Flow measurement system based on pressure pulsation method

Abstract

The invention relates to the technical field of indirect measurement by means of physical properties, in particular to a flow measurement system based on a pressure pulsation method. The system comprises a data acquisition module which is installed at an outlet of a plunger pump and used for acquiring periodic pressure pulsation signals generated by the plunger pump in a fluid pipeline in real time; the signal processing module is used for receiving and processing the collected pressure pulse signals and extracting frequency characteristics and amplitude characteristics of pressure pulsation by using a characteristic extraction algorithm; and the calculation module is used for establishing a flow calculation model according to the pumping frequency, the liquid volume pumped each time, the pump efficiency, the pipeline resistance, the medium viscosity, the pipeline length and the pipeline diameter, and calculating the flow of the solid-liquid mixture. According to the flow measurement system based on the pressure pulsation method, the pressure pulsation method is utilized, the pumping characteristics of a plunger pump are combined, pressure signals are converted into flow signals by capturing and analyzing pressure pulsation signals, and therefore indirect measurement of the flow of a solid-liquid mixture is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of indirect measurement by means of physical properties, and in particular to a flow measurement system based on a pressure pulsation method. Background Art

[0002] In the prior art, for the flow measurement of solid-liquid mixtures, especially for highly viscous media such as asphalt concrete, traditional flow meters such as orifice flow meters, turbine flow meters, and electromagnetic flow meters cannot directly measure the flow of such mixtures, resulting in inaccurate or impossible measurements. In particular, when using a plunger pump for pumping, due to the high viscosity of the medium, traditional flow meters cannot effectively measure its flow.

[0003] For example, in the patent publication number CN118883622A, entitled "A thermal feedback phase content and flow measurement method and device", the method includes: obtaining the voltage, current and first temperature measurement value of the first thermometer; obtaining the heat transfer of the fluid according to the voltage and the current; obtaining the second temperature measurement value of the second thermometer; obtaining the local heat transfer coefficient according to the first temperature measurement value, the second temperature measurement value and the heat transfer; fitting the local heat transfer coefficient to obtain the relationship between the phase content and the flow; obtaining the phase content and the flow according to the relationship between the phase content and the flow. The disadvantage is that it indirectly measures the flow by means of the physical properties of thermal feedback, and its scope of application is limited to gas-liquid mixing. It cannot effectively measure solid-liquid mixtures such as asphalt and concrete that are common in the engineering industry. Summary of the invention

[0004] In view of the problem that flow meters in the prior art cannot effectively measure the flow rate of solid-liquid mixtures, the present invention provides a flow measurement system based on the pressure pulsation method. The pressure pulsation method is used in combination with the pumping characteristics of the plunger pump to capture and analyze the pressure pulsation signal, and convert the pressure signal into a flow signal, thereby realizing indirect measurement of the flow rate of the solid-liquid mixture.

[0005] In order to achieve the above technical purpose, the present invention provides a technical solution, which is a flow measurement system based on a pressure pulsation method, comprising: A data acquisition module is installed at the outlet of the plunger pump and is used to collect the periodic pressure pulsation signal generated by the plunger pump in the fluid pipeline in real time; The signal processing module is connected to the data acquisition module and is used to receive and process the collected pressure pulse signal and extract the frequency and amplitude characteristics of the pressure pulsation using a feature extraction algorithm; The calculation module establishes a flow calculation model based on the pumping frequency, the volume of liquid pumped each time, the efficiency of the pump, the pipeline resistance, the medium viscosity, the pipeline length and the pipeline diameter, and calculates the flow rate of the solid-liquid mixture; The calibration feedback module is used to calibrate the flow calculation results output by the calculation module. It uses a machine learning algorithm to train and optimize based on historical measurement data and actual flow data, and provides feedback when the flow is abnormal.

[0006] In this technical solution, by integrating the pumping characteristics of the plunger pump and the signal processing technology, efficient and indirect measurement of the flow rate of the solid-liquid mixture is achieved. The data acquisition module is installed at the outlet of the plunger pump to capture the periodic pressure pulsation signal generated by the pump in real time. Subsequently, these original signals are transmitted to the signal processing module, which is not only responsible for receiving the signal, but also uses the feature extraction algorithm to deeply explore the frequency and amplitude characteristics of the pressure pulsation. At the same time, the built-in filtering unit effectively eliminates noise interference to ensure the purity and accuracy of the signal quality. Based on signal processing, the calculation module uses a series of key parameters, including pumping frequency, single pumping liquid volume, pump efficiency, pipeline resistance, medium viscosity, pipeline length and diameter, to construct a flow calculation model. The flow calculation model can accurately convert the pressure signal into intuitive flow data, providing solid technical support for the quantitative analysis of the flow rate of solid-liquid mixtures. The system's built-in calibration feedback module introduces machine learning algorithms to conduct deep learning and optimization of historical measurement data and actual flow data, which not only greatly improves the accuracy of flow calculations, but also can respond quickly when flow abnormalities occur and provide instant feedback. This not only ensures the continued reliability of measurement results, but also reflects the system's flexibility and intelligence in dealing with complex working conditions.

