CN104110251A - Pumping unit indicator diagram identification method based on ART2 - Google Patents
Pumping unit indicator diagram identification method based on ART2 Download PDFInfo
- Publication number
- CN104110251A CN104110251A CN201410283957.1A CN201410283957A CN104110251A CN 104110251 A CN104110251 A CN 104110251A CN 201410283957 A CN201410283957 A CN 201410283957A CN 104110251 A CN104110251 A CN 104110251A
- Authority
- CN
- China
- Prior art keywords
- pumping unit
- indicator diagram
- sample
- piston
- oil
- 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
- 238000005086 pumping Methods 0.000 title claims abstract description 98
- 238000010586 diagram Methods 0.000 title claims abstract description 84
- 238000000034 method Methods 0.000 title claims abstract description 44
- 102100024405 GPI-linked NAD(P)(+)-arginine ADP-ribosyltransferase 1 Human genes 0.000 title claims abstract 10
- 101000981252 Homo sapiens GPI-linked NAD(P)(+)-arginine ADP-ribosyltransferase 1 Proteins 0.000 title claims abstract 10
- 238000013528 artificial neural network Methods 0.000 claims abstract description 29
- 238000003745 diagnosis Methods 0.000 claims abstract description 13
- 239000003129 oil well Substances 0.000 claims description 20
- 239000013598 vector Substances 0.000 claims description 16
- 241001023788 Cyttus traversi Species 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 14
- 238000007781 pre-processing Methods 0.000 claims description 13
- 239000007788 liquid Substances 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 12
- 238000000605 extraction Methods 0.000 claims description 11
- 230000001174 ascending effect Effects 0.000 claims description 9
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 241001331845 Equus asinus x caballus Species 0.000 claims description 3
- 239000012530 fluid Substances 0.000 claims description 3
- 230000003068 static effect Effects 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 abstract description 5
- 238000003909 pattern recognition Methods 0.000 description 11
- 238000005516 engineering process Methods 0.000 description 8
- 238000004519 manufacturing process Methods 0.000 description 5
- 230000007787 long-term memory Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000011426 transformation method Methods 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000004576 sand Substances 0.000 description 2
- 238000011425 standardization method Methods 0.000 description 2
- 239000000725 suspension Substances 0.000 description 2
- 235000017166 Bambusa arundinacea Nutrition 0.000 description 1
- 235000017491 Bambusa tulda Nutrition 0.000 description 1
- 241001330002 Bambuseae Species 0.000 description 1
- 235000015334 Phyllostachys viridis Nutrition 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 239000011425 bamboo Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000008021 deposition Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 101150050759 outI gene Proteins 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000005507 spraying Methods 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 230000008719 thickening Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Landscapes
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a pumping unit indicator diagram identification method based on an ART2. By collecting load and suspension center displacement signals of a pumping unit, a closed curve formed along with time changes in a pumping period is drawn with polished rod displacement signals serving as an x-coordinate and loading signals serving as a y-coordinate, and a pumping unit indicator diagram sample is obtained. If the category of the pumping unit indicator diagram sample is known, the pumping unit indicator diagram sample is identified through a mode with a supervision module. If the sample category of the pumping unit indicator diagram sample is unknown in advance, the pumping unit indicator diagram sample is identified through a mode without the supervision module. The pumping unit indicator diagram sample is identified through an ART2 neural network, according to an identification result, fault diagnosis and underground work condition judgment are performed on the pumping unit, and a judgment result is displayed. The pumping unit indicator diagram identification method based on the ART2 has the advantages of performing fault diagnosis and underground work condition judgment on the pumping unit, identifying the indicator diagram stably, efficiently and accurately, and the like.
Description
Technical Field
The invention relates to an energy-saving method of a pumping unit, in particular to an ART2 (Adaptive resonance theory) -based pumping unit indicator diagram identification method.
Background
With the continuous progress of science and technology, the fact that the requirement for oil field exploitation is higher and higher, and the 'digital oil field' is no longer a slogan is achieved. Oil fields increasingly need automatic oil measuring technology with high automation degree and strong real-time performance. The traditional oil measuring technology (such as oil measuring at a metering station, oil measuring at a tipping bucket and oil measuring at a glass tube) has the problems of complex process flow, more devices, high investment cost, high labor intensity, low efficiency and the like, and is difficult to adapt to the production management requirements of simplifying ground flow and continuously calculating yield.
Abundant working condition information is contained in the indicator diagram of the pumping unit, and the working condition and the underground working condition of the pumping unit can be judged by analyzing the indicator diagram. A method for diagnosing the fault of the oil pumping system of the sucker-rod pump by utilizing the indicator diagram is long-standing. The fault of the oil pumping unit is judged by using a method with manual identification. Later, according to the research results of the research institute of the midwestern united states, an API simulation indicator diagram comparison diagnosis technology appears. By 1966, after american scientists s s.g.gibbs created a one-dimensional wave equation for solving pump work diagrams, pump work diagram-based fault diagnosis techniques have rapidly developed.
