Ultrasonic water meter flow measuring method and system based on dynamic temperature compensation

Abstract

The invention discloses an ultrasonic water meter flow measurement method and system based on dynamic temperature compensation, and relates to the technical field of ultrasonic water meter flow measurement. According to the ultrasonic water meter flow measurement method based on dynamic temperature compensation, flow measurement time sequence data in a set area are continuously obtained, feature extraction processing is carried out respectively, an initial flow measurement value, a temperature compensation factor and a measurement interference suppression factor of each sampling period in the set area are obtained, comprehensive analysis is carried out, and the flow measurement time sequence data are obtained. According to the method, intelligent alarm is carried out on the set area based on the actual flow measurement value of each sampling period, so that the original measurement result is dynamically corrected, and a more accurate actual flow value is formed; the reliability and the long-term stability of flow measurement in complex environments such as severe temperature difference and fluctuating flow state are improved, and then the accuracy and the usability of flow monitoring in scenes such as hot spring areas and industrial heat exchange pipelines are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic water meter flow measurement, and in particular to an ultrasonic water meter flow measurement method and system based on dynamic temperature compensation. Background Art

[0002] As a metering device that measures flow based on the principle of sound wave propagation time difference, ultrasonic water meter has been widely used in many fields. It is especially suitable for commercial or industrial water use scenarios with high measurement accuracy requirements, no mechanical wear, and strong adaptability. Traditional ultrasonic water meters arrange transducers at both ends of the pipeline to send ultrasonic signals in the downstream and upstream directions respectively, and calculate the water velocity and flow rate based on the propagation time difference in the two directions combined with the pipeline structure parameters. However, in high-temperature water source scenarios such as hot spring areas and geothermal wells, the environment where the water meter is located has physical characteristics that are significantly different from those of the normal temperature water supply system. On the one hand, the water temperature is high all year round, and the sound velocity is significantly affected by temperature changes, resulting in systematic deviations in the flow results calculated based on the fixed sound velocity model. On the other hand, high-temperature fluids are often accompanied by factors such as increased turbulence and thermal expansion of equipment in pipelines, which further reduces the credibility of traditional measurement results. In addition, affected by the thermal inertia of the environment, the response of the temperature sensor is delayed, and the self-heating effect of the transducer body may also introduce measurement errors. These factors are difficult to effectively correct through traditional single temperature compensation strategies.

[0003] The prior art, such as the method, device, electronic device and storage medium for measuring the flow rate of the fluid in the pipeline disclosed in the patent application with the publication number CN116295678A, obtains the corresponding forward and reverse flow transmission time of the ultrasonic wave in the fluid to be measured in the pipeline through the ultrasonic transducer in the ultrasonic water meter, and obtains the target time difference and the target time and; based on the target time difference and the motion state of the fluid to be measured, the surface velocity of the corresponding cross section of the fluid to be measured is determined; the surface velocity represents the average flow velocity of the corresponding cross section of the fluid to be measured; based on the preset temperature compensation table, the target time and the corresponding temperature compensation coefficient are used as the target compensation coefficient; the surface velocity is adjusted according to the target compensation coefficient to obtain the target surface velocity; based on the target surface velocity, the flow rate corresponding to the fluid to be measured is determined. Through the above measurement method, it is not necessary to measure the current temperature, avoiding the problem of equipment resource occupation caused by setting a temperature sensor, and improving work efficiency to a certain extent while ensuring measurement accuracy.

[0004] Based on the above scheme, it is found that the limitations of the existing technology include at least the following problems: the existing technology fails to take into account the dynamic and multi-source disturbance characteristics of temperature changes in hot spring areas, resulting in systematic deviations in sound speed estimation under high temperature and variable temperature scenarios, which seriously affects the accuracy of flow calculation. That is, in high temperature environments such as hot spring areas, the outlet water temperature is high all year round, and is affected by factors such as water source supply fluctuations, pipeline opening and closing status, and surrounding thermal environment. The water temperature fluctuates greatly on a short time scale, and has significant thermal inertia. The pipeline structure is easily affected by thermal expansion. The static compensation table is difficult to truly reflect the thermal physical state under each sampling period, which can easily cause sound speed estimation deviations, thereby affecting the accuracy of flow calculation. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides an ultrasonic water meter flow measurement method and system based on dynamic temperature compensation, which solves the problem that the compensation mechanism of the prior art is insufficiently accurate in a temperature disturbance environment and is difficult to effectively support the flow measurement stability in high temperature scenarios.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an ultrasonic water meter flow measurement method based on dynamic temperature compensation, comprising the following steps: continuously acquiring flow measurement time series data within a set area, including ultrasonic sampling time series data, temperature compensation time series data, and measurement interference time series data; performing feature extraction processing on the flow measurement time series data within the set area to obtain the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period in the set area; and performing a comprehensive analysis on the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period in the set area to obtain the actual flow measurement value of each sampling period in the set area; and performing intelligent alarm on the set area based on the actual flow measurement value of each sampling period.

[0007] Furthermore, the specific formula for calculating the actual flow measurement value of a certain sampling period in the set area is as follows: ;in, The actual flow measurement value of a certain sampling period in the set area. The initial flow measurement value of a certain sampling period in the set area. is the temperature compensation factor for a certain sampling period in the set area, It is the temperature compensation correction adjustment coefficient stored in the database. is the temperature deviation adjustment coefficient stored in the database, is the measurement interference suppression factor of a certain sampling period in the set area, is the measurement interference adjustment coefficient stored in the database, It is the nonlinear adjustment coefficient of the measured disturbance stored in the database.

[0008] Furthermore, the ultrasonic sampling timing data includes ultrasonic signal data of each sampling period, and the ultrasonic signal data is specifically the signal voltage amplitude at each time point. The specific steps for obtaining the initial flow measurement value of each sampling period in the set area are as follows: obtaining the downstream propagation time value, upstream propagation time value, propagation path length value, and pipeline cross-sectional area value of each sampling period in the set area, and performing comprehensive analysis to obtain the initial flow estimation value of each sampling period in the set area; inputting the ultrasonic signal data of each sampling period in the set area into a pre-trained ultrasonic prediction model for comprehensive analysis to obtain the signal flow value and signal credibility index of each sampling period in the set area; and performing comprehensive analysis on the initial flow estimation value, signal flow value, and signal credibility index of each sampling period in the set area to obtain the initial flow measurement value of each sampling period in the set area.

