Weighing type flow metering device

Abstract

According to the weighing type flow metering device, liquid flow is metered in a weighing mode and a volume mode, double metering is achieved through the device, the accuracy of data is remarkably improved, and interference of gas-liquid mixing on a traditional flow meter is effectively avoided. Especially under the conditions of low oil extraction flow velocity and small flow, a traditional flow meter often generates large errors due to flow velocity fluctuation or gas mixing, and the device can ensure the metering stability and accuracy. In addition, a closed tank body design is adopted, gas and liquid are separated through an internal gas-liquid separation device, and gas pressure in the tank is guided back to a pipeline again by utilizing a gas internal circulation mechanism. By means of the design, the gas discharging problem in the liquid discharging process is solved, environmental pollution caused by direct gas discharging is effectively prevented, and meanwhile the safety risk caused by overpressure in the tank body is reduced.

Description

Technical Field

[0001] The invention belongs to the field of oil field production, and in particular relates to a weighing type flow metering device. Background Art

[0002] At present, the oilfield production industry has significant deficiencies in single well monitoring, especially in my country's oilfield environment, this problem is particularly prominent. Since the crude oil content in oil wells is generally low, the gas-liquid mixing ratio is high, and the oil production flow rate is small, the existing flowmeters are difficult to provide accurate metering data under such complex working conditions. Traditional flowmeters are susceptible to gas interference under gas-liquid mixing conditions, resulting in large errors in liquid flow measurement. In addition, although multi-well combined metering can obtain overall production data, it masks the differences between single wells, making it difficult for managers to grasp the specific production performance of each well. This model not only fails to identify inefficient wells or abnormal wells in a timely manner, affecting the optimization of single well production, but may also lead to irrational resource allocation, further exacerbating ineffective production and waste of resources.

[0003] In view of this, the present invention is proposed. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art and provide a weighing flow metering device to solve the problems raised in the above-mentioned background technology.

[0005] In order to solve the above technical problems, the basic concept of the technical solution adopted by the present invention is:

[0006] A weighing flow metering device comprises: at least two oil storage tank units, each of which comprises an oil storage tank, a bracket with a weighing function, an automatic pressure relief valve, a liquid level detection device and a connecting pipe;

[0007] The two oil storage tanks are divided into tank A and tank B. Tank A and tank B are connected in parallel through pipelines. Automatic pressure relief valves are provided on the tops of tanks A and B to release the gas pressure in the tanks. Tank A and tank B are respectively installed on brackets with weighing functions to measure the weight of the liquid in the tanks in real time. A liquid level detection device is installed in tank A and tank B to monitor the liquid level height in the tanks. Tank A and tank B are connected by an air pipe to balance the pressure between the two tanks.

[0008] A weighing flow metering device also includes a liquid inlet branch pipe, the liquid outlet of the liquid inlet branch pipe is provided with a three-way valve, the liquid outlets at the upper and lower ends of the three-way valve are connected to liquid outlet pipes, and the liquid outlets of the two liquid outlet pipes are respectively connected to the liquid inlets of tank A and tank B, so as to control the liquid to flow into tank A or tank B.

[0009] A weighing flow metering device also includes a liquid outlet branch pipe. The liquid outlets of tank A and tank B are both connected to the discharge pipe. Solenoid valve A and solenoid valve B are respectively arranged at the two discharge pipes. The liquid outlets of solenoid valve A and solenoid valve B are both connected to the liquid outlet branch pipe.

[0010] Optionally, a connecting branch pipe is provided at the end of the liquid inlet of the liquid inlet branch pipe, a solenoid valve C is provided at the connecting branch pipe, and the liquid outlet of the connecting branch pipe is connected to the liquid outlet of the liquid outlet branch pipe.

[0011] Optionally, the liquid level detection device is in contact with the liquid inside the tank.

[0012] Optionally, the gas in tank B enters tank A through the air pipe, and the liquid in tank A is discharged through the solenoid valve. When the three-way valve is switched to allow tank A to inlet liquid, the gas in tank A enters tank B through the air pipe, and the liquid in tank B is discharged through the solenoid valve.