The actually measured indicator diagram is used as a curve carrier for recording the working condition of the well pump of the pumping unit, and various abnormal phenomena of the deep well pump can be reflected. The indicator diagram is analyzed, so that people can be helped to find out the main reasons influencing the failure of the deep-well pump, whether the working system of the pumping well is reasonable, whether the combination of the parameters of the pump, the rod and the pump pumping oil is adaptive to the underground liquid supply condition or not is evaluated, whether the oil well is sand, gas, wax and the like or not is indirectly reflected, whether different media in the well have negative influence on an oil well pump, namely ground equipment or not is indirectly judged, and finally, the failure of the oil well is pertinently removed according to the diagnosis and analysis result of the indicator diagram, so that the method has important significance for ensuring the normal production of the oil well or improving the yield of the.
In the engineering practice, the indicator diagram of the pumping unit is very complicated, and the typical working conditions are more than 18, such as insufficient liquid supply, piston falling out of a working barrel, gas influence, air lock, sucker rod breaking, oil pipe leakage, sand production of an oil well, pump bottom collision, pump top collision, continuous pumping belt spraying, oil thickening, fixed valve leakage, floating valve leakage, double valve simultaneous leakage, liquid impact, wax deposition, piston seizure, stroke loss and the like.
Pattern recognition is a scientific discipline aimed at classifying objects. Depending on the application, the object may be an image, a waveform signal, or any other quantity that needs to be classified. In recent years, with the development of computer technology, people have achieved fruitful results in the field of pattern recognition, and these results in turn expand the application field of computer technology and promote the development of artificial intelligence technology. At present, an effective method for efficiently and accurately judging and identifying indicator diagrams of all oil pumping units does not exist.
Disclosure of Invention
The invention provides an ART 2-based oil pumping unit indicator diagram identification method for avoiding the defects in the prior ART, so as to carry out fault diagnosis and downhole working condition judgment on the oil pumping unit through the identification of the indicator diagram.
The invention adopts the following technical scheme to solve the technical problem.
An ART 2-based oil pumping unit indicator diagram identification method is characterized in that the identification process is shown in figure 1 and comprises the following steps:
step 1: drawing a theoretical indicator diagram of the oil pumping unit (the theoretical indicator diagram is shown in figure 2) to obtain a sample of the indicator diagram of the oil pumping unit;
step 2: carrying out data preprocessing on the sample of the indicator diagram of the pumping unit obtained in the step 1;
and step 3: performing feature extraction on the sample of the pumping unit indicator diagram subjected to data preprocessing in the step 2;
and 4, step 4: if the category number of the samples of the indicator diagram of the oil pumping unit obtained in the step 1 is known, firstly, a set of known categories is used as a training set to train the ART2 neural network, a discriminant model is established, and then the established model is used for identifying unknown samples according to the similarity principle; if the category number of the sample of the pumping unit indicator diagram obtained in the step 1 is unknown, directly identifying the sample by depending on the natural characteristics of the sample; during identification, inputting a sample of the indicator diagram of the pumping unit into a preset ART2 neural network;
and 5: the ART2 neural network is used for identifying the sample of the indicator diagram of the pumping unit, carrying out fault diagnosis and underground working condition judgment on the pumping unit according to the identification result, and displaying the judgment result.
As shown in fig. 2, in step 1, the process of drawing the theoretical indicator diagram of the pumping unit is as follows:
step 101: starting from a horse head bottom dead center A of the oil pumping unit, enabling a polish rod of the oil pumping unit to ascend, closing a traveling valve and a fixed valve of the oil pumping unit, enabling the polish rod to bear the mass of a liquid column on the upper part of a piston of the oil pumping unit, enabling an AB interval to be a load increasing process, and finishing load increase when reaching a point B;
step 102: starting from a point B where the piston starts to ascend, the piston ascends, the fixed valve is opened, the traveling valve is closed, the BC section is an ascending line of the polish rod, and the polish rod finishes ascending after reaching the upper dead point C of the horse head;
step 103: starting from the upper dead point C of the horse head, the polished rod descends, the traveling valve and the fixed valve are both closed, liquid in the oil well pump begins to be discharged, the CD interval is the unloading process, and when the D point is reached, the unloading is finished;
step 104: starting from a piston descending point D, descending the piston, closing the fixed valve, opening the traveling valve, descending the polished rod in a DA interval to a mule head bottom dead center A, finishing the descending of the polished rod, and finishing a pumping cycle;
in the theoretical indicator diagram, the abscissa S is the stroke of the polished rod, the starting point of the coordinate is the lower dead center A of the horse head, and the end point is the upper dead center C of the horse head; the ordinate P is the load of the polish rod; b is the starting ascending point of the piston, and D is the starting descending point of the piston;the mass of the liquid column above the piston of the oil well pump;the mass of the sucker rod string immersed in the well fluid;is the stroke of the piston;the stroke loss of the oil well pump;the length of the sucker rod is the telescopic length;the length of the oil pipe is the telescopic length; the OA section is the minimum static load the polished rod is subjected to on the downstroke.
The data preprocessing in the step 2 adopts a central transformation method, a logarithmic transformation method or a standardization method.