[0007] The present invention is further configured as follows: the calculation formula of the flow rate is: Q final =η×f×V / (1+α×R+β×μ+γ×L / D); Among them, η is the actual working efficiency of the plunger pump, f is the pumping frequency, V is the volume of liquid pumped each time, R is the pipeline resistance, μ is the medium viscosity, L is the pipeline length, D is the pipeline diameter, and α, β, and γ are empirical coefficients.

[0008] This formula fully considers the key factors in fluid dynamics, such as pipeline resistance, medium viscosity and pipeline geometry. By taking these factors into consideration, the formula can more accurately calculate the actual flow rate of solid-liquid mixtures, providing strong technical support for industrial production and scientific research experiments.

[0009] The present invention is further configured as follows: the system also includes a power and resistance analysis module, which is used to analyze the fluid flow condition by using the one-way check characteristic of the plunger pump under the condition that the power is less than the resistance; The power and resistance analysis module includes: The pressure accumulation analysis unit is used to monitor and analyze the pressure increase and release process in the pipeline after each plunger pumping, and identify the impact of the pressure accumulation effect on the fluid flow; The one-way check feature utilizes the one-way check valve of the plunger pump to ensure the one-way flow of the fluid during the pumping process, and maintains the slow flow of the fluid in the pipeline through pressure accumulation during the pumping interval; The flow resistance dynamic evaluation unit monitors and evaluates the flow resistance encountered by the fluid in the pipeline in real time, and dynamically adjusts the value of the pipeline resistance in combination with the pressure accumulation effect, and dynamically adjusts the prediction model and flow calculation model for the fluid fluidity.

[0010] In this technical solution, the core function of the power and resistance analysis module is that when the system detects that the power is less than the resistance, it can use the one-way check characteristic of the plunger pump to analyze the fluid flow. The pressure accumulation analysis unit is responsible for finely monitoring and analyzing the change process of the pressure in the pipeline after each plunger pumping, including the pressure increase and subsequent release. By deeply analyzing the pressure accumulation effect, the unit can accurately identify its specific impact on the fluid flow, providing an important reference for the accuracy of flow calculation. The one-way check characteristic utilization unit cleverly uses the one-way check valve built into the plunger pump to ensure that the fluid always maintains a one-way flow during the pumping process, effectively avoiding the backflow phenomenon. During the pumping interval, the unit also maintains the slow but stable flow of the fluid in the pipeline by reasonably utilizing the pressure accumulation effect, thereby ensuring the continuity of the flow measurement. The flow resistance dynamic evaluation unit has the ability to monitor the flow resistance encountered by the fluid in the pipeline in real time, and can flexibly adjust the prediction model of the fluid flow according to the dynamic changes of the pressure accumulation effect. This dynamic evaluation mechanism ensures the adaptability and accuracy of the flow calculation model under different working conditions.

[0011] The present invention is further configured as follows: the pressure accumulation analysis unit comprises: By continuously monitoring the pressure changes in the pipeline, the pressure pulse caused by each pumping of the plunger pump is identified, and the pressure accumulation after the pressure pulse is calculated; The relationship between pressure accumulation and time is analyzed, and a pressure accumulation model is established to predict the slow flow state of the fluid maintained by pressure accumulation when the power is less than the resistance.

[0012] In this technical solution, the core of the pressure accumulation analysis unit lies in its ability to continuously monitor the pressure changes in the pipeline. It can keenly identify the pressure pulses caused by each pumping action of the plunger pump and accurately calculate the pressure accumulation after these pulses. This process requires not only high sensitivity, but also advanced algorithm support to ensure the accuracy and real-time nature of the data. More importantly, the unit is not only satisfied with the simple recording of pressure changes, but further analyzes the dynamic relationship between pressure accumulation and time. Through this analysis, the unit can establish an accurate pressure accumulation model. This model plays a vital role in predicting the flow state of the fluid under the condition that the power is less than the resistance. It can simulate the slow flow state of the fluid maintained by the pressure accumulation, which provides strong support for the continuity and accuracy of flow measurement.

[0013] The present invention is further configured as follows: the one-way non-return characteristic utilizing unit comprises: After each pumping of the plunger pump, the one-way check valve automatically closes to prevent the fluid from flowing back and ensure that the pressure in the pipeline accumulates gradually; During the pumping interval, the accumulated pressure is used to push the fluid to flow slowly in the pipeline until the next pumping starts, forming a continuous but non-uniform fluid delivery process.

[0014] In this technical solution, the one-way check characteristic utilization unit can automatically close the one-way check valve after each plunger pumping, effectively preventing the backflow of the fluid, ensuring that the pressure in the pipeline can gradually accumulate, and providing the necessary power support for subsequent fluid transportation. During the pumping interval, the one-way check characteristic utilization unit cleverly uses the accumulated pressure as a driving force to keep the fluid in the pipeline in a slow but stable flow state. This process not only avoids stagnation and sedimentation of the fluid in the pipeline, but also ensures the continuity of flow measurement. At the same time, since the fluid transportation process is non-uniform, the design of the unit also fully considers the complexity of the actual working conditions and improves the accuracy and reliability of flow measurement.

