CN104867166A - Oil well indicator diagram compression storage method based on sparse dictionary learning - Google Patents
Oil well indicator diagram compression storage method based on sparse dictionary learning Download PDFInfo
- Publication number
- CN104867166A CN104867166A CN201510263006.2A CN201510263006A CN104867166A CN 104867166 A CN104867166 A CN 104867166A CN 201510263006 A CN201510263006 A CN 201510263006A CN 104867166 A CN104867166 A CN 104867166A
- Authority
- CN
- China
- Prior art keywords
- load
- data
- sparse
- displacement
- dictionary
- 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.)
- Granted
Links
- 238000010586 diagram Methods 0.000 title claims abstract description 74
- 239000003129 oil well Substances 0.000 title claims abstract description 37
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000007906 compression Methods 0.000 title claims abstract description 17
- 230000006835 compression Effects 0.000 title claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 16
- 238000006073 displacement reaction Methods 0.000 claims description 54
- 238000005070 sampling Methods 0.000 claims 1
- 238000013144 data compression Methods 0.000 abstract description 3
- 230000000694 effects Effects 0.000 abstract description 3
- 238000007781 pre-processing Methods 0.000 abstract 1
- 238000012545 processing Methods 0.000 description 6
- 238000000354 decomposition reaction Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000000875 corresponding effect Effects 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000005086 pumping Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Landscapes
- Complex Calculations (AREA)
Abstract
The invention discloses an oil well indicator diagram compression storage method based on sparse dictionary learning, especially relates to the method based on the dictionary learning and aims at solving a storage problem of a lot of rod pumped well indicator diagram data. The method is based on a historical oil well indicator diagram database. Typical oil well indicator diagram data is selected. After data preprocessing, a training sample database is formed. A K-SVD algorithm in a signal sparse field is used to train the sample database so as to obtain a sparse dictionary and store the dictionary. An original indicator diagram is combined with the sparse dictionary so as to acquire a sparse coefficient through an OMP algorithm and store the sparse coefficient. Oil well indicator diagram data compression is realized. The sparse coefficient is combined with the sparse dictionary so as to carry out indicator diagram reconstruction and then the original indicator diagram is obtained. By using the dictionary through the method, a sparse effect of the oil well indicator diagram is good and precision of reconstruction oil well indicator diagram is high. By using the method, prior information of a target is not needed and a wide application scope is possessed.
Description
Technical field
The invention belongs to signal processing applications field, be specifically related to a kind of oil well load-position diagram compression and storage method based on sparse dictionary study.
Background technology
Oil is as a kind of important national energy, and the efficiency of Petroleum Production is most important.Along with oil field digitizing, informationalized development, can produce in a large number and the data message of complexity in production run, these data messages need collection, transmission, Storage and Processing.In miscellaneous data message, oil well load-position diagram data message is the most important.Oil well load-position diagram be pumping unit hanging point in a reciprocal process, load, with the figure of its change in displacement rule, is called polished rod (ground) load-position diagram.It is the closed curve that indicator measures in oil pumper suction period.Oil well load-position diagram can reflect the working condition of oil well intuitively, as the important means of fault diagnosis, also can provide raw data for production metering of oil wells.The oil well load-position diagram data obtained are more, more can Timeliness coverage fault process in time, and it is also more accurate that follow-up oil well output calculates.But the generation of a large amount of load-position diagram data brings great challenge also to the process of the operations such as the storage of data and inquiry.Consider the practical working situation of oil well, with a bite oil well when parameter constant, the load-position diagram difference of acquisition is little; The load-position diagram data of different oil well are also likely similar in shape.Therefore the mode compressed can be taked to reduce data space, improve access speed.The mode of record load-position diagram data source point and subsequent point difference is adopted to compress in a kind of pumping-unit workdone graphic data compression storage method of patent of invention and device (patent No. CN104484476A), it is not high to there is ratio of compression in the method, if mistake appears in intermediate data points, the shortcoming that subsequent data point is made mistakes can be caused.
Sparse representation theory, as the brand-new theory be born in signal transacting field, more and more causes the concern of association area researchist.Its basic thought is that signal can carry out approximate representation by the linear combination of atom a small amount of in dictionary, and the signal originally without sparse characteristic is become sparse signal.Key challenge in this model is, how to choose base or the dictionary of rarefaction representation.Nearest research shows that the dictionary by learning to obtain can obtain better effect than using predefined dictionary.Sparse representation theory is used for the data compression of oil well load-position diagram, can ensures in Hi-Fi situation, significantly amount of compressed data, on average can reach the ratio of compression of 90%, compress storage space greatly, greatly improve the speed of the data processings such as inquiry.
Summary of the invention
The object of the present invention is to provide a kind of oil well load-position diagram compression and storage method based on sparse dictionary study, in order to solve the problem that a large amount of load-position diagram data storage capacity is large and processing speed is slow, compression load-position diagram data storage capacity, improves data processing speed.