And the feature extraction in the step 3 adopts a invariant moment feature extraction method.
The ART2 neural network training algorithm for training the ART2 neural network in the step 4 is as follows:
step 401: parameters to initialize ART2 neural networksa，b，c，d，θ，eAlarm value(ii) a Connection matrix from F1 layer to F2 layerAnd a connection matrix from the F2 layer to the F1 layerThen the feature vector is addedThe input of the network is carried out,∈[0,1], i∈[1,n]。
step 402: calculate the vectors in layer F1:x，w，u，v，q，p；
step 403: computing input vectors in layer F2Calculating winning nodesJ(ii) a When the F2 layer is not excited, all；
Step 404: information feedback is carried out; by winning nodes at level F2JSend back the top-down weight vectorAnd calculate a value；
Step 405: detecting a warning line; if it isThen receiveJA winning node, step 406; otherwise, a Reset signal is sent to(not allowing it to participate in the competition), start the search phase, step 402;
step 406: according to the weight value adjustment formula:the bottom-up and top-down weight vectors are adjusted. The parameters in the ART2 neural network in the invention belong to common parameters, and the ART2 neural network is described in detail in the paper: ART2 neural network study of slow weight update (dawn, bamboo forest; ART2 neural network study of slow weight update; computer engineering and application, 2010,46 (24): 146-.
Compared with the prior art, the invention has the beneficial effects that:
according to the oil pumping unit indicator diagram identification method based on ART2, the oil pumping unit indicator diagram sample is obtained by collecting oil pumping unit load and suspension point displacement signals, drawing a closed curve formed by time change in a pumping period by taking a polished rod displacement signal as an abscissa and a load signal as an ordinate. For the collected oil pumping unit indicator diagram samples, if the number of the classes of the samples is known, a set of known classes is used as a training set to establish a discriminant model, and then the established model is used for identifying unknown samples according to the similarity principle, so that the supervised pattern identification is called. If the sample class is not known in advance, the method of completely relying on the natural characteristics of the sample for identification is called unsupervised pattern identification.
The ART2 neural network is adopted to identify a plurality of invariant moment characteristics of the indicator diagram, and 18 typical working conditions of the oil pumping unit can be efficiently and accurately identified, so that fault diagnosis and underground working condition judgment are carried out on the oil pumping unit, and further the oil well fault can be pertinently relieved. The invention has important significance for ensuring the normal production of the oil well or improving the yield of the oil well.
The method for identifying the indicator diagram of the pumping unit based on the ART2 has the advantages that fault diagnosis and underground working condition judgment can be carried out on the pumping unit, and the indicator diagram can be stably, efficiently and accurately identified.
Drawings
Fig. 1 is a schematic diagram of a pattern recognition process of the oil pumping unit indicator diagram recognition method.
Fig. 2 is a theoretical indicator diagram of the oil pumping unit indicator diagram identification method of the invention.
Fig. 3 is an ART2 network structure of the oil pumping unit indicator diagram identification method.
The present invention will be further described with reference to the following detailed description and accompanying drawings.
Detailed Description
Referring to fig. 1 to 3, an ART 2-based oil pumping unit indicator diagram identification method (the identification process is shown in fig. 1) comprises the following steps:
step 1: drawing a theoretical indicator diagram of the oil pumping unit (the theoretical indicator diagram is shown in figure 2) to obtain a sample of the indicator diagram of the oil pumping unit;
step 2: carrying out data preprocessing on the sample of the indicator diagram of the pumping unit obtained in the step 1;
and step 3: performing feature extraction on the sample of the pumping unit indicator diagram subjected to data preprocessing in the step 2;
and 4, step 4: if the category number of the samples of the indicator diagram of the oil pumping unit obtained in the step 1 is known, firstly, a set of known categories is used as a training set to train the ART2 neural network, a discriminant model is established, and then the established model is used for identifying unknown samples according to the similarity principle; if the category number of the sample of the pumping unit indicator diagram obtained in the step 1 is unknown, directly identifying the sample by depending on the natural characteristics of the sample; during identification, inputting a sample of the indicator diagram of the pumping unit into a preset ART2 neural network;
and 5: the ART2 neural network is used for identifying the sample of the indicator diagram of the pumping unit, carrying out fault diagnosis and underground working condition judgment on the pumping unit according to the identification result, and displaying the judgment result.
