Separated layer water injection real-time data monitoring and management system based on DTU
By using DTU network and edge computing technology in the oil field, combined with machine learning and real-time data monitoring, the problems of data lag and low management efficiency under traditional manual measurement and adjustment methods are solved, and remote automated management and refined control of water injection wells are realized. Improve the management efficiency and stability of the oil field water injection process.
Patent Information
- Application Number
- CN202410961201.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-18
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The traditional stratified water injection management method relies on manual measurement and adjustment, resulting in inaccurate and lagging data recording, which cannot promptly reflect the actual situation of the oil field. As the scale of the oil field expands and the number of water injection wells increases, manual methods are difficult to deal with and cannot meet production needs.
Using a layered water injection real-time data monitoring and management system based on DTU, combined with edge computing technology, by placing water injection devices and wellhead sensors downhole and connecting them with edge computing hosts, real-time monitoring and management of the oil layer water injection state is realized, and machine learning is used The training module and the data processing module perform data processing and analysis, provide remote management and control functions, and adopt polling mode and real-time measurement and adjustment mode to alleviate the pressure on network nodes.
It improves data accuracy and processing efficiency, realizes remote automated management of water injection wells, reduces on-site operation and management work, improves the overall performance and stability of the system, and supports the refined and personalized management of the oil field water injection process.
Smart Images

Figure CN118945187A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of petroleum stratified water injection, and in particular to a stratified water injection real-time data monitoring and management system. Background Art
[0002] Layered water injection technology is an important means in oil field development, especially in the secondary development stage. Its principle is to separate the oil layers through packers so as to carry out fine water injection according to the characteristics of each oil layer. This technology controls the water injection volume of each oil layer by adjusting the water nozzle of the water distributor, so that the water absorption of the high permeability layer is reduced, while the water absorption of the medium and low permeability layers is increased, so that each oil layer benefits evenly, improves the overall recovery rate of the oil field, and ensures the stable output of the oil field. However, there are a series of problems in the traditional layered water injection management method. Among them, it relies on manual measurement and adjustment, resulting in inaccurate data recording and serious lag. On-site measurement and adjustment work requires measurement and adjustment workers to monitor and adjust the water injection wells in real time, and the recording and uploading of data rely entirely on manual work. This method often cannot reflect the actual situation of the oil field in a timely and accurate manner, resulting in managers being unable to flexibly manage and supervise the status of water injection wells.
[0003] With the continuous advancement of oil extraction technology and the continuous expansion of oil field development, the importance of stratified water injection has become increasingly prominent. However, the traditional manual measurement and adjustment method can no longer meet the growing production needs. With the expansion of oil field scale and the increase in the number of water injection wells, the workload of measurement and adjustment has shown an exponential growth trend, and manual methods have become difficult to cope with. Therefore, there is an urgent need for a more intelligent and automated stratified water injection management system to improve production efficiency, reduce costs, and ensure the stable development of oil fields.
[0004] The document with the publication number CN 116752935 A discloses a ground control device for high-speed pulse communication and a control method thereof. The invention is mainly used in the intelligent stratified water injection system of oil fields. The device includes a flow regulating component and a fast code sending component, which are connected through a water injection pipeline and connected to a downhole water distributor. The main control board of the ground control device controls the fast code sending component through an electrical connection to continuously open and close it, thereby generating high and low pressure waveforms in the water injection pipeline. The downhole water distributor detects these changing pressure waveforms and uses pressure wave codes to communicate pressure pulses with the ground control device. This design enables the main control board to finely adjust the flow and fast code sending by adding a fast code sending function, thereby improving the management efficiency of the stratified water injection system. The patent improves the downhole wave code communication speed through structural design, but while improving the communication speed, it still requires manual on-site connection to measure and adjust the water injection wells, which improves the communication efficiency of the water injection device, but faces the problem of an increase in water injection wells and a large workload of on-site measurement and adjustment.
[0005] The document with the publication number CN 116537756 A discloses a high temperature and high pressure resistant intelligent stratified water injection system based on DSP, which is composed of key components such as an outer tube, a central tube, a receiving chamber, a water injection inlet, multiple water injection outlets, an adjustable valve, a driving mechanism and a DSP processor. The central tube is located inside the outer tube, and both ends are connected to the water injection pipe to form a water injection channel. The water injection inlet is located at the lower end of the central tube and is connected to the receiving chamber, and multiple water injection outlets are arranged at intervals along the circumference of the outer tube. Each water injection outlet is equipped with an adjustable valve, which is driven by a corresponding driving mechanism and can move axially along the outer tube to adjust the water injection amount. The DSP processor is connected to the driving mechanism and is responsible for controlling its operation to ensure the accuracy of the water injection process. In addition, the system also includes a flow meter for real-time monitoring of the water injection flow rate, thereby achieving the purpose of intelligent stratified water injection. This patent realizes sensor integration and real-time data monitoring through DSP, but it still requires manual remote control and monitoring of the stratified water injection device through the host computer on site. Therefore, the problem of heavy workload in on-site measurement and adjustment of water injection wells cannot be solved simply from the perspective of sensor integration and data real-time performance. Summary of the invention
[0006] In order to solve the technical problems existing in the background technology, the present disclosure proposes a real-time data monitoring and management system for stratified water injection based on DTU. The technical solution provided by the present disclosure, combined with edge computing technology, realizes real-time monitoring and management of the water injection status of the oil layer by placing water injection devices and wellhead sensors underground in the oil field and connecting them to the edge computing host. The system includes a data layer, a service layer and a management layer. It processes and analyzes data through a machine learning training module and a data processing module, and provides remote management and control functions. The implementation methods adopted by the edge computer include polling mode and real-time measurement and adjustment mode, which can effectively alleviate the pressure on network nodes and realize remote management and monitoring of water injection wells. The system has the advantages of improving data accuracy and realizing intelligent management of water injection wells, and has important practical application value for the refined and personalized management of the oil field water injection process.
