CN116084921A - Working fluid level prediction method, device, apparatus, and readable storage medium - Google Patents
Working fluid level prediction method, device, apparatus, and readable storage medium Download PDFInfo
- Publication number
- CN116084921A CN116084921A CN202111308480.4A CN202111308480A CN116084921A CN 116084921 A CN116084921 A CN 116084921A CN 202111308480 A CN202111308480 A CN 202111308480A CN 116084921 A CN116084921 A CN 116084921A
- Authority
- CN
- China
- Prior art keywords
- data
- working fluid
- fluid level
- historical
- single well
- 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
- 239000012530 fluid Substances 0.000 title claims abstract description 218
- 238000000034 method Methods 0.000 title claims abstract description 68
- 238000003860 storage Methods 0.000 title claims abstract description 36
- 238000012549 training Methods 0.000 claims abstract description 75
- 238000004519 manufacturing process Methods 0.000 claims abstract description 60
- 230000008859 change Effects 0.000 claims abstract description 29
- 230000003068 static effect Effects 0.000 claims abstract description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 46
- 230000005499 meniscus Effects 0.000 claims description 28
- 239000007788 liquid Substances 0.000 claims description 25
- 238000009826 distribution Methods 0.000 claims description 11
- 238000007637 random forest analysis Methods 0.000 claims description 11
- 230000004927 fusion Effects 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 5
- 230000004044 response Effects 0.000 claims description 5
- 238000011161 development Methods 0.000 abstract description 18
- 230000008569 process Effects 0.000 abstract description 15
- 239000003208 petroleum Substances 0.000 abstract description 6
- 239000003921 oil Substances 0.000 description 80
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 22
- 238000005086 pumping Methods 0.000 description 18
- 238000004364 calculation method Methods 0.000 description 13
- 239000003129 oil well Substances 0.000 description 12
- 238000002347 injection Methods 0.000 description 11
- 239000007924 injection Substances 0.000 description 11
- 238000000691 measurement method Methods 0.000 description 11
- 239000010779 crude oil Substances 0.000 description 10
- 238000003066 decision tree Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 10
- 239000007789 gas Substances 0.000 description 9
- 238000012545 processing Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 238000000605 extraction Methods 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 6
- 230000008901 benefit Effects 0.000 description 6
- 238000004590 computer program Methods 0.000 description 6
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 238000005259 measurement Methods 0.000 description 5
- 230000008878 coupling Effects 0.000 description 4
- 238000010168 coupling process Methods 0.000 description 4
- 238000005859 coupling reaction Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 230000002354 daily effect Effects 0.000 description 3
- 239000003345 natural gas Substances 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000035699 permeability Effects 0.000 description 3
- 239000007787 solid Substances 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 239000000047 product Substances 0.000 description 2
- 239000000243 solution Substances 0.000 description 2
- 238000012952 Resampling Methods 0.000 description 1
- 239000002253 acid Substances 0.000 description 1
- 230000002238 attenuated effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000010924 continuous production Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000000994 depressogenic effect Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000020477 pH reduction Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 239000002244 precipitate Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 238000009991 scouring Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000005406 washing Methods 0.000 description 1
Images
Classifications
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH OR ROCK DRILLING; MINING
- E21B—EARTH OR ROCK DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
- E21B47/00—Survey of boreholes or wells
- E21B47/04—Measuring depth or liquid level
- E21B47/047—Liquid level
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH OR ROCK DRILLING; MINING
- E21B—EARTH OR ROCK DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
- E21B2200/00—Special features related to earth drilling for obtaining oil, gas or water
- E21B2200/20—Computer models or simulations, e.g. for reservoirs under production, drill bits
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Geology (AREA)
- Mining & Mineral Resources (AREA)
- Geophysics (AREA)
- Environmental & Geological Engineering (AREA)
- Fluid Mechanics (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geochemistry & Mineralogy (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a working fluid level prediction method, a device, equipment and a readable storage medium, and relates to the field of petroleum development. The method comprises the following steps: acquiring basic data of a single well and historical dynamic data of the single well, wherein the basic data are static characteristic data of the single well, and the historical dynamic data are dynamic change characteristic data of the single well in a target historical time period; acquiring reference working fluid level data corresponding to historical dynamic data; model training is carried out based on the basic data, the historical dynamic data and the reference working fluid level data, so as to obtain a working fluid level prediction model; acquiring current dynamic data of a single well, wherein the current dynamic data corresponds to the data type of historical dynamic data; and predicting the working fluid level of the current dynamic data through the working fluid level prediction model to obtain the working fluid level depth corresponding to the current dynamic data. Real-time prediction of the working fluid level is realized through establishment of a working fluid level prediction model, and real-time adjustment of working conditions and measures in the oil gas development process is guided, so that the efficiency of a production system is further improved.
Description
Technical Field
The embodiment of the application relates to the field of petroleum development, in particular to a working fluid level prediction method, a device, equipment and a readable storage medium.
Background
The working fluid level is an important parameter for the production of the pumping well, and can reflect the formation energy change and the change of the working state of the oil pump.
In the related art, since the working fluid level is located under the ground of thousands of meters, direct measurement is difficult, and usually, a pulse acoustic echo method is used for indirectly calculating the working fluid level, a sound bullet is used as a sound source, compressed natural gas in an annular space between an oil pipe and a casing pipe is transmitted to the underground, an echo is generated when the sound wave encounters a coupling in the transmission process, and finally, a strong echo is reflected when the sound wave reaches the fluid level, and the depth of the working fluid level of the oil well is calculated according to the number of reflected coupling before the sound pulse reaches the fluid level of the well and the average distance between the coupling of the oil pipe.
However, in the above-described method, firstly, echo energy generated from the sound source to the coupling and then to the liquid surface is sequentially attenuated, so that accuracy of the obtained waveform is lowered; and secondly, when an actual oil well acquires an echo signal, the waveform is more easily affected by the environment, so that the obtained signal has larger noise and is not easy to identify, and the actual efficiency of measuring the working fluid level is affected.
Disclosure of Invention
The embodiment of the application provides a working fluid level prediction method, device, equipment and readable storage medium, which can effectively guide the real-time adjustment of working conditions and measures in the oil gas development process and further improve the efficiency of a production system. The technical scheme is as follows:
in one aspect, a method of predicting a working fluid level is provided, the method comprising:
acquiring basic data of a single well and historical dynamic data of the single well, wherein the basic data are static characteristic data of the single well, and the historical dynamic data are dynamic change characteristic data of the single well in a target historical time period;
acquiring reference working fluid level data corresponding to the historical dynamic data;
model training is carried out based on the basic data, the historical dynamic data and the reference working fluid level data, so as to obtain a working fluid level prediction model;
acquiring current dynamic data of the single well, wherein the current dynamic data corresponds to the data type of the historical dynamic data;
and carrying out the dynamic liquid level prediction on the current dynamic data through the dynamic liquid level prediction model to obtain the dynamic liquid level depth corresponding to the current dynamic data.