As shown in fig. 2, in step 1, the process of drawing the theoretical indicator diagram of the pumping unit is as follows:
step 101: starting from a horse head bottom dead center A of the oil pumping unit, enabling a polish rod of the oil pumping unit to ascend, closing a traveling valve and a fixed valve of the oil pumping unit, enabling the polish rod to bear the mass of a liquid column on the upper part of a piston of the oil pumping unit, enabling an AB interval to be a load increasing process, and finishing load increase when reaching a point B;
step 102: starting from a point B where the piston starts to ascend, the piston ascends, the fixed valve is opened, the traveling valve is closed, the BC section is an ascending line of the polish rod, and the polish rod finishes ascending after reaching the upper dead point C of the horse head;
step 103: starting from the upper dead point C of the horse head, the polished rod descends, the traveling valve and the fixed valve are both closed, liquid in the oil well pump begins to be discharged, the CD interval is the unloading process, and when the D point is reached, the unloading is finished;
step 104: starting from a piston descending point D, descending the piston, closing the fixed valve, opening the traveling valve, descending the polished rod in a DA interval to a mule head bottom dead center A, finishing the descending of the polished rod, and finishing a pumping cycle;
in the theoretical indicator diagram, the abscissa S is the stroke of the polished rod, the starting point of the coordinate is the lower dead center A of the horse head, and the end point is the upper dead center C of the horse head; the ordinate P is the load of the polish rod; b is the starting ascending point of the piston, and D is the starting descending point of the piston;the mass of the liquid column above the piston of the oil well pump;the mass of the sucker rod string immersed in the well fluid;is the stroke of the piston;the stroke loss of the oil well pump;the length of the sucker rod is the telescopic length;the length of the oil pipe is the telescopic length; the OA section is the minimum static load the polished rod is subjected to on the downstroke. Fig. 2 is a theoretical indicator diagram of the pumping unit. The indicator diagram is a closed curve formed by the change of the displacement of the suspension point of the pumping unit and the load along with the time in a pumping period. The theoretical indicator diagram is in the form of a parallelogram in a rectangular coordinate system, as shown in FIG. 1, and is analyzed and shownThe work diagram can judge the working condition of the pumping unit and the underground working condition. The worker can easily confirm the fault of the oil pumping unit through the identification result of the invention, can judge the underground working condition, can specifically remove the fault of the oil well, and can ensure the normal production of the oil well or improve the yield of the oil well.
The data preprocessing in the step 2 adopts a central transformation method, a logarithmic transformation method or a standardization method. The purpose of data preprocessing in pattern recognition is to remove noise and reduce data dimensionality. The original data may have large data size and a plurality of dimensions, the time complexity of computer processing is high, and the data dimensions can be reduced by preprocessing.
And the feature extraction in the step 3 adopts a invariant moment feature extraction method.
The ART2 neural network training algorithm for training the ART2 neural network in the step 4 is as follows:
step 401: parameters to initialize ART2 neural networksa，b，c，d，θ，eAlarm value(ii) a Connection matrix from F1 layer to F2 layerAnd a connection matrix from the F2 layer to the F1 layerThen the feature vector is addedThe input of the network is carried out,∈[0,1], i∈[1,n]。
step 402: calculate the vectors in layer F1:x，w，u，v，q，p；
step 403: computing input vectors in layer F2Calculating winning nodesJ(ii) a When the F2 layer is not excited, all；
Step 404: information feedback is carried out; by winning nodes at level F2JSend back the top-down weight vectorAnd calculate a value；
Step 405: detecting a warning line; if it isThen receiveJA winning node, step 406; otherwise, a Reset signal is sent to(not allowing it to participate in the competition), start the search phase, step 402;
step 406: according to the weight value adjustment formula:the bottom-up and top-down weight vectors are adjusted.
The introduction of the mode recognition technology opens up a way for the diagnosis automation and scientific quantification of the working condition of the pumping system of the sucker-rod pump. At present, various feature value extraction methods and pattern recognition algorithms can be used for recognizing the indicator diagram. For example, an ART or ART2 neural network is adopted to identify a plurality of invariant moment characteristics of the indicator diagram, so that fault diagnosis, underground working condition judgment and the like are carried out on the pumping unit.
The indicator diagram identification method based on the ART2 neural network has the following advantages:
(1) the ART2 neural network adopts a teacher-free, competitive and self-stabilization learning mechanism, and can obtain high identification capability without a complete training set.
(2) The warning value is adjustable, which means that the identification precision is adjustable, and the identification precision can be adjusted according to the specific identification requirement, so that the best effect is obtained.
(3) The network has the characteristics of real-time work and real-time learning, and has high learning speed and lower requirements on hardware performance.
(4) The learned approximate indicator diagram can be quickly and directly accessed from the memory bank, and the indicator diagram of new contact can be deeply memorized and stored in a long-term memory bank (LTM), so that the method has stable and quick classification capability.
The structure of the ART2 neural network adopted by the invention is shown in fig. 3, and the whole network consists of two parts, namely an attention subsystem and an orientation subsystem. The subsystem comprises a preprocessing layer F0, a comparison layer F1, a presentation layer F2 and an adaptive Long Term Memory (LTM) connected between the F1 layer and the F2 layer, wherein the comparison layer F1 consists of 6 sublayers (L), (L) and (L)w、x、v、u、p、q) And 3 gain control modules (filled circles in FIG. 2), the pre-processing layer F0 being composed of 4 sub-layers (C:)w 0、x 0、v 0、u 0) And 2 gain control modules. The orientation subsystem consists of a reset mechanism.
In the context of figure 3, it is shown,in order to be the input mode,response for layer F2 as input, vectorMonitoringAnddegree of matching with a system alarm valueAnd comparing to judge whether the current reaction result of the system F2 layer is correct.Andrepresenting the connection weights of the F1 layer to the F2 layer and the F2 layer to the F1 layer, respectively.
