Single-tank oil drawing well risk early warning method based on Internet of Things

Abstract

The invention discloses a single-tank oil drawing well risk early warning method based on the Internet of Things, which comprises the steps of data acquisition, instrument diagnosis, well state diagnosis, yield evaluation and remaining time judgment, and is characterized in that working condition abnormity and parameter change of the well state are analyzed through a data analysis technology, mechanism modeling and data-driven modeling are adopted, and a single-tank oil drawing well risk early warning result is obtained. Real-time data of an oil well are collected, well-tank linkage is achieved, abnormal events are early warned, oil pulling time is predicted in real time based on a production state analysis result, production scheduling is helped to master field conditions, and the pulling efficiency is improved. And meanwhile, the transportation progress is monitored, transportation and oil stealing events are distinguished, and the conditions of tank overflow and production halt are avoided.

Description

Technical Field

[0001] This invention belongs to the field of oil and gas production diagnosis, specifically involving a risk early warning method for single-tank oil wells based on the Internet of Things. Background Technology

[0002] In many low-permeability blocks of my country's oilfields, there are numerous non-gathering and transportation wells. These wells typically employ a single-well tank for oil collection and periodic tanker truck hauling of oil. Each well is equipped with a known-capacity storage tank, and several large storage tanks are located at the unloading point. Tanker trucks travel between the wells and unloading points from a garage to transport crude oil. Single-tank hauling wells generally have low production and are geographically dispersed and remote. With the development of the Internet of Things (IoT) in oilfields, single-tank hauling wells have achieved remote monitoring and unmanned operation, collecting various production data, including wellhead pressure and tank level. Some oilfields and institutions have begun researching production hauling scheduling systems for single-tank hauling wells to improve production efficiency. However, anomaly warnings for these wells still rely primarily on manual analysis, which suffers from delays, poor effectiveness, and high personnel requirements. The main problems include: Oil theft incidents still occur; Production status relies on manual analysis; Delayed hauling poses a risk of tank overflow and production shutdown. Using thresholds generates a large number of invalid alarms.

[0003] There are few relevant technologies and solutions in the industry. A search of publicly published professional journals and national patents revealed very few patents and documents related to this achievement, mainly including the following typical technologies: Patent 1: CN202010263016 Large-scale non-gathering and transportation well group production and transportation scheduling collaborative optimization system and method This invention relates to a large-scale non-gathering and transportation well cluster production and transportation scheduling collaborative optimization system, comprising: a production capacity acquisition module for collecting production capacity parameters of individual tankers, an onboard monitoring module for collecting operating parameters of tank trucks, and a dispatch cloud platform located in the oilfield command center; the dispatch cloud platform is communicatively connected to both the production capacity acquisition module and the onboard monitoring module, enabling it to output a transportation plan based on production capacity parameters and / or operating parameters for tank trucks to transport crude oil from geographically dispersed individual tankers to unloading points; the onboard monitoring module is connected to the unloading module located in the well area via a near-field communication module. This invention enables safe and reliable operation of tank trucks and controllable production from individual tankers.

[0004] Patent 2: CN202010263017 A Non-Gathering and Transportation Well Group Transportation and Dispatch System and Method This invention relates to a non-gathering and transportation oil well cluster transportation scheduling system, comprising: a monitoring module on oil well storage tanks in different geographical locations, used to collect at least one production capacity parameter of the oil well storage tanks and intermittently send at least one production capacity parameter to a data service module; a data service module: which connects oil well storage tanks in different geographical locations into an IoT network according to the method of acquiring production capacity parameters, and outputs a transportation plan for oil tank trucks to transport crude oil from oil well storage tanks in different geographical locations to unloading points, and can dynamically update the transportation plan according to at least one production capacity parameter and send it to an on-board module during the execution of the transportation plan by the oil tank trucks; and an on-board module, which is installed on the oil tank trucks and communicates with the data service module, so that the data service module can dynamically update the transportation plan based on the driving parameters of the oil tank trucks during the execution of the transportation plan and feed it back to the on-board module.