In some embodiments, the training of the model based on the base data, the historical dynamic data, and the reference working fluid level data to obtain a working fluid level prediction model includes:
Acquiring at least two basic models, wherein the at least two basic models are models with different prediction algorithms;
and performing fusion training on the at least two basic models by taking the basic data, the historical dynamic data and the reference working fluid level data as training data to obtain the working fluid level prediction model.
In some embodiments, the performing fusion training on the at least two basic models by using the basic data, the historical dynamic data and the reference working fluid level data as training data to obtain the working fluid level prediction model includes:
dividing the training data to obtain at least two training data sets;
training the at least two basic models in the ith cycle through the at least two training data sets respectively to obtain an ith prediction result of the at least two training data sets;
splicing the ith prediction results of the at least two models to serve as training data sets in the (i+1) th cycle until the at least two basic models are converged;
and fusing the at least two converged basic models to obtain the dynamic liquid level prediction model.
In some embodiments, the at least two models include at least two of a linear model, a random forest model, an iterative algorithm model, a gradient lifting algorithm model, and an extreme gradient lifting algorithm model.
In some embodiments, the training of the model based on the base data, the historical dynamic data, and the reference working fluid level data to obtain a working fluid level prediction model includes:
normalizing the basic data and the historical dynamic data to obtain normalized basic data and historical dynamic data;
and carrying out model training through the normalized basic data and the historical dynamic data and the reference working fluid level data to obtain a working fluid level prediction model.
In some embodiments, before the model training based on the base data, the historical dynamic data, and the reference working fluid level data, the method further includes:
performing data filtering based on a data value range of the reference working fluid level data, wherein the data value range is used for indicating an effective range of the working fluid level depth distribution;
in response to the missing data being present, interpolation is used to fill in the missing data.
In some embodiments, before the model training based on the base data, the historical dynamic data, and the reference working fluid level data, the method further includes:
acquiring n single well production data in a target historical time period before the reference working fluid level data, wherein n is a positive integer;
acquiring the yield change rate of the n single well production data according to the change condition of the n single well production data;
adding the yield rate of change to the historical dynamic data.
In another aspect, a dynamic liquid level prediction apparatus is provided, the apparatus comprising:
the acquisition module is used for acquiring basic data of a single well and historical dynamic data of the single well, wherein the basic data are static characteristic data of the single well, and the historical dynamic data are dynamic change characteristic data of the single well in a target historical time period;
the acquisition module is also used for acquiring reference working fluid level data corresponding to the historical dynamic data;
the determining module is used for carrying out model training based on the basic data, the historical dynamic data and the reference working fluid level data to obtain a working fluid level prediction model;
the acquisition module is further used for acquiring current dynamic data of the single well, wherein the current dynamic data corresponds to the data type of the historical dynamic data;
And the prediction module is used for predicting the working fluid level of the current dynamic data through the working fluid level prediction model to obtain the working fluid level depth corresponding to the current dynamic data.
In another aspect, a computer device is provided, the computer device including a processor and a memory, where the memory stores at least one instruction, at least one program, a set of codes, or a set of instructions, the at least one instruction, the at least one program, the set of codes, or the set of instructions being loaded and executed by the processor to implement a method for predicting a meniscus as in any one of the embodiments of the application described above.
In another aspect, a computer readable storage medium having stored therein at least one instruction, at least one program, code set, or instruction set loaded and executed by a processor to implement a method of meniscus prediction as described in any of the embodiments of the present application.
In another aspect, a computer program product or computer program is provided, the computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the method of predicting the meniscus as described in any one of the above embodiments.
The beneficial effects that technical scheme that this application embodiment provided include at least:
according to the basic data and the historical dynamic characteristics of a single well, training and modeling are carried out on the acquired data, real-time calculation of the working fluid level is realized through the working fluid level prediction model, and the real-time calculation result of the working fluid level is compared with the actual test result, so that the working fluid level prediction model is optimized, the working fluid level is rapidly, simply, conveniently and efficiently acquired, the real-time adjustment of working conditions and measures in the oil gas development process is guided, the efficiency of a production system is further improved, and the benefit development level is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of a pumping unit oil extraction apparatus according to an exemplary embodiment of the present application;
FIG. 2 is a schematic illustration of an implementation environment provided by an exemplary embodiment of the present application;
FIG. 3 is a flow chart of a method of predicting a meniscus according to one exemplary embodiment of the present application;
FIG. 4 is a flow chart of a method of meniscus prediction provided based on FIG. 3;
FIG. 5 is a schematic view of a meniscus predicting device provided in an exemplary embodiment of the present application;
FIG. 6 is a schematic view of a meniscus prediction apparatus provided in another exemplary embodiment of the present application;
fig. 7 shows a schematic structural diagram of a server according to an exemplary embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
With continuous exploitation of crude oil, the energy of an underground oil layer is gradually reduced when the crude oil enters an oil field at the middle and later stages of exploitation, and a self-injection oil extraction mode of oil well production is gradually converted into a mechanical oil extraction mode by means of external energy. The most commonly used rod pump is adopted for oil extraction (namely oil extraction by an oil pumping machine), and the plunger type oil pump is driven to work by the up-and-down reciprocating motion of the oil pumping rod. Specifically, referring to the schematic diagram of the oil extraction device of the pumping unit shown in fig. 1, a typical sucker rod pumping device mainly comprises three parts, namely, a ground driving device (i.e., a pumping unit) 10, an oil pump 20 installed at the lower part of an oil pipe column, and a sucker rod column 30, wherein the sucker rod column 30 mainly transmits the motion and power of the ground driving device 10 to the underground oil pump 20 to enable the underground oil pump 20 to reciprocate up and down, so that the liquid in the oil pipe column is pressurized, and the produced liquid of an oil layer is pumped to the ground.
The working fluid level 40 is the liquid level depth in the annular space of the oil pipe 211 and the sleeve 212 when the ground driving device 10 works normally, the position of the oil pump 20 relative to the working fluid level 40 has a larger influence on the production efficiency of the pumping well, the oil pump 20 is positioned above the working fluid level 40, the condition that the ground driving device 10 is empty or burns the pump is easily caused, the oil pump 20 is submerged in the working fluid level 40 deeply, the load of the oil pump 20 is increased, the energy consumption is serious, a large amount of manpower resources are input in the process, the working strength of workers is large, the measurement accuracy is also easily influenced by environmental factors, personnel factors and the like, and more potential safety hazards exist.
Therefore, in order to solve the problems faced by the conventional measurement method, new methods and techniques are needed, and a hard measurement technique (i.e. a conventional measurement method) and a soft measurement technique are needed to be effectively combined, so that the problem that certain key parameters (such as a real-time calculation result of a working fluid level) are difficult to directly measure in the production process is solved. The oil field production is a complex industrial process, if the working fluid level measurement of the pumping well is realized only by a theoretical modeling mode, the satisfactory technical effect cannot be achieved, the working fluid level prediction mode provided by the application can be used for generating a soft measurement method based on oil well (referred to as a single well in the application) data drive based on the production data under the current oil field digital platform, so that the real-time prediction of the working fluid level is realized, and the accuracy of a prediction result can be ensured.