[0009] Furthermore, the ultrasonic prediction model is specifically a convolutional neural network, which includes an input layer, a feature extraction convolutional layer, a feature compression layer, and a task output layer. The specific steps of obtaining the signal flow value and the signal credibility index of each sampling period in the set area are as follows: in the input layer of the convolutional neural network, the ultrasonic signal data of each sampling period in the set area is received and preprocessed; in the feature extraction convolutional layer of the convolutional neural network, multi-scale convolution feature extraction is performed on the preprocessed ultrasonic signal data of each sampling period in the set area to obtain a high-dimensional feature vector of each sampling period in the set area; in the feature compression layer of the convolutional neural network, the high-dimensional feature vector of each sampling period in the set area is compressed and aggregated to obtain a waveform structure feature vector of each sampling period in the set area; in the task output layer of the convolutional neural network, branch prediction processing is performed on the waveform structure feature vector of each sampling period in the set area to obtain a signal flow value and a signal credibility index of each sampling period in the set area.

[0010] Furthermore, the temperature compensation time series data includes the water temperature value, ambient temperature value, pipe wall conduction temperature hysteresis value, equipment self-heating compensation index, pipeline thermal expansion length change value, and water body compressibility temperature response index of each sampling period. The specific steps for obtaining the temperature compensation factor of each sampling period in the set area are as follows: comprehensively analyzing the temperature compensation time series data in the set area to obtain a set of temperature perception correction indexes for each sampling period in the set area, including a thermal field credible perception index and a thermally induced structural response index; and comprehensively analyzing the set of temperature perception correction indexes for each sampling period in the set area to obtain the temperature compensation factor for each sampling period in the set area.

[0011] Furthermore, the specific steps for obtaining the temperature perception correction index set for each sampling period in the set area are as follows: read the water temperature value, ambient temperature value, pipe wall conduction temperature hysteresis value, and equipment self-heating compensation index for each sampling period in the set area, and perform a comprehensive analysis to obtain the thermal field credible perception index for each sampling period in the set area; read the pipeline thermal expansion length change value and water body compressibility temperature response index for each sampling period in the set area, and perform a comprehensive analysis to obtain the thermally induced structural response index for each sampling period in the set area.

[0012] Furthermore, the specific formula for calculating the temperature compensation factor of a certain sampling period in the set area is as follows: ;in, is the temperature compensation factor for a certain sampling period in the set area, is the thermally induced structural response index of a certain sampling period in the set area, is the thermally induced structure adjustment coefficient stored in the database, It is the thermal field credible perception index of a certain sampling period in the set area. is the thermal field credible adjustment coefficient stored in the database, is the interaction adjustment coefficient stored in the database.

[0013] Furthermore, the measurement interference time series data includes turbulence index, micro-vibration index, waveform stability entropy value, and transduction coupling index. The specific steps for obtaining the measurement interference suppression factor of each sampling period in the set area are as follows: standardizing the turbulence index, micro-vibration index, waveform stability entropy value, and transduction coupling index of each sampling period in the set area; and comprehensively analyzing the turbulence index, micro-vibration index, waveform stability entropy value, and transduction coupling index of each sampling period in the set area after standardization based on the particle optimization algorithm to obtain the measurement interference suppression factor of each sampling period in the set area.

[0014] Furthermore, the specific steps for intelligent alarm of the set area based on the actual flow measurement value of each sampling period are as follows: compare and analyze the actual flow measurement value of each sampling period in the set area with the preset flow threshold; if the actual flow measurement value of each sampling period in the set area is lower than or equal to the preset flow threshold, no alarm is issued; if the actual flow measurement value of each sampling period in the set area is higher than the preset flow threshold, an alarm is issued and corresponding warning measures are taken.

[0015] The ultrasonic water meter flow measurement system based on dynamic temperature compensation includes: a data acquisition module, which is used to continuously acquire flow measurement time series data within a setting, including ultrasonic sampling time series data, temperature compensation time series data, and measurement interference time series data; a feature extraction module, which is used to perform feature extraction processing on the flow measurement time series data in a set area, and obtain the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period in the set area; a comprehensive correction module, which is used to perform comprehensive analysis on the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period in the set area, and obtain the actual flow measurement value of each sampling period in the set area; and a measurement feedback module, which is used to perform intelligent alarm on the set area based on the actual flow measurement value of each sampling period.

[0016] The present invention has the following beneficial effects: (1) The ultrasonic water meter flow measurement method based on dynamic temperature compensation comprehensively considers the actual influencing factors of temperature environment and measurement interference in each sampling period, dynamically corrects the original measurement results, and forms a more accurate actual flow value. In terms of thermal disturbance and interference influence, multi-dimensional parameter modeling and analysis are performed respectively to obtain temperature compensation factor and measurement interference suppression factor, and the initial flow measurement value obtained by ultrasonic sampling time series data analysis is analyzed to generate actual flow measurement value, thereby improving the reliability and long-term stability of flow measurement in complex environments such as drastic temperature difference and fluctuating flow state, thereby ensuring the accuracy and availability of flow monitoring in scenes such as hot spring areas and industrial heat exchange pipelines.

[0017] (2) The ultrasonic water meter flow measurement method based on dynamic temperature compensation introduces a convolutional neural network structure to extract features from the complete ultrasonic waveform sequence within the sampling period, thereby constructing a dual-branch architecture of a signal flow prediction model and a waveform credibility assessment model. The predicted signal flow value and the signal credibility index are weightedly fused to finally output the initial flow measurement value. This allows the method to perceive complex information such as local interference, structural deformation, and reflected echo based on the overall waveform morphology, automatically reduce the weight of abnormal signal periods, and effectively improve the anti-interference ability and discrimination accuracy under complex working conditions, thereby achieving a cycle-level flow measurement result with higher confidence.