[0007] The present disclosure first provides a DTU-based stratified water injection real-time data monitoring and management system, including a data layer, a service layer, and a management layer, as well as:
[0008] The system also includes an access layer and a DTU network;
[0009] The access layer includes a water injection device, a wellhead sensor and an edge computing host; the water injection device is arranged underground and is used to inject water into a designated oil layer;
[0010] The wellhead sensor includes a wellhead pressure sensor, a wellhead temperature sensor and a wellhead flow sensor; the wellhead sensor is arranged at the wellhead and is used to display the wellhead pressure, temperature and flow data so that the operator can perceive the displayed data, and the operator can adjust the flow and pressure of the water injection well through the wellhead valve according to the on-site working conditions of the oil field;
[0011] The edge computing host is connected to the water injection device and the wellhead sensor respectively; the edge computing host includes a machine learning training module, a data processing module and a terminal control module;
[0012] The data processing module is used to collect and process the data transmitted by the water injection device and the wellhead sensor;
[0013] The machine learning training module has computing capabilities, is responsible for the training and calculation of the water injection well intelligent algorithm, and can calculate the data transmitted by the data processing module;
[0014] The terminal control module has a terminal interface for operating the internal system of the edge computing host;
[0015] The terminal control module can generate remote well stopping and sealing detection instructions, which are pushed by the edge host to the web layered water injection remote monitoring client in the management layer through the DTU network in sequence;
[0016] The data layer is provided with a stratified water injection database, which includes a stratified water injection relational database and a stratified water injection real-time database;
[0017] The hierarchical water injection relational database is used to clean, process, associate, integrate and classify the information of factories, mines, wells, teams and users; the hierarchical water injection relational database stores static data of water injection wells, and the static data of water injection wells includes the mining area where the well is located, the well location coordinates, the number of well layers and the injection volume; the hierarchical water injection relational database can exchange data with the data analysis module, and the data exchanged includes the layer information, location information and well affiliation information of the water injection wells; the data analysis module can extract the basic data of water injection wells in the hierarchical water injection relational database for statistical analysis;
[0018] The stratified water injection real-time database can be used to implement the following functions:
[0019] The water injection well data obtained by the water injection device and the wellhead sensor are sent to the DTU network via the edge computing host, and then sent to the stratified water injection real-time database for storage through the DTU network. The stored data include wellhead pressure, temperature and flow data; the stratified water injection real-time database can exchange data with the data analysis module, and the exchanged data include historical data of temperature and flow, and historical data of the valve sensor of the water injection device; the data analysis module can perform statistical analysis on the data in the stratified water injection real-time database;
[0020] The service layer includes a data analysis module, a water injection real-time data module and a measurement and adjustment module;
[0021] The data analysis module is used to extract well basic information from the stratified water injection relational database in the stratified water injection database, and output effective data analysis results in combination with historical data in the stratified water injection real-time database, wherein the effective data analysis results include historical statistical analysis of water injection well flow, historical statistical analysis of cumulative water injection volume, historical statistical analysis of water injection well pressure parameters, historical statistical analysis of water injection well temperature parameters, historical statistical analysis of water injection well nozzle opening data, historical statistical analysis of wellhead flow data, historical statistical analysis of wellhead pressure data, and historical statistical analysis of wellhead valve opening data; the effective data analysis results are provided to the operator via the web stratified water injection remote monitoring client in the management layer;
[0022] The water injection real-time data module can be used to achieve the following functions:
[0023] The edge computing host transmits the sensor parameters collected from the water injection device and the wellhead sensor to the water injection real-time data module through the DTU network, and the water injection real-time data module transmits the collected data to the web layered water injection remote management client and the layered water injection remote control client in real time;
[0024] The measurement and adjustment module is used to remotely control the valve sensor of the water injection device and the wellhead valve. The measurement and adjustment module sends the water injection device valve sensor control command and the wellhead valve control command to the edge computing host based on the DTU network to adjust the downhole and wellhead flow rates;
[0025] The management layer includes a web layered water injection remote management client and a layered water injection remote control client;
[0026] The web layered water injection remote management client exchanges data with the data analysis module and the water injection real-time data module respectively;
[0027] The web-based layered water injection remote management client realizes real-time analysis of water injection well data and real-time monitoring of water injection wells through a data analysis module and a water injection real-time data module;
[0028] The stratified water injection remote control client realizes the real-time monitoring and real-time measurement and adjustment functions of water injection wells through the real-time water injection data module and the measurement and adjustment module, and then sends the feedback results to the stratified water injection database through the DTU network for storage;
[0029] The DTU network is used to implement data communication between the access layer, the data layer and the management layer.
[0030] Furthermore, the water injection device comprises a water injection device pressure sensor, a water injection device flow sensor, a water injection device temperature sensor and a water injection device valve sensor, which are respectively used to detect the pressure, flow and temperature during the oil layer water injection process.
[0031] Further, the edge computing host is configured to adopt a polling mode and a real-time measurement and adjustment mode;
[0032] In the polling mode, the edge computing host reads data from the water injection device pressure sensor, water injection device flow sensor, water injection device temperature sensor, wellhead pressure sensor, wellhead flow sensor and wellhead valve at a determined time interval;
[0033] The edge computing host collects water injection well data at a determined frequency and sends it to the layered water injection database through the DTU network for storage. The amount of data collected is selected by the user according to the type of oil well.
[0034] In the real-time measurement and adjustment mode, the edge computing host processes the collected water injection device signals and wellhead sensor signals and sends them to the stratified water injection real-time database, the stratified water injection relational database, the web stratified water injection remote management client and the stratified water injection remote control client respectively;
[0035] The web layered water injection remote management client and the layered water injection remote control client monitor and control the water injection wells.
[0036] Furthermore, the web stratified water injection remote management client and stratified water injection remote control client can receive remote well stop and sealing detection instructions generated by the terminal control module in the edge computing host sent through the DTU network for monitoring and control, and the feedback results are then delivered to the stratified water injection database through the edge computing host and the DTU network for storage;
[0037] After receiving the water injection data transmitted by the edge computing host, the stratified water injection real-time database stores the real-time data of the water injection wells;
[0038] The data analysis module extracts the required storage data or real-time data for analysis through the layered water injection database.