The F0 layer was similar in structure to the F1 layer. The addition of the F0 layer enables input modesAs well as an input to the orientation subsystem.
Fig. 1 is a schematic diagram of a pattern recognition process. From the viewpoint of the nature of the processing problem and the method of solving the problem, pattern recognition can be classified into supervised pattern recognition and unsupervised pattern recognition. If the class number of the sample is known, firstly, a set of known classes is used as a training set to establish a discriminant model, and then the established model is used for identifying the unknown sample according to the similarity principle, so that the supervised pattern identification is called. If the sample class is not known in advance, the method of completely relying on the natural characteristics of the sample for identification is called unsupervised pattern identification.
The data preprocessing in the pattern recognition aims at removing noise and reducing data dimensionality, and the method mainly comprises center transformation, logarithmic transformation, standardization and the like.
Feature extraction is carried outIs to make a set containing n metric valuesBy some transformation, a set of m eigenvalues is generatedThe purpose of the new classification feature (or called quadratic feature) is to reduce the feature space dimension as much as possible and perform effective classification on the premise of retaining the identification information.
The moment features mainly characterize the geometric features of the image region, which are also called geometric moments and are also called invariant moments because of invariant features of rotation, translation, scale and the like. In image processing, the geometric invariant moment can be used as an important feature to represent an object, and the image can be classified according to the feature.
The invariant moment is a highly concentrated image feature, and has the characteristics of clear concept, stable recognition rate, inconvenient rotation and scaling and the like, because the essential features of the image can be effectively reflected.
In the continuous case, let the image function beOf the imageThe order geometric moment (standard moment) is defined as:
the order center moment is defined as:
wherein:representing the center of the image.
For discrete digital images:
wherein,NandMrepresenting the height and width of the image, respectively.
The normalized central moment is defined as:
wherein,。
the following 7 invariant moments can be constructed using the second and third order normalized central moments：
In order to reduce the distribution range of moment values and facilitate comparison, an evolution method can be adopted for data compression, and the actually adopted moment is as follows in consideration of the possibility of negative values of the moment:。
in the present invention, the ART and ART2 neural network are referred to as follows:
[1].G. A. Carpenter，S. Grossberg，A massively parallel architecture for a self-organizing neural pattern recognition machine[J]，Computer Vision, Graphics, and Image Processing，1987，37(1)：54-115；
[2].G.A.Carpenter，S.Grossberg，ART2: Self_organization of stable category recognition codes for analog input patterns[J]，Applied Optics，1987，01(26)： 4919~4930。
Claims (4)
1. An ART 2-based oil pumping unit indicator diagram identification method is characterized by comprising the following steps:
step 1: drawing a theoretical indicator diagram of the pumping unit to obtain a sample of the indicator diagram of the pumping unit;
step 2: carrying out data preprocessing on the sample of the indicator diagram of the pumping unit obtained in the step 1;
and step 3: performing feature extraction on the sample of the pumping unit indicator diagram subjected to data preprocessing in the step 2;
and 4, step 4: if the category number of the samples of the indicator diagram of the oil pumping unit obtained in the step 1 is known, firstly, a set of known categories is used as a training set to train the ART2 neural network, a discriminant model is established, and then the established model is used for identifying unknown samples according to the similarity principle; if the category number of the sample of the pumping unit indicator diagram obtained in the step 1 is unknown, directly identifying the sample by depending on the natural characteristics of the sample; during identification, inputting a sample of the indicator diagram of the pumping unit into a preset ART2 neural network;
and 5: the ART2 neural network is used for identifying the sample of the indicator diagram of the pumping unit, carrying out fault diagnosis and underground working condition judgment on the pumping unit according to the identification result, and displaying the judgment result.
2. The ART 2-based pumping unit indicator diagram identification method as claimed in claim 1, wherein in the step 1, the process of drawing the theoretical indicator diagram of the pumping unit is as follows:
step 101: starting from a horse head bottom dead center A of the oil pumping unit, enabling a polish rod of the oil pumping unit to ascend, closing a traveling valve and a fixed valve of the oil pumping unit, enabling the polish rod to bear the mass of a liquid column on the upper part of a piston of the oil pumping unit, enabling an AB interval to be a load increasing process, and finishing load increase when reaching a point B;
step 102: starting from a point B where the piston starts to ascend, the piston ascends, the fixed valve is opened, the traveling valve is closed, the BC section is an ascending line of the polish rod, and the polish rod finishes ascending after reaching the upper dead point C of the horse head;
step 103: starting from the upper dead point C of the horse head, the polished rod descends, the traveling valve and the fixed valve are both closed, liquid in the oil well pump begins to be discharged, the CD interval is the unloading process, and when the D point is reached, the unloading is finished;
step 104: starting from a piston descending point D, descending the piston, closing the fixed valve, opening the traveling valve, descending the polished rod in a DA interval to a mule head bottom dead center A, finishing the descending of the polished rod, and finishing a pumping cycle;
in the theoretical indicator diagram, the abscissa S is the stroke of the polished rod, the starting point of the coordinate is the lower dead center A of the horse head, and the end point is the upper dead center C of the horse head; the ordinate P is the load of the polish rod; b is the starting ascending point of the piston, D is the starting descending point of the piston； The mass of the liquid column above the piston of the oil well pump;the mass of the sucker rod string immersed in the well fluid;is the stroke of the piston;the stroke loss of the oil well pump;the length of the sucker rod is the telescopic length;the length of the oil pipe is the telescopic length; the OA section is the minimum static load the polished rod is subjected to on the downstroke.
