CN102418518A - Method for identifying water logging grades of oil reservoir by using neural network analogue cross plot - Google Patents
Method for identifying water logging grades of oil reservoir by using neural network analogue cross plot Download PDFInfo
- Publication number
- CN102418518A CN102418518A CN2011100903423A CN201110090342A CN102418518A CN 102418518 A CN102418518 A CN 102418518A CN 2011100903423 A CN2011100903423 A CN 2011100903423A CN 201110090342 A CN201110090342 A CN 201110090342A CN 102418518 A CN102418518 A CN 102418518A
- Authority
- CN
- China
- Prior art keywords
- cross plot
- network
- neural network
- oil
- layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention discloses a method for identifying water logging grades of an oil reservoir by using a neural network analogue cross plot. In the method, the conventional cross plot technology is improved by a neural network algorithm, so nonlinear identification and quantitative analysis functions of the cross plot are realized, and a back propagation (BP) neural network algorithm is used and the method comprises the following steps of screening object characteristic parameters, selecting network structure parameters, training a neural network model, testing the network model, and establishing a neural network analogue cross plot layout. The method specifically comprises the following steps of: according to various characteristics of oil, gas and water layers in reservoirs, accurately selecting parameter samples which can best reflect the characteristics of the oil, gas and water layers in the reservoirs from parameters calculated during well logging or the well logging curves relevant to oil and gas interpretation by a statistics method; selecting appropriate weight values and threshold values by the BP neural network algorithm to establish the network model, and training the model and checking errors; and judging the fluid type or water logging degree of the reservoir with the depth according to projective points of identification vectors which are obtained by network output on a plane.
Description
Technical field:
The present invention relates to other method of a kind of neuron network simulation cross plot identification reservoir water flooding level, promptly utilize the network topology structure of neutral net to make up the method that cross plot carries out Water Flooding Layer classification that the oil reservoir logging Water Flooding Layer explains and well logging oil gas, water layer identification.
Background technology:
The cross plot recognition technology is widely used in oil-gas exploration, it in the inspection logging data quality, select interpretation parameters, confirm lithology, check explanation results and estimate and bringing into play important effect aspect the formation fluid type.Aspect petrophysics, can make the rock physics template through cross plot, utilize template to carry out lithology prediction; At earthquake AVO (Amplitude Versus Offset, amplitude is with the variation of offset distance) technical application aspect,, be about to AVO attribute (λ ρ-μ ρ, I through cros splot technique
P-I
SDeng) project on the cross plot, utilize the lithology of different reservoir and fluid type occupies zones of different unusually on the cross plot plane characteristics, divide unusually; Aspect reservoir and fluid explanation, cros splot technique not only can be used for attribute optimization, can be applied to the qualitative identification of the division of reservoir type, the evaluation of reservoir logging in water flooded layer and reservoir fluid simultaneously.
The tradition cross plot generally is to select relevant with research object two kinds of parameters or attribute, on the XY coordinate plane, makes up interpretation chart, and it is unable to do what one wishes that this method just seems when the attribute of object is more, shortage operability and accuracy.Meanwhile; In traditional crossplot analysis and its; Generally adopt rough description or manual work to sketch to the division of cross plot territory inner region, the very big uncertainty of existence of method itself is particularly under the more situation of data point; The sample point of different attribute overlaps, and is difficult to quickly and accurately the classification under the sample point judged and discerned.The existence of these problems becomes the new highlight of cross plot Study of recognition.
The BP neutral net is a kind of multilayer feedforward neural network, and the main feature of this network is the transmission of signal forward direction, error back propagation.In forward direction transmitted, input signal was successively handled through hidden layer from input layer, until output layer.The neuron state of each layer only influences one deck neuron state down.If output layer can not get desired output, then change backpropagation over to, according to predicated error adjustment network weight and threshold value, thereby make network prediction output constantly approach desired output.
The topological structure of BP neutral net as shown above, X among the figure
1, X
2..., X
nBe the input value of BP neutral net, Y
1, Y
2Be the predicted value of BP neutral net, ω
Ij, ω
JkBe the BP neural network weight.As can be seen from the figure, when the BP neutral net can be regarded a nonlinear function m as, the BP neutral net had just been expressed the Function Mapping relation from n independent variable to m dependent variable.Therefore, can solve the type identification problem of multi-parameter (attribute) object with it.The identification vector projection in the plane that combines network output to obtain again, the distance at the classification center that obtains through measuring and calculating subpoint and model training can realize the quantitative recognition function of cross plot.