In this embodiment, fig. 2 is a schematic diagram of an implementation environment provided in an exemplary embodiment of the present application, as shown in fig. 2, where the implementation environment includes a terminal 210, a server 220, and a working fluid level prediction model 230, where the terminal 210 and the server are connected through a communication network, where the communication network may be a wired network or a wireless network, which is not limited in this application.
The relevant parameters of the single well are stored in the terminal 210, and include, but are not limited to, basic data, historical dynamic data of the single well and reference working fluid level data corresponding to the historical dynamic data, and the server 220 obtains the relevant parameters of the single well from the terminal 210 (S1), and performs model training based on the relevant parameters to obtain a working fluid level prediction model 230 (S2).
Optionally, after determining the meniscus prediction model 230, if a real-time meniscus depth of the single well is to be predicted, the server 220 obtains the current dynamic data of the single well from the terminal 210 (S3), and performs meniscus prediction on the current dynamic data through the meniscus prediction model 230 (S4).
The working fluid level prediction model 230 receives current dynamic data of a single well to obtain a current working fluid level depth of the single well. Alternatively, the meniscus prediction model 230 may be stored in the server 220 for use alone as a monolithic model, or may be embedded in an existing model as a separate model, not limited by this comparison.
Optionally, after the working fluid level prediction model 230 obtains the prediction result, the prediction result is fed back to the server 220, and the server 220 feeds back the prediction result to the terminal 210 (S5).
Alternatively, the foregoing execution body may be executed by the terminal 210, or may be executed by the server 220, or may be executed by both the server 220 and the terminal 210, which is not limited in this embodiment of the present application.
The server 220 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), basic cloud computing services such as big data and artificial intelligence platforms, and the like.
In some embodiments, the server 220 described above may also be implemented as a node in a blockchain system. Blockchain (Blockchain) is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanisms, encryption algorithms, and the like. The blockchain is essentially a decentralised database, and is a series of data blocks which are generated by association by using a cryptography method, and each data block contains information of a batch of network transactions and is used for verifying the validity (anti-counterfeiting) of the information and generating a next block. The blockchain may include a blockchain underlying platform, a platform product services layer, and an application services layer.
It should be noted that, the server 220 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or may be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, a content delivery network (Content Delivery Network, CDN), and basic cloud computing services such as big data and an artificial intelligence platform.
In connection with the above implementation environment, the method for predicting the working fluid level according to the embodiment of the present application will be described, and fig. 3 is a flowchart of the method for predicting the working fluid level according to an exemplary embodiment of the present application, where the method is applied to the server shown in fig. 2, for example, as shown in fig. 3, and the method includes:
The working fluid level directly reflects the fluid supply capacity of an underground reservoir, and the higher the collection frequency of the working fluid level is, the stronger the guiding significance for dynamic analysis of an oil well is. Therefore, the method is comprehensively influenced by the pumping unit and the peripheral water injection wells in the actual production process of the pumping unit.
According to the actual production service characteristics of the oil pumping well, the change of the working fluid level of the oil pumping well is mainly influenced by the following factors:
Single well production rate: the method is used for indicating the percentage of the ratio of the accumulated oil production of the single well to the reserve of the mobile address, reflecting the production condition of the oil reserve of the single well, and if the production degree is higher, the method represents that the loss of the single well is larger and the working fluid level is relatively lower.
Water injection rate of water injection well: the device is used for indicating the water quantity actually injected into an underground oil layer every day by the water injection well, and the higher the water injection quantity is, the higher the working fluid level is.
And (3) communicating the water injection well: the device is used for indicating the communication degree between the water injection wells, and the higher the communication degree is, the higher the working fluid level is.
Casing gas: the device is used for indicating the oil well separation gas led out to the ground through the casing, and the larger the casing gas is, the larger the casing pressure is, and the easier the working fluid level is to be depressed.
Pump diameter: the device is used for indicating the inner diameter of the working cylinder of the oil pump, and the larger the pump diameter is, the lower the corresponding working fluid level is.
Pump depth: the device is used for indicating the depth from a wellhead to an oil well pump, and the working fluid level corresponding to the deeper pump depth is lower, wherein the length of the pump is unchanged in the continuous production process, but the length of the pump is also changed under the conditions of well stopping and well shutting.
And (3) the stroke frequency: indicating the number of up and down strokes per minute of the sucker rod in the rod string. The stroke frequency of the pumping unit applied to the existing oil field is generally more than 3. The oil well with high yield is suitable for the pumping unit with high sprint and the oil well with low yield is suitable for the pumping unit with lower sprint. The stroke frequency is used as the output quantity of closed-loop control of the whole oil well and is adjusted according to the working fluid level. Wherein the more the stroke frequency is, the lower the dynamic liquid level is.
Pump failure: and the device is used for indicating whether the oil pump has mechanical faults or not, and if the mechanical faults occur, the working fluid level is influenced to be increased.
Oil pipe leakage: the oil pipe is used for indicating whether oil leakage occurs or not, and if the oil leakage occurs, the working fluid level is influenced to be increased.
Gas-oil ratio: for indicating the production of an oil well, oil and gas are simultaneously discharged from a single well, and the natural volume (cubic meters) of oil is produced per ton of crude oil. Under the condition of underground oil layer, a certain amount of natural gas is dissolved in crude oil, and the natural gas is dissolved in the petroleum to cause the volume expansion of the petroleum, the specific gravity and the viscosity to be reduced, the pressure of a fluid liquid column to enable the oil well to be easier to self-blow, and the petroleum exploitation is facilitated. The higher the gas-oil ratio, the lower the working fluid level.
Reservoir state: the method is used for indicating the distribution state of the reservoirs corresponding to the underground of the single well, wherein the reduction of the working fluid level is affected if the sand buried reservoirs exist.
Yield increasing measures: the method is used for indicating the implementation means for improving the permeability of the underground reservoir of the single well by means of chemical corrosion, and is mainly used in exploitation of the single well (petroleum well) and geothermal resource development, and the effect of manually injecting acid liquid into the stratum is mainly achieved, wherein the higher the acidification degree is, the lower the profile control is represented.
Factors influencing working fluid level depth (height) above, wherein some of the factors pertain to basic data corresponding to a single well (rod-pumped well), wherein the basic data are used for indicating static characteristic data of the single well (rod-pumped well), including, but not limited to, daily fluid production, first oil pressure, first set pressure, water cut, stroke frequency, crude oil relative density, reservoir middle depth, oil inlet depth, pump diameter data, well number, well deviation and production horizon corresponding to the single well; in the embodiment of the application, liquid production amount, first oil pressure, first set pressure, water content, stroke frequency, crude oil relative density, oil reservoir middle depth, oil inlet hole depth and pump diameter data are adopted as basic data of a single well.