3. The ART 2-based pumping unit indicator diagram identification method as claimed in claim 1, wherein the feature extraction in step 3 is a invariant moment feature extraction method.
4. The ART 2-based oil pumping unit indicator diagram recognition method as claimed in claim 1, wherein the ART2 neural network training algorithm for training the ART2 neural network in the step 4 is as follows:
step 401: parameters to initialize ART2 neural networksa，b，c，d，θ，eAlarm value(ii) a Connection matrix from F1 layer to F2 layerAnd a connection matrix from the F2 layer to the F1 layerThen the feature vector is addedThe input of the network is carried out,∈[0,1], i∈[1,n]。
step 402: calculate the vectors in layer F1:x，w，u，v，q，p；
step 403: computing input vectors in layer F2Calculating winning nodesJ(ii) a When the F2 layer is not excited, all；
Step 404: information feedback is carried out; by winning nodes at level F2JSend back the top-down weight vectorAnd calculate a value；
Step 405: detecting a warning line; if it isThen receiveJA winning node, step 406; otherwise, a Reset signal is sent to(not allowing it to participate in the competition), start the search phase, step 402;
step 406: according to the weight value adjustment formula:the bottom-up and top-down weight vectors are adjusted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410283957.1A CN104110251A (en) | 2014-06-24 | 2014-06-24 | Pumping unit indicator diagram identification method based on ART2 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410283957.1A CN104110251A (en) | 2014-06-24 | 2014-06-24 | Pumping unit indicator diagram identification method based on ART2 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104110251A true CN104110251A (en) | 2014-10-22 |
Family
ID=51707224
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410283957.1A Pending CN104110251A (en) | 2014-06-24 | 2014-06-24 | Pumping unit indicator diagram identification method based on ART2 |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104110251A (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104729773A (en) * | 2015-03-24 | 2015-06-24 | 沈阳理工大学 | On-line soft measuring method and device for indicator diagram of beam-pumping unit based on RBF neural network |
| CN104732063A (en) * | 2015-02-10 | 2015-06-24 | 中国石油天然气股份有限公司 | Narrow strip-shaped indicator diagram working condition analysis method |
| CN105484730A (en) * | 2015-12-28 | 2016-04-13 | 江西飞尚科技有限公司 | Upper and lower dead point accurate extraction method based on sucker rod indicator diagram |
| CN105525913A (en) * | 2015-12-21 | 2016-04-27 | 浙江中控技术股份有限公司 | Oil sucking machine suspension point displacement determining method and indicator |
| CN106401561A (en) * | 2016-11-10 | 2017-02-15 | 中国石油化工股份有限公司 | Real-time working condition diagnostic method for belt type oil pumping unit |
| CN106844328A (en) * | 2016-08-23 | 2017-06-13 | 华南师范大学 | A kind of new extensive document subject matter semantic analysis and system |
| CN106930751A (en) * | 2017-04-20 | 2017-07-07 | 东北大学 | A kind of Dlagnosis of Sucker Rod Pumping Well fault separating method |
| CN108153925A (en) * | 2016-12-06 | 2018-06-12 | 中国石油天然气股份有限公司 | Method and device for evaluating efficiency of oil pumping unit |
| CN108710920A (en) * | 2018-06-05 | 2018-10-26 | 北京中油瑞飞信息技术有限责任公司 | Indicator card recognition methods and device |
| CN108764361A (en) * | 2018-06-01 | 2018-11-06 | 北京中油瑞飞信息技术有限责任公司 | The operating mode's switch method of beam type oil pumping machine indicating diagram based on integrated study |
| CN109934239A (en) * | 2019-03-07 | 2019-06-25 | 上海皓桦科技股份有限公司 | Image characteristic extracting method |
| CN110439537A (en) * | 2019-08-13 | 2019-11-12 | 北京助创科技有限公司 | The method of pumping unit electric work figure inverting ground function figure |
| CN111980664A (en) * | 2020-07-28 | 2020-11-24 | 中国石油天然气股份有限公司 | Narrow strip-shaped indicator diagram analysis method for oil well under special working condition |
| CN112228042A (en) * | 2020-10-22 | 2021-01-15 | 南京富岛信息工程有限公司 | Cloud edge cooperative computing-based rod-pumped well working condition similarity judgment method |
| CN112360427A (en) * | 2019-07-26 | 2021-02-12 | 中国石油天然气股份有限公司 | Alarm method for variation trend of oil well pump supply and discharge relation |
| CN112651165A (en) * | 2019-09-25 | 2021-04-13 | 北京国双科技有限公司 | Wax precipitation initial stage identification model obtaining method, wax precipitation initial stage identification method and device |
| CN112766301A (en) * | 2020-12-11 | 2021-05-07 | 南京富岛信息工程有限公司 | Similarity judgment method for indicator diagram of oil extraction machine |