[0015] The present invention is further configured as follows: the flow resistance dynamic evaluation unit comprises: Real-time monitoring of the flow rate, pressure and temperature parameters of the fluid in the pipeline, combined with the fluid properties and pipeline material, dynamically calculate the current flow resistance; The dynamically evaluated flow resistance value is fed back to the flow calculation module, and the pipeline resistance term in the flow calculation formula is adjusted in real time to improve the accuracy of flow calculation.

[0016] In this technical solution, firstly, the dynamic evaluation unit of flow resistance can monitor the key parameters such as flow velocity, pressure and temperature of the fluid in the pipeline in real time. These data are important bases for evaluating flow resistance. Through high-precision sensors and real-time data acquisition technology, the accuracy and real-time nature of the data are ensured. Secondly, the dynamic evaluation unit of flow resistance combines the physical properties of the fluid and the material of the pipeline, and uses advanced calculation models and algorithms to dynamically calculate the current flow resistance value. This process fully considers the interaction between the fluid and the pipeline, as well as the influence of environmental factors on the flow resistance, making the evaluation results closer to the actual situation. More importantly, the dynamic evaluation unit of flow resistance will immediately feed back the real-time evaluated flow resistance value to the flow calculation module, so that the resistance term in the flow calculation formula can be dynamically adjusted as the flow resistance changes, thereby effectively improving the accuracy of flow calculation. Whether it is under complex working conditions such as changes in fluid properties, pipeline wear or ambient temperature fluctuations, the unit can ensure the accuracy and stability of flow measurement.

[0017] The present invention is further configured as follows: the establishment of the flow calculation model includes: Based on the historical working data of the plunger pump, an initial prediction model is established according to the machine learning algorithm; During actual operation, new working condition data and corresponding flow measurement values ​​are continuously collected, and the prediction model is updated in real time based on the online learning algorithm; According to the updated prediction model output, the actual working efficiency and pumping frequency of the plunger pump are initialized.

[0018] In this technical solution, firstly, the system uses the historical working data of the plunger pump and combines it with advanced machine learning algorithms to establish an initial prediction model. It not only makes full use of the valuable information in the historical data, but also uses the powerful ability of machine learning to dig out the regularity and trend behind the data, providing a solid foundation for subsequent flow prediction. Secondly, during the actual operation process, the system continuously collects new working condition data and corresponding flow measurement values. These data reflect the actual performance of the system under different working conditions and are an important basis for optimizing the prediction model. The system uses an online learning algorithm to update and adjust the prediction model in real time, so that it can better adapt to the current working conditions and environmental changes. Finally, according to the updated model output, the system can automatically adjust the relevant parameters in the flow calculation model to ensure that the flow calculation model is always highly consistent with the actual working conditions, thereby greatly improving the accuracy of flow measurement. At the same time, the automatic adjustment of parameters also reduces the need for manual intervention and improves the automation and operation efficiency of the system.

[0019] The present invention is further configured as follows: the calibration feedback module also includes an anomaly detection unit, which constructs a feature space of normal flow based on a deep learning anomaly detection method, detects real-time flow data, and when flow data that deviates from the normal range is identified, it is determined to be abnormal and triggers flow anomaly feedback.

[0020] In this technical solution, firstly, the anomaly detection unit constructs the feature space of normal traffic using deep learning technology. This process involves deep analysis and learning of a large amount of historical traffic data, thereby extracting the inherent laws and characteristics of normal traffic data. This feature space provides a benchmark and reference for subsequent anomaly detection. Secondly, the anomaly detection unit continuously monitors the real-time traffic data. By comparing the difference between the real-time traffic data and the normal traffic feature space, the system can quickly identify the traffic data that deviates from the normal range. This process fully utilizes the advantages of deep learning algorithms in pattern recognition and anomaly detection, and realizes real-time and accurate monitoring of traffic data. When the system identifies abnormal traffic data, it will immediately trigger the traffic anomaly feedback mechanism. This mechanism includes sending alarms to operators, recording abnormal events for subsequent analysis, and possible automatic corrective measures. By responding to abnormal traffic events in a timely manner, the system can effectively prevent the occurrence of potential faults and ensure the continuity and accuracy of traffic measurement.

[0021] The present invention is further configured as follows: the machine learning algorithm comprises: Combining random forest and gradient boosted tree learners, the prediction performance of the prediction model is improved through weighted average method; Use the model pre-trained on different flow measurement tasks as a starting point and adapt it to the current task by fine-tuning the parameters; Extract representative features from raw traffic data through unsupervised learning; Generate synthetic data with similar distribution to real traffic data to generate adversarial networks.