[0013] Optionally, the liquid level data and the weighing data are simultaneously transmitted to a data processor to calculate the volume of the liquid in the tank and to calculate the liquid flow rate per unit time in combination with the weight data, the steps are as follows:

[0014] Acquire multi-dimensional data in real time from liquid level detection devices, weight measuring devices and external environmental sensors;

[0015] According to the characteristics of traffic data, a prediction model is selected. After the model is selected, the data after feature extraction and time windowing is input into the model for training, so that the model can learn the historical laws and trends of traffic changes;

[0016] The prediction performance of the model is evaluated through cross-validation, the prediction error is calculated, the model parameters are adjusted according to the verification results, and the latest real-time data is input into the trained model to predict the traffic trend for a period of time in the future. At the same time, the model will be continuously updated to adapt to the dynamic changes of traffic data, which is expressed as: in, is the predicted flow value at the future time t+h, which is predicted based on the data at the current time t. f(·) is the function of the main prediction model, which is responsible for making basic flow prediction based on the current input and parameters. t is the input feature set at the current time t, θ t is the parameter set of the model at time t, α is the compensation coefficient, which is used to adjust the impact of historical prediction errors on the current prediction. is the average prediction error of cross-validation, calculated based on the past k samples, reflecting the performance of the model on the validation set. β is a dynamic adjustment coefficient used to balance the impact of new data on model updates. ΔX t =X t -X t-1is the amount of change in feature data, g(·) is an online learning function that uses feature changes and the parameter θ at the previous moment t-1 , dynamically adjust model parameters;

[0017] The predicted flow trend is presented to the operator in the form of a chart or alarm, providing intuitive flow change information. According to the predicted results, the opening and closing strategies of the three-way valve and solenoid valve are dynamically adjusted to optimize the oil production process. When abnormal flow is predicted, the alarm mechanism is triggered.

[0018] Optionally, the prediction model has an input layer, a hidden layer, and an output layer, the number of neural units in the input layer is 5, the number of hidden units in the hidden layer is 100, and the number of neural units in the output layer is 1;

[0019] After feature extraction, the generated key feature data is organized according to time windows, and the time-windowed feature data is divided into a training set and a validation set. The training set is used for model parameter learning, and the validation set is used to evaluate model performance. According to the selected LSTM model and the LSTM model structure is initialized, the time-windowed input data is gradually passed to the LSTM model to enable it to learn the historical traffic change rules.

[0020] Continue to train the LSTM model until the prediction error of the model tends to be stable, that is, the preset convergence condition is reached. At this time, the model can better reflect the historical trend of traffic changes. Save the trained model for subsequent real-time prediction.

[0021] Optionally, the data processor receives the prediction results of future flow trends from the time series prediction model and analyzes the predicted data. If the prediction shows that the flow will reach or approach the set abnormal threshold at a certain time in the future, the device immediately evaluates the flow status and determines whether valve adjustment or alarm operation is required.

[0022] After adopting the above technical solution, the present invention has the following beneficial effects compared with the prior art. Of course, any product implementing the present invention does not necessarily need to achieve all the advantages described below at the same time:

[0023] The present invention measures the liquid flow rate by weighing and volume. The device realizes dual measurement, which not only significantly improves the accuracy of the data, but also effectively avoids the interference of gas-liquid mixing on the traditional flow meter. Especially in the case of low oil production flow rate and small flow rate, the traditional flow meter often produces large errors due to flow rate fluctuations or gas mixing, while the present device can ensure the stability and accuracy of measurement. In addition, a closed tank design is adopted, and the gas and liquid are separated by an internal gas-liquid separation device, and the gas pressure in the tank is redirected back to the pipeline by using the gas internal circulation mechanism. This design not only solves the problem of gas discharge during liquid discharge, but also effectively prevents environmental pollution caused by direct gas discharge, while reducing the safety risks caused by overpressure inside the tank. The device is particularly suitable for scenarios with small single-well oil production flow rates. It can accurately count the oil production of each oil well, provide reliable data support for production management, thereby improving overall production efficiency and avoiding ineffective production and waste of resources.