| CN115434690A (en) * | 2021-06-04 | 2022-12-06 | 中国科学院沈阳自动化研究所 | Unsupervised online mutation point detection and fusion method for pumping unit based on Bayes |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5220640A (en) * | 1990-09-20 | 1993-06-15 | Motorola, Inc. | Neural net architecture for rate-varying inputs |
| CN102011576A (en) * | 2010-11-24 | 2011-04-13 | 河海大学 | Method for hierarchically diagnosing fault of rod-comprising oil pumping system based on indicator diagram |
| CN102043900A (en) * | 2010-11-24 | 2011-05-04 | 河海大学 | Failure prediction method of rod pumping system based on indicator diagram |
| CN102184414A (en) * | 2011-05-16 | 2011-09-14 | 中国石油天然气股份有限公司 | Method and system for identifying and judging pump indicator diagram |
| CN103177180A (en) * | 2011-12-20 | 2013-06-26 | 郑芳田 | Method for screening modeling samples of prediction model |
-
2014
- 2014-06-24 CN CN201410283957.1A patent/CN104110251A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5220640A (en) * | 1990-09-20 | 1993-06-15 | Motorola, Inc. | Neural net architecture for rate-varying inputs |
| CN102011576A (en) * | 2010-11-24 | 2011-04-13 | 河海大学 | Method for hierarchically diagnosing fault of rod-comprising oil pumping system based on indicator diagram |
| CN102043900A (en) * | 2010-11-24 | 2011-05-04 | 河海大学 | Failure prediction method of rod pumping system based on indicator diagram |
| CN102184414A (en) * | 2011-05-16 | 2011-09-14 | 中国石油天然气股份有限公司 | Method and system for identifying and judging pump indicator diagram |
| CN103177180A (en) * | 2011-12-20 | 2013-06-26 | 郑芳田 | Method for screening modeling samples of prediction model |
Non-Patent Citations (2)
| Title |
|---|
| YUEHUI PENG，ET.AL: "Indicator Diagram Identification Based on ART2 Neural Network and Features of Moment Invariant", 《INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS，COMMUNICATIONS AND NETWORKS (CECNET2012)》 * |
| 潘志坚等: "自适应神经网络示功图识别", 《石油学报》 * |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104732063A (en) * | 2015-02-10 | 2015-06-24 | 中国石油天然气股份有限公司 | Narrow strip-shaped indicator diagram working condition analysis method |
| CN104729773A (en) * | 2015-03-24 | 2015-06-24 | 沈阳理工大学 | On-line soft measuring method and device for indicator diagram of beam-pumping unit based on RBF neural network |
| CN105525913A (en) * | 2015-12-21 | 2016-04-27 | 浙江中控技术股份有限公司 | Oil sucking machine suspension point displacement determining method and indicator |
| CN105484730A (en) * | 2015-12-28 | 2016-04-13 | 江西飞尚科技有限公司 | Upper and lower dead point accurate extraction method based on sucker rod indicator diagram |
| CN106844328A (en) * | 2016-08-23 | 2017-06-13 | 华南师范大学 | A kind of new extensive document subject matter semantic analysis and system |
| CN106844328B (en) * | 2016-08-23 | 2020-04-21 | 华南师范大学 | Large-scale document theme semantic analysis method and system |
| CN106401561B (en) * | 2016-11-10 | 2019-10-25 | 中国石油化工股份有限公司 | A kind of belt type pumping unit real-time working condition diagnostic method |
| CN106401561A (en) * | 2016-11-10 | 2017-02-15 | 中国石油化工股份有限公司 | Real-time working condition diagnostic method for belt type oil pumping unit |
| CN108153925A (en) * | 2016-12-06 | 2018-06-12 | 中国石油天然气股份有限公司 | Method and device for evaluating efficiency of oil pumping unit |
| CN106930751B (en) * | 2017-04-20 | 2020-02-11 | 东北大学 | Fault separation method for sucker-rod pump pumping well |
| CN106930751A (en) * | 2017-04-20 | 2017-07-07 | 东北大学 | A kind of Dlagnosis of Sucker Rod Pumping Well fault separating method |
| CN108764361B (en) * | 2018-06-01 | 2020-12-22 | 北京中油瑞飞信息技术有限责任公司 | Working condition identification method of indicator diagram of beam-pumping unit based on integrated learning |
| CN108764361A (en) * | 2018-06-01 | 2018-11-06 | 北京中油瑞飞信息技术有限责任公司 | The operating mode's switch method of beam type oil pumping machine indicating diagram based on integrated study |
| CN108710920A (en) * | 2018-06-05 | 2018-10-26 | 北京中油瑞飞信息技术有限责任公司 | Indicator card recognition methods and device |
| CN109934239A (en) * | 2019-03-07 | 2019-06-25 | 上海皓桦科技股份有限公司 | Image characteristic extracting method |
| CN109934239B (en) * | 2019-03-07 | 2021-09-21 | 上海皓桦科技股份有限公司 | Image feature extraction method |
| CN112360427B (en) * | 2019-07-26 | 2023-11-28 | 中国石油天然气股份有限公司 | Alarm method for change trend of oil pump supply and discharge relation |
| CN112360427A (en) * | 2019-07-26 | 2021-02-12 | 中国石油天然气股份有限公司 | Alarm method for variation trend of oil well pump supply and discharge relation |
| CN110439537A (en) * | 2019-08-13 | 2019-11-12 | 北京助创科技有限公司 | The method of pumping unit electric work figure inverting ground function figure |