[0022] In this technical solution, by combining random forest and gradient boosting tree learners and adopting a weighted average method, the prediction performance of the model is significantly improved, combining the advantages of two powerful machine learning algorithms, making the prediction results more accurate and reliable. Using the model pre-trained on different traffic measurement tasks as a starting point, the parameters are fine-tuned to adapt to the current task, which improves the generalization ability of the model, enabling it to quickly adapt to new traffic measurement tasks and reducing the time and cost of retraining. Representative features are extracted from raw traffic data through unsupervised learning. Unsupervised learning can automatically discover hidden structures and patterns in the data and extract features that are useful for prediction tasks, thereby improving the efficiency and accuracy of the model. Generate synthetic data that is similar to the distribution of real traffic data and enhance the data set in a generative adversarial network manner. This method increases the diversity of training samples by generating synthetic data similar to real data, which helps the model learn more comprehensive data features and improves the robustness and generalization ability of the model.

[0023] The present invention is further configured as follows: the feature extraction algorithm includes a Fourier transform algorithm and a peak detection algorithm, the Fourier transform algorithm is used to extract frequency features, and the peak detection algorithm is used to extract amplitude features; The Fourier transform algorithm includes: Preprocessing the collected pressure pulsation signal; According to the Fourier transform algorithm, the time domain signal is converted into a frequency domain signal; Extract key frequency components from frequency domain signals as pumping frequency input features in flow calculation models; The peak detection algorithm includes: obtaining the pressure change caused by each pumping, and calculating the liquid volume of each pumping.

[0024] In this technical solution, by clearly pointing out that the feature extraction algorithm includes the Fourier transform algorithm and the peak detection algorithm, the respective functions of the Fourier transform algorithm and the peak detection algorithm are elaborated in detail. The Fourier transform algorithm is used to extract frequency features and convert the time domain signal into the frequency domain signal to obtain the frequency components in the signal; the peak detection algorithm is used to extract amplitude features and focus on the peak information in the signal. This functional division makes the feature extraction more comprehensive and targeted. The Fourier transform algorithm is explained in detail, including the steps of preprocessing the pressure pulsation signal, converting the time domain signal into the frequency domain signal, and extracting the key frequency components. It can deeply understand the specific application and operation process of the Fourier transform algorithm in feature extraction. By pointing out that the key frequency components in the extracted frequency domain signal will be used as the input features of the flow calculation model, the importance of feature extraction in the flow measurement system is emphasized.

[0025] The beneficial effects of the present invention are as follows: (1) the pressure pulsation method is used in combination with the pumping characteristics of the plunger pump, and the pressure signal is converted into a flow signal by capturing and analyzing the pressure pulsation signal, thereby realizing indirect measurement of the flow rate of the solid-liquid mixture; (2) the data acquisition module is installed at the outlet of the plunger pump to capture the periodic pressure pulsation signal generated by the pump in real time. Subsequently, these original signals are transmitted to the signal processing module. The signal processing module is not only responsible for receiving the signal, but also uses the feature extraction algorithm to deeply explore the frequency and amplitude characteristics of the pressure pulsation. At the same time, the built-in filtering unit effectively eliminates noise interference to ensure the purity and accuracy of the signal quality; (3) based on signal processing, the calculation module uses a series of key parameters, including pumping frequency, single pumping liquid volume, pump efficiency, pipeline resistance, medium viscosity, pipeline length and diameter, to construct a flow calculation model. The flow calculation model can accurately convert the pressure signal into intuitive flow data, providing a solid technical support for the quantitative analysis of the flow rate of the solid-liquid mixture. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a structural schematic diagram of the flow measurement system based on the pressure pulsation method. DETAILED DESCRIPTION

[0027] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0028] like Figure 1 As shown, as the first embodiment of the present invention, a flow measurement system based on a pressure pulsation method includes: A data acquisition module is installed at the outlet of the plunger pump and is used to collect the periodic pressure pulsation signal generated by the plunger pump in the fluid pipeline in real time; The signal processing module is connected to the data acquisition module and is used to receive and process the collected pressure pulse signal and extract the frequency and amplitude characteristics of the pressure pulsation using a feature extraction algorithm; The calculation module establishes a flow calculation model based on the pumping frequency, the volume of liquid pumped each time, the efficiency of the pump, the pipeline resistance, the medium viscosity, the pipeline length and the pipeline diameter, and calculates the flow rate of the solid-liquid mixture; The calibration feedback module is used to calibrate the flow calculation results output by the calculation module. It uses a machine learning algorithm to train and optimize based on historical measurement data and actual flow data, and provides feedback when the flow is abnormal.

[0029] In this embodiment, by integrating the pumping characteristics of the plunger pump and the signal processing technology, efficient and indirect measurement of the flow rate of the solid-liquid mixture is achieved. The data acquisition module is installed at the outlet of the plunger pump to capture the periodic pressure pulsation signal generated by the pump in real time. Subsequently, these original signals are transmitted to the signal processing module, which is not only responsible for receiving the signal, but also uses the feature extraction algorithm to deeply explore the frequency and amplitude characteristics of the pressure pulsation. At the same time, the built-in filtering unit effectively eliminates noise interference to ensure the purity and accuracy of the signal quality. On the basis of signal processing, the calculation module uses a series of key parameters, including pumping frequency, single pumping liquid volume, pump efficiency, pipeline resistance, medium viscosity, pipeline length and diameter, to construct a flow calculation model. The flow calculation model can accurately convert the pressure signal into intuitive flow data, providing solid technical support for the quantitative analysis of the flow rate of solid-liquid mixtures. The system's built-in calibration feedback module introduces machine learning algorithms to conduct deep learning and optimization of historical measurement data and actual flow data, which not only greatly improves the accuracy of flow calculations, but also can respond quickly when flow abnormalities occur and provide instant feedback. This not only ensures the continued reliability of measurement results, but also reflects the system's flexibility and intelligence in dealing with complex working conditions.