Summary of the invention:
Applying neural network technology of the present invention is improved traditional cros splot technique; Thereby realize the non-linear identification and the quantitative analysis function of cross plot; What utilize is the BP neural network algorithm, comprises the foundation of network model, the structure of the training of network and network output simulation cross plot; Make the input of network directly obtain an identification cross plot, and the error precision of measuring and calculating cross plot.
Comprising the screening of characteristics of objects parameter, the selection of network architecture parameters, the training of neural network model, the test of network model and five steps of formulation of neuron network simulation intersection plate.According to the multiple character of reservoir hydrocarbons layer, water layer, the utilization statistical method is explained relevant log from the well logging parameters calculated or with oil gas, comes accurate selection can reflect the parameter sample of oil-gas Layer in the reservoir, water layer characteristics; Use the BP neural network algorithm then, select proper weight and threshold value to come the building network model, and to model training and error-tested; The identification vector subpoint in the plane that obtains through output at last to network; Judge the distance between these distributed points and each the fluid type central point with Euclidean distance, wherein the fluid type apart from the shortest central point representative is exactly the fluid type or the water flooding degree of this degree of depth section reservoir.
The key technology main points comprise:
(1) setting of network input/output terminal;
The utilization statistical method selects to explain relevant log with oil gas or the input as network is unified after the dimension in the calculating parameter normalization of logging well, like natural gamma curve, resistivity curve, degree of porosity, oil saturation; The output layer of network is made as two-layer, so that be projected on the two dimensional surface coordinate system, wherein the oil reservoir code is made as (0.25,0.25), and weak Water Flooding Layer code is made as (0.75,0.25), and middle Water Flooding Layer is made as (0.75,0.75), and strong Water Flooding Layer is made as (0.25,0.75).
(2) realization of BP neural network model;
Through research to sample number strong point characteristic, according to the algorithm building network model of BP neutral net, and model carried out training and testing, error is lower than and presets the error criterion model and be available.
(3) neuron network simulation cross plot Quantitative Classification Method;
The subpoint that utilization Euclidean distance method measuring and calculating disaggregated model obtains obtains the water flooded grade of this layer position apart from the distance at each fluid center, realizes the quantitative classification of cross plot.
Description of drawings
Figure 1B P neural network topology structure figure
Fig. 2 neuron network simulation cross plot algorithm sketch map
Fig. 3 neuron network simulation cross plot water logging identification plate
Fig. 4 BP neural metwork training error-detecting figure
Fig. 5 neuron network simulation intersection plate classification reservoir water flooded grade is figure as a result
The specific embodiment:
Method is formed:
Neural network algorithm and well logging cross plot recognition technology are merged in the present invention, and traditional cross plot method is transformed, and realize multi-parameter research and quantitatively identification to oil field Water Flooding Layer well logging recognition and well logging oil gas, water layer identification.
See that from the statistics viewpoint problems such as well logging Water Flooding Layer evaluation or fluid identification are actually a classification problem, therefore can utilize neural net method to set up disaggregated model and accomplish.The neural network model that utilizes sample training to obtain, various features parameter that can the comprehensive study object accurately divides the classification of object, remedied the limitation that cross plot can only comprehensive two kinds of characteristic parameters carries out category division.
The cognizance code that the utilization of cross plot then obtains neural network model is projected in the plane coordinate system; Distribution situation through further analysis subpoint; The measuring and calculating subpoint just can quantitatively be divided the characteristic parameter of identifying object apart from the distance of classification center on cross plot.
Research method:
The recognition methods of neuron network simulation cross plot combines the basic principle of neural network algorithm and cross plot identification, and the key of method is the distance measuring and calculating of the accuracy and the cross plot data point of network model structure.Its detailed technology path classificating introduction is following:
(1) sample is chosen, and in well log interpretation, the authenticity of learning sample, representativeness and generalization are the keys of decision classifying quality, mainly is made up of formation testing result and one group of corresponding with it log response value
(2) sample normalization causes negative effect for avoiding each parameter dimension difference to predicting the outcome, the property value of learning sample is carried out normalization handle, and property value is the conventional normalization of carrying out of normal distribution, and what property value was Non-Gaussian Distribution carries out logarithm normalization.