| CN110439537B (en) * | 2019-08-13 | 2020-04-28 | 北京助创科技有限公司 | Method for inverting ground indicator diagram by using electric indicator diagram of pumping unit |
| CN112651165A (en) * | 2019-09-25 | 2021-04-13 | 北京国双科技有限公司 | Wax precipitation initial stage identification model obtaining method, wax precipitation initial stage identification method and device |
| CN112651165B (en) * | 2019-09-25 | 2024-05-03 | 北京国双科技有限公司 | Method for obtaining wax deposition initial identification model, wax deposition initial identification method and device |
| CN111980664A (en) * | 2020-07-28 | 2020-11-24 | 中国石油天然气股份有限公司 | Narrow strip-shaped indicator diagram analysis method for oil well under special working condition |
| CN112228042A (en) * | 2020-10-22 | 2021-01-15 | 南京富岛信息工程有限公司 | Cloud edge cooperative computing-based rod-pumped well working condition similarity judgment method |
| CN112766301A (en) * | 2020-12-11 | 2021-05-07 | 南京富岛信息工程有限公司 | Similarity judgment method for indicator diagram of oil extraction machine |
| CN112766301B (en) * | 2020-12-11 | 2024-04-12 | 南京富岛信息工程有限公司 | Oil extraction machine indicator diagram similarity judging method |
| CN115434690A (en) * | 2021-06-04 | 2022-12-06 | 中国科学院沈阳自动化研究所 | Unsupervised online mutation point detection and fusion method for pumping unit based on Bayes |
| CN115434690B (en) * | 2021-06-04 | 2024-05-14 | 中国科学院沈阳自动化研究所 | Bayesian-based oil pumping unit unsupervised online mutation point detection and fusion method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104110251A (en) | Pumping unit indicator diagram identification method based on ART2 | |
| Luo et al. | Automated visual defect classification for flat steel surface: a survey | |
| WO2020125668A1 (en) | Method and system for automatically identifying surrounding rock level by applying while-drilling parameters | |
| CN109948841B (en) | Prediction method for residual oil distribution of water flooding oil field based on deep learning | |
| CN109614488B (en) | Text classification and image recognition-based distribution network live working condition judgment method | |
| Han et al. | Online fault diagnosis for sucker rod pumping well by optimized density peak clustering | |
| CN112112631A (en) | Intelligent identification method for indicator diagram of oil pumping unit | |
| CN112861912A (en) | Deep learning-based method and system for identifying indicator diagram of complex working condition of pumping well | |
| CN111089632A (en) | Method and device for detecting liquid level of resin solution tank | |
| Lv et al. | Adaptive fault diagnosis of sucker rod pump systems based on optimal perceptron and simulation data | |
| CN110334508A (en) | A kind of host sequence intrusion detection method | |
| CN113435486A (en) | Coal gangue identification method based on PCA-IFOA-SVM combined with gray level-texture fusion features | |
| CN113866455A (en) | Bridge acceleration monitoring data anomaly detection method, system and device based on deep learning | |
| Lv et al. | Quantitative diagnosis method of the sucker rod pump system based on the fault mechanism and inversion algorithm | |
| Yin et al. | Recognition of grape leaf diseases using MobileNetV3 and deep transfer learning | |
| CN110516526A (en) | A kind of small sample target identification method based on Feature prototype metric learning | |
| CN104615635B (en) | Palm vein classified index construction method based on direction character | |
| Wei et al. | Fault diagnosis of sucker rod pump based on deep-broad learning using motor data | |
| Liu et al. | A lightweight fault diagnosis method of beam pumping units based on dynamic warping matching and parallel deep network | |
| He et al. | Working condition recognition of sucker rod pumping system based on 4-segment time-frequency signature matrix and deep learning | |
| CN108979624B (en) | Rod pumping system friction factor identification method based on indicator diagram moment characteristics | |
| Hao et al. | Research on fault diagnosis of indicator diagram based on BP neural network optimized by iterative learning control | |
| CN116011351A (en) | Oil well reasonable sinking degree determining method based on clustering algorithm and Widedeep network | |
| CN110197120A (en) | Personnel target identification method for unattended sensor system | |
| CN112699954B (en) | Closed loop detection method based on deep learning and bag-of-word model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20141022 |
|
| RJ01 | Rejection of invention patent application after publication |