[0039] Furthermore, the data processing module in the edge computing host processes the historical data of the water injection device flow sensor, water injection device temperature sensor, water injection device valve sensor, wellhead pressure sensor, wellhead flow sensor, and wellhead valve in the stratified water injection real-time database in the following manner, specifically:
[0040] Step 1: Data collection and preprocessing, obtaining corresponding historical data from the stratified water injection real-time database;
[0041] Step 2: Use box plot to detect outliers, sort the data from large to small, Q1 represents the data point at the 25% position, Q2 represents the data point at the 50% position, Q3 represents the data point at the 75% position, IQR = Q3-Q1; determine the outlier boundary, the upper boundary is Q1-1.5×IQR, the lower boundary is Q3+1.5×IQR, and the data outside this range is determined as an outlier and deleted;
[0042] Step 3: Convert data of different dimensions to the same dimension. The historical data is represented by x. The data is scaled to the interval [0, 1]. The public representation is:
[0043]
[0044] The machine learning training module in the edge computing host uses a reinforcement learning algorithm to generate an intelligent optimization model for water injection wells by analyzing and learning the historical data of water injection wells. This is specifically achieved through the following steps:
[0045] Step 1: Construct the state space, Among them i represents the temperature of the i-th layer underground, x i represents the water nozzle opening of the i-th layer in the well, h i represents the depth of the i-th layer in the well, p bi represents the pressure of the i-th layer in the well, q i represents the flow rate of the i-th layer in the well, p o Indicates the wellhead pressure, te o represents the wellhead temperature, x o Indicates the wellhead valve opening, q tai represents the target injection volume of the i-th layer;
[0046] Step 2: Action space, define the action space according to the operation requirements of the water injection well where t i represents the valve movement time unit s, t of the i-th layer i ∈[-30,30], when t i >0, the valve rotates in the direction of increasing opening. iWhen <0, the valve rotates in the direction of decreasing opening;
[0047] Step 3: Setting the reward function, giving the water spout movement direction reward based on the target injection amount and the actual input amount. The movement reward is expressed as:
[0048]
[0049] Reduce the number of faucet adjustment steps and give rewards based on the number of faucet adjustments. k represents the number of faucet adjustments, and the adjustment step reward is expressed as:
[0050] R2=-0.5×k
[0051] The injection volume error is used to measure the accuracy of the water nozzle, and the error reward is expressed as:
[0052]
[0053] The total reward is expressed as the movement reward, and the adjustment step reward and error reward are expressed as follows:
[0054] R = R1 + R2 + R3;
[0055] Step 4: Reinforcement learning training. The model is trained through the proximal policy optimization (PPO) reinforcement learning algorithm. The machine learning training module uses the historical data preprocessed by the data acquisition and preprocessing module and the current real-time data to iteratively update the model parameters. The number of training steps is 500 and the stopping condition is that the error is less than 5%. The maximum deployment per round is 100 steps. The number of hidden layer neurons is 128×3, the number of input layer neurons is 8×n+1, and the number of output layer neurons is n.
[0056] Furthermore, the terminal control module generates remote well stopping and sealing detection instructions according to the following path:
[0057] The remote well stopping instruction is implemented by the following steps:
[0058] Step 1: The water injection well management user issues a well stop command through the stratified water injection remote control client;
[0059] Step 2: After receiving the command, the measurement and adjustment module sends a well stop command to the edge computing host through the DTU network;
[0060] Step 3: Close the wellhead sensor valve to 0 opening;
[0061] Step 4: Close the water injection device valve sensor and set the opening to 0;
[0062] The seal detection instruction is implemented by the following steps:
[0063] Step 1: The water injection well management user issues a sealing detection command through the stratified water injection remote control client;
[0064] Step 2: After receiving the command, the test and adjustment module sends a sealing detection command to the edge computing host through the DTU network;
[0065] Step 3: Close the valve sensor of the water injection device of the odd-numbered layers;
[0066] Step 4: The wellhead valve cycle executes the valve opening and closing commands.
[0067] At least one of the above technical solutions adopted in one or more embodiments of this specification can achieve the following beneficial effects:
[0068] First of all, the technical solution provided by the present invention combines edge computing technology with the water injection control system, improves the processing and transmission efficiency of water injection data, realizes automatic adjustment of water injection well layers based on reinforcement learning technology, greatly reduces on-site operations and management work, and truly realizes remote automated management of water injection wells.
[0069] Secondly, the real-time water injection data monitoring and management system provided by the present disclosure builds a remote wireless two-way communication network for water injection wells based on the DTU network, making remote management and monitoring of water injection wells possible. Through the web-based layered water injection remote management client and the layered water injection remote control client, operators can easily adjust the flow rate, test and monitor the water injection wells in real time, making the management of water injection wells more convenient and efficient.
[0070] In addition, the real-time water injection data monitoring and management system provided by the present disclosure relies on the intelligent characteristics of the edge computing host, realizes the automatic control technology of the water injection device based on the reinforcement learning module, realizes the data collection and processing technology of the water injection well based on the data module, and realizes the remote well stopping and sealing detection technology based on the control terminal. Through the application of the edge computing host, the data processing pressure of the network node center is effectively reduced, and the real-time monitoring and management of the data are realized, thereby improving the overall performance and stability of the system.
[0071] In summary, the technical solution provided by the present invention realizes real-time data monitoring and management of stratified water injection based on the DTU network, can realize the scheduled automatic upload of water injection data, ensures the accuracy and reliability of the data, and uses the edge computing host of the ARM architecture with certain computing capabilities, which lays a good foundation for the intelligent management of water injection wells, further improves the intelligence level of the system, and can support the refined and personalized management of the oilfield water injection process, solving the problem that the existing water injection management system cannot meet production needs in terms of water injection management and automation.
[0072] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, rather than limiting the present disclosure. According to the following detailed description of the embodiments and with reference to the accompanying drawings, other features of the present disclosure will become more apparent. BRIEF DESCRIPTION OF THE DRAWINGS
[0073] The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and are used to illustrate the technical solutions of the present disclosure together with the specification.
[0074] Figure 1 Schematic diagram of the structure of the real-time data monitoring and management system for stratified water injection based on DTU.
[0075] Figure 2 A schematic structural diagram of a water injection device used in a specific implementation of the present disclosure.
[0076] Figure 3 A schematic diagram of the structure of a wellhead sensor used in a specific implementation of the present disclosure.
[0077] Figure 4 A schematic diagram of the three module structures in the edge computing host used in the specific implementation of the present disclosure.