(3) classification results digitlization shows in the plane that for the ease of network model output setting the oil reservoir code is (0.25; 0.25), weak Water Flooding Layer code is made as (0.75,0.25); Middle Water Flooding Layer is made as (0.75,0.75), and strong Water Flooding Layer is made as (0.25; 0.75), four water logging centers are as the disaggregated model output.
(4) making up the BP neural network model, according to the structure of system's inputoutput data characteristics research network, mainly is that the node number of hidden layer and the excitation function of hidden layer reasonably are set.
(5) training BP neural network model, the sample that screening obtains in core analysis or the oil test data be to the training of BP neural network classification model, in training process according to the weights and the threshold value of network predicated error adjustment network.
(6) test network model, the disaggregated model that utilization is set up is predicted the Water Flooding Layer sample that is obtained by formation testing result and well-log information and is returned and declare, the classifying quality of test model.
(7) make neuron network simulation cross plot plate; The BP neural network classification model that utilization trains, the water logging parameter of unknown water logging type layer position of classifying, and be projected in plane coordinates and fasten; Utilization Euclidean distance measuring and calculating subpoint obtains the water flooded grade of this layer position apart from the distance at each water logging center.
Claims (3)
1. the present invention relates to a kind of cross plot Classification and Identification technology based on neural network algorithm; Be used for the identification of the Water Flooding Layer classification of oil reservoir logging Water Flooding Layer explanation and log well oil-gas Layer, water layer; It is characterized in that the nerual network technique that will have Error Feedback and non-linear recognition function combines with traditional cross plot recognition technology; Apply in the identification of oil-gas Layer in the stratum, water layer and go, and with two dimension (also can the be three-dimensional) output of the neutral net network topology structure that training obtains to sample learning, as the input of traditional cross plot; Through the distribution characteristics of comparative analysis sample point on cross plot, confirm the classification that this sample point should belong to.
Step 1; Multiple character according to the research sample; The utilization statistical method selects to reflect the parameter of oil-water-layer characteristics and fluid type that oil test data obtains or water flooding degree input vector and the object vector as neutral net from well logging parameters calculated or log;
Step 2 after input vector being carried out normalization and object vector carried out digitlization, utilizes the BP neutral net that network is trained; In training process according to the weights and the threshold value of network predicated error adjustment network; Finally obtain the network model in the error allowed band, owing to require the network output vector on two-dimentional cross plot, to show, so setting network is the dual output layer; And the coordinate of several kinds of reservoir fluids is made as fixed value and is presented on the cross plot, is called the fluid type central point;
Step 3; With the neural network classification characteristic of fluid parameter that trains; Obtain one group of bivector data point; It is projected on the two-dimentional cross plot plane form distributed points, judge the distance between these distributed points and each the fluid type central point with Euclidean distance again, wherein the fluid type of the shortest central point representative of distance is exactly the corresponding fluid type of this characteristic of fluid parameter.
2. what neutral net according to claim 1 was selected for use is the BP neural network algorithm; It is characterized in that the transmission of signal forward direction; Error back propagation, it is to instruct a kind of learning algorithm that is suitable for the multilayer neuroid down the tutor, is based upon on the basis of gradient algorithm.