[0024] The specific implementation modes of the present invention are further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings described below are only some embodiments. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0026] In the figure:

[0027] Figure 1 It is a three-dimensional diagram of the flow metering device;

[0028] Figure 2 It is a front view of the flow metering device;

[0029] Figure 3 It is a side view of the flow metering device.

[0030] In the accompanying drawings, the components represented by the reference numerals are listed as follows:

[0031] 1. Oil storage tank; 2. Bracket with weighing function; 3. Automatic pressure relief valve; 4. Liquid level detection device; 5. Connecting pipe; 6. Liquid inlet branch pipe; 7. Three-way valve; 8. Liquid outlet pipe; 9. Liquid drain pipe; 10. Solenoid valve A; 11. Solenoid valve B; 12. Liquid outlet branch pipe; 13. Connecting branch pipe; 14. Solenoid valve C.

[0032] It should be noted that these drawings and textual descriptions are not intended to limit the conceptual scope of the present invention in any way, but are intended to illustrate the concept of the present invention for those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0033] The present invention will now be described in further detail with reference to the accompanying drawings.

[0034] See also Figure 1-3 As shown, in this embodiment, a weighing flow metering device is provided, comprising at least two oil storage tank units, each of which comprises an oil storage tank 1, a bracket 2 with a weighing function, an automatic pressure relief valve 3, a liquid level detection device 4 and a pipe 5;

[0035] The two oil storage tanks 1 are divided into tank A and tank B, which are connected in parallel through pipelines. The tops of tanks A and B are both provided with automatic pressure relief valves 3 for releasing the gas pressure in the tanks. Tanks A and B are respectively installed on brackets 2 with weighing functions for real-time measurement of the weight of the liquid in the tanks. Liquid level detection devices 4 are installed in tanks A and B for monitoring the liquid level in the tanks. Tanks A and B are connected through air pipes for balancing the pressure between the two tanks. The brackets 2 with weighing functions are weighed using weighing scales.

[0036] A weighing flow metering device also includes a liquid inlet branch pipe 6, the liquid outlet of the liquid inlet branch pipe 6 is provided with a three-way valve 7, the liquid outlets at the upper and lower ends of the three-way valve 7 are connected to liquid outlet pipes 8, and the liquid outlets of the two liquid outlet pipes 8 are respectively connected to the liquid inlets of tank A and tank B, which are used to control the liquid to flow into tank A or tank B. By setting the three-way valve 7, the liquid flow into tank A or tank B can be flexibly controlled to achieve alternating operation of the two tanks, thereby ensuring the continuity and efficiency of the oil production process. The configuration of the three-way valve 7 enables the device to flexibly allocate the liquid inlet path according to the actual flow demand, thereby avoiding flow interruption or inaccurate data caused by the operation of a single tank.

[0037] A weighing flow metering device also includes a liquid outlet branch pipe 12, the liquid outlets of tank A and tank B are both connected to a liquid discharge pipe 9, and the two liquid discharge pipes 9 are respectively provided with a solenoid valve A10 and a solenoid valve B11, and the liquid outlets of the solenoid valve A10 and the solenoid valve B11 are both connected to the liquid outlet branch pipe 12. The solenoid valve A10 and the solenoid valve B11 are provided on the liquid outlet pipe 8 of each tank, so that the liquid outlet path can be accurately controlled. This design not only helps to optimize the liquid outlet process, but also can adjust the discharge amount and flow rate of the liquid through the solenoid valve when switching between different tank bodies, thereby improving the flexibility and adaptability of the overall device.