In the process of water injection development of oilfield reservoirs, due to water washing and scouring of injected water, the reservoirs are transformed to a certain extent, so that the characteristics of the reservoirs after water injection development are different from the original state, and the characteristics of the reservoirs are changed more remarkably along with the extension of development time. In the embodiment of the application, proper development adjustment is made for different exploitation stages of the oil field so as to improve the development effect. The water flooding index can intuitively reflect the water injection utilization rate and is used for judging the effect obtained by various measures taken by the oil field. The water drive index can be found in equation 1.
Equation 1:
in formula 1, R wo For indicating water drive index, w i For indicating the accumulated water volume, the unit is: a square; w (w) p Used for indicating the accumulated water yield, unit: a square; n (N) p For indicating cumulative oil production, unit: a square; b (B) 0 /ρ 0 For indicating conversion factor, B 0 Is the volume coefficient of oil phase, ρ 0 For indicating the oil phase density.
In addition to the static characteristic data described above, factors affecting the depth of the working fluid level include historical dynamic data for the individual well indicating the dynamic change characteristics of the individual well over a target historical time period, including, but not limited to, perforation mode, perforation thickness, perforation bottom depth, perforation top depth perforation depth (perforation depth is used to indicate weighted average results in the present embodiment), control jacket production, second oil pressure, second jacket pressure; the second oil pressure is used for indicating the oil pressure data measured in the target historical time period, can also be used for indicating the oil pressure data during the measurement of the working fluid level, and the second casing pressure is used for indicating the casing pressure data measured in the target time period, and can also be used for indicating the casing pressure data during the measurement of the working fluid level; in the embodiment of the application, the first oil pressure is used for indicating oil pressure data during measuring the working fluid level, and the second casing pressure is used for indicating casing pressure data during measuring the working fluid level.
Optionally, analyzing data (including basic data and historical dynamic data) with a high degree of association with the working fluid level according to a production principle, such as oil inlet hole depth; and carrying out characteristic enhancement processing on the data with high correlation degree with the working fluid level, wherein the specific characteristic enhancement processing comprises square change and opening change on the data, so that the data is further enhanced, and the obtained data has high correlation with the working fluid level.
Optionally, the target historical time period is used for indicating a time period with high correlation with current production data (basic data and/or historical dynamic data) of the single well, and in this embodiment of the application, the target historical time period is selected to be five days, that is, the basic data and/or the historical dynamic data in five days of the single well are acquired for performing subsequent model training, so that the obtained working fluid level prediction model is ensured to have high prediction accuracy.
In some embodiments, when the production data (base data and/or historical dynamic data) is unstable for a target historical period of time, the production data is changed to rate of change data, i.e., a production rate of change is calculated, the production rate of change being equal to the ratio of the production on the day to the production on the day before; the output of unstable change of production data is amplified to a certain extent, and the important data of partial unstable data characteristics of the production data can be obviously increased when the subsequent input model is trained. In the embodiment of the present application, the instability of the production data may be used to indicate that the production data does not increase/decrease linearly, or may be used to indicate that the occurrence of the continuous/intermittent fine increase/decrease of the production data, which is not limited in this application.
Nouns appearing in the above-described base data and history dynamic data are described as follows:
the reciprocating of the piston of the oil pump is called a stroke, and the stroke frequency is the number of strokes performed by the oil pump per minute. The stroke and the stroke frequency directly affect the pumping capacity of the pumping unit, and in the embodiment of the application, the stroke is a fixed preset value during oil extraction.
Casing pressure is used to indicate the pressure developed at the casing annulus during well production due to formation pressure changes that precipitate oil out of the crude oil liquid. The oil pressure is used for indicating the residual pressure of crude oil flowing from the bottom of a well to the top of the well, and is divided into static pressure and flowing pressure, the oil pressure during well closing is called static pressure, and the oil pressure during production is called flowing pressure.
The perforation depth is used for indicating the depth of an oil reservoir, multiple sections of perforation are generated in the production process of a single well, the number of perforation holes is weighted by the intermediate value of each section of perforation depth to obtain the weighted average perforation depth, and the number of perforation holes is determined by the product of perforation density and perforation thickness, wherein the perforation holes and perforation density influence permeability, and therefore, the perforation holes and perforation density are listed in historical dynamic data.
In some embodiments, after the staff collects the basic data and the historical dynamic data corresponding to the single well, the data is stored in the terminal, and the terminal may select to store the data locally or upload the data to a database corresponding to the server for storage.
In some embodiments, the staff sends a first acquiring instruction to the server through the terminal, where the first acquiring instruction is used to instruct the server to acquire, from the database, the basic data and the historical dynamic data corresponding to the target single well, where the number of the target single wells may be one or several, which is not limited in this application.
In some embodiments, after the server obtains the historical dynamic data corresponding to the target individual well, the server obtains reference working fluid level data corresponding to the target individual well from a database, where the reference working fluid level data is used to indicate working fluid level depth data calculated by a worker based on the prior art, and the prior art includes a indicator diagram measurement method, a pontoon measurement method, and an echo measurement method.
In some embodiments, the server performs data analysis based on features contained in the historical dynamic data and the reference meniscus data and draws a corresponding analysis map, which may be a histogram, a pie chart, or a line chart, which is not limited in this application.
The analysis graph is used to indicate the degree of linear correlation between features in the target single well historical dynamic data and the reference meniscus data. In the embodiment of the application, the analysis chart analysis shows that the relative density of the water content, stroke frequency, oil pressure, casing pressure and crude oil has no linear relation with the depth of the working fluid level, and the characteristics of the depth of the oil inlet hole, daily liquid yield, pump diameter and the like have linear relation with the depth of the working fluid level.
And step 303, performing model training based on the basic data, the historical dynamic data and the reference working fluid level data to obtain a working fluid level prediction model.
In some embodiments, the server obtains at least two base models, the at least two base models being models with different prediction algorithms; the at least two basic models comprise at least two of a linear model, a random forest model, an iterative algorithm model, a gradient lifting algorithm model and an extreme gradient lifting algorithm model.
Optionally, the linear model is completed by using a scikit-learn series linear regression Linear Regression method in Python, and finally a first prediction graph is obtained, where the first prediction graph is used for indicating a linear relation graph corresponding to reference working fluid level data and predicted working fluid level data after training by using the linear model.