3. cross plot recognition technology according to claim 1 is a kind of mapping interpretation technique of well-log information.It is two kinds of log data intersections on plan view, makes the numerical value or the scope of the parameter of asking according to the coordinate of plotted point, is a kind of method that extensively adopts when confirming lithology, identification Water Flooding Layer and hydrocarbon saturation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100903423A CN102418518A (en) | 2011-04-12 | 2011-04-12 | Method for identifying water logging grades of oil reservoir by using neural network analogue cross plot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100903423A CN102418518A (en) | 2011-04-12 | 2011-04-12 | Method for identifying water logging grades of oil reservoir by using neural network analogue cross plot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102418518A true CN102418518A (en) | 2012-04-18 |
Family
ID=45943138
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011100903423A Pending CN102418518A (en) | 2011-04-12 | 2011-04-12 | Method for identifying water logging grades of oil reservoir by using neural network analogue cross plot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102418518A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103197348A (en) * | 2013-03-26 | 2013-07-10 | 西北大学 | Method using internal samples at reservoirs to carry out weighting and compile logging crossplot |
| CN103324981A (en) * | 2013-05-20 | 2013-09-25 | 浙江大学 | Chemoradiotherapy standardized quality control quantitative method based on neural network |
| CN103867196A (en) * | 2014-04-01 | 2014-06-18 | 北京师范大学 | Method for recognizing petrographic rhythm change in siltstone and mudstone alternate stratum through imaging logging image |
| CN104380143A (en) * | 2012-06-18 | 2015-02-25 | 悉尼大学 | Systems and methods for processing geophysical data |
| CN104376361A (en) * | 2014-10-15 | 2015-02-25 | 南京航空航天大学 | Nuclear accident source item inversion method based on BP neural network algorithm |
| CN104453875A (en) * | 2014-10-29 | 2015-03-25 | 中国石油集团川庆钻探工程有限公司 | Shale gas reservoir recognition method based on self-organizing competitive neural network |
| CN106062310A (en) * | 2014-02-28 | 2016-10-26 | 界标制图有限公司 | Facies definition using unsupervised classification procedures |
| CN106250984A (en) * | 2016-07-29 | 2016-12-21 | 中国石油天然气股份有限公司 | The determination methods of the oil water relation pattern of oil well and device |
| CN109583333A (en) * | 2018-11-16 | 2019-04-05 | 中证信用增进股份有限公司 | Image-recognizing method based on water logging method and convolutional neural networks |
| CN109727238A (en) * | 2018-12-27 | 2019-05-07 | 贵阳朗玛信息技术股份有限公司 | The recognition methods of x-ray chest radiograph and device |
| CN109800521A (en) * | 2019-01-28 | 2019-05-24 | 中国石油大学(华东) | A kind of oil-water relative permeability curve calculation method based on machine learning |
| CN110284873A (en) * | 2019-06-27 | 2019-09-27 | 中国石油集团东方地球物理勘探有限责任公司 | A kind of oil well preserves the detection method and detection device of property |
| CN110458169A (en) * | 2019-07-22 | 2019-11-15 | 中海油信息科技有限公司 | A kind of landwaste CT characteristics of image recognition methods |
| CN110552693A (en) * | 2019-09-19 | 2019-12-10 | 中国科学院声学研究所 | layer interface identification method of induction logging curve based on deep neural network |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1811306A (en) * | 2006-02-22 | 2006-08-02 | 天津大学 | Automatic volume regulating and controlling method for gas-burning machine heat pump |
| CN101487898A (en) * | 2009-02-27 | 2009-07-22 | 中国石油集团川庆钻探工程有限公司 | Method for oil gas water recognition by employing longitudinal wave seismic exploration post-stack data |
-
2011
- 2011-04-12 CN CN2011100903423A patent/CN102418518A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1811306A (en) * | 2006-02-22 | 2006-08-02 | 天津大学 | Automatic volume regulating and controlling method for gas-burning machine heat pump |
| CN101487898A (en) * | 2009-02-27 | 2009-07-22 | 中国石油集团川庆钻探工程有限公司 | Method for oil gas water recognition by employing longitudinal wave seismic exploration post-stack data |
Non-Patent Citations (1)
| Title |
|---|
| 梁丽梅等: "神经网络模拟交会图在低阻油层流体识别中的应用", 《石油工业计算机应用》 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104380143A (en) * | 2012-06-18 | 2015-02-25 | 悉尼大学 | Systems and methods for processing geophysical data |
| CN103197348A (en) * | 2013-03-26 | 2013-07-10 | 西北大学 | Method using internal samples at reservoirs to carry out weighting and compile logging crossplot |
| CN103197348B (en) * | 2013-03-26 | 2015-04-15 | 西北大学 | Method using internal samples at reservoirs to carry out weighting and compile logging crossplot |
| CN103324981A (en) * | 2013-05-20 | 2013-09-25 | 浙江大学 | Chemoradiotherapy standardized quality control quantitative method based on neural network |