[0005] Paper 3: Research on Energy Saving and Consumption Reduction Technology for Single-Well Gathering and Transportation under Information Technology Conditions To address the current issues of high energy consumption and low data accuracy in single-well operations in oilfields, an information-based upgrade was implemented for these wells, leveraging the production information platform. This resulted in the construction of a comprehensive operation and management platform integrating monitoring, evaluation, and assessment. This platform allows for real-time tracking, evaluation, and optimization of single-well heating periods, durations, temperature rise quality, and power consumption. The total investment cost was approximately 415,600 yuan. After the information-based upgrade, 46 workers were saved, totaling 7.08 million yuan in labor costs. Annual electricity savings for on-duty personnel reached 270,000 kWh, and annual electricity savings for tank heating reached 308,000 kWh, resulting in annual electricity cost savings of 224,000 yuan. The energy-saving and consumption-reducing effects are significant.

[0006] Paper 4: Exploration and Innovation of Remote Management Mode for Single-Well Pulling The above four methods all focus on scheduling and collaborative analysis algorithms and energy consumption optimization. For anomalies in the production process and oil theft issues at single-tank oil wells, threshold alarms are still used. This method requires manual analysis and has problems such as difficulty in location and large alarm volume. Summary of the Invention The purpose of this invention is to provide a risk early warning method for single-tank oil wells based on the Internet of Things, so as to construct a production status analysis model for single-tank oil wells, realize well-tank linkage, provide early warning of abnormal events, help production scheduling control the on-site situation, and improve transportation efficiency.

[0007] The objective of this invention is achieved through the following technical means: a risk early warning method for single-tank oil wells based on the Internet of Things, comprising the following steps: Data acquisition involves collecting real-time data from oil wells, as well as basic equipment information. Instrument diagnostics: Performs instrument diagnostics based on real-time data and issues alarms when abnormal data is detected by the instrument. Well status diagnosis: Diagnoses the operating conditions of oil wells based on real-time data and issues an alarm when an abnormality occurs; For production assessment, determine the calculation period t weeks, retrieve liquid level data from two periods and one period, and calculate the upward slopes k1 and k2 respectively. If K1*80%≤K2≤K1*120%, or K2 is 0, the predicted slope k value is K1; otherwise, it is K2. The remaining time is determined by the formula: t_remaining = (L_safe_level - L_current_level) / k, where L_safe_level is the safe level of the storage tank and L_current_level is the current level of the storage tank. An alarm will be triggered when the time t_vehicle arrives at the oil well exceeds t_remaining. The real-time data includes oil well operating status, electrical parameters, force parameters, wellhead pressure, tank liquid level, and acquisition frequency. Basic equipment information includes the bottom area of ​​the large tank, the coordinates of the storage tank, and the information of the oil truck.

[0008] The specific method for instrument diagnostics is as follows: If the liquid level remains constant within the set time, it indicates that the liquid level gauge is stuck. If data collection exceeds a preset ratio within a set time period, it indicates a communication problem. If a single parameter is dead, negative, or exceeds its range, while other parameters are normal, it indicates that the parameter with the dead, negative, or excessive range is abnormal. When electrical parameters, force parameters, pressure, and tank data are all dead values, it indicates that the RTU is stuck.

[0009] In the well condition diagnosis, the specific abnormal manifestations of each operating condition are as follows: The belt breaks, and at the same time, the number of strokes during well operation is 0, the average A-phase current in the last 30 minutes is less than 85% of the average A-phase current in the previous 135 minutes, and the fluctuation is reduced, and the load tends to be linear; and when there is fluctuation, the load span drops significantly and is lower than the normal value. The pump simultaneously meets the trends of increasing average current, increasing current fluctuation, increasing average load, and increasing difference between maximum and minimum load; When the pump is in contact with the oil well, the maximum load remains unchanged, but the minimum load drops significantly or falls below the limit. When the bare rod breaks off, the current increases, and both the maximum and minimum loads decrease significantly. When the load drops below the minimum load, the fluctuations decrease significantly and tend to be a straight line. The stroke rate is normal and does not drop to 0. Abnormal shutdown, electrical parameters: ABC current >= 0, A-phase current <= equipment current threshold, meter reading > 0, voltage >= equipment voltage threshold. Force parameters: The load fluctuates but tends to be a straight line, the number of strokes changes from greater than 0 to equal to 0, and the stroke remains unchanged; Motor phase loss, current phase loss: one or two phases of current are abnormal, while other currents are normal; Voltage phase loss; one or two phases exhibit abnormal current and voltage, while the rest are normal. If the tank pipeline is blocked, within a time period T, the oil pressure and back pressure will increase by A%, but the tank level will not rise.