Optionally, the random forest model mainly uses an integrated learning idea to integrate multiple trees, and the basic unit of the random forest model is a decision tree, and the random forest model essentially belongs to a large branch of machine learning, namely an integrated learning (Ensemble Learning) method. The importance of each feature in a tree can be calculated once for each feature in the tree, and in some embodiments, the average of the importance of a feature value in multiple trees can be selected as the importance of the feature in the forest. The specific algorithm flow is as follows:
1. Generating n samples from a sample set (basic data, historical dynamic data and reference working fluid level data) by resampling, wherein the sample set can also comprise the basic data, the historical dynamic data and the reference working fluid level data in a target historical time period optionally, and the application is not limited to the above;
2. assuming that the number of sample features is a, k features in a are selected for n samples, and an optimal division point is obtained by establishing a decision tree, wherein the number of sample features a comprises basic data and historical dynamic data, in the embodiment of the application, the number of sample features of the basic data is 10, and the number of sample features of the historical dynamic data is 8, so that the number of sample features a is all the sample feature data, and a is 18;
3. repeating m times to generate m decision trees;
4. and predicting the depth of the working fluid level by adopting a majority voting mechanism to obtain a second prediction graph and a first characteristic importance distribution graph.
The second prediction graph is used for indicating a linear relation graph corresponding to reference working fluid level data and predicted working fluid level data after training by using a random forest model, the first feature importance distribution graph is used for indicating importance degrees of all features in a sample set on influencing the working fluid level depth, and the importance degrees of the sample feature data can be calculated by using a formula 2.
Equation 2:
in equation 2, ntree is used to indicate the number of trees present in the forest, errOOBt 1 When the data is used for indicating the establishment of the decision tree, the data obtained by repeated sampling is used for training the decision tree, and the data which does not participate in the decision tree establishment process is used for evaluating the performance of the decision tree and calculating the preset error rate of the model; errOOBt 2 The data are used for indicating that noise interference is added to the characteristics of all samples of the data outside the bag, and error calculation is carried out on the data outside the bag added with the noise again; a is that i For indicating the ith tree in the forest.
Optionally, the iterative algorithm model (Adaboost algorithm model) mainly trains different classifiers (weak classifiers) aiming at the same training set, and then integrates the weak classifiers to form a stronger final classifier (strong classifier); the training process is essentially a process of continuously improving a weak classification algorithm, and can improve the classification capability of data, and the specific algorithm flow is as follows:
1. the method comprises the steps of firstly, obtaining a first weak classifier by classifying and learning a first sample set (basic data, historical dynamic data and reference working fluid level data);
2. Forming a new second sample set by the error-divided samples and other new data, and obtaining a second weak classifier by classifying and learning the second sample set again;
3. adding other new samples to the samples with the errors of 1 and 2 to form another new third sample set, and performing classification learning again on the third sample set to obtain a third weak classifier;
4. and finally obtaining the lifted strong classifier. It can be understood that a certain data feature is classified into which class is determined by the weight corresponding to each classifier obtained above.
Optionally, a third prediction graph and a second feature importance distribution graph are finally obtained based on the Adaboost algorithm model, wherein the third prediction graph is used for indicating a linear relation graph corresponding to reference working fluid level data and predicted working fluid level data after training by using the Adaboost algorithm model, and the second feature importance distribution graph is used for indicating importance degrees of all features in the first sample set on influencing the working fluid level depth.
Optionally, the gradient lifting algorithm model (Gradient Boosting algorithm model) predicts the working fluid level depth by using a gradient enhancement regressor Gradient Boosting Regressor in scikit-learn package of Python and a function interface of a gradient lifting decision tree (Gradient Boosting Decision Tree, abbreviated as GBDT), wherein the Gradient Boosting algorithm model includes a loss function, and the loss function is used for describing the degree of "leaning to the spectrum" of the model, and the loss function is subjected to a descent operation in the gradient direction during the training process.
Optionally, the extreme gradient lifting algorithm model (Extreme Gradient Boosting, abbreviated as XGboost algorithm model) is an advanced gradient enhancement algorithm, and has enough capability of mechanically learning various irregular features corresponding to various data. The specific algorithm flow is as follows: constructing T trees, wherein T is an integer greater than or equal to 1; when constructing the T-th tree, fitting the residual errors generated by the training sample data set and the classification regression of the previous T-1 tree, wherein T is an integer greater than or equal to 1 and less than or equal to T; every time a fit produces a new tree, all possible trees are traversed and the tree that minimizes the objective function value (cost) is selected. In the embodiment of the application, only one branch is generated each time when a new tree is constructed, and the optimal branch is selected; if the objective function value (cost) of the branch is greater than when not generated or the improvement effect is not obvious, then the branch is aborted.
Alternatively, the XGboost algorithm model is a classification regression tree (CART), where the meaning of the classification is used for discrete value decision, optimizing the objective function with training errors, and constructing the decision tree with the objective function.
Optionally, the server obtains a fourth prediction graph and a third feature importance distribution graph by using the Gradient Boosting algorithm model, wherein the fourth prediction graph is used for indicating a linear relation graph corresponding to the reference working fluid level data and the predicted working fluid level data after training by using the Gradient Boosting algorithm model, and the third feature importance distribution graph is used for indicating importance degrees of all features in the training sample set on influencing the working fluid level depth.
Optionally, the server selects at least two models of a linear model, a random forest model, an iterative algorithm model, a gradient lifting algorithm model and an extreme gradient lifting algorithm model, and performs fusion training on training data (basic data, historical dynamic data and reference working fluid level data) to obtain a working fluid level prediction model.
In this embodiment of the present application, the working fluid level prediction model may use an average absolute error (Mean Absolute Error, abbreviated as MAE) as an evaluation index to predict working fluid level data corresponding to current dynamic data.
The process of fusion training will be described in further detail below, and the specific contents are as follows:
dividing training data (basic data, historical dynamic data and reference working fluid level data) to obtain at least two training data sets;
training at least two basic models in the ith cycle through at least two training data sets respectively to obtain an ith prediction result of the at least two training data sets;
splicing the ith prediction results of at least two models to serve as training data sets in the (i+1) th cycle until at least two basic models are converged;
and fusing the at least two converged basic models to obtain a working fluid level prediction model.
In some embodiments, the training of the five models is that each algorithm corresponding to the model has an emphasis point, and if the algorithm is used alone, the reliability of the predicted value of the algorithm cannot be ensured in the process of predicting the depth data of the working fluid level.
In some embodiments, after the server obtains the working fluid level prediction model, the staff sends a working fluid level prediction request of the single well through the terminal, wherein the working fluid level prediction request comprises a well number uniquely identifying the single well, and the server obtains the well number corresponding to the target single well from the working fluid level prediction request.
In some embodiments, current dynamic data corresponding to the target single well is obtained from the database based on the well number, and the current data corresponding to the target single well is input into the working fluid level prediction model to obtain a working fluid level depth predicted value corresponding to the single well.
And 305, carrying out the dynamic liquid level prediction on the current dynamic data through a dynamic liquid level prediction model to obtain the dynamic liquid level depth corresponding to the current dynamic data.
Optionally, the server predicts the working fluid level of the current dynamic data of the target single well based on the fusion trained working fluid level prediction model to obtain the working fluid level depth corresponding to the current dynamic data corresponding to the target single well.