[0030] It can be understood that the signal processing module includes a filtering unit for filtering the collected pressure pulse signal to remove noise interference and improve signal quality.

[0031] The flow rate is calculated as follows: Q final =η×f×V / (1+α×R+β×μ+γ×L / D); Among them, η is the actual working efficiency of the plunger pump, f is the pumping frequency, V is the volume of liquid pumped each time, R is the pipeline resistance, μ is the medium viscosity, L is the pipeline length, D is the pipeline diameter, and α, β, and γ are empirical coefficients.

[0032] This formula fully considers the key factors in fluid dynamics, such as pipeline resistance, medium viscosity and pipeline geometry. By taking these factors into consideration, the formula can more accurately calculate the actual flow rate of solid-liquid mixtures, providing strong technical support for industrial production and scientific research experiments.

[0033] It will be appreciated that the empirical coefficients α, β, γ and the pump efficiency η are determined through experiments or historical data.

[0034] In one embodiment of the present invention, the system further comprises a power and resistance analysis module for analyzing the fluid flow condition by using the one-way check characteristic of the plunger pump under the condition that the power is less than the resistance; The power and resistance analysis module includes: The pressure accumulation analysis unit is used to monitor and analyze the pressure increase and release process in the pipeline after each plunger pumping, and identify the impact of the pressure accumulation effect on the fluid flow; The one-way check feature utilizes the one-way check valve of the plunger pump to ensure the one-way flow of the fluid during the pumping process, and maintains the slow flow of the fluid in the pipeline through pressure accumulation during the pumping interval; The flow resistance dynamic evaluation unit monitors and evaluates the flow resistance encountered by the fluid in the pipeline in real time, and dynamically adjusts the value of the pipeline resistance in combination with the pressure accumulation effect, and dynamically adjusts the prediction model and flow calculation model for the fluid fluidity.

[0035] In this technical solution, the core function of the power and resistance analysis module is that when the system detects that the power is less than the resistance, it can use the one-way check characteristic of the plunger pump to analyze the fluid flow. The pressure accumulation analysis unit is responsible for finely monitoring and analyzing the change process of the pressure in the pipeline after each pumping of the plunger pump, including the pressure increase and subsequent release. By deeply analyzing the pressure accumulation effect, the unit can accurately identify its specific impact on the fluid flow, providing an important reference for the accuracy of flow calculation. The one-way check characteristic utilization unit cleverly uses the one-way check valve built into the plunger pump to ensure that the fluid always maintains a one-way flow during the pumping process, effectively avoiding the backflow phenomenon. During the pumping interval, the unit also maintains the slow but stable flow of the fluid in the pipeline by reasonably utilizing the pressure accumulation effect, thereby ensuring the continuity of the flow measurement. The flow resistance dynamic evaluation unit has the ability to monitor the flow resistance encountered by the fluid in the pipeline in real time, and can flexibly adjust the prediction model of the fluid flow according to the dynamic changes of the pressure accumulation effect. This dynamic evaluation mechanism ensures the adaptability and accuracy of the flow calculation model under different working conditions.

[0036] In one embodiment of the present invention, the pressure accumulation analysis unit comprises: By continuously monitoring the pressure changes in the pipeline, the pressure pulse caused by each pumping of the plunger pump is identified, and the pressure accumulation after the pressure pulse is calculated; The relationship between pressure accumulation and time is analyzed, and a pressure accumulation model is established to predict the slow flow state of the fluid maintained by pressure accumulation when the power is less than the resistance.

[0037] In this technical solution, the core of the pressure accumulation analysis unit lies in its ability to continuously monitor the pressure changes in the pipeline. It can keenly identify the pressure pulses caused by each pumping action of the plunger pump and accurately calculate the pressure accumulation after these pulses. This process requires not only high sensitivity, but also advanced algorithm support to ensure the accuracy and real-time nature of the data. More importantly, the unit is not only satisfied with the simple recording of pressure changes, but further analyzes the dynamic relationship between pressure accumulation and time. Through this analysis, the unit can establish an accurate pressure accumulation model. This model plays a vital role in predicting the flow state of the fluid under the condition that the power is less than the resistance. It can simulate the slow flow state of the fluid maintained by the pressure accumulation, which provides strong support for the continuity and accuracy of flow measurement.