[0010] When the current liquid level in the storage tank drops, the oil-pulling status is also judged. If the distance between the oil-pulling truck and the storage tank is within the error range, it is considered normal oil pulling, and the remaining time judgment is paused. If the distance between the oil truck and the storage tank is outside the error range, an oil theft alarm will be triggered.

[0011] d is the distance between the oil truck and the storage tank, θ1 and λ1 are the longitude and latitude of the well site, θ2 and λ2 are the longitude and latitude of the oil truck, and r is 6371.

[0012] Based on the abnormal operating conditions, an abnormal operating condition diagnostic model is established to perform instrument diagnostics.

[0013] Based on abnormal operating conditions, a well condition diagnostic model is established to perform well condition diagnosis.

[0014] The beneficial effects of this invention are as follows: Through data analysis technology, it analyzes abnormal operating conditions and changes in well parameters; employs mechanistic modeling and data-driven modeling; collects real-time well data; achieves well-tank linkage; provides early warning of abnormal events; and, based on production status analysis results, predicts oil hauling time in real time, helping production scheduling to control the on-site situation and improve hauling efficiency. Simultaneously, it monitors hauling progress, distinguishes between hauling and oil theft incidents, and prevents tank overflows and production shutdowns. Attached Figure Description

[0015] Figure 1 A flowchart of a single-tank oil well risk early warning method based on the Internet of Things; The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Detailed Implementation

[0016]

Example 1

[0017] Instrument diagnostics: Performs instrument diagnostics based on real-time data and issues alarms when abnormal data is detected by the instrument. The specific method for instrument diagnostics is as follows: If the liquid level remains constant within the set time, it indicates that the liquid level gauge is stuck. If data collection exceeds a preset ratio within a set time period, it indicates a communication problem. If a single parameter is dead, negative, or exceeds its range, while other parameters are normal, it indicates that the dead, negative, or over-range parameter is abnormal. Since some parameters are on a single measuring instrument, and multiple instruments transmit data to the monitoring room through a single RTU device, if a certain parameter is dead, negative, or exceeds its range, but its associated parameters are normal, it indicates that the aforementioned parameter has a problem.

[0018] If the electrical, force, pressure, and main tank data are all dead values, it indicates that the RTU is stuck. If all parameters corresponding to the RTU are dead values, it indicates that the corresponding RTU is experiencing a stuck condition.

[0019] The system uses real-time data to comprehensively determine whether the instrument data is normal. If an anomaly is detected, the data becomes unavailable, and an alarm is triggered. The main issues include: The level gauge is stuck, and the level value remains constant for a certain period of time. Communication issues; a certain percentage of data collection is abnormal within a certain timeframe. Single parameter anomaly: dead value, negative value, or over-range; other parameters normal. The RTU is stuck; electrical parameters, force parameters, pressure, and tank data are all dead values.

[0020] The specific duration of the anomaly can be determined by selecting 1-2 acquisition cycles as the judgment standard based on the acquisition frequency of the well site. For example, if the level gauge is stuck, the level value remains constant within one consecutive acquisition cycle.

[0021] Based on the abnormal operating conditions, an abnormal operating condition diagnostic model is established to perform instrument diagnostics.

[0022] Well status diagnosis: Diagnoses the operating conditions of oil wells based on real-time data and issues an alarm when an abnormality occurs; In the well condition diagnosis, the specific abnormal manifestations of each operating condition are as follows: The belt breaks, and at the same time, the number of strokes during well operation is 0, the average A-phase current in the last 30 minutes is less than 85% of the average A-phase current in the previous 135 minutes, and the fluctuation is reduced, and the load tends to be linear; and when there is fluctuation, the load span drops significantly and is lower than the normal value. The pump simultaneously meets the trends of increasing average current, increasing current fluctuation, increasing average load, and increasing difference between maximum and minimum load; When the pump is in operation and the oil well is running, the maximum load remains unchanged, but the minimum load drops significantly or falls below the limit; specifically, it drops by 50% or falls below 50% of the theoretical value.