[0038] In this embodiment, a connecting branch pipe 13 is provided at the end of the liquid inlet of the liquid inlet branch pipe 6, and a solenoid valve C14 is provided at the connecting branch pipe 13, and the liquid outlet of the connecting branch pipe 13 is connected to the liquid outlet of the liquid outlet branch pipe 12. The liquid flow path is switched by the connecting branch pipe 13 and the solenoid valve C14 to ensure the continuity of the operation of the device. In addition, this design provides more operation options under special working conditions, and enhances the safety and fault tolerance of the device.

[0039] In this embodiment, the liquid level detection device 4 is in contact with the liquid inside the tank. The liquid level detection device 4 directly contacts the liquid and can monitor the liquid level in the tank in real time and accurately. This configuration can be combined with the data of the weighing device to provide volume and quality information of the tank liquid, ensure multi-dimensional monitoring of flow data, and help improve the accuracy of measurement.

[0040] In this embodiment, the gas in tank B enters tank A through the air pipe, and the liquid in tank A is discharged through the solenoid valve. When the three-way valve 7 is switched to tank A to enter liquid, the gas in tank A enters tank B through the air pipe, and the liquid in tank B is discharged through the solenoid valve. The dynamic balance of gas is achieved through the air pipe and the solenoid valve. When one tank body enters liquid, the other tank body uses air pressure to discharge liquid. This design avoids the flow measurement error caused by gas-liquid mixing, realizes gas circulation in the closed device, reduces the impact of external emissions on the environment, and improves the environmental protection and safety of the device.

[0041] In actual application, when the oil pumped out of the oil well enters the pipeline Figure 1-3 At the liquid inlet shown, the liquid is controlled by the three-way valve 7 to flow into tank B first, and tank A is closed. The two tanks are connected in parallel in the pipeline. The pipeline where the solenoid valve C is located belongs to the emergency bypass and is normally closed. At this time, the solenoid valve C is closed and the solenoid valve A is opened. After the liquid enters tank B, the gas in the tank enters tank A through the air pipe and enters the liquid outlet through the solenoid valve A. After tank B reaches the set weight or height, the three-way valve 7 switches to liquid flowing into tank A, and tank B is closed. At this time, the solenoid valve A is closed and the solenoid valve B is opened. The gas pressure generated by tank A after the liquid enters tank B through the air pipe, thereby pressing the liquid in tank B into the pipeline through the compressed gas entering through the solenoid valve B. At the same time, every time the tank weight or volume reaches the set value, the pressure sensor and the liquid level measurement sensor respectively transmit the monitoring data to the data processor for data statistics and storage, so as to calculate the flow rate of the liquid in the pipeline per unit time through the weight weighed by the two tanks and the data monitored by the liquid level sensor per unit time.

[0042] In this embodiment, the liquid level data and the weighing data are simultaneously transmitted to the data processor, the liquid volume in the tank is calculated and the liquid flow rate per unit time is counted in combination with the weight data, and the steps are as follows:

[0043] Acquire multi-dimensional data in real time from liquid level detection devices, weight measuring devices and external environmental sensors;

[0044] According to the characteristics of traffic data, a prediction model is selected. After the model is selected, the data after feature extraction and time windowing is input into the model for training, so that the model can learn the historical laws and trends of traffic changes;

[0045] The prediction performance of the model is evaluated through cross-validation, the prediction error is calculated, the model parameters are adjusted according to the verification results, and the latest real-time data is input into the trained model to predict the traffic trend for a period of time in the future. At the same time, the model will be continuously updated to adapt to the dynamic changes of traffic data, which is expressed as: in, is the predicted flow value at the future time t+h, which is predicted based on the data at the current time t. f(·) is the function of the main prediction model, which is responsible for making basic flow prediction based on the current input and parameters. t is the input feature set at the current time t, θ t is the parameter set of the model at time t, α is the compensation coefficient, which is used to adjust the impact of historical prediction errors on the current prediction. is the average prediction error of cross-validation, calculated based on the past k samples, reflecting the performance of the model on the validation set. β is a dynamic adjustment coefficient used to balance the impact of new data on model updates. ΔX t =X t -X t-1 is the amount of change in feature data, g(·) is an online learning function that uses feature changes and the parameter θ at the previous moment t-1 , dynamically adjust model parameters; evaluate model performance through cross-validation and error calculation, and dynamically adjust model parameters based on the validation results. This design ensures that the model is continuously optimized during the training process and improves the accuracy of predictions. Input the latest data in real time and update the model in a rolling manner so that it can always adapt to the dynamic changes in traffic data. This process clearly reflects the closed-loop mechanism of the prediction process, from data input to model optimization to prediction output, ensuring the reliability of the prediction results.