[0078] Figure 5 A schematic diagram of the implementation process of the data processing module used in the specific implementation of the present disclosure.
[0079] Figure 6 A schematic diagram of the implementation process of the machine learning training module used in the specific implementation of the present disclosure.
[0080] Figure 7 A schematic diagram of the remote well shutdown implementation process used in the specific implementation of the present disclosure.
[0081] Figure 8 A schematic diagram of the sealing detection implementation process used in the specific implementation of the present disclosure.
[0082] In the figure, 1-access layer, 11-water injection device, 111-water injection device pressure sensor, 112-water injection device flow sensor, 113-water injection device temperature sensor, 114-injection device valve sensor, 12-wellhead sensor, 121-wellhead pressure sensor, 122-wellhead flow sensor, 123-wellhead valve, 13-edge computing host, 131-machine learning training module, 132-data processing module, 133-terminal control module, 2-data layer, 21-layered water injection data, 22-layered water injection relational database, 23-layered water injection real-time database, 3-service layer, 31-data analysis module, 32-water injection real-time data module, 33-measurement and adjustment module, 4-management layer, 41-web layered water injection remote management client, 42-layered water injection remote control client, 5-DTU network. DETAILED DESCRIPTION
[0083] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.
[0084] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, methods and means well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present disclosure.
[0085] This disclosure provides a DTU-based stratified water injection real-time data monitoring and management system, including a data layer, a service layer, and a management layer. The uniqueness of the system is that:
[0086] The system also includes an access layer and a DTU network;
[0087] The access layer includes a water injection device 11, a wellhead sensor 12 and an edge computing host 13; the water injection device 11 is arranged underground and is used to inject water into a designated oil layer;
[0088] The water injection device 11 is as follows Figure 2 As shown, it includes a water injection device pressure sensor 111, a water injection device flow sensor 112, a water injection device temperature sensor 113, and a water injection device valve sensor 114. The operator controls the layer flow by adjusting the water injection device valve sensor 114. The water injection device pressure sensor 111, the water injection device flow sensor 112, the water injection device temperature sensor 113, and the water injection device valve sensor 114 are used to detect the pressure, flow and temperature during the oil layer water injection process and the water injection device water nozzle opening information.
[0089] The wellhead sensor 12 is as follows Figure 3 As shown, it includes a wellhead pressure sensor 121 and a wellhead flow sensor 122; the wellhead pressure sensor 121 and the wellhead flow sensor 122 are arranged at the wellhead, and are used to display the wellhead pressure and flow data, so that the operator can perceive the displayed data, and the operator can adjust the flow and pressure of the water injection well through the wellhead valve 123 according to the on-site working conditions of the oil field.
[0090] The edge computing host 13 is connected to the water injection device 11 and the wellhead sensor 12 respectively; Figure 4 The edge computing host 13 shown includes a machine learning training module 131, a data processing module 132 and a terminal control module 133;
[0091] The data processing module 132 is used to collect and process the data transmitted by the water injection device 11 and the wellhead sensor 12 .
[0092] The machine learning training module 131 has computing capabilities, is responsible for the training and calculation of the water injection well intelligent algorithm, and can calculate the data transmitted by the data processing module 132.
[0093] The terminal control module 133 has a terminal interface for operating the internal system of the edge computing host 13.
[0094] The terminal control module 133 can generate remote well stopping and sealing detection instructions, and the edge host 13 pushes them to the stratified water injection remote control client 42 and the web stratified water injection remote management client 41 in the management layer through the DTU network 5.
[0095] The data layer stratified water injection database 21 includes a stratified water injection relational database 22 and a stratified water injection real-time database 23 .
[0096] The stratified water injection relation database 22 and other units are related in the following mode: the stratified water injection relation database 22 stores the static data of water injection wells, including the mining area, well location coordinates, well layer number and injection amount. It exchanges data with the data analysis module 31, and the exchanged data includes the water injection well layer information, location information, and well belonging relationship information. The data analysis module 31 extracts the basic data of water injection wells in the stratified water injection relation database 22 for statistical analysis.
[0097] The stratified water injection real-time database 23 and other units are related in the following mode: on the one hand, the water injection well data obtained by the water injection device 11 and the wellhead sensor 12 are sent to the DTU network 5 via the edge computing host 13, and sent to the stratified water injection real-time database 23 for storage through the DTU network 5. The stored data includes the water injection device pressure sensor 111 data, the water injection device flow sensor 112 data, the water injection device temperature sensor 113 data, the water injection device valve sensor 114 data, the wellhead pressure sensor 121 data, and the wellhead flow sensor 122 data. On the other hand, the stratified water injection real-time database 23 exchanges data with the data analysis module 31, and the exchanged data includes the historical data of the pressure sensor 111 of the water injection device of the water injection well, the historical data of the flow sensor 112 of the water injection device, the historical data of the temperature sensor 113 of the water injection device, the historical data of the valve sensor 114 of the water injection device, the historical data of the wellhead pressure sensor 121, the historical data of the wellhead flow sensor 122, and the historical data of the wellhead valve 123. The data analysis module 31 performs statistical analysis on the data in the stratified water injection real-time database 23. The stratified water injection real-time database 23 is used to store the historical data of the pressure sensor 111 of the water injection well, the flow sensor 112 of the water injection device, the temperature sensor 113 of the water injection device, the valve sensor 114 of the water injection device, the wellhead pressure sensor 121, the wellhead flow sensor 122, and the wellhead valve 123.
[0098] The service layer includes a data analysis module, a water injection real-time data module and a measurement and adjustment module.
[0099] The data analysis module 31 and other units have a relationship according to the following mode: extracting well basic information from the stratified water injection relationship database 22 in the stratified water injection database 21, combining the historical data of the water injection device pressure sensor 111, the historical data of the water injection device flow sensor 112, the historical data of the water injection device temperature sensor 113, the historical data of the water injection device valve sensor 114, the historical data of the wellhead pressure sensor 121, the historical data of the wellhead flow sensor 122, and the historical data of the wellhead valve 123 in the stratified water injection real-time database 23, and outputting data analysis results, the effective data analysis results include historical statistical analysis of water injection well flow, historical statistical analysis of cumulative water injection volume, historical statistical analysis of water injection well pressure parameters, historical statistical analysis of water injection well temperature parameters, historical statistical analysis of water injection well nozzle opening data, historical statistical analysis of wellhead flow data, historical statistical analysis of wellhead pressure data, and historical statistical analysis of wellhead valve opening data; the data analysis results are provided to the operator via the web stratified water injection remote management client 41 in the management layer.