[0018] (3) The ultrasonic water meter flow measurement method based on dynamic temperature compensation introduces the particle swarm optimization algorithm to adaptively learn the interference parameter weights, thereby significantly enhancing the adaptability and optimization ability of the interference suppression mechanism in complex environments, so that the measurement interference suppression factor finally constructed can more accurately reflect the degree of abnormal disturbance in the current cycle. The particle swarm algorithm has the characteristics of global search and rapid convergence, and can complete the efficient optimization of multi-parameter weights at a limited computational cost, thereby effectively avoiding the subjective errors and local optimal problems of artificial weight setting, and then can still stably output reliable flow results in scenarios with variable interference and frequent signal quality fluctuations, thereby improving the measurement robustness and parameter adaptation ability under dynamic non-steady-state conditions.

[0019] (4) The ultrasonic water meter flow measurement system based on dynamic temperature compensation improves the real-time response efficiency and data closed-loop capability of flow monitoring results through process hierarchical design and functional module deconstruction. The system adopts a continuous process of data acquisition, feature extraction, comprehensive correction, and feedback decision-making. It automatically completes the linkage processing of ultrasonic signal data and multi-dimensional environmental factors in each sampling cycle, and outputs the corrected flow results in real time on the local end, thereby avoiding the feedback lag caused by batch calculations in the cloud. At the same time, the abnormal data is immediately transmitted back in the form of an alarm through the measurement feedback module, forming a closed-loop control between front-end collection and back-end response. It is especially suitable for high-security scenarios that require rapid response and dynamic adjustment, such as hot spring water volume monitoring and instantaneous leakage identification in smart water services, thereby effectively improving the timeliness, robustness and on-site control capabilities of the system.

[0020] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of the ultrasonic water meter flow measurement method based on dynamic temperature compensation of the present invention.

[0022] Figure 2 The present invention is a flowchart of the specific steps of obtaining the initial flow measurement value of each sampling period in a set area in the ultrasonic water meter flow measurement method based on dynamic temperature compensation.

[0023] Figure 3 This is an example diagram of a temperature perception correction timing index set within a set area in the ultrasonic water meter flow measurement method based on dynamic temperature compensation of the present invention.

[0024] Figure 4 This is a block diagram of the ultrasonic water meter flow measurement system based on dynamic temperature compensation of the present invention. DETAILED DESCRIPTION

[0025] See also Figure 1 The embodiment of the present invention provides a technical solution: an ultrasonic water meter flow measurement method based on dynamic temperature compensation, comprising the following steps: continuously acquiring flow measurement time series data in a set area (such as a water pipe for commercial water use in a hot spring area), including ultrasonic sampling time series data, temperature compensation time series data, and measurement interference time series data; performing feature extraction processing on the flow measurement time series data in the set area to obtain the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period (i.e., the ultrasonic water meter completes a complete flow measurement action) in the set area; and performing a comprehensive analysis on the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period in the set area to obtain the actual flow measurement value of each sampling period in the set area; and performing intelligent alarm on the set area based on the actual flow measurement value of each sampling period.

[0026] The specific formula for calculating the actual flow measurement value of a certain sampling period in the set area is as follows: ;in, The actual flow measurement value of a certain sampling period in the set area. The initial flow measurement value of a sampling period in the set area. is the temperature compensation factor for a certain sampling period in the set area, It is the temperature compensation correction adjustment coefficient stored in the database. is the temperature deviation adjustment coefficient stored in the database, is the measurement interference suppression factor of a certain sampling period in the set area, is the measurement interference adjustment coefficient stored in the database, It is the nonlinear adjustment coefficient of the measured disturbance stored in the database.

[0027] It needs to be explained that the specific form of the tanh function is: ,in, is a natural constant and can be taken as 2.71 in this implementation example, with a domain of (−∞, +∞) and a range of (−1, +1).

[0028] , , , It can be obtained through the following steps: using historical monitoring data, combined with initial flow measurement values, temperature compensation factors, and measurement interference suppression factors, statistical regression analysis is performed to quantify the specific impact of each factor on the actual flow measurement value, thereby fitting the initial weight value. Secondly, the sensitivity analysis method is used to adjust the value range of each coefficient, and observe its impact on the evaluation results of the actual flow measurement value to ensure the stability and rationality of the model. Based on regional characteristics and actual conditions, the preliminary fitting coefficients are corrected and optimized to finally determine the coefficient value applicable to the specific area.

[0029] Specifically, if Figure 2 As shown, the ultrasonic sampling timing data includes ultrasonic signal data of each sampling period, and the ultrasonic signal data is specifically the signal voltage amplitude at each time point. The specific steps for obtaining the initial flow measurement value of each sampling period in the set area are as follows: obtain the downstream propagation time value, upstream propagation time value, propagation path length value, and pipeline cross-sectional area value of each sampling period in the set area, and perform a comprehensive analysis, that is, pipeline cross-sectional area value × ((propagation path length value × (downstream propagation time value-upstream propagation time value)) / (2×downstream propagation time value×upstream propagation time value)), and obtain each sampling period in the set area. The initial flow estimation value is obtained; the ultrasonic signal data of each sampling period in the set area is input into the pre-trained ultrasonic prediction model for comprehensive analysis to obtain the signal flow value and signal credibility index of each sampling period in the set area; and the initial flow estimation value, signal flow value and signal credibility index of each sampling period in the set area are comprehensively analyzed, that is, weighted processing is performed, and the signal credibility index is used as the weight coefficient of the signal flow value, specifically: signal flow value × signal credibility index + (1-signal credibility index) × initial flow estimation value, to obtain the initial flow measurement value of each sampling period in the set area.

[0030] Among them, the downstream propagation duration value is the difference in time required for the ultrasonic signal to propagate in the downstream direction in the water body, which is recorded in real time by the receiver timing module in the ultrasonic water meter.

[0031] The countercurrent propagation duration value is the time required for the ultrasonic signal to propagate in the water body in the countercurrent direction, which is recorded in real time by the receiver timing module in the ultrasonic water meter.

[0032] The propagation path length value is a fixed straight-line distance for the ultrasonic wave to propagate between sensors, and can be obtained through a structural design table stored in a database.

[0033] The pipe cross-sectional area value can be obtained from the pipe specifications stored in the database.