[0038] In one embodiment of the present invention, the one-way non-return characteristic utilizing unit comprises: After each pumping of the plunger pump, the one-way check valve automatically closes to prevent the fluid from flowing back and ensure that the pressure in the pipeline accumulates gradually; During the pumping interval, the accumulated pressure is used to push the fluid to flow slowly in the pipeline until the next pumping starts, forming a continuous but non-uniform fluid delivery process.

[0039] In this technical solution, the one-way check characteristic utilization unit can automatically close the one-way check valve after each plunger pumping, effectively preventing the backflow of the fluid, ensuring that the pressure in the pipeline can gradually accumulate, and providing the necessary power support for subsequent fluid transportation. During the pumping interval, the one-way check characteristic utilization unit cleverly uses the accumulated pressure as a driving force to keep the fluid in the pipeline in a slow but stable flow state. This process not only avoids stagnation and sedimentation of the fluid in the pipeline, but also ensures the continuity of flow measurement. At the same time, since the fluid transportation process is non-uniform, the design of the unit also fully considers the complexity of the actual working conditions and improves the accuracy and reliability of flow measurement.

[0040] Furthermore, the flow resistance dynamic evaluation unit includes: Real-time monitoring of the flow rate, pressure and temperature parameters of the fluid in the pipeline, combined with the fluid properties and pipeline material, dynamically calculate the current flow resistance; The dynamically evaluated flow resistance value is fed back to the flow calculation module, and the pipeline resistance term in the flow calculation formula is adjusted in real time to improve the accuracy of flow calculation.

[0041] In this technical solution, firstly, the dynamic evaluation unit of flow resistance can monitor the key parameters such as flow velocity, pressure and temperature of the fluid in the pipeline in real time. These data are important bases for evaluating flow resistance. Through high-precision sensors and real-time data acquisition technology, the accuracy and real-time nature of the data are ensured. Secondly, the dynamic evaluation unit of flow resistance combines the physical properties of the fluid and the material of the pipeline, and uses advanced calculation models and algorithms to dynamically calculate the current flow resistance value. This process fully considers the interaction between the fluid and the pipeline, as well as the influence of environmental factors on the flow resistance, making the evaluation results closer to the actual situation. More importantly, the dynamic evaluation unit of flow resistance will immediately feed back the real-time evaluated flow resistance value to the flow calculation module, so that the resistance term in the flow calculation formula can be dynamically adjusted as the flow resistance changes, thereby effectively improving the accuracy of flow calculation. Whether it is under complex working conditions such as changes in fluid properties, pipeline wear or ambient temperature fluctuations, the unit can ensure the accuracy and stability of flow measurement.

[0042] In one embodiment of the present invention, the establishment of the flow calculation model includes: Based on the historical working data of the plunger pump, an initial prediction model is established according to the machine learning algorithm; During actual operation, new working condition data and corresponding flow measurement values ​​are continuously collected, and the prediction model is updated in real time based on the online learning algorithm; According to the updated prediction model output, the actual working efficiency and pumping frequency of the plunger pump are initialized.

[0043] In this technical solution, firstly, the system uses the historical working data of the plunger pump and combines it with advanced machine learning algorithms to establish an initial prediction model. It not only makes full use of the valuable information in the historical data, but also uses the powerful ability of machine learning to dig out the regularity and trend behind the data, providing a solid foundation for subsequent flow prediction. Secondly, during the actual operation process, the system continuously collects new working condition data and corresponding flow measurement values. These data reflect the actual performance of the system under different working conditions and are an important basis for optimizing the prediction model. The system uses an online learning algorithm to update and adjust the prediction model in real time, so that it can better adapt to the current working conditions and environmental changes. Finally, according to the updated model output, the system can automatically adjust the relevant parameters in the flow calculation model to ensure that the flow calculation model is always highly consistent with the actual working conditions, thereby greatly improving the accuracy of flow measurement. At the same time, the automatic adjustment of parameters also reduces the need for manual intervention and improves the automation and operation efficiency of the system.

[0044] It can be understood that the relevant parameters include pumping frequency, liquid volume pumped each time, pump efficiency, pipeline resistance, medium viscosity, pipeline length and pipeline diameter, and empirical coefficients.

[0045] It can be understood that the online learning algorithm can be a stochastic gradient descent algorithm, an online support vector machine algorithm, a recursive least squares algorithm, or a reinforcement learning algorithm.

[0046] The calibration feedback module also includes an anomaly detection unit, which constructs a feature space of normal flow based on a deep learning anomaly detection method, detects real-time flow data, and when flow data deviating from the normal range is identified, it is determined to be abnormal and triggers flow anomaly feedback.

[0047] In this technical solution, firstly, the anomaly detection unit constructs the feature space of normal traffic using deep learning technology. This process involves deep analysis and learning of a large amount of historical traffic data, thereby extracting the inherent laws and characteristics of normal traffic data. This feature space provides a benchmark and reference for subsequent anomaly detection. Secondly, the anomaly detection unit continuously monitors the real-time traffic data. By comparing the difference between the real-time traffic data and the normal traffic feature space, the system can quickly identify the traffic data that deviates from the normal range. This process fully utilizes the advantages of deep learning algorithms in pattern recognition and anomaly detection, and realizes real-time and accurate monitoring of traffic data. When the system identifies abnormal traffic data, it will immediately trigger the traffic anomaly feedback mechanism. This mechanism includes sending alarms to operators, recording abnormal events for subsequent analysis, and possible automatic corrective measures. By responding to abnormal traffic events in a timely manner, the system can effectively prevent the occurrence of potential faults and ensure the continuity and accuracy of traffic measurement.