Based on an oil well load-position diagram compression and storage method for sparse dictionary study, the method comprises:
(1) history oil well load-position diagram data are collected, the typical load-position diagram in classifying and selecting oil well load-position diagram data, composition load-position diagram data training sample database; (2) respectively displacement data and load data are adjusted according to displacement reference position in load-position diagram data, load-position diagram data training sample database is decomposed into displacement data Sample Storehouse M and load data Sample Storehouse L; (3) initialization displacement sparse dictionary D
mwith load sparse dictionary D
lfor DCT dictionary; (4) displacement training sample database M is utilized to train displacement sparse dictionary D
m, utilize load training sample database L to train load sparse dictionary D
l; (5) displacement sparse dictionary D is preserved
mwith load sparse dictionary D
l; (6) for original oil well load-position diagram data, displacement data and load data is decomposed into; (7) respectively displacement data and load data are adjusted according to displacement data reference position, obtain displacement data vector x and load data vector y; (8) displacement sparse coefficient vector α is obtained by OMP algorithm
mwith load sparse coefficient vector α
l; (9) displacement sparse coefficient vector α is preserved
mwith load sparse coefficient vector α
l; (10) displacement sparse coefficient vector α is extracted
mwith load sparse coefficient vector α
l; (11) by displacement sparse coefficient vector α
mwith displacement sparse dictionary D
mcombine, obtain displacement data vector x, by load sparse coefficient vector α
lwith load sparse dictionary D
lcombine, obtain load data vector y; (12) load-position diagram is reconstructed by displacement data vector x and load data vector y.
Accompanying drawing explanation
Fig. 1 is total process flow diagram.
Fig. 2 is the process flow diagram of history load-position diagram Sample Storehouse training dictionary.
Fig. 3 is the process flow diagram of original load-position diagram rarefaction representation.
Fig. 4 is the process flow diagram reconstructing original load-position diagram.
Fig. 5 is the load-position diagram that the load-position diagram Sparse chosen in Sample Storehouse represents rear reconstruct.
Fig. 6,7,8 for the load-position diagram Sparse chosen not in Sample Storehouse represent after reconstruct load-position diagram.
Embodiment
The present invention proposes a kind of oil well load-position diagram compression and storage method based on sparse dictionary study, and concrete steps are as follows:
(1) oil well load-position diagram historical data is collected, according to different characteristic (different oil pumper types, the shape of load-position diagram under different operating situation, different strokes, jig frequency scope, different maximum, minimum load etc.) choose typical load-position diagram, composition training sample database.
(2) reference position adjustment is carried out to load-position diagram data.Find in displacement data numerical value be 0 point, with this point for starting point, ring shift left is until the point that numerical value is 0 moves to the most starting position of data.Corresponding load data also carries out the left position of circulation of same number.
(3) the load-position diagram data after adjustment reference position are divided into displacement data Sample Storehouse M and load data Sample Storehouse L,
M={x
j,j=1,2,...,N}
L={y
j,j=1,2,...,N}
Wherein, x is displacement data vector, and y is load data vector, and j is the sequence number of sample, and N is the quantity of sample in Sample Storehouse.
(4) initialization displacement dictionary D
mwith load dictionary D
lfor DCT dictionary;
(5) sparse coding (at this for displacement data, load data similarly) is carried out, for displacement dictionary D to displacement training sample
m, carry out approximate solution by following formula
X ∈ R in formula
nfor original signal, D
m∈ R
n × m(n < m) is sparse dictionary, and α is rarefaction representation coefficient, and ε represents error margin and ε>=0, || ||
0for l
0norm, represents the number of nonzero element in vector.Find each sample x
jrarefaction representation α
j, wherein α
jit is a sparse coefficient vector.
(6) dictionary updating is carried out (at this for displacement data, load data similarly), with K-SVD algorithm (Aharon, M.and Elad, M.and Bruckstein, A, K-SVD:An Algorithm for Designing Overcomplete Dictionaries for Sparse Representation, Signal Processing, IEEE Transactions on 2006) upgrade dictionary D
lin atom.
K-SVD algorithm is put forward by the people such as Michal Aharon, Michael Elad of the Institute of Technology of Israel for 2006, is a kind of very classical dictionary training algorithm, and reaches good training effect.K-SVD algorithm, according to error minimum principle, carries out SVD decomposition to error term, selects the decomposition item making error minimum as the dictionary atom upgraded and corresponding atom coefficient, through continuous iteration thus the solution be optimized.
(7) sparse coding and dictionary updating are constantly repeated to all samples, obtain sparse dictionary D from training sample database study
mand D
l.
(8) by sparse dictionary D
mand D
lpreserve.
(9) choose oil well load-position diagram data in Sample Storehouse, first carry out the reference position adjustment as described in step (2).Obtain displacement data vector x and load data vector y.
(10) difference displacement calculating data vector x and load data vector y is at sparse dictionary D
mand D
lon rarefaction representation, obtain sparse coefficient vector by OMP Algorithm for Solving.
OMP (Orthogonal Matching Pursuit) orthogonal matching pursuit algorithm, as one of method of carrying out Its Sparse Decomposition to signal, decomposes signal on complete dictionary library.Its basic thought is: with the row in the method choice dictionary matrix of greedy iteration, and the row selected by making in each iteration are relevant to current redundancy vector maximum degree, deducts and be correlated with and iterate, until meet stopping criterion for iteration from measurement vector.
(11) sparse coefficient obtained in step (10) combined with the sparse dictionary obtained in step (7), reconstruct displacement and load data, reconstruct original load-position diagram data.Result as shown in Figure 5.
(12) choose three oil well load-position diagram data not in Sample Storehouse, first carry out the reference position adjustment as described in step (2).Obtain displacement data vector x and load data vector y.
(13) difference displacement calculating data vector x and load data vector y is at sparse dictionary D
mand D
lon rarefaction representation, obtain sparse coefficient by OMP Algorithm for Solving.
(14) sparse coefficient obtained in step (13) combined with the sparse dictionary obtained in step (7), reconstruct displacement and load data, reconstruct original load-position diagram data.Result as Fig. 6,7, shown in 8.
Table 1 is the contrast of above-mentioned experimental result, lists four mouthfuls of oil well true load-position diagram compression comparison that is front and compression effectiveness after compressing in table.
Experimental result shows, this method can in guarantee in the Hi-Fi situation of original load-position diagram data, and mean pressure shrinkage can reach more than 90%, and the error of other relevant load-position diagram parameter is within 1%.Therefore be a kind of load-position diagram compression method of high efficient and reliable.
Claims (4)
1. the oil well load-position diagram compression and storage method based on sparse dictionary study, it is characterized in that, the method comprises: (1) collects history oil well load-position diagram data, the typical load-position diagram in classifying and selecting oil well load-position diagram data, composition load-position diagram data training sample database; (2) respectively displacement data and load data are adjusted according to displacement reference position in load-position diagram data, load-position diagram data training sample database is decomposed into displacement data Sample Storehouse M and load data Sample Storehouse L; (3) initialization displacement sparse dictionary D
mwith load sparse dictionary D
lfor DCT dictionary; (4) displacement training sample database M is utilized to train displacement sparse dictionary D
m, utilize load training sample database L to train load sparse dictionary D
l; (5) displacement sparse dictionary D is preserved
mwith load sparse dictionary D
l; (6) for original oil well load-position diagram data, displacement data and load data is decomposed into; (7) respectively displacement data and load data are adjusted according to displacement data reference position, obtain displacement data vector x and load data vector y; (8) displacement sparse coefficient vector α is obtained by 0MP algorithm
mwith load sparse coefficient vector α
l; (9) displacement sparse coefficient vector α is preserved
mwith load sparse coefficient vector α
l; (10) displacement sparse coefficient vector α is extracted
mwith load sparse coefficient vector α
l; (11) by displacement sparse coefficient vector α
mwith displacement sparse dictionary D
mcombine, obtain displacement data vector x, by load sparse coefficient vector α
lwith load sparse dictionary D
lcombine, obtain load data vector y; (12) load-position diagram is reconstructed by displacement data vector x and load data vector y.
2. according to claim 1 a kind of based on sparse dictionary study oil well load-position diagram compression and storage method, it is characterized in that step (3), (4), (5) and the sparse dictionary described in (11) are matrixes, its line number is load-position diagram sampling number, and columns is determined according to reconstruction accuracy and arithmetic speed.
3. a kind of oil well load-position diagram compression and storage method based on sparse dictionary study according to claim 1, when it is characterized in that training sparse dictionary, should be divided into displacement sparse dictionary and load sparse dictionary is trained respectively.