[0034] The ultrasonic prediction model is specifically a (one-dimensional) convolutional neural network. The convolutional neural network includes an input layer, a feature extraction convolution layer, a feature compression layer, and a task output layer. The specific steps for obtaining the signal flow value and the signal credibility index of each sampling period in the set area are as follows: In the input layer of the convolutional neural network, the ultrasonic signal data of each sampling period in the set area, that is, the signal voltage amplitude at each time point, is received and preprocessed; in the feature extraction convolution layer of the convolutional neural network, multi-scale convolution feature extraction is performed on the ultrasonic signal data of each sampling period in the set area after preprocessing (that is, multi-scale structural feature extraction and nonlinear activation processing are performed on the input ultrasonic signal data, and the ultrasonic signal in each sampling period is input). The signal sequence is input into multiple one-dimensional convolution channels with different receptive fields for parallel processing. Each convolution channel uses different convolution kernel sizes, for example: 3, 5, 7, to extract local waveform structure features at different time scales. The convolution kernel is used to extract feature information such as signal rising edge, peak, local disturbance, etc., to adapt to the different amplitude and frequency components in the waveform. Then each convolution channel outputs a corresponding time series feature map to represent the structural response at the current scale, and splices the feature maps of the above multiple scales in the channel dimension to form a unified intermediate multi-scale feature vector for comprehensively expressing the different scale attributes of the waveform structure. Finally, the spliced ​​feature vector is processed by activation function to introduce nonlinear expression capability. This processing is used In order to enhance the network's perception of complex waveform changes, such as asymmetric oscillations, spike interference, and rapid attenuation, the feature vector after activation processing is the structural feature representation result of the current sampling period), and obtain a high-dimensional feature vector for each sampling period in the set area (characterizing the local structural characteristics and disturbance patterns of the signal, and being a collection of multiple feature vectors); in the feature compression layer of the convolutional neural network, the high-dimensional feature vectors of each sampling period in the set area are compressed and aggregated (the high-dimensional feature vectors output by the feature extraction convolution layer are pooled along the time dimension, that is, the global average pooling method is adopted to average all time step features of each convolution channel to obtain a preliminary channel summary vector, so that the model aggregates the input signal in the entire sampling period The global structural pattern within the image is extracted to remove the interference of local fluctuations on the overall structural expression, and the globally pooled feature vector is input into one or more fully connected layers. Through linear transformation and nonlinear activation functions such as ReLU, the compressed global features are further fused and transformed to enhance the correlation and distinction between features, forming a more discriminative structural expression. In the feature vector fusion stage, a channel attention mechanism is introduced, such as a channel weighting strategy, to dynamically adjust the feature importance of different channels to enhance the contribution of key channel features to the final structural representation), and the waveform structure feature vector of each sampling period in the set area is obtained (representing the global structural characteristics, disturbance pattern and key feature distribution of the ultrasonic signal in the period);In the task output layer of the convolutional neural network, the waveform structure feature vector of each sampling period in the set area is subjected to branch prediction processing (the structure feature vector is taken as input and input into two parallel fully connected output branches, corresponding to the signal flow prediction task and the signal credibility prediction task respectively. For the signal flow value, the input structure feature vector is subjected to linear mapping processing, the weighted sum is calculated through the fully connected layer, and the bias term is added, and then it is input into the nonlinear activation function for nonlinear transformation, that is, the ReLU function is used. This processing can extract pattern features closely related to the water flow motion characteristics from the input features, such as peak density, symmetry, signal change rate, waveform attenuation trend, etc., and the prediction branch can further include one or more intermediate hidden layers, each of which is also composed of fully connected operations and nonlinear activation functions, so as to help simulate the possible nonlinear mapping relationship between the waveform structure and the flow value, and after all the intermediate mapping processing is completed, it finally enters a linear output layer without an activation function, which is used to map the features of the previous layer. The result of the projection is converted into a scalar form, and the signal flow value of the current sampling period is directly output as a continuous real number, which is used to represent the volume of water passing through the set pipe section in unit time. For the signal credibility index, the input structural feature vector is input into the first fully connected layer of the branch for linear transformation processing, and the nonlinear mapping is completed through the ReLU activation function to extract features related to signal quality, such as envelope integrity, waveform jitter amplitude, structural repeatability, local noise energy, etc., and this branch can set one or more additional hidden layers to enhance the ability to judge the reliability of the signal structure. This layer also uses a combination of a fully connected structure and a nonlinear activation function to allow the model to learn deeper structural properties, such as the combined influence of reflection overlap, signal symmetry deviation, asynchronous mismatch and other factors. The last layer is a single neuron output layer with a Sigmoid activation function, which is used to normalize the network prediction value to the range of 0-1 and output the signal credibility index of the current sampling period), and obtain the signal flow value and signal credibility index of each sampling period in the set area. ;

[0035] Among them, the input layer is used to receive the ultrasonic signal waveform sequence data of each sampling period in the set area. The waveform sequence consists of the signal amplitude value at each time point, reflecting the complete received signal structure within the period during the ultrasonic propagation process.

[0036] The feature extraction convolution layer is used to perform multi-scale convolution feature extraction and nonlinear activation processing on the input waveform sequence, extract typical local features of the signal in the time domain, such as local disturbance patterns, edge structure features, peak density, rising edge changes, etc., to form a high-dimensional feature tensor containing multi-channel timing features.

[0037] The feature compression layer is used to perform time dimension compression and channel aggregation processing on high-dimensional feature tensors, including global average pooling operations and fully connected mapping processes, so as to compress the multi-scale structural information of the entire waveform into a structural feature vector of fixed length as a unified input expression for subsequent prediction tasks of the model.

[0038] The task output layer is used to output the signal flow value and signal credibility index of the corresponding sampling period.

[0039] And the pre-training process of the convolutional neural network is as follows: A labeled signal data set is obtained, including the original ultrasonic signal waveform sequence and its corresponding target label information within multiple sampling periods. The signal waveform sequence is a time series composed of the received signal amplitude values ​​at each time point. The target label information at least includes the actual flow value label and signal quality score label of each sampling period. The actual flow value label is obtained by synchronous measurement of a high-precision standard flow meter, and the signal quality score label is automatically generated based on manual scoring or a preset rule algorithm to evaluate the structural stability and interference degree of the current ultrasonic signal. Then the labeled signal data set is divided into a signal training set and a signal verification set.

[0040] Initialize the convolutional neural network, that is, initialize the weights of all layers in the network, such as using He initialization or Xavier initialization, and bind the loss function module (MSE for traffic, BCE for credibility), then set the multi-task loss weighting coefficient, and configure the optimizer (such as Adam, initial learning rate).