[0046] It should be noted that the formula takes into account both the compensation of historical errors and the rolling optimization of the model by real-time data, providing more accurate and stable predictions. t , dynamically updates model parameters, improving the model's adaptability to sudden traffic changes. In addition, the α and β coefficients adjust the weights of historical errors and real-time updates respectively, achieving a balanced optimization of long-term trends and short-term fluctuations.

[0047] The predicted flow trend is presented to the operator in the form of a chart or alarm, providing intuitive flow change information. According to the predicted results, the opening and closing strategies of the three-way valve and solenoid valve are dynamically adjusted to optimize the oil production process. When abnormal flow is predicted, the alarm mechanism is triggered.

[0048] In this embodiment, the prediction model has an input layer, a hidden layer, and an output layer, the number of neural units in the input layer is 5, the number of hidden units in the hidden layer is 100, and the number of neural units in the output layer is 1;

[0049] After feature extraction, the generated key feature data is organized according to time windows, and the time-windowed feature data is divided into a training set and a validation set. The training set is used for model parameter learning, and the validation set is used to evaluate model performance. According to the selected LSTM model and the LSTM model structure is initialized, the time-windowed input data is gradually passed to the LSTM model to enable it to learn the historical traffic change rules.

[0050] Continue to train the LSTM model until the prediction error of the model tends to be stable, that is, the preset convergence condition is reached. At this time, the model can better reflect the historical trend of traffic changes. Save the trained model for subsequent real-time prediction.

[0051] In this embodiment, the data processor receives the prediction results of future flow trends from the time series prediction model and analyzes the prediction data. If the prediction shows that the flow will reach or approach the set abnormal threshold at a certain time in the future, the device immediately evaluates the flow state and determines whether valve adjustment or alarm operation is required.

[0052] The present invention is not limited to the above-mentioned embodiments. Anyone should be aware that any structural changes made under the enlightenment of the present invention, and any technical solutions that are the same or similar to the present invention, fall within the protection scope of the present invention. The technology, shape, and structural parts not described in detail in the present invention are all well-known technologies.

Claims

1. A weighing flow metering device, characterized in that: It comprises at least two oil storage tank units, each of which comprises an oil storage tank (1), a bracket with a weighing function (2), an automatic pressure relief valve (3), a liquid level detection device (4) and a connecting pipe (5); The two oil storage tanks (1) are divided into tank A and tank B, which are connected in parallel via a pipeline. Tanks A and B are each provided with an automatic pressure relief valve (3) on the top to release the gas pressure in the tank. Tanks A and B are respectively mounted on a bracket (2) with a weighing function to measure the weight of the liquid in the tank in real time. A liquid level detection device (4) is installed in tank A and tank B to monitor the height of the liquid level in the tank. Tanks A and B are connected via an air pipe to balance the pressure between the two tanks.

2. A weighing flow metering device according to claim 1, characterized in that: The invention also comprises a liquid inlet branch pipe (6), the liquid outlet of the liquid inlet branch pipe (6) is provided with a three-way valve (7), the liquid outlets at the upper and lower ends of the three-way valve (7) are both connected to liquid outlet pipes (8), and the liquid outlets of the two liquid outlet pipes (8) are respectively connected to the liquid inlets of tank A and tank B, so as to control the liquid to flow into tank A or tank B.