| CN103324981B (en) * | 2013-05-20 | 2015-12-02 | 浙江大学 | Based on the quantization method that the chemicotherapy normalized quality of neural network controls |
| CN106062310A (en) * | 2014-02-28 | 2016-10-26 | 界标制图有限公司 | Facies definition using unsupervised classification procedures |
| CN103867196A (en) * | 2014-04-01 | 2014-06-18 | 北京师范大学 | Method for recognizing petrographic rhythm change in siltstone and mudstone alternate stratum through imaging logging image |
| CN103867196B (en) * | 2014-04-01 | 2019-03-22 | 北京师范大学 | A method of replacing the lithofacies rhythm in stratum with mud stone using imaging logging image identification siltstone and changes |
| CN104376361A (en) * | 2014-10-15 | 2015-02-25 | 南京航空航天大学 | Nuclear accident source item inversion method based on BP neural network algorithm |
| CN104453875B (en) * | 2014-10-29 | 2018-05-01 | 中国石油集团川庆钻探工程有限公司 | Shale gas reservoir stratum identification method based on Self-organizing Competitive Neutral Net |
| CN104453875A (en) * | 2014-10-29 | 2015-03-25 | 中国石油集团川庆钻探工程有限公司 | Shale gas reservoir recognition method based on self-organizing competitive neural network |
| CN106250984A (en) * | 2016-07-29 | 2016-12-21 | 中国石油天然气股份有限公司 | The determination methods of the oil water relation pattern of oil well and device |
| CN109583333A (en) * | 2018-11-16 | 2019-04-05 | 中证信用增进股份有限公司 | Image-recognizing method based on water logging method and convolutional neural networks |
| CN109727238A (en) * | 2018-12-27 | 2019-05-07 | 贵阳朗玛信息技术股份有限公司 | The recognition methods of x-ray chest radiograph and device |
| CN109800521A (en) * | 2019-01-28 | 2019-05-24 | 中国石油大学(华东) | A kind of oil-water relative permeability curve calculation method based on machine learning |
| CN110284873A (en) * | 2019-06-27 | 2019-09-27 | 中国石油集团东方地球物理勘探有限责任公司 | A kind of oil well preserves the detection method and detection device of property |
| CN110458169A (en) * | 2019-07-22 | 2019-11-15 | 中海油信息科技有限公司 | A kind of landwaste CT characteristics of image recognition methods |
| CN110552693A (en) * | 2019-09-19 | 2019-12-10 | 中国科学院声学研究所 | layer interface identification method of induction logging curve based on deep neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102418518A (en) | Method for identifying water logging grades of oil reservoir by using neural network analogue cross plot | |
| Abbaszadeh et al. | Permeability prediction by hydraulic flow units—theory and applications | |
| AU668903B2 (en) | Method for estimating formation permeability from wireline logs using neural networks | |
| CN112505778B (en) | Three-dimensional in-situ characterization method for heterogeneity of shale storage and generation performance | |
| CN108897066B (en) | Carbonate rock crack density quantitative prediction method and device | |
| AU664047B2 (en) | Neural network interpretation of aeromagnetic data | |
| EP1151326A1 (en) | Uncertainty constrained subsurface modeling | |
| CN111489034B (en) | Construction method and application of oil and gas reservoir permeability prediction model | |
| KR101625660B1 (en) | Method for making secondary data using observed data in geostatistics | |
| US11126694B2 (en) | Automatic calibration for modeling a field | |
| CN108376295A (en) | A kind of oil gas dessert prediction technique and storage medium | |
| CN111766635A (en) | Sand body communication degree analysis method and system | |
| CN107316341A (en) | A kind of Multiple-Point Geostatistics facies modelization method | |
| CN110988997A (en) | Hydrocarbon source rock three-dimensional space distribution quantitative prediction technology based on machine learning | |
| Alameedy et al. | Evaluating machine learning techniques for carbonate formation permeability prediction using well log data | |
| CN108665545B (en) | Logging parameter three-dimensional geological model establishing method | |
| CN108765562B (en) | Oil-gas productivity evaluation method based on three-dimensional geological model | |
| US11927717B2 (en) | Optimized methodology for automatic history matching of a petroleum reservoir model with Ensemble Kalman Filter (EnKF) | |
| US11340381B2 (en) | Systems and methods to validate petrophysical models using reservoir simulations | |
| Amanipoor | Providing a subsurface reservoir quality maps in oil fields by geostatistical methods | |
| CN111856569B (en) | Stratum sand body prediction method and device | |
| CN117174203B (en) | Logging curve response analysis method for sandstone uranium ores | |
| Ebinesan et al. | Rapid history matching of petroleum production from well logs and 4D seismic via Machine Learning techniques in the Norne Field, offshore Norway | |
| Li et al. | Application of neural network technique for logging fluid identification in low resistance reservoir | |
| CN115857015A (en) | Method for quantitatively predicting distribution of tuff in volcanic formation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20120418 |