[0041] Training is performed based on the signal training set, the number of training cycles is set (such as 50 times), and forward propagation calculation is performed in each training cycle (the ultrasonic waveform data of each small batch is input into the convolutional neural network model, and passes through the input layer, feature extraction convolution layer, feature compression layer and task output layer in turn, and the predicted signal flow value and signal credibility index of the current small batch samples are output respectively), loss function calculation (the loss value between the current small batch prediction result and the label is calculated respectively, including the regression loss of the signal flow value, such as the mean square error MSE, and the classification loss of the signal credibility index, such as the binary cross entropy BCE, and the total loss value of the current small batch is obtained by weighting according to the multi-task loss combination method), back propagation and gradient update (back propagation operation is performed based on the current total loss value, the gradient value of the weight parameters of each layer is calculated, and the model parameters are updated through the selected optimization algorithm, such as Adam, to gradually minimize the training error).

[0042] After each training session, an evaluation and analysis is performed based on the signal validation set. All waveform sequences in the signal validation set that have not participated in the training are input batch by batch into the network model of the current training round, and a complete forward propagation process is performed to obtain the corresponding signal traffic prediction value and signal credibility index prediction value, respectively. Based on the corresponding actual traffic labels and scoring labels in the validation set, the following are calculated: signal traffic prediction error (such as mean square error MSE or mean absolute error MAE), signal credibility score accuracy (such as binary classification accuracy, AUC or cross entropy loss), and then the prediction performance results of the current training round on the validation set are recorded to draw the model performance change trend curve during the training process (such as loss vs. epoch, etc.). If the validation loss of the current round is lower than the historical best record, the current model parameters are saved as the optimal model version for subsequent deployment or for triggering judgment of the Early Stopping mechanism.

[0043] When the training is completed and the loss on the validation set reaches the expected standard, the training process ends and a trained network model is obtained.

[0044] In this implementation scheme, by introducing a one-dimensional convolutional neural network model, multi-scale structural feature extraction and semantic modeling are performed on the complete ultrasonic waveform sequence within each sampling period, which can effectively identify microstructural abnormal features such as weak disturbances, reflection interference, waveform asymmetry, noise echo, etc. contained in the signal. On this basis, the network outputs the signal flow prediction value and the credibility index in parallel, and then weightedly fuses the prediction result with the traditional physical valuation (such as the forward and reverse propagation time difference calculation method) based on the credibility to form a more robust initial flow measurement value, especially in the presence of periodic interference such as bubble disturbances, pipe wall echoes, local distortion, etc., thereby automatically reducing the impact of the abnormal signal period on the overall flow output, enhancing the adaptability to non-ideal sampling periods, and thus ensuring that the period-level flow estimation results have higher confidence and lower drift errors, thereby effectively improving the accuracy and anti-interference of the system under complex field conditions.

[0045] Specifically, the temperature compensation time series data include the water temperature value, ambient temperature value, pipe wall conduction temperature hysteresis value, equipment self-heating compensation index, pipeline thermal expansion length change value, and water body compressibility temperature response index of each sampling period. The specific steps for obtaining the temperature compensation factor of each sampling period in the set area are as follows: comprehensively analyze the temperature compensation time series data in the set area to obtain a set of temperature perception correction indexes for each sampling period in the set area, including a thermal field credible perception index (characterizing the perception credibility and measurement accuracy of the actual temperature state of the water body) and a thermally induced structural response index (characterizing the nonlinear response intensity of temperature to the ultrasonic propagation path structure and the sound velocity of the water body); and comprehensively analyze the set of temperature perception correction indexes for each sampling period in the set area to obtain the temperature compensation factor for each sampling period in the set area.

[0046] The water temperature value is the average of the actual temperature of the flowing water in the pipeline at each time point during the sampling period, and the actual temperature at each time point can be obtained through the high-precision temperature sensor built into the water meter.

[0047] The ambient temperature value is the average of the air temperature at each time point in the operating environment of the ultrasonic water meter during the sampling period, and the air temperature at each time point can be obtained by an independent ambient temperature sensor in the device cavity.

[0048] The hysteresis value of the pipe wall conduction temperature is the hysteresis deviation between the temperature value measured by the high-precision temperature sensor built into the water meter and the true center temperature of the water body. It can be obtained through the temperature change rate within the cycle (i.e., the difference between the temperature value at the first time and the temperature value at the last time point / cycle duration), convection heat transfer coefficient (which can be obtained through the convection heat transfer table stored in the database), pipe wall thickness (which can be obtained through the pipe specifications stored in the database), pipe material thermal conductivity (which can be obtained through the material technical specifications stored in the database), and effective thermal contact area (the wall area of ​​the sensor bottom structure, which can be obtained through the installation data stored in the database). A comprehensive analysis is performed, i.e., temperature change rate × ((convection heat transfer coefficient × pipe wall thickness) / (pipe material thermal conductivity × effective thermal contact area)) = pipe wall conduction temperature hysteresis value.

[0049] The equipment self-heating compensation index is the compensation for the water temperature offset caused by self-heating when the equipment is in operation during the sampling period. It can be obtained by obtaining the operating power (i.e., the average of the operating power at each time of the period, which can be obtained through a power meter), the thickness of the heat conduction path between the heating element and the water body (the equivalent physical thickness of the heat conduction from the heating unit to the water body through a continuous solid structure, which can be obtained through the technical specifications stored in the database), the thermal conductivity of the pipeline material, and the effective thermal contact area, and performing a comprehensive analysis, i.e., the equipment self-heating compensation index = operating power × (the thickness of the heat conduction path between the heating element and the water body / (the thermal conductivity of the structural material × the effective thermal contact area)).

[0050] The change in the thermal expansion length of the pipeline is the actual change in the ultrasonic propagation path length caused by the thermal expansion of the pipeline material during the sampling period. It can be obtained through the material thermal expansion coefficient (obtained through the standard thermal expansion coefficient of the material stored in the database), the initial path length (obtained through the structural design data stored in the database), the current water temperature (which can be obtained through the temperature sensor), and the water temperature reference temperature (obtained through the installation setting value stored in the database). After a comprehensive analysis, that is, the change in the thermal expansion length of the pipeline = material thermal expansion coefficient × initial path length × (current water temperature - water temperature reference temperature), and the calculation result is the change in the thermal expansion length of the pipeline.