[0023] When the bare rod breaks off, the current increases, and both the maximum and minimum loads decrease significantly. When the load drops below the minimum load, the fluctuations decrease significantly and tend to be a straight line. The stroke rate is normal and does not drop to 0. Abnormal shutdown, electrical parameters: ABC current >= 0, A-phase current <= equipment current threshold, meter reading > 0, voltage >= equipment voltage threshold. The current threshold varies for different equipment, and the current threshold is usually set to 0.5, indicating that the motor stops.

[0024] Force parameters: The load fluctuates but tends to be a straight line, the number of strokes changes from greater than 0 to equal to 0, and the stroke remains unchanged; Motor phase loss, current phase loss: one or two phases of current are abnormal, while other currents are normal; Voltage phase loss; one or two phases exhibit abnormal current and voltage, while the rest are normal. If the tank pipeline is blocked, within time T, the oil pressure and back pressure will increase by A%, but the tank level will not rise. A% is obtained based on historical well site data; for example, based on historical experience, it can be set to 30%.

[0025] Example: The oil pressure of a certain well increased by 30% within 5 minutes compared to the maximum value in the previous 30 minutes.

[0026] Based on abnormal operating conditions, a well condition diagnostic model is established for well condition diagnosis. A multi-parameter diagnostic model is built using data collected from oil wells and storage tanks to determine the on-site production status. When applying this method, appropriate parameters can be selected for modeling based on the actual on-site conditions.

[0027] Common anomalies are shown in the table below:

[0028] For production assessment, determine the calculation period t weeks, retrieve liquid level data from two periods and one period, and calculate the upward slopes k1 and k2 respectively. If K1*80%≤K2≤K1*120%, or K2 is 0, the predicted slope k value is K1; otherwise, it is K2. The remaining time is determined by the formula: t_remaining = (L_safe_level - L_current_level) / k, where L_safe_level is the safe level of the storage tank and L_current_level is the current level of the storage tank. An alarm will be triggered when the time t_vehicle arrives at the oil well exceeds t_remaining. The calculation cycle can be selected as 30 minutes, depending on the oil well's production capacity. Retrieve liquid level data from the past 1 hour and the past 30 minutes (depending on production capacity and tank size), and calculate the upward slopes k1 and K2 respectively.

[0029] k1 = L current level - L1, where L1 is the level 1 hour ago; k2 = (L current level - L2) * 2, where L2 is the level 30 minutes ago. When K1*80%≤K2≤K1*120%, or K2 is 0, the predicted slope k is K1; otherwise, it is K2.

[0030] Calculate the remaining oil hauling time: t_remaining = (L_safe level - L_current level) / k. When the time it takes for the oil truck to arrive at the well, t_truck > t_remaining, trigger an alarm.

[0031] The status of oil transportation and oil theft is determined based on the GPS coordinates of the oil truck.

[0032] When the current liquid level in the storage tank drops, the oil-pulling status is also judged. If the distance between the oil-pulling truck and the storage tank is within the error range, it is considered normal oil pulling, and the remaining time judgment is paused. If the distance between the oil truck and the storage tank is outside the error range, an oil theft alarm will be triggered.

[0033] Specifically, the distance calculation formula is as follows: d is the distance between the oil truck and the storage tank, θ1 and λ1 are the longitude and latitude of the well site, θ2 and λ2 are the longitude and latitude of the oil truck, and r is 6371.

[0034] The following explanation will be based on a single-tank oil well at a plant in Xinjiang Oilfield.