[0048] It can be understood that the anomaly detection method includes: Use deep learning technology to learn from a large amount of historical normal traffic data; Through learning, the model extracts the features of normal traffic data and constructs a feature space representing normal traffic behavior, which includes frequency features, amplitude features, and time series patterns; When new real-time traffic data is generated, the system inputs it into the already trained deep learning model; The model extracts features from new data and compares the extracted features with the feature space of normal traffic; If the characteristics of real-time data deviate significantly from the characteristic space of normal traffic, that is, exceed the preset threshold or range, the model will determine that these data are abnormal, and the abnormality is manifested as a sudden increase or decrease in traffic, a significant change in frequency, and abnormal fluctuations in amplitude; If an abnormality is identified, the system immediately triggers a flow abnormality feedback mechanism, which includes sending an alarm message to an operator, automatically adjusting the parameters of the flow calculation model to adapt to the abnormal condition, or suspending flow measurement.

[0049] The machine learning algorithm includes: Combining random forest and gradient boosted tree learners, the prediction performance of the prediction model is improved through weighted average method; Use the model pre-trained on different flow measurement tasks as a starting point and adapt it to the current task by fine-tuning the parameters; Extract representative features from raw traffic data through unsupervised learning; Generate synthetic data with similar distribution to real traffic data to generate adversarial networks.

[0050] By combining the random forest and gradient boosting tree learners and using a weighted average method, the prediction performance of the model is significantly improved, combining the advantages of two powerful machine learning algorithms to make the prediction results more accurate and reliable. Using the pre-trained model on different traffic measurement tasks as a starting point, the generalization ability of the model is improved by fine-tuning the parameters to adapt to the current task, so that it can quickly adapt to new traffic measurement tasks and reduce the time and cost of retraining. Extract representative features from the raw traffic data through unsupervised learning. Unsupervised learning can automatically discover hidden structures and patterns in the data and extract features that are useful for prediction tasks, thereby improving the efficiency and accuracy of the model. Generate synthetic data with a distribution similar to that of real traffic data and enhance the dataset in a generative adversarial network manner. This method increases the diversity of training samples by generating synthetic data similar to real data, which helps the model learn more comprehensive data features and improves the robustness and generalization ability of the model.

[0051] The feature extraction algorithm includes a Fourier transform algorithm and a peak detection algorithm, wherein the Fourier transform algorithm is used to extract frequency features, and the peak detection algorithm is used to extract amplitude features; The Fourier transform algorithm includes: Preprocessing the collected pressure pulsation signal; According to the Fourier transform algorithm, the time domain signal is converted into a frequency domain signal; Extract key frequency components from frequency domain signals as pumping frequency input features in flow calculation models; The peak detection algorithm includes: obtaining the pressure change caused by each pumping, and calculating the liquid volume of each pumping.

[0052] In this technical solution, by clearly pointing out that the feature extraction algorithm includes the Fourier transform algorithm and the peak detection algorithm, the respective functions of the Fourier transform algorithm and the peak detection algorithm are elaborated in detail. The Fourier transform algorithm is used to extract frequency features and convert the time domain signal into the frequency domain signal to obtain the frequency components in the signal; the peak detection algorithm is used to extract amplitude features and focus on the peak information in the signal. This functional division makes the feature extraction more comprehensive and targeted. The Fourier transform algorithm is explained in detail, including the steps of preprocessing the pressure pulsation signal, converting the time domain signal into the frequency domain signal, and extracting the key frequency components. It can deeply understand the specific application and operation process of the Fourier transform algorithm in feature extraction. By pointing out that the key frequency components in the extracted frequency domain signal will be used as the input features of the flow calculation model, the importance of feature extraction in the flow measurement system is emphasized.

[0053] It can be understood that the key frequency components include primary frequency and secondary frequency.

[0054] The specific description of the present invention in the above embodiments is only used to further illustrate the present invention and cannot be understood as limiting the scope of protection of the present invention. Technical engineers in this field may make some non-essential improvements and adjustments to the present invention based on the contents of the above invention, which fall within the scope of protection of the present invention.

Claims

1. A flow measurement system based on a pressure pulsation method, characterized in that: include: A data acquisition module is installed at the outlet of the plunger pump and is used to collect the periodic pressure pulsation signal generated by the plunger pump in the fluid pipeline in real time; The signal processing module is connected to the data acquisition module and is used to receive and process the collected pressure pulse signal and extract the frequency and amplitude characteristics of the pressure pulsation using a feature extraction algorithm; The calculation module establishes a flow calculation model based on the pumping frequency, the volume of liquid pumped each time, the efficiency of the pump, the pipeline resistance, the medium viscosity, the pipeline length and the pipeline diameter, and calculates the flow rate of the solid-liquid mixture; The calibration feedback module is used to calibrate the flow calculation results output by the calculation module. It uses a machine learning algorithm to train and optimize based on historical measurement data and actual flow data, and provides feedback when the flow is abnormal.