[0051] The water body compressibility temperature response index is the response degree of the nonlinear change of the water body sound velocity with temperature during the sampling period. It can be obtained by extracting the water sound velocity data under multiple temperature conditions based on the international standard database (such as IAPWS), and taking the sound velocity at a certain reference temperature (such as 25°C) as the standard value; then the sound velocity values ​​at different temperatures are normalized with the reference value to form the sound velocity change ratio; then, by calculating the change of each temperature relative to the reference temperature, the fitting relationship between the sound velocity change ratio and the temperature increment is constructed, and the primary response coefficient is extracted from the fitting curve as the water body compressibility temperature response index.

[0052] The specific steps for obtaining the temperature perception correction index set for each sampling period in the set area are as follows: read the water temperature value, ambient temperature value, pipe wall conduction temperature hysteresis value, and equipment self-heating compensation index of each sampling period in the set area, and perform a comprehensive analysis (i.e., first perform standardization processing, and perform weighted processing based on the standardization processing result, and apply the Sigmoid function to map the result to between 0 and 1), and obtain the thermal field credible perception index of each sampling period in the set area; read the pipeline thermal expansion length change value and water body compressibility temperature response index of each sampling period in the set area, and perform a comprehensive analysis (i.e., first perform standardization processing, and perform weighted processing based on the standardization processing result, and apply the Sigmoid function to map the result to between 0 and 1), and obtain the thermally induced structural response index of each sampling period in the set area.

[0053] The specific formula for calculating the temperature compensation factor for a certain sampling period within the set area is as follows: ;in, is the temperature compensation factor for a certain sampling period in the set area, is the thermally induced structural response index of a certain sampling period in the set area, is the thermally induced structure adjustment coefficient stored in the database, It is the thermal field credible perception index of a certain sampling period in the set area. is the thermal field credible adjustment coefficient stored in the database, is the interaction adjustment coefficient stored in the database.

[0054] It needs to be explained that , , It can be obtained through the following steps: Based on historical data, determine the initial influence weight of each variable (thermal field credible perception index, thermal-induced structural response index) on the temperature compensation factor through statistical regression analysis, and then use the sensitivity analysis method to adjust the value range of the coefficient to evaluate the stability and applicability of these parameters to the formula output. Next, further fit the weights through model optimization (such as machine learning algorithms or multi-objective optimization) to ensure that the formula can accurately reflect the actual temperature compensation effect.

[0055] The specific implementation example of calculating the temperature compensation factor of a certain sampling period in the set area is as follows. The existing data is as follows: including the thermal field credible perception index and the thermally induced structural response index of 5 sampling periods in the set area, as shown in Table 1 and Figure 3 As shown: Table 1 Temperature perception correction time series index set in the set area Thermally induced structural adjustment coefficients stored in the database Approximately: 0.572; Thermal field credible adjustment coefficients stored in the database Approximately: 0.493; Interaction adjustment coefficients stored in the database Approximately: 0.413; Substituting the data in Table 1 and the above adjustment coefficient into the specific formula for calculating the temperature compensation factor of a certain sampling period in the set area, we get: The temperature compensation factor of the first sampling period in the set area = ln (1 + 0.712 0.572 ×|1-√0.683| 0.493 )×(1+exp(-0.413×0.712×0.683))≈0.542; The temperature compensation factor of the second sampling period in the set area = ln (1 + 0.694 0.572 ×|1-√0.665| 0.493 )×(1+exp(-0.413×0.694×0.665))≈0.551; The temperature compensation factor of the third sampling period in the set area = ln (1 + 0.709 0.572 ×|1-√0.697| 0.493 )×(1+exp(-0.413×0.709×0.697))≈0.551≈0.569; The temperature compensation factor of the fourth sampling period in the set area = ln (1 + 0.698 0.572 ×|1-√0.681| 0.493 )×(1+exp(-0.413×0.698×0.681))≈0.499; The temperature compensation factor of the fifth sampling period in the set area = ln (1 + 0.685 0.572 ×|1-√0.672| 0.493 )×(1+exp(-0.413×0.685×0.672))≈0.541; In this implementation scheme, a complete temperature perception feature set is constructed based on thermal parameters with actual physical significance in terms of temperature, and the thermal field credible perception index and the thermally induced structural response index are further calculated, and finally a periodic temperature compensation factor is generated, thereby comprehensively covering the multiple impacts of thermal disturbances on the sound velocity and propagation path of water bodies, and achieving parameter normalization and stability enhancement through standardized processing, weighted fusion and Sigmoid mapping. In the calculation link, interactive adjustment items and historical data-driven weight fitting mechanisms are introduced to ensure that the composite effects between different thermal variables are reasonably modeled. Finally, in the face of complex working conditions such as industrial high-temperature heat exchange equipment, pipelines with drastic water temperature fluctuations, or diversified pipe deployment, the measurement offset caused by temperature disturbance can still be accurately corrected at the periodic level, thereby significantly enhancing the practicality and adaptability of flow measurement.

[0056] Specifically, the measurement interference time series data includes turbulence index, micro-vibration index, waveform stability entropy value, and transducer coupling index. The specific steps of obtaining the measurement interference suppression factor of each sampling period in the set area are as follows: the turbulence index, micro-vibration index, waveform stability entropy value, and transducer coupling index of each sampling period in the set area are standardized; and based on the particle optimization algorithm, the turbulence index, micro-vibration index, waveform stability entropy value, and transducer coupling index of each sampling period in the set area after the standardized processing are comprehensively analyzed (that is, weighted processing, and the weighting coefficients corresponding to the turbulence index, micro-vibration index, waveform stability entropy value, and transducer coupling index can be obtained based on the particle optimization algorithm, that is, by constructing a flow measurement error, that is, the actual flow measurement value of the period is compared with the true standard The error between the standard values ​​is the optimization function of the target, and the true standard value can be obtained by obtaining the average of the flow measurement values ​​of several historical cycles and marking it as the true standard value. The particle swarm optimization algorithm is used to search and optimize the weight coefficients of each interference parameter in the comprehensive interference analysis, and each group of particles is represented as a group of candidate weighted coefficient combinations. In each round of iteration, the fitness value is calculated according to the fitting residual between the current weighted result and the actual error, and then the optimal coefficient solution is continuously approached through the particle update mechanism; after reaching the preset number of iterations or convergence conditions, the weighted coefficient group with the best fitness in the current group is extracted as the final weighted coefficient of the turbulence index, micro-vibration index, waveform stability entropy value and transducer coupling index), and the measurement interference suppression factor of each sampling period in the set area is obtained.