[0035] (1) Data collection The Xinjiang Oilfield has a high coverage rate of IoT infrastructure, so it directly calls the real-time database interface or connects to field instruments to complete the real-time data acquisition. The main parameters to be acquired include:

[0036] Basic information is obtained through the equipment management department, stored in a relational database, and retrieved during calculations, including: Tank bottom area, maximum liquid level in the tank, oil well coordinates, and information on oil transport vehicles. (2) Instrument diagnostics Based on actual well site data, an instrument diagnostic model was constructed to perform real-time diagnostics and monitor instrument status. JavaScript functions were used for this purpose. Level gauge stuck: Liquid level remains constant for 30 minutes, pumping unit operates normally. Communication error: Communication failed within 1 hour Abnormal parameters: Electrical parameters are within the normal range but not constant; force parameters are abnormal or constant. Force parameters are within the normal range but not constant; electrical parameters are abnormal or constant. Electrical and force parameters are within normal values, but wellhead pressure is negative or constant. RTU stuck: Electrical parameters, force parameters, pressure, and main tank data remain constant for 20 minutes. During normal production at a certain well, communication was suddenly interrupted for more than 1 hour, which triggered an alarm and was then manually repaired.

[0037] The abnormal working condition modeling is the same as before, taking belt breakage as an example.

[0038] The well model shows that the maximum current value in the most recent 30 minutes has decreased by more than 30% compared to the previous maximum value, the range is less than 2, and the minimum value is greater than 1.

[0039] After the alarm was triggered, on-site verification confirmed that the well belt had broken.

[0040] This paper uses a normal producing oil well as an example to illustrate the prediction of oil extraction time and the judgment of its status.

[0041] ① During production, the well is monitored through the aforementioned steps to determine if the well is operating normally, and a level prediction is performed every 30 minutes.

[0042] ②At 13:00, the liquid level is 1.52m. Retrieve the liquid level data for the past 1 hour and the liquid level data for the past 30 minutes, and calculate the upward slope k1 and K2 respectively.

[0043] k1 = L current liquid level - L1, where L1 is the liquid level 1 hour ago. k1 = L - L1 = 1.52 - 1.50 = 0.02 k2 = (L current level - L2) * 2, where L2 is the level 30 minutes ago. K2=(L-L2)*2=(1.52-1.51)*2=0.02 Then the value of k is 0.02. ③ Calculate the remaining oil-pulling time t_remaining = (L_safe_liquid_level - L_current_liquid_level) / k The safe liquid level in the well is 2.6m, therefore t = (2.6 - 1.52) / 0.02 = 54h The latest arrival time, T_vehicle, is calculated as follows: T_current time + t = April 4th, 13:00 + 54 hours = April 6th, 19:00. This time is dynamically updated. At this time, 54 hours is much shorter than the fastest travel time from the garage to the well site (30 minutes), so no alarm is triggered. On April 6, a drop in the liquid level was detected. At this time, the GPS coordinates of the associated oil truck were retrieved, and the straight-line distance between the storage tank and the oil truck was calculated according to the coordinate formula.

[0044] The calculated value is d=0.3km. At this point, the oil spill is considered normal and no alarm will be triggered. The predicted time for stopping oil spill is then determined.

[0045] The beneficial effects of the proposed solution are: improved operational efficiency: using automatic diagnosis to replace manual analysis significantly reduces the amount of manual work and improves employee work efficiency. It can reduce the daily workload of employees in relevant positions by more than 2 hours. Taking a million-ton-level oilfield as an example, it can save at least one person and reduce costs by 250,000 yuan per year.

[0046] Reduce operating costs: By predicting refueling time, the phenomenon of vehicles running empty can be effectively reduced, which helps to improve dispatching efficiency and reduce vehicle operating costs. It is estimated that vehicle maintenance and fuel costs can be reduced by 100,000 yuan per year.

[0047] Reduce production losses: Timely detection of abnormal well shutdowns and oil theft incidents can prevent tank overflows and production stoppages, thereby reducing production losses and saving oil production plants approximately 210,000 yuan annually.