2. The flow measurement system based on the pressure pulsation method according to claim 1, characterized in that: The flow rate is calculated as follows: Q final =η×f×V / (1+α×R+β×μ+γ×L / D); Among them, η is the actual working efficiency of the plunger pump, f is the pumping frequency, V is the volume of liquid pumped each time, R is the pipeline resistance, μ is the medium viscosity, L is the pipeline length, D is the pipeline diameter, and α, β, and γ are empirical coefficients.

3. The flow measurement system based on the pressure pulsation method according to claim 1, characterized in that: The system also includes a power and resistance analysis module, which is used to analyze the fluid flow condition by using the one-way check characteristic of the plunger pump when the power is less than the resistance; The power and resistance analysis module includes: The pressure accumulation analysis unit is used to monitor and analyze the pressure increase and release process in the pipeline after each plunger pumping, and identify the impact of the pressure accumulation effect on the fluid flow; The one-way check feature utilizes the one-way check valve of the plunger pump to ensure the one-way flow of the fluid during the pumping process, and maintains the slow flow of the fluid in the pipeline through pressure accumulation during the pumping interval; The flow resistance dynamic evaluation unit monitors and evaluates the flow resistance encountered by the fluid in the pipeline in real time, and dynamically adjusts the value of the pipeline resistance in combination with the pressure accumulation effect, and dynamically adjusts the prediction model and flow calculation model for the fluid fluidity.

4. The flow measurement system based on the pressure pulsation method according to claim 3, characterized in that: The pressure accumulation analysis unit comprises: By continuously monitoring the pressure changes in the pipeline, the pressure pulse caused by each pumping of the plunger pump is identified, and the pressure accumulation after the pressure pulse is calculated; The relationship between pressure accumulation and time is analyzed, and a pressure accumulation model is established to predict the slow flow state of the fluid maintained by pressure accumulation when the power is less than the resistance.

5. The flow measurement system based on the pressure pulsation method according to claim 3, characterized in that: The one-way non-return characteristic utilizing unit comprises: After each pumping of the plunger pump, the one-way check valve automatically closes to prevent the fluid from flowing back and ensure that the pressure in the pipeline accumulates gradually; During the pumping interval, the accumulated pressure is used to push the fluid to flow slowly in the pipeline until the next pumping starts, forming a continuous but non-uniform fluid delivery process.

6. The flow measurement system based on the pressure pulsation method according to claim 3, characterized in that: The flow resistance dynamic evaluation unit comprises: Real-time monitoring of the flow rate, pressure and temperature parameters of the fluid in the pipeline, combined with the fluid properties and pipeline material, dynamically calculate the current flow resistance; The dynamically evaluated flow resistance value is fed back to the flow calculation module, and the pipeline resistance term in the flow calculation formula is adjusted in real time to improve the accuracy of flow calculation.

7. The flow measurement system based on the pressure pulsation method according to claim 1 or 3, characterized in that: The establishment of the flow calculation model includes: Based on the historical working data of the plunger pump, an initial prediction model is established according to the machine learning algorithm; During actual operation, new working condition data and corresponding flow measurement values ​​are continuously collected, and the prediction model is updated in real time based on the online learning algorithm; According to the updated prediction model output, the actual working efficiency and pumping frequency of the plunger pump are initialized.

8. The flow measurement system based on the pressure pulsation method according to claim 1, characterized in that: The calibration feedback module also includes an anomaly detection unit, which constructs a feature space of normal flow based on a deep learning anomaly detection method, detects real-time flow data, and when flow data deviating from the normal range is identified, it is determined to be abnormal and triggers flow anomaly feedback.

9. The flow measurement system based on the pressure pulsation method according to claim 7, characterized in that: The machine learning algorithm includes: Combining random forest and gradient boosted tree learners, the prediction performance of the prediction model is improved through weighted average method; Use the model pre-trained on different flow measurement tasks as a starting point and adapt it to the current task by fine-tuning the parameters; Extract representative features from raw traffic data through unsupervised learning; Generate synthetic data with similar distribution to real traffic data to generate adversarial networks.

10. The flow measurement system based on the pressure pulsation method according to claim 1, characterized in that: The feature extraction algorithm includes a Fourier transform algorithm and a peak detection algorithm, wherein the Fourier transform algorithm is used to extract frequency features, and the peak detection algorithm is used to extract amplitude features; The Fourier transform algorithm includes: Preprocessing the collected pressure pulsation signal; According to the Fourier transform algorithm, the time domain signal is converted into a frequency domain signal; Extract key frequency components from frequency domain signals as pumping frequency input features in flow calculation models; The peak detection algorithm includes: obtaining the pressure change caused by each pumping, and calculating the liquid volume of each pumping.

Images