[0100] The water injection real-time data module 32 and other units have a relationship in the following mode: the edge computing host 13 transmits the sensor parameters collected from the water injection device 11 and the wellhead sensor 12 to the water injection real-time data module 32 through the DTU network 5. The sensor data includes: real-time data of the water injection device pressure sensor 111, real-time data of the water injection device flow sensor 112, real-time data of the water injection device temperature sensor 113, real-time data of the water injection device valve sensor 114, real-time data of the wellhead pressure sensor 121, real-time data of the wellhead flow sensor 122, and real-time data of the wellhead valve 123. The collected data are transmitted in real time to the web layered water injection remote management client 41 and the layered water injection remote control client 42.
[0101] The measurement and adjustment module 33 and other units interact in the following mode: the measurement and adjustment module 33 remotely controls the water injection device valve sensor 114 and the wellhead valve 123. The measurement and adjustment module 33 sends the water injection device valve sensor 114 control command and the wellhead valve 123 control command to the edge computing host 13 based on the DTU network 5 to adjust the downhole and wellhead flow rates.
[0102] The management layer includes a web layered water injection remote management client and a layered water injection remote control client.
[0103] The web layered water injection remote management client 41 and other units have a relationship according to the following mode: the web layered water injection remote management client 41 exchanges data with the data analysis module 31 and the water injection real-time data module 32 respectively, and the exchanged data includes water injection well layer information, location information, well affiliation information, well water injection device pressure sensor 111 historical data, water injection device flow sensor 112 historical data, water injection device temperature sensor 113 historical data, water injection device valve sensor 114 historical data, wellhead pressure sensor 121 historical data, wellhead flow sensor 122 historical data, wellhead valve 123 historical data, water injection device pressure sensor 111 real-time data, water injection device flow sensor 112 real-time data, water injection device temperature sensor 113 real-time data, water injection device valve sensor 114 real-time data, wellhead pressure sensor 121 real-time data, wellhead flow sensor 122 real-time data, wellhead valve 123 real-time data
[0104] The edge computing host 13 queries the water injection device 11 and the wellhead sensor 12 at a time interval of 1 minute. The data collected by the edge computing host 13 include the data of the water injection device pressure sensor 111, the data of the water injection device flow sensor 112, the data of the water injection device temperature sensor 113, the data of the water injection device valve sensor 114, the data of the wellhead pressure sensor 121, the data of the wellhead flow sensor 122, and the data of the wellhead valve 123. The water injection device 11 data and the wellhead sensor 12 data collected by the edge computing host 13 are stored in the stratified water injection real-time database 23 based on the DTU network 5 transmission.
[0105] The data analysis module 31 provides the web layered water injection remote management client 41 with historical statistical analysis services for water injection well flow, historical statistical analysis services for cumulative water injection volume, historical statistical analysis services for water injection well pressure parameters, historical statistical analysis services for water injection well temperature parameters, historical statistical analysis services for water injection well nozzle opening data, historical statistical analysis services for wellhead flow data, historical statistical analysis services for wellhead pressure data, and historical statistical analysis services for wellhead valve opening data. According to the service requirements, the layer information, location information, and well affiliation information of the water injection well are extracted from the layered water injection relational database 22, and the historical data of the water injection device pressure sensor 111, the historical data of the water injection device flow sensor 112, the historical data of the water injection device temperature sensor 113, the historical data of the water injection device valve sensor 114, the historical data of the wellhead pressure sensor 121, the historical data of the wellhead flow sensor 122, and the historical data of the wellhead valve 123 are extracted from the layered water injection real-time database 23. The data analysis module 31 displays the historical statistical analysis results of the injection well flow rate, the historical statistical analysis results of the cumulative water injection volume, the historical statistical analysis results of the injection well pressure parameters, the historical statistical analysis results of the injection well temperature parameters, the historical statistical analysis results of the injection well nozzle opening data, the historical statistical analysis results of the wellhead flow data, the historical statistical analysis results of the wellhead pressure data and the historical statistical analysis results of the wellhead valve opening data on the web layered water injection remote management client 41 interface.
[0106] The water injection real-time data module 32 provides the web layered water injection remote management client 41 with real-time data services of the water injection device pressure sensor 111, the water injection device flow sensor 112, the water injection device temperature sensor 113, the water injection device valve sensor 114, the wellhead pressure sensor 121, the wellhead flow sensor 122, and the wellhead valve 123. The edge computing host queries the water injection device 11 and the wellhead sensor 12 at a time interval of 1 second. The data collected by the edge computing host 13 include the data of the water injection device pressure sensor 111, the water injection device flow sensor 112, the water injection device temperature sensor 113, the water injection device valve sensor 114, the wellhead pressure sensor 121, the wellhead flow sensor 122, and the wellhead valve 123. The water injection real-time data module 32 transmits data with the edge computing host 13 through the DTU network 5. The transmitted data includes the real-time data of the water injection device pressure sensor 111, the real-time data of the water injection device flow sensor 112, the real-time data of the water injection device temperature sensor 113, the real-time data of the water injection device valve sensor 114, the real-time data of the wellhead pressure sensor 121, the real-time data of the wellhead flow sensor 122, and the real-time data of the wellhead valve 123. The water injection real-time data module displays the real-time data results of the water injection device pressure sensor 111, the real-time data results of the water injection device flow sensor 112, the real-time data results of the water injection device temperature sensor 113, the real-time data results of the water injection device valve sensor 114, the real-time data results of the wellhead pressure sensor 121, the real-time data results of the wellhead flow sensor 122, and the real-time data results of the wellhead valve 123 on the web layered water injection remote management client 41 interface.