According to the working fluid level prediction method, training modeling is conducted on collected data according to basic data and historical dynamic characteristics of a single well, real-time calculation of the working fluid level is achieved through the working fluid level prediction model, and the real-time calculation result of the working fluid level is compared with an actual test result, so that the working fluid level prediction model is optimized, the on-line real-time calculation accuracy of the working fluid level is improved, quick, simple and efficient acquisition of the working fluid level is achieved, real-time adjustment of working conditions and measures in the oil gas development process is guided, the efficiency of a production system is further improved, and the benefit development level is improved.
Referring to fig. 4, fig. 4 is a flowchart of a method for predicting a working fluid level according to another embodiment of the present application, where the method is applied to a server in the implementation environment shown in fig. 2, and specifically includes the following steps.
Optionally, the base data is used to indicate static characteristic data of a single well (rod-pumped well), including, but not limited to, a corresponding well number, well deviation, casing inside diameter, tubing outside diameter, production horizon, perforation thickness, perforation density, porosity, permeability, raw gas-to-oil ratio, crude oil relative density, first oil pressure, first casing pressure, reservoir middle depth, and oil intake depth.
In addition to the static characteristic data described above, factors affecting the depth of the working fluid level include historical dynamic data for the individual well indicating dynamic change characteristics of the individual well over a target historical time period including, but not limited to, daily fluid production, perforation depth, water cut, stroke, number of strokes, perforation number, control casing production, second oil pressure, second casing pressure.
This step is the same as the specific flow of step 301, and will not be described here again.
In some embodiments, after the server obtains the historical dynamic data corresponding to the target individual well, the server obtains reference working fluid level data corresponding to the target individual well from a database, where the reference working fluid level data is used to indicate working fluid level depth data calculated by a worker based on the prior art, and the prior art includes a indicator diagram measurement method, a pontoon measurement method, and an echo measurement method.
In some embodiments, data features corresponding to production data (basic data and/or historical dynamic data) in a historical time period of a single well are analyzed according to a production principle, data features with high correlation degree with a working fluid level are analyzed, and data features with high correlation degree are enhanced.
In some embodiments, the server obtains n single well production data for a target historical period of time prior to the reference meniscus, n being a positive integer; obtaining the yield change rate of the n single well production data according to the change condition of the n single well production data; the yield rate of change is added to the historical dynamic data.
In some embodiments, the server performs further model training based on changes in production data for n individual wells over a historical period of time.
In some embodiments, after the server obtains the basic data, the historical dynamic data and the reference working fluid level data of the single well, the data needs to be preprocessed, and a specific preprocessing flow is as follows:
1. drawing an analysis chart between the single well base data and the historical dynamic data and the reference working fluid level data;
2. performing data filtering based on a data value range of reference working fluid level data, wherein the data value range is used for indicating an effective range of the depth distribution of the working fluid level; in response to missing data, the server fills the missing data by adopting an interpolation method, so that the data integrity is ensured; in response to the existence of the extreme data, performing a removal operation on the existing extreme value, wherein the extreme data is used for indicating that the reference working fluid level data is larger than a data value range; in addition, for some redundant data, a mode of manually selecting transformation is adopted for processing.
Step 404, performing model training based on the basic data, the historical dynamic data and the reference working fluid level data to obtain a working fluid level prediction model.
In some embodiments, the server normalizes the base data and the historical dynamic data of the single well to obtain normalized base data and historical dynamic data;
and carrying out model training through the normalized basic data and the historical dynamic data and referring to the working fluid level data to obtain a working fluid level prediction model.
In some embodiments, the server obtains at least two base models, the at least two base models being models with different prediction algorithms; the at least two basic models comprise at least two of a linear model, a random forest model, an iterative algorithm model, a gradient lifting algorithm model and an extreme gradient lifting algorithm model.
Optionally, the server selects at least two models of a linear model, a random forest model, an iterative algorithm model, a gradient lifting algorithm model and an extreme gradient lifting algorithm model, and performs fusion training on the normalized basic data, the historical dynamic data and the reference working fluid level data to obtain a working fluid level prediction model.
The specific flow of this step is the same as that of step 303, and will not be described here again.
In some embodiments, after the server obtains the working fluid level prediction model, the staff sends a working fluid level prediction request of the single well through the terminal, wherein the working fluid level prediction request comprises a well number uniquely identifying the single well, and the server obtains the well number corresponding to the target single well from the working fluid level prediction request.
In some embodiments, current dynamic data corresponding to the target single well is obtained from the database based on the well number, and the current data corresponding to the target single well is input into the working fluid level prediction model to obtain a working fluid level depth predicted value corresponding to the single well.
Optionally, the server predicts the working fluid level of the current dynamic data of the target single well based on the fusion trained working fluid level prediction model to obtain the working fluid level depth corresponding to the current dynamic data corresponding to the target single well.
According to the working fluid level prediction method, training modeling is conducted on collected data according to basic data and historical dynamic characteristics of a single well, real-time calculation of the working fluid level is achieved through the working fluid level prediction model, and the real-time calculation result of the working fluid level is compared with an actual test result, so that the working fluid level prediction model is optimized, the on-line real-time calculation accuracy of the working fluid level is improved, quick, simple and efficient acquisition of the working fluid level is achieved, real-time adjustment of working conditions and measures in the oil gas development process is guided, the efficiency of a production system is further improved, and the benefit development level is improved.
FIG. 5 is a block diagram of a device for predicting a working fluid level according to an exemplary embodiment of the present application, and as shown in FIG. 5, the device includes:
the obtaining module 510 is configured to obtain basic data of a single well and historical dynamic data of the single well, where the basic data is static feature data of the single well, and the historical dynamic data is dynamic change feature data of the single well in a target historical time period;
the obtaining module 510 is further configured to obtain reference working fluid level data corresponding to the historical dynamic data;
the determining module 520 is configured to perform model training based on the base data, the historical dynamic data, and the reference working fluid level data, to obtain a working fluid level prediction model;
the obtaining module 510 is further configured to obtain current dynamic data of the single well, where the current dynamic data corresponds to a data type of the historical dynamic data;
and the prediction module 530 is configured to predict the working fluid level of the current dynamic data according to the working fluid level prediction model, so as to obtain a working fluid level depth corresponding to the current dynamic data.
In an optional embodiment, the obtaining module 510 is further configured to obtain at least two base models, where the at least two base models are models with different prediction algorithms;
The determining module 520 is further configured to perform fusion training on the at least two basic models by using the basic data, the historical dynamic data, and the reference working fluid level data as training data, so as to obtain the working fluid level prediction model.
In an alternative embodiment, as depicted in fig. 6, the apparatus comprises:
the determining module 520 is further configured to divide the training data to obtain at least two training data sets;
the determining module 520 is further configured to train the at least two basic models in the ith cycle through the at least two training data sets, to obtain an ith prediction result of the at least two training data sets;
a stitching module 540, configured to stitch the ith prediction result of the at least two models, as a training data set in the (i+1) th cycle, until the at least two base models converge;
the determining module 520 is further configured to fuse the at least two converged base models to obtain the working fluid level prediction model.