Claims

1. A risk early warning method for single-tank oil wells based on the Internet of Things, characterized in that: Includes the following steps, Data acquisition involves collecting real-time data from oil wells, as well as basic equipment information. Instrument diagnostics: Performs instrument diagnostics based on real-time data and issues alarms when abnormal data is detected by the instrument. Well status diagnosis: Diagnoses the operating conditions of oil wells based on real-time data and issues an alarm when an abnormality occurs; For production assessment, determine the calculation period t weeks, retrieve liquid level data from two periods and one period, and calculate the upward slopes k1 and k2 respectively. If K1*80%≤K2≤K1*120%, or K2 is 0, the predicted slope k value is K1; otherwise, it is K2. The remaining time is determined by the formula: t_remaining = (L_safe_level - L_current_level) / k, where L_safe_level is the safe level of the storage tank and L_current_level is the current level of the storage tank. An alarm will be triggered when the time t_vehicle arrives at the oil well exceeds t_remaining.

2. The method for risk early warning of a single-tank oil well based on the Internet of Things as described in claim 1, characterized in that: The real-time data includes oil well operating status, electrical parameters, force parameters, wellhead pressure, tank liquid level, and acquisition frequency. Basic equipment information includes the bottom area of ​​the large tank, the coordinates of the storage tank, and the information of the oil truck.

3. The method for risk early warning of a single-tank oil well based on the Internet of Things as described in claim 1, characterized in that: The specific method for instrument diagnostics is as follows: If the liquid level remains constant within the set time, it indicates that the liquid level gauge is stuck. If data collection exceeds a preset ratio within a set time period, it indicates a communication problem. If a single parameter is dead, negative, or exceeds its range, while other parameters are normal, it indicates that the parameter with the dead, negative, or excessive range is abnormal. When electrical parameters, force parameters, pressure, and tank data are all dead values, it indicates that the RTU is stuck.

4. The method for risk early warning of a single-tank oil well based on the Internet of Things as described in claim 1, characterized in that: In the well condition diagnosis, the specific abnormal manifestations of each operating condition are as follows: The belt breaks, and at the same time, the number of strokes during well operation is 0, the average A-phase current in the last 30 minutes is less than 85% of the average A-phase current in the previous 135 minutes, and the fluctuation is reduced, and the load tends to be linear; and when there is fluctuation, the load span drops significantly and is lower than the normal value. The pump simultaneously meets the trends of increasing average current, increasing current fluctuation, increasing average load, and increasing difference between maximum and minimum load; When the pump is in contact with the oil well, the maximum load remains unchanged, but the minimum load drops significantly or falls below the limit. When the bare rod breaks off, the current increases, and both the maximum and minimum loads decrease significantly. When the load drops below the minimum load, the fluctuations decrease significantly and tend to be a straight line. The stroke rate is normal and does not drop to 0. Abnormal shutdown, electrical parameters: ABC current >= 0, A-phase current <= equipment current threshold, meter reading > 0, voltage >= equipment voltage threshold. Force parameters: The load fluctuates but tends to be a straight line, the number of strokes changes from greater than 0 to equal to 0, and the stroke remains unchanged; Motor phase loss, current phase loss: one or two phases of current are abnormal, while other currents are normal; Voltage phase loss; one or two phases exhibit abnormal current and voltage, while the rest are normal. If the inlet pipeline is blocked, within a time period T, the oil pressure and back pressure will increase by A%, but the tank level will not rise.

5. The method for risk early warning of a single-tank oil well based on the Internet of Things as described in claim 1, characterized in that: When the current liquid level in the storage tank drops, the oil-pulling status is also judged. If the distance between the oil-pulling truck and the storage tank is within the error range, it is considered normal oil pulling, and the remaining time judgment is paused. If the distance between the oil truck and the storage tank is outside the error range, an oil theft alarm will be triggered.

6. The method for risk early warning of a single-tank oil well based on the Internet of Things as described in claim 5, characterized in that: d is the distance between the oil truck and the storage tank, θ1 and λ1 are the longitude and latitude of the well site, θ2 and λ2 are the longitude and latitude of the oil truck, and r is 6371.

7. The method for risk early warning of a single-tank oil well based on the Internet of Things as described in claim 3, characterized in that: Based on the abnormal operating conditions, an abnormal operating condition diagnostic model is established to perform instrument diagnostics.

8. A method for risk early warning of a single-tank oil well based on the Internet of Things as described in claim 4, characterized in that: Based on abnormal operating conditions, a well condition diagnostic model is established to perform well condition diagnosis.

Images