[0107] The stratified water injection remote control client 42 and other units have a relationship in the following mode: the stratified water injection remote control client 42 exchanges data with the water injection real-time data module 32 and the measurement and adjustment module 33 respectively, and the exchanged data includes the real-time data of the water injection device pressure sensor 111, the real-time data of the water injection device flow sensor 112, the real-time data of the water injection device temperature sensor 113, the real-time data of the water injection device valve sensor 114, the real-time data of the wellhead pressure sensor 121, the real-time data of the wellhead flow sensor 122, and the real-time data of the wellhead valve 123.
[0108] The water injection real-time data module 32 provides the stratified water injection remote control client 42 with real-time data services of the water injection device pressure sensor 111, the water injection device flow sensor 112, the water injection device temperature sensor 113, the water injection device valve sensor 114, the wellhead pressure sensor 121, the wellhead flow sensor 122, and the wellhead valve 123. The edge computing host queries the water injection device 11 and the wellhead sensor 12 at a time interval of 1 second. The data collected by the edge computing host 13 include the data of the water injection device pressure sensor 111, the water injection device flow sensor 112, the water injection device temperature sensor 113, the water injection device valve sensor 114, the wellhead pressure sensor 121, the wellhead flow sensor 122, and the wellhead valve 123. The water injection real-time data module 32 transmits data with the edge computing host 13 through the DTU network 5. The transmitted data includes the real-time data of the water injection device pressure sensor 111, the real-time data of the water injection device flow sensor 112, the real-time data of the water injection device temperature sensor 113, the real-time data of the water injection device valve sensor 114, the real-time data of the wellhead pressure sensor 121, the real-time data of the wellhead flow sensor 122, and the real-time data of the wellhead valve 123. The water injection real-time data module 32 displays the real-time data results of the water injection device pressure sensor 111, the real-time data results of the water injection device flow sensor 112, the real-time data results of the water injection device temperature sensor 113, the real-time data results of the water injection device valve sensor 114, the real-time data results of the wellhead pressure sensor 121, the real-time data results of the wellhead flow sensor 122, and the real-time data results of the wellhead valve 123 on the interface of the stratified water injection remote control client 42.
[0109] The measurement and adjustment module 31 provides water injection device flow regulation services and wellhead flow regulation services to the stratified water injection remote control client 42. The stratified water injection remote control client 42 issues control commands to the measurement and adjustment module, and the measurement and adjustment module 33 and the edge computing host 13 transmit the water injection device valve sensor 114 control command and the wellhead valve 123 control command through the DTU network 5. After receiving the command, the edge computing host 13 issues control commands to the water injection device valve sensor 114 and the wellhead valve 123 to change the wellhead and water injection device flow.
[0110] The DTU network 5 is used to implement data communication between the access layer, the data layer and the management layer.
[0111] Preferably, the edge computing host 13 is configured to adopt a polling mode and a real-time measurement and adjustment mode.
[0112] In the polling mode:
[0113] The edge computing host 13 reads data from the water injection device pressure sensor 111, the water injection device flow sensor 112, the water injection device temperature sensor 113, the water injection device valve sensor 114, the wellhead pressure sensor 121, the wellhead flow sensor 122 and the wellhead valve 123 at a determined time interval of 1 minute, and sends it to the stratified water injection real-time database 23 for storage through the DTU network 5.
[0114] In the real-time test and adjustment mode:
[0115] The edge computing host 13 reads data from the water injection device pressure sensor 111, the water injection device flow sensor 112, the water injection device temperature sensor 113, the water injection device valve sensor 114, the wellhead pressure sensor 121, the wellhead flow sensor 122 and the wellhead valve 123 at a determined time interval of 1 second, and sends the data to the stratified water injection real-time database 23, the web stratified water injection remote management client 41 and the stratified water injection remote control client 42 through the DTU network 5;
[0116] The stratified water injection remote control client 42 and the web stratified water injection remote management client 41 monitor and control the water injection wells.
[0117] The stratified water injection remote control client 42 and the web stratified water injection remote management client 41 can receive remote well stopping and sealing detection instructions generated by the terminal control module 133 in the edge computing host 13 sent through the DTU network 5 for monitoring and control.
[0118] The data processing module is specifically implemented as follows Figure 5 As shown. The historical data of the water injection device flow sensor 112, the historical data of the water injection device temperature sensor 113, the historical data of the water injection device valve sensor 114, the wellhead pressure sensor 121, the wellhead flow sensor 122, and the historical data of the wellhead valve 123 in the stratified water injection real-time database 23 are processed. This module is specifically implemented by the following steps:
[0119] Step 1: Data collection and preprocessing, obtain historical data of the water injection device pressure sensor 111, historical data of the water injection device flow sensor 112, historical data of the water injection device temperature sensor 113, historical data of the water injection device valve sensor 114, wellhead pressure sensor 121, wellhead flow sensor 122, and historical data of the wellhead valve 123 from the stratified water injection real-time database 23.
[0120] Step 2: Use box plot to detect outliers, sort the data from large to small, Q1 represents the data point at the 25% position, Q2 represents the data point at the 50% position, Q3 represents the data point at the 75% position, IQR = Q3-Q1. Determine the outlier boundary, the upper boundary is Q1-1.5×IQR, the lower boundary is Q3+1.5×IQR, and the data outside this range is determined as an outlier and deleted.
[0121] Step 3: Convert data of different dimensions to the same dimension. The historical data is represented by x. The data is scaled to the interval [0, 1]. The public representation is:
[0122]
[0123] The machine learning training module uses a reinforcement learning algorithm to generate an intelligent optimization model for water injection wells by analyzing and learning the historical data of water injection wells. The specific implementation method is as follows: Figure 6 This module is implemented through the following steps:
[0124] Step 1: Construct the state space, Among them i represents the temperature of the i-th layer underground, x i represents the water nozzle opening of the i-th layer in the well, h i represents the depth of the i-th layer in the well, p bi represents the pressure of the i-th layer in the well, q i represents the flow rate of the i-th layer in the well, p o Indicates the wellhead pressure, te o represents the wellhead temperature, x o Indicates the wellhead valve opening, q tai represents the target injection volume of the i-th layer.
[0125] Step 2: Action space, define the action space according to the operation requirements of the water injection well where t i represents the valve movement time unit s, t of the i-th layer i ∈[-30,30], when t i >0, the valve rotates in the direction of increasing opening. i When <0, the valve rotates in the direction of decreasing opening.