In an alternative embodiment, the at least two base models include at least two of a linear model, a random forest model, an iterative algorithm model, a gradient lifting algorithm model, and an extreme gradient lifting algorithm model.
In an optional embodiment, the determining module 520 is further configured to normalize the base data and the historical dynamic data to obtain normalized base data and historical dynamic data;
the determining module 520 is configured to perform model training through the normalized basic data and the historical dynamic data, and the reference working fluid level data, so as to obtain a working fluid level prediction model.
In an alternative embodiment, as shown in fig. 6, the apparatus further comprises:
a filtering module 550, configured to perform data filtering based on a data value range of the reference working fluid level data, where the data value range is used to indicate an effective range of the working fluid level depth distribution;
and a filling module 560, configured to, in response to the missing data, fill in the missing data by using interpolation.
In an alternative embodiment, as shown in fig. 6, the apparatus includes:
the acquiring module 510 is further configured to acquire n single well production data in a target historical period before the reference working fluid level data, where n is a positive integer;
the obtaining module 510 is further configured to obtain a yield change rate of the n single well production data according to a change condition of the n single well production data;
The adding module 570 is further configured to add the yield change rate to the historical dynamic data.
According to the working fluid level prediction device provided by the embodiment of the application, the collected data is trained and modeled according to the basic data and the historical dynamic characteristics of a single well, the real-time calculation of the working fluid level is realized through the working fluid level prediction model, and the real-time calculation result of the working fluid level is compared with the actual test result, so that the working fluid level prediction model is optimized, the online real-time calculation accuracy of the working fluid level is improved, the working fluid level is rapidly, simply, conveniently and efficiently obtained, the real-time adjustment of working conditions and measures in the oil gas development process is guided, the efficiency of a production system is further improved, and the benefit development level is improved.
Referring to fig. 7, fig. 7 is a schematic structural diagram of a server according to an exemplary embodiment of the present application. The server may be the server 220 shown in fig. 2. Specifically, the present invention relates to a method for manufacturing a semiconductor device.
The server 220 includes a central processing unit (CPU, central Processing Unit) 701, a system Memory 704 including a random access Memory (RAM, random Access Memory) 702 and a Read Only Memory (ROM) 703, and a system bus 705 connecting the system Memory 704 and the central processing unit 701. The server 220 further includes a basic input/output system (I/O system, input Output System) 706 for facilitating the transfer of information between various devices within the computer, and a mass storage device 707 for storing an operating system 713, application programs 714, and other program modules 715.
The basic input/output system 706 includes a display 708 for displaying information and an input device 709, such as a mouse, keyboard, or the like, for a user to input information. Wherein both the display 708 and the input device 709 are coupled to the central processing unit 701 through an input output controller 710 coupled to the system bus 705. The basic input/output system 706 may also include an input/output controller 710 for receiving and processing input from a number of other devices, such as a keyboard, mouse, or electronic stylus. Similarly, the input output controller 710 also provides output to a display screen, a printer, or other type of output device.
The mass storage device 707 is connected to the central processing unit 701 through a mass storage controller (not shown) connected to the system bus 705. The mass storage device 707 and its associated computer readable media provide non-volatile storage for the server 220. That is, the mass storage device 707 may include a computer readable medium (not shown) such as a hard disk or compact disc read only memory (CD-ROM, compact Disc Read Only Memory) drive.
Computer readable media may include computer storage media and communication media without loss of generality. Computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes RAM, ROM, erasable programmable read-only memory (EPROM, erasable Programmable Read Only Memory), electrically erasable programmable read-only memory (EEPROM, electrically Erasable Programmable Read Only Memory), flash memory or other solid state storage, CD-ROM, digital versatile disks (DVD, digital Versatile Disc) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices. Of course, those skilled in the art will recognize that computer storage media are not limited to the ones described above. The system memory 704 and mass storage device 707 described above may be collectively referred to as memory.
According to various embodiments of the present application, server 220 may also operate by being connected to remote computers on a network, such as the Internet. That is, the server 220 may be connected to the network 712 via a network interface unit 711 coupled to the system bus 705, or the network interface unit 711 may be used to connect to other types of networks or remote computer systems (not shown).
The memory also includes one or more programs, one or more programs stored in the memory and configured to be executed by the CPU.
Embodiments of the present application also provide a computer device, where the computer device includes a processor and a memory, where the memory stores at least one instruction, at least one program, a code set, or an instruction set, and where the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement the method for predicting a working fluid level provided by the above method embodiments.
Embodiments of the present application further provide a computer readable storage medium having at least one instruction, at least one program, a code set, or an instruction set stored thereon, where the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the method for predicting a working fluid level provided by the above method embodiments.
Embodiments of the present application also provide a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the method of predicting the meniscus as described in any one of the above embodiments.
Alternatively, the computer-readable storage medium may include: read Only Memory (ROM), random access Memory (RAM, random Access Memory), solid state disk (SSD, solid State Drives), or optical disk, etc. The random access memory may include resistive random access memory (ReRAM, resistance Random Access Memory) and dynamic random access memory (DRAM, dynamic Random Access Memory), among others. The foregoing embodiment numbers of the present application are merely for describing, and do not represent advantages or disadvantages of the embodiments.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program for instructing relevant hardware, where the program may be stored in a computer readable storage medium, and the storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The foregoing description of the preferred embodiments is merely exemplary in nature and is in no way intended to limit the invention, since it is intended that all modifications, equivalents, improvements, etc. that fall within the spirit and scope of the invention.
Claims (10)
1. A method of predicting a meniscus, the method comprising:
acquiring basic data of a single well and historical dynamic data of the single well, wherein the basic data are static characteristic data of the single well, and the historical dynamic data are dynamic change characteristic data of the single well in a target historical time period;
acquiring reference working fluid level data corresponding to the historical dynamic data;
model training is carried out based on the basic data, the historical dynamic data and the reference working fluid level data, so as to obtain a working fluid level prediction model;
acquiring current dynamic data of the single well, wherein the current dynamic data corresponds to the data type of the historical dynamic data;
and carrying out the dynamic liquid level prediction on the current dynamic data through the dynamic liquid level prediction model to obtain the dynamic liquid level depth corresponding to the current dynamic data.
2. The method of claim 1, wherein the model training based on the base data, the historical dynamic data, and the reference meniscus data to obtain a meniscus prediction model comprises:
Acquiring at least two basic models, wherein the at least two basic models are models with different prediction algorithms;
and performing fusion training on the at least two basic models by taking the basic data, the historical dynamic data and the reference working fluid level data as training data to obtain the working fluid level prediction model.