[0057] Among them, the turbulence index is the turbulence intensity of the fluid in the ultrasonic measurement area. Turbulence will cause instability in the sound wave propagation path (multipath propagation, sound path deviation). It can be obtained by obtaining the downstream propagation time value, the upstream propagation time value, and the symmetry reference value of the ultrasonic propagation time (measure the normal proportional structure between the ultrasonic propagation time in the downstream and upstream directions under the design conditions of the system, which can be obtained through the equipment design data stored in the data), and performing calculation and analysis, that is, turbulence index = |((upstream propagation time value-downstream propagation time value) / (upstream propagation time value+downstream propagation time value))-symmetry reference value of ultrasonic propagation time|.

[0058] The micro-vibration index is the intensity of tiny mechanical vibrations to which the device is subjected. Vibration can affect the fit of the transducer and cause unstable signal energy. The vibration data can be collected in real time through the MEMS accelerometer installed in the device, and the root mean square value of the vibration in the key frequency band (such as 50~300Hz) can be extracted. The result is the micro-vibration index.

[0059] The waveform stability entropy value is the stability of the shape of the ultrasonic waveform received in each cycle. The ultrasonic signal is obtained and discretized into a signal point sequence. The normalized energy probability distribution is constructed based on the amplitude of each sampling point, and its Shannon entropy is calculated and used as the waveform stability entropy value.

[0060] The transducer coupling index is the coupling stability between the ultrasonic transducer and the pipe surface (i.e., the degree of close fit). It can be obtained by obtaining the driving voltage and driving current at each time point in the sampling period, and performing ratio processing, i.e., driving voltage / driving current, and performing mean analysis based on the ratio processing result to obtain the average excitation impedance, and obtain the transducer reference coupling impedance (obtained through the technical specifications of the transducer stored in the database), and then performing ratio processing on the average excitation impedance and the transducer reference coupling impedance to obtain the transducer coupling index.

[0061] In this implementation scheme, the turbulence index, micro-vibration index, waveform stability entropy value and transduction coupling index are taken as key interference factors, and are jointly modeled through standardization processing and particle swarm optimization algorithm, so as to realize dynamic adaptive adjustment of interference suppression capability in each sampling period, and construct the error between the actual measured value and the historical standard value as the objective function, and use the swarm intelligence mechanism to iteratively optimize the parameter weight combination, so as to automatically approach the optimal suppression strategy. For example, when some periodic turbulence fluctuations are significant and other periodic structural coupling mismatches are more serious, the algorithm can flexibly adjust the weights of the corresponding parameters in the interference factors, thereby improving the overall recognition and suppression capabilities of multi-dimensional interference factors such as waveform anomalies, vibration disturbances and poor structural fit, and then significantly enhancing the robustness of flow measurement and error control capabilities under complex working conditions.

[0062] Specifically, the specific steps for intelligent alarm of the set area based on the actual flow measurement value of each sampling period are as follows: compare and analyze the actual flow measurement value of each sampling period in the set area with the preset flow threshold; if the actual flow measurement value of each sampling period in the set area is lower than or equal to the preset flow threshold, no alarm is issued, and the result is uploaded to the local data storage module; if the actual flow measurement value of each sampling period in the set area is higher than the preset flow threshold, an alarm is issued and corresponding warning measures are taken, that is, pushing remote alarm information, activating local visual sound and light alarms, closing valves in linkage or recording abnormal logs, etc.

[0063] In this implementation scheme, based on the actual flow measurement value dynamically generated in each sampling period, a real-time comparison mechanism for the preset flow threshold is constructed, so that sudden or continuous abnormal flow behavior can be identified with second-level accuracy. When the measurement value exceeds the threshold, not only can the remote alarm signal be pushed immediately, but also the local sound and light alarm device can be triggered and physical control measures (such as emergency valve closure) can be linked to form a data-driven active response chain. At the same time, the alarm-free data of the normal period will also be fully recorded in the local data storage module, which is convenient for subsequent trend tracing and risk modeling analysis, so as to be suitable for scenarios that require highly stable operation, such as hot spring water meters and industrial heat flow systems, so as to effectively prevent equipment rupture, water waste or illegal water extraction, thereby significantly improving the safety defense and event handling efficiency of flow monitoring.

[0064] See also Figure 4 The embodiment of the present invention provides a technical solution: an ultrasonic water meter flow measurement system based on dynamic temperature compensation, comprising: a data acquisition module, used to continuously acquire flow measurement time series data within a setting, including ultrasonic sampling time series data, temperature compensation time series data, and measurement interference time series data; a feature extraction module, used to perform feature extraction processing on the flow measurement time series data in a set area, and obtain the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period in the set area; a comprehensive correction module, used to perform comprehensive analysis on the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period in the set area, and obtain the actual flow measurement value of each sampling period in the set area; a measurement feedback module, used to perform intelligent alarm on the set area based on the actual flow measurement value of each sampling period.

[0065] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0066] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An ultrasonic water meter flow measurement method based on dynamic temperature compensation, characterized in that: The following steps are involved: Continuously obtain flow measurement time series data within the set area, including ultrasonic sampling time series data, temperature compensation time series data, and measurement interference time series data; Perform feature extraction processing on the flow measurement time series data in the set area to obtain the initial flow measurement value, temperature compensation factor and measurement interference suppression factor of each sampling period in the set area; And conduct a comprehensive analysis of the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period in the set area to obtain the actual flow measurement value of each sampling period in the set area; Intelligent alarm for set areas based on actual flow measurement values ​​in each sampling period.