[0126] Step 3: Setting the reward function, giving the water spout movement direction reward based on the target injection amount and the actual input amount. The movement reward is expressed as:
[0127]
[0128] In order to reduce the number of steps to adjust the faucet, a reward is given for the number of times the faucet is adjusted. k represents the number of times the faucet is adjusted, and the reward for the number of adjustment steps is expressed as:
[0129] R2=-0.5×k
[0130] The injection volume error is used to measure the accuracy of the water nozzle, and the error reward is expressed as:
[0131]
[0132] The total reward is expressed as the movement reward, and the adjustment step reward and error reward are expressed as follows:
[0133] R=R1+R2+R3
[0134] Step 4: Reinforcement learning training, the model is trained by optimizing the PPO reinforcement learning algorithm through the proximal strategy. The machine learning training module uses the historical data and current real-time data preprocessed by the data acquisition and preprocessing module to iteratively update the model parameters. The training steps are 500 steps and the stopping condition is that the error is less than 5%. The maximum deployment per round is 100 steps. The number of neurons in the hidden layer is 128×3, the number of neurons in the input layer is 8×n+1, and the number of neurons in the output layer is n.
[0135] The terminal control module 133 generates remote well stopping and sealing detection instructions according to the following path.
[0136] The remote well shutdown implementation process is as follows: Figure 7 This is achieved by following the steps shown below:
[0137] Step 1: The water injection well management user issues a well stop command through the stratified water injection remote control client 42 .
[0138] Step 2: After receiving the command, the measurement and adjustment module 33 sends a well stop command to the edge computing host 13 through the DTU network 5.
[0139] Step 3: Close the wellhead sensor valve 123 to a degree of opening of 0.
[0140] Step 4: Close the water injection device valve sensor 114 and set the opening to 0.
[0141] The sealing detection implementation process is as follows Figure 8 This is achieved by following the steps shown below:
[0142] Step 1: The water injection well management user issues a sealing detection command through the stratified water injection remote control client 42.
[0143] Step 2: After receiving the command, the detection and adjustment module 33 sends a sealing detection command to the edge computing host 13 through the DTU network 5.
[0144] Step 3: Close the valve sensor 114 of the odd-numbered layer water injection device.
[0145] Step 4: The wellhead valve 123 cyclically executes the valve opening and closing commands.
[0146] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. It can be understood by a person of ordinary skill in the art that various modifications can be made in form and details without departing from the spirit and scope of the present invention defined in the appended claims.
Claims
1. A DTU-based stratified water injection real-time data monitoring and management system, including a data layer, a service layer and a management layer, characterized in that: The system also includes an access layer and a DTU network; The access layer includes a water injection device, a wellhead sensor and an edge computing host; the water injection device is arranged underground and is used to inject water into a designated oil layer; The wellhead sensor includes a wellhead pressure sensor, a wellhead temperature sensor and a wellhead flow sensor; the wellhead sensor is arranged at the wellhead and is used to display the wellhead pressure, temperature and flow data so that the operator can perceive the displayed data, and the operator can adjust the flow and pressure of the water injection well through the wellhead valve according to the on-site working conditions of the oil field; The edge computing host is connected to the water injection device and the wellhead sensor respectively; the edge computing host includes a machine learning training module, a data processing module and a terminal control module; The data processing module is used to collect and process the data transmitted by the water injection device and the wellhead sensor; The machine learning training module has computing capabilities, is responsible for the training and calculation of the water injection well intelligent algorithm, and can calculate the data transmitted by the data processing module; The terminal control module has a terminal interface for operating the internal system of the edge computing host; The terminal control module can generate remote well stop, real-time water injection data trajectory tracking and sealing detection instructions, which are pushed by the edge host to the web layered water injection remote monitoring client in the management layer through the DTU network; The data layer is provided with a stratified water injection database, and the stratified water injection database includes a stratified water injection relational database and a stratified water injection real-time database; The hierarchical water injection relational database is used to clean, process, associate, integrate and classify the information of factories, mines, wells, teams and users; the hierarchical water injection relational database stores static data of water injection wells, which includes the mining area where the well is located, the well location coordinates, the number of well layers and the injection volume; The layered water injection relation database can exchange data with the data analysis module, and the data exchanged includes water injection well layer information, location information and well affiliation information; The data analysis module can extract the basic data of water injection wells in the layered water injection relational database for statistical analysis; The stratified water injection real-time database can be used to implement the following functions: The water injection well data obtained by the water injection device and the wellhead sensor are sent to the DTU network via the edge computing host, and then sent to the stratified water injection real-time database for storage through the DTU network. The stored data include wellhead pressure, temperature and flow data; the stratified water injection real-time database can exchange data with the data analysis module, and the exchanged data include historical data of temperature and flow, and historical data of the valve sensor of the water injection device; the data analysis module can perform statistical analysis on the data in the stratified water injection real-time database; The service layer includes a data analysis module, a water injection real-time data module and a measurement and adjustment module; The data analysis module is used to extract well basic information from the stratified water injection relational database in the stratified water injection database, and output effective data analysis results in combination with historical data in the stratified water injection real-time database, wherein the effective data analysis results include historical statistical analysis of water injection well flow, historical statistical analysis of cumulative water injection volume, historical statistical analysis of water injection well pressure parameters, historical statistical analysis of water injection well temperature parameters, historical statistical analysis of water injection well nozzle opening data, historical statistical analysis of wellhead flow data, historical statistical analysis of wellhead pressure data, and historical statistical analysis of wellhead valve opening data; the effective data analysis results are provided to the operator via the web stratified water injection remote monitoring client in the management layer; The water injection real-time data module can be used to achieve the following functions: The edge computing host transmits the sensor parameters collected from the water injection device and the wellhead sensor to the water injection real-time data module through the DTU network, and the water injection real-time data module transmits the collected data to the web layered water injection remote management client and the layered water injection remote control client in real time; The measurement and adjustment module is used to remotely control the valve sensor of the water injection device and the wellhead valve. The measurement and adjustment module sends the water injection device valve sensor control command and the wellhead valve control command to the edge computing host based on the DTU network to adjust the downhole and wellhead flow rates; The management layer includes a web layered water injection remote management client and a layered water injection remote control client; The web layered water injection remote management client exchanges data with the data analysis module and the water injection real-time data module respectively; The web-based layered water injection remote management client realizes real-time analysis of water injection well data and real-time monitoring of water injection wells through a data analysis module and a water injection real-time data module; The stratified water injection remote control client realizes the real-time monitoring and real-time measurement and adjustment functions of water injection wells through the real-time water injection data module and the measurement and adjustment module, and then sends the feedback results to the stratified water injection database through the DTU network for storage; The DTU network is used to implement data communication between the access layer, the data layer and the management layer.