3. The method according to claim 2, wherein the performing fusion training on the at least two basic models by using the basic data, the historical dynamic data and the reference working fluid level data as training data to obtain the working fluid level prediction model includes:
dividing the training data to obtain at least two training data sets;
training the at least two basic models in the ith cycle through the at least two training data sets respectively to obtain an ith prediction result of the at least two training data sets;
splicing the ith prediction results of the at least two models to serve as training data sets in the (i+1) th cycle until the at least two basic models are converged;
and fusing the at least two converged basic models to obtain the dynamic liquid level prediction model.
4. The method of claim 2, wherein the at least two base models comprise at least two of a linear model, a random forest model, an iterative algorithm model, a gradient lifting algorithm model, and an extreme gradient lifting algorithm model.
5. The method of any one of claims 1 to 4, wherein the model training based on the base data, the historical dynamic data, and the reference meniscus data to obtain a meniscus prediction model comprises:
normalizing the basic data and the historical dynamic data to obtain normalized basic data and historical dynamic data;
and carrying out model training through the normalized basic data and the historical dynamic data and the reference working fluid level data to obtain a working fluid level prediction model.
6. The method of any one of claims 1 to 4, wherein the model training based on the base data, the historical dynamic data, and the reference meniscus data further comprises, prior to obtaining a meniscus prediction model:
performing data filtering based on a data value range of the reference working fluid level data, wherein the data value range is used for indicating an effective range of the working fluid level depth distribution;
In response to the missing data being present, interpolation is used to fill in the missing data.
7. The method of any one of claims 1 to 4, wherein the model training based on the base data, the historical dynamic data, and the reference meniscus data further comprises, prior to obtaining a meniscus prediction model:
acquiring n single well production data in a target historical time period before the reference working fluid level data, wherein n is a positive integer;
acquiring the yield change rate of the n single well production data according to the change condition of the n single well production data;
adding the yield rate of change to the historical dynamic data.
8. A meniscus predicting device, the device comprising:
the system comprises an acquisition module, a storage module and a storage module, wherein the acquisition module is used for acquiring basic data of a single well and historical dynamic data of the single well, the basic data are static characteristic data of the single well, and the historical dynamic data are dynamic change characteristic data of the single well in a target historical time period;
the acquisition module is also used for acquiring reference working fluid level data corresponding to the historical dynamic data;
the determining module is used for carrying out model training based on the basic data, the historical dynamic data and the reference working fluid level data to obtain a working fluid level prediction model;
The acquisition module is further used for acquiring current dynamic data of the single well, wherein the current dynamic data corresponds to the data type of the historical dynamic data;
and the prediction module is used for predicting the working fluid level of the current dynamic data through the working fluid level prediction model to obtain the working fluid level depth corresponding to the current dynamic data.
9. A computer device comprising a processor and a memory, wherein the memory has stored therein at least one program that is loaded and executed by the processor to implement the method of predicting a meniscus as claimed in any one of claims 1 to 8.
10. A computer readable storage medium having stored therein at least one instruction, at least one program, code set, or instruction set, the at least one instruction, the at least one program, the code set, or instruction set being loaded and executed by a processor to implement the meniscus predicting method of any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111308480.4A CN116084921A (en) | 2021-11-05 | 2021-11-05 | Working fluid level prediction method, device, apparatus, and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111308480.4A CN116084921A (en) | 2021-11-05 | 2021-11-05 | Working fluid level prediction method, device, apparatus, and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116084921A true CN116084921A (en) | 2023-05-09 |
Family
ID=86212544
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111308480.4A Pending CN116084921A (en) | 2021-11-05 | 2021-11-05 | Working fluid level prediction method, device, apparatus, and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116084921A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117514148A (en) * | 2024-01-05 | 2024-02-06 | 贵州航天凯山石油仪器有限公司 | Oil-gas well working fluid level identification and diagnosis method based on multidimensional credibility fusion |
-
2021
- 2021-11-05 CN CN202111308480.4A patent/CN116084921A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117514148A (en) * | 2024-01-05 | 2024-02-06 | 贵州航天凯山石油仪器有限公司 | Oil-gas well working fluid level identification and diagnosis method based on multidimensional credibility fusion |
| CN117514148B (en) * | 2024-01-05 | 2024-03-26 | 贵州航天凯山石油仪器有限公司 | Oil-gas well working fluid level identification and diagnosis method based on multidimensional credibility fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10060247B2 (en) | Hydrocarbon well performance monitoring system | |
| RU2600497C2 (en) | Methods and related system of constructing models and predicting operational results of drilling operation | |
| US20190251460A1 (en) | Method for predicting oil and gas reservoir production | |
| US9810212B2 (en) | Fluid load line calculation and concavity test for downhole pump card | |
| US20200018151A1 (en) | Diagnostics of downhole dynamometer data for control and troubleshooting of reciprocating rod lift systems | |
| CN107288617A (en) | A kind of method and system for improving rod-pumped well indicator card gauging precision | |
| US20090182694A1 (en) | Evaluation of acid fracturing treatments in an oilfield | |
| Okoro et al. | Application of artificial intelligence in predicting the dynamics of bottom hole pressure for under-balanced drilling: Extra tree compared with feed forward neural network model | |
| US20220027538A1 (en) | Hydraulic fracturing job plan real-time revisions utilizing detected response feature data | |
| RU2598261C1 (en) | Monitoring and diagnostics of flooded reservoirs data using operational data | |
| CN103510940B (en) | Mechanical oil well operating mode integrated diagnosing and analyzing method and device | |
| US9200509B2 (en) | System and method for measuring well flow rate | |
| US11972183B2 (en) | Reduced physics well production monitoring | |
| US20230184239A1 (en) | System and method for rod pump autonomous optimization without a continued use of both load cell and electric power sensor | |
| Deng et al. | Rapid and comprehensive artificial lift systems performance analysis through data analytics, diagnostics and solution evaluation | |
| Molinari et al. | A Reduced Physics Modeling Approach to Understand Multiphase Well Production Performance for Unconventional Reservoirs | |
| CN116084921A (en) | Working fluid level prediction method, device, apparatus, and readable storage medium | |
| US11767750B1 (en) | Gas-oil ratio forecasting in unconventional reservoirs | |
| US20220074291A1 (en) | System and method for reservoired oil production based on calculated composition of natural tracers | |
| US20220195861A1 (en) | Identifying operation anomalies of subterranean drilling equipment | |
| Tran | A Data-Driven Approach For Monitoring And Predictive Diagnosis Of Sucker Rod Pump System | |
| Simanjuntak et al. | Applying artificial neural network and xgboost to improve data analytics in oil and gas industry | |
| Min et al. | Prediction of single well production rate in water-flooding oil fields driven by the fusion of static, temporal and spatial information | |
| US20230193736A1 (en) | Infill development prediction system | |
| US20240060405A1 (en) | Method and system for generating predictive logic and query reasoning in knowledge graphs for petroleum systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |