seismic data to realize the quantitative prediction of physical parameters.
[0030] Described method specifically comprises the steps:
[0031] a. Collect the physical parameters of wells drilled in various regions or geological structures, select the category of parameters to be predicted (such as porosity, permeability), and determine the value of this parameter;
[0032] b. Normalize the known physical parameters, the formula is: vv=(v-vmin) / (vmax-vmin);
[0033] (V is the size of the parameter itself, vv is the normalized size, vmin and vmax are the mi...
Embodiment 2
[0041] figure 1 It is a schematic diagram of a commonly used BP network structure. The network consists of an input layer, one or several intermediate layers, and an output layer. Each layer contains multiple nodes, and the nodes between layers are connected to each other to form a network. The known information X=(x1, x2,...,xn) is input from the input terminal, and reaches the output layer through network operations, and the output value Y=(y1, y2,...,ym) is obtained. In pattern recognition, the output value expresses the classification result, such as 1 for the first category, 0.5 for the second category, 0 for the third category, etc.
[0042] As a quantitative prediction of physical parameters, each node at the output end has the same type of parameters, but the values are different, such as y1=10, y2=4, y3=0, etc.
[0043] For example, suppose there are 3 wells in a certain area or geological structure, the porosity of well 1 is 13, the porosity of well 2 is 6, and ...
Embodiment 3
[0045] This embodiment is a quantitative prediction of gas field permeability.
[0046] 1. Add the following permeability parameters of known well reservoir intervals to the output of the neural network.
[0047]
[0048] 2. Extract multiple attribute parameters for each channel of the seismic data reservoir section, and input them to the input terminal of the neural network.
[0049] 3. Establish the network relationship (weight) between input and output.
[0050] 4. Apply weights to all seismic data to obtain prediction results.
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 