2. The ultrasonic water meter flow measurement method based on dynamic temperature compensation according to claim 1 is characterized in that: The specific formula for calculating the actual flow measurement value of a certain sampling period in the set area is as follows: ； in, , , , The actual flow measurement value of a certain sampling period in the set area, the initial flow measurement value, the temperature compensation factor, and the measurement interference suppression factor are listed in turn. , , , They are the temperature compensation correction adjustment coefficient, the temperature deviation adjustment coefficient, the measurement interference adjustment coefficient, and the measurement interference nonlinear adjustment coefficient stored in the database.

3. The ultrasonic water meter flow measurement method based on dynamic temperature compensation according to claim 1 is characterized in that: The ultrasonic sampling time series data includes ultrasonic signal data of each sampling period, and the ultrasonic signal data is specifically the signal voltage amplitude at each time point. The specific steps of obtaining the initial flow measurement value of each sampling period in the set area are as follows: Obtain the downstream propagation time value, upstream propagation time value, propagation path length value, and pipeline cross-sectional area value of each sampling period in the set area, and conduct a comprehensive analysis to obtain the initial flow estimation value of each sampling period in the set area; The ultrasonic signal data of each sampling period in the set area is input into the pre-trained ultrasonic prediction model for comprehensive analysis to obtain the signal flow value and signal credibility index of each sampling period in the set area; The initial flow estimation value, signal flow value, and signal credibility index of each sampling period in the set area are comprehensively analyzed to obtain the initial flow measurement value of each sampling period in the set area.

4. The ultrasonic water meter flow measurement method based on dynamic temperature compensation according to claim 3 is characterized in that: The ultrasonic prediction model is specifically a convolutional neural network, which includes an input layer, a feature extraction convolution layer, a feature compression layer, and a task output layer. The specific steps of obtaining the signal flow value and the signal credibility index of each sampling period in the set area are as follows: In the input layer of the convolutional neural network, the ultrasonic signal data of each sampling period in the set area is received and preprocessed; In the feature extraction convolution layer of the convolutional neural network, multi-scale convolution feature extraction is performed on the ultrasonic signal data of each sampling period in the set area after preprocessing to obtain a high-dimensional feature vector of each sampling period in the set area; In the feature compression layer of the convolutional neural network, the high-dimensional feature vector of each sampling period in the set area is compressed and aggregated to obtain the waveform structure feature vector of each sampling period in the set area; In the task output layer of the convolutional neural network, branch prediction processing is performed on the waveform structure feature vector of each sampling period in the set area to obtain the signal flow value and signal credibility index of each sampling period in the set area.

5. The ultrasonic water meter flow measurement method based on dynamic temperature compensation according to claim 1 is characterized in that: The temperature compensation time series data includes the water temperature value, ambient temperature value, pipe wall conduction temperature hysteresis value, equipment self-heating compensation index, pipeline thermal expansion length change value, water body compressibility temperature response index of each sampling period. The specific steps for obtaining the temperature compensation factor of each sampling period in the set area are as follows: The temperature compensation time series data in the set area are comprehensively analyzed to obtain a set of temperature perception correction indexes for each sampling period in the set area, including a thermal field credible perception index and a thermally induced structural response index; A comprehensive analysis is performed on the temperature perception correction index set for each sampling period in the set area to obtain the temperature compensation factor for each sampling period in the set area.

6. The ultrasonic water meter flow measurement method based on dynamic temperature compensation according to claim 5 is characterized in that: The specific steps to obtain the temperature perception correction index set for each sampling period in the set area are as follows: Read the water temperature value, ambient temperature value, pipe wall conduction temperature hysteresis value, and equipment self-heating compensation index of each sampling period in the set area, and conduct a comprehensive analysis to obtain the thermal field credible perception index of each sampling period in the set area; The change value of the pipeline thermal expansion length and the water body compressibility temperature response index of each sampling period in the set area are read, and a comprehensive analysis is performed to obtain the thermally induced structural response index of each sampling period in the set area.

7. The ultrasonic water meter flow measurement method based on dynamic temperature compensation according to claim 5 is characterized in that: The specific formula for calculating the temperature compensation factor for a certain sampling period within the set area is as follows: ； in, , , They are the temperature compensation factor, thermal structure response index, and thermal field credible perception index of a certain sampling period in the set area. , , They are the thermally induced structure adjustment coefficient, thermal field credibility adjustment coefficient, and interaction adjustment coefficient stored in the database respectively.

8. The ultrasonic water meter flow measurement method based on dynamic temperature compensation according to claim 1 is characterized in that: The measurement interference time series data includes turbulence index, micro-vibration index, waveform stability entropy value, and transducer coupling index. The specific steps of obtaining the measurement interference suppression factor of each sampling period in the set area are as follows: Standardize the turbulence index, micro-vibration index, waveform stability entropy value, and transducer coupling index of each sampling period in the set area; Based on the particle optimization algorithm, the turbulence index, micro-vibration index, waveform stability entropy value and transducer coupling index of each sampling period in the set area after standardization are comprehensively analyzed to obtain the measurement interference suppression factor of each sampling period in the set area.

9. The ultrasonic water meter flow measurement method based on dynamic temperature compensation according to claim 1, characterized in that: The specific steps for intelligent alarming of the set area based on the actual flow measurement value of each sampling period are as follows: Compare and analyze the actual flow measurement value of each sampling period in the set area with the preset flow threshold; If the actual flow measurement value of each sampling period in the set area is lower than or equal to the preset flow threshold, no alarm will be issued; If the actual flow measurement value of each sampling period in the set area is higher than the preset flow threshold, an alarm will be issued and corresponding warning measures will be taken.

10. An ultrasonic water meter flow measurement system based on dynamic temperature compensation, applying the ultrasonic water meter flow measurement method based on dynamic temperature compensation according to any one of claims 1 to 9, characterized in that: include: A data acquisition module is used to continuously acquire flow measurement time series data within the setting, including ultrasonic sampling time series data, temperature compensation time series data, and measurement interference time series data; The feature extraction module is used to perform feature extraction processing on the flow measurement time series data in the set area to obtain the initial flow measurement value, temperature compensation factor and measurement interference suppression factor of each sampling period in the set area; A comprehensive correction module is used to comprehensively analyze the initial flow measurement value, temperature compensation factor, and measurement interference suppression factor of each sampling period in the set area to obtain the actual flow measurement value of each sampling period in the set area; The measurement feedback module is used to provide intelligent alarm for the set area based on the actual flow measurement value of each sampling period.

Images