2. According to a DTU-based stratified water injection real-time data monitoring and management system according to claim 1, it is characterized by: The water injection device comprises a water injection device pressure sensor, a water injection device flow sensor, a water injection device temperature sensor and a water injection device valve sensor, which are respectively used to detect the pressure, flow and temperature during the oil layer water injection process.
3. A DTU-based stratified water injection real-time data monitoring and management system according to claim 2, characterized in that: The edge computing host is configured to adopt a polling mode and a real-time measurement and adjustment mode; In the polling mode, The edge computing host reads data from the water injection device pressure sensor, the water injection device flow sensor, the water injection device temperature sensor, the water injection device valve sensor, the wellhead pressure sensor, the wellhead flow sensor and the wellhead valve at a determined time interval; The edge computing host collects water injection well data at a determined frequency and sends it to the layered water injection database through the DTU network for storage. The amount of data collected is selected by the user according to the type of oil well. In the real-time measurement and adjustment mode, The edge computing host processes the collected water injection device signals and wellhead sensor signals and sends them to the stratified water injection real-time database, stratified water injection relational database, web stratified water injection remote management client and stratified water injection remote control client respectively; The web layered water injection remote management client and the layered water injection remote control client monitor and control the water injection wells.
4. A DTU-based stratified water injection real-time data monitoring and management system according to claim 3, characterized in that: The web stratified water injection remote management client and stratified water injection remote control client can receive remote well stop, real-time water injection data trajectory tracking, and sealing detection instructions generated by the terminal control module in the edge computing host sent through the DTU network for monitoring and control, and the feedback results are then delivered to the stratified water injection database through the edge computing host and the DTU network for storage; After receiving the water injection data transmitted by the edge computing host, the stratified water injection real-time database stores the real-time data of the water injection wells; The data analysis module extracts the required storage data or real-time data for analysis through the layered water injection database.
5. A DTU-based stratified water injection real-time data monitoring and management system according to claim 4, characterized in that: The data processing module in the edge computing host processes the historical data of the water injection device flow sensor, water injection device temperature sensor, water injection device valve sensor, wellhead pressure sensor, wellhead flow sensor, and wellhead valve in the stratified water injection real-time database in the following manner, specifically: Step 1: Data collection and preprocessing, obtaining corresponding historical data from the stratified water injection real-time database; Step 2: Use box plot to detect outliers, sort the data from large to small, Q1 represents the data point at the 25% position, Q2 represents the data point at the 50% position, Q3 represents the data point at the 75% position, IQR = Q3-Q1; determine the outlier boundary, the upper boundary is Q1-1.5×IQR, the lower boundary is Q3+1.5×IQR, and the data outside this range is determined as an outlier and deleted; Step 3: Convert data of different dimensions to the same dimension. The historical data is represented by x. The data is scaled to the interval [0,1]. The public representation is: The machine learning training module in the edge computing host uses a reinforcement learning algorithm to generate an intelligent optimization model for water injection wells by analyzing and learning the historical data of water injection wells. This is specifically achieved through the following steps: Step 1: Construct the state space, Among them i represents the temperature of the i-th layer underground, x i represents the water nozzle opening of the i-th layer in the well, h i represents the depth of the i-th layer in the well, p bi represents the pressure of the i-th layer in the well, q i represents the flow rate of the i-th layer in the well, p o Indicates the wellhead pressure, te o represents the wellhead temperature, x o Indicates the wellhead valve opening, q tai represents the target injection volume of the i-th layer; Step 2: Action space, define the action space according to the operation requirements of the water injection well where t i The time unit s, t represents the valve movement time of the i-th layer i ∈[-30,30], when t i >0, the valve rotates in the direction of increasing opening. i When <0, the valve rotates in the direction of decreasing opening; Step 3: Setting the reward function, giving the water spout movement direction reward based on the target injection amount and the actual input amount. The movement reward is expressed as: Reduce the number of faucet adjustment steps and give rewards based on the number of faucet adjustments. k represents the number of faucet adjustments, and the adjustment step reward is expressed as: R2=-0.5×k The injection volume error is used to measure the accuracy of the water nozzle, and the error reward is expressed as: The total reward is expressed as the movement reward, and the adjustment step reward and error reward are expressed as follows: R = R1 + R2 + R3; Step 4: Reinforcement learning training: train the model through the proximal policy optimization (PPO) reinforcement learning algorithm; The machine learning training module uses the historical data preprocessed by the data acquisition and preprocessing module and the current real-time data to iteratively update the model parameters. The number of training steps is 500 and the stopping condition is that the error is less than 5%. The maximum deployment per round is 100 steps. The number of hidden layer neurons is 128×3, the number of input layer neurons is 8×n+1, and the number of output layer neurons is n.
6. A DTU-based stratified water injection real-time data monitoring and management system according to claim 5, characterized in that: The terminal control module generates remote well stopping and sealing detection instructions according to the following path: The remote well stopping instruction is implemented by the following steps: Step 1: The water injection well management user issues a well stop command through the stratified water injection remote control client; Step 2: After receiving the command, the measurement and adjustment module sends a well stop command to the edge computing host through the DTU network; Step 3: Close the wellhead sensor valve to 0 opening; Step 4: Close the water injection device valve sensor and set the opening to 0; The seal detection instruction is implemented by the following steps: Step 1: The water injection well management user issues a sealing detection command through the stratified water injection remote control client; Step 2: After receiving the command, the test and adjustment module sends a sealing detection command to the edge computing host through the DTU network; Step 3: Close the valve sensor of the water injection device of the odd-numbered layers; Step 4: The wellhead valve cycle executes the valve opening and closing commands.