4. according to claim 1 a kind of based on sparse dictionary study oil well load-position diagram compression and storage method, it is characterized in that adjusting displacement and load data respectively according to displacement data reference position in load-position diagram data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510263006.2A CN104867166B (en) | 2015-05-22 | 2015-05-22 | A kind of oil well indicator card compression and storage method based on sparse dictionary study |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510263006.2A CN104867166B (en) | 2015-05-22 | 2015-05-22 | A kind of oil well indicator card compression and storage method based on sparse dictionary study |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104867166A true CN104867166A (en) | 2015-08-26 |
CN104867166B CN104867166B (en) | 2018-04-06 |
Family
ID=53912981
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510263006.2A Expired - Fee Related CN104867166B (en) | 2015-05-22 | 2015-05-22 | A kind of oil well indicator card compression and storage method based on sparse dictionary study |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104867166B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108798641A (en) * | 2018-06-19 | 2018-11-13 | 东北大学 | A kind of Diagnosing The Faults of Sucker Rod Pumping System method based on subspace transfer learning |
CN111064472A (en) * | 2020-01-15 | 2020-04-24 | 洛阳乾禾仪器有限公司 | Method and device for compressing and decompressing indicator diagram data and computer readable storage medium |
CN112905551A (en) * | 2019-12-04 | 2021-06-04 | 阿里巴巴集团控股有限公司 | Data compression method and device, electronic equipment and computer readable storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1786939A (en) * | 2005-11-10 | 2006-06-14 | 浙江中控技术有限公司 | Real-time data compression method |
CN104484476A (en) * | 2014-12-31 | 2015-04-01 | 中国石油天然气股份有限公司 | Method and device for compressing and storing indicator diagram data of oil pumping unit |
-
2015
- 2015-05-22 CN CN201510263006.2A patent/CN104867166B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1786939A (en) * | 2005-11-10 | 2006-06-14 | 浙江中控技术有限公司 | Real-time data compression method |
CN104484476A (en) * | 2014-12-31 | 2015-04-01 | 中国石油天然气股份有限公司 | Method and device for compressing and storing indicator diagram data of oil pumping unit |
Non-Patent Citations (1)
Title |
---|
王智慧: "基于压缩感知的图像重构算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108798641A (en) * | 2018-06-19 | 2018-11-13 | 东北大学 | A kind of Diagnosing The Faults of Sucker Rod Pumping System method based on subspace transfer learning |
CN108798641B (en) * | 2018-06-19 | 2021-06-11 | 东北大学 | Rod pump pumping well fault diagnosis method based on subspace migration learning |
CN112905551A (en) * | 2019-12-04 | 2021-06-04 | 阿里巴巴集团控股有限公司 | Data compression method and device, electronic equipment and computer readable storage medium |
CN112905551B (en) * | 2019-12-04 | 2022-04-29 | 阿里巴巴集团控股有限公司 | Data compression method and device, electronic equipment and computer readable storage medium |
CN111064472A (en) * | 2020-01-15 | 2020-04-24 | 洛阳乾禾仪器有限公司 | Method and device for compressing and decompressing indicator diagram data and computer readable storage medium |
CN111064472B (en) * | 2020-01-15 | 2023-09-29 | 洛阳乾禾仪器有限公司 | Method, apparatus and computer readable storage medium for compressing and decompressing work diagram data |
Also Published As
Publication number | Publication date |
---|---|
CN104867166B (en) | 2018-04-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102804230B (en) | Use repeat patterns compression 3D grid | |
CN105163121B (en) | Big compression ratio satellite remote sensing images compression method based on depth autoencoder network | |
CN105354800B (en) | The non-convex compressed sensing image reconstructing method of particle group optimizing based on picture structure | |
CN106096066A (en) | The Text Clustering Method embedded based on random neighbor | |
Xie et al. | Localization-aware channel pruning for object detection | |
CN107192878A (en) | A kind of trend of harmonic detection method of power and device based on compressed sensing | |
CN112991472B (en) | Image compressed sensing reconstruction method based on residual error dense threshold network | |
CN103023510B (en) | A kind of movement data compression method based on sparse expression | |
CN106301385A (en) | The method and apparatus carrying out reasonable compression and decompression for logarithm | |
CN106961104B (en) | Wind power prediction method based on data analysis and combined basis function neural network | |
CN104867166A (en) | Oil well indicator diagram compression storage method based on sparse dictionary learning | |
CN102075749B (en) | Image compression reconstruction method under compressed sensing frame based on non-convex model | |
CN105894547A (en) | Image processing method based on group-wave transformation compressed sensing | |
CN105574265B (en) | Entire assembly model quantitative description towards model index | |
CN116881665A (en) | CMOA optimization-based TimesNet-BiLSTM photovoltaic output prediction method | |
CN109408765A (en) | Intelligent Matching based on quasi-Newton method tracks sparse reconstruction method | |
CN104156768A (en) | Small data size chaos identifying method through fuzzy C-means cluster | |
CN103269429B (en) | A kind of ultraphotic spectrum signal rapid vector quantization coding method | |
CN109977961A (en) | Binary feature learning method and system based on layering attention mechanism | |
CN116743182B (en) | Lossless data compression method | |
CN107133348B (en) | Approximate searching method based on semantic consistency in large-scale picture set | |
Abdelwahab et al. | LiDAR data compression challenges and difficulties | |
CN113362405A (en) | StOMP (static latent image processing) -based compressed sensing image reconstruction system construction method | |
CN101877005A (en) | Document mode-based GML compression method | |
Mamatha et al. | Significance of pre-processing and its impact on lossless hyperspectral image compression |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180406 |