3. A weighing flow metering device according to claim 1, characterized in that: The invention also comprises a liquid outlet branch pipe (12), the liquid outlets of tank A and tank B are both connected to a liquid outlet pipe (9), the two liquid outlet pipes (9) are respectively provided with a solenoid valve A (10) and a solenoid valve B (11), and the liquid outlets of the solenoid valve A (10) and the solenoid valve B (11) are both connected to the liquid outlet branch pipe (12).

4. A weighing flow metering device according to claim 1, characterized in that: A connecting branch pipe (13) is provided at the end of the liquid inlet of the liquid inlet branch pipe (6), a solenoid valve C (14) is provided at the connecting branch pipe (13), and the liquid outlet of the connecting branch pipe (13) is connected to the liquid outlet of the liquid outlet branch pipe (12).

5. A weighing flow metering device according to claim 1, characterized in that: The liquid level detection device (4) is in contact with the liquid inside the tank.

6. A weighing flow metering device according to claim 1, characterized in that: The gas in tank B enters tank A through the air pipe and discharges the liquid in tank A through the solenoid valve. When the three-way valve (7) is switched to tank A to allow liquid to flow in, the gas in tank A enters tank B through the air pipe and discharges the liquid in tank B through the solenoid valve.

7. A weighing flow metering device according to claim 1, characterized in that: The liquid level data and weighing data are simultaneously transmitted to the data processor to calculate the volume of the liquid in the tank and combine the weight data to count the liquid flow rate per unit time. The steps are as follows: Acquire multi-dimensional data in real time from liquid level detection devices, weight measuring devices and external environmental sensors; According to the characteristics of traffic data, a prediction model is selected. After the model is selected, the data after feature extraction and time windowing is input into the model for training, so that the model can learn the historical laws and trends of traffic changes; The prediction performance of the model is evaluated through cross-validation, the prediction error is calculated, the model parameters are adjusted according to the verification results, and the latest real-time data is input into the trained model to predict the traffic trend for a period of time in the future. At the same time, the model will be continuously updated to adapt to the dynamic changes of traffic data, which is expressed as: in, is the predicted flow value at the future time t+h, based on the data at the current time t. f(·) is the function of the main prediction model, which is responsible for making basic flow prediction based on the current input and parameters. t is the input feature set at the current time t, θ t is the parameter set of the model at time t, α is the compensation coefficient, which is used to adjust the impact of historical prediction errors on the current prediction. is the average prediction error of cross-validation, calculated based on the past k samples, reflecting the performance of the model on the validation set. β is the dynamic adjustment coefficient used to balance the impact of new data on model updates. ΔX t =X t -X t-1 is the amount of change in feature data, g(·) is an online learning function that uses feature changes and the parameter θ at the previous moment t-1 , dynamically adjust model parameters; The predicted flow trend is presented to the operator in the form of a chart or alarm, providing intuitive flow change information. According to the predicted results, the opening and closing strategies of the three-way valve and solenoid valve are dynamically adjusted to optimize the oil production process. When abnormal flow is predicted, the alarm mechanism is triggered.

8. A weighing flow metering device according to claim 7, characterized in that: The prediction model has an input layer, a hidden layer, and an output layer. The number of neural units in the input layer is 5, the number of hidden units in the hidden layer is 100, and the number of neural units in the output layer is 1; After feature extraction, the generated key feature data is organized according to time windows, and the time-windowed feature data is divided into a training set and a validation set. The training set is used for model parameter learning, and the validation set is used to evaluate model performance. According to the selected LSTM model and the LSTM model structure is initialized, the time-windowed input data is gradually passed to the LSTM model to enable it to learn the historical traffic change rules. Continue to train the LSTM model until the prediction error of the model tends to be stable, that is, the preset convergence condition is reached. At this time, the model can better reflect the historical trend of traffic changes. Save the trained model for subsequent real-time prediction.

9. A weighing flow metering device according to claim 8, characterized in that: The data processor receives the forecast results of future flow trends from the time series forecast model and analyzes the forecast data. If the forecast shows that the flow will reach or approach the set abnormal threshold at a certain time in the future, the device immediately evaluates the flow status and determines whether valve adjustment or alarm operation is required.

Images