Flowmeter rapid in-situ calibration processing method and system based on weighing method and medium
By combining the weighing method with Kalman filtering and differential fitting, the problem of switch action error in the in-situ calibration of flowmeters was solved, achieving higher accuracy flowmeter calibration and meeting the accuracy requirements of pump-turbine model tests.
Patent Information
- Application Number
- CN202511825213.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-05
- Publication Date
- 2026-03-27
AI Technical Summary
In existing in-situ calibration techniques for flow meters, errors caused by switch operation and related calibration methods have not been effectively eliminated, resulting in high uncertainty in in-situ sensor calibration, which cannot meet the accuracy requirements of pump-turbine model tests.
A rapid in-situ calibration method for flow meters based on weighing is adopted. The flow rate and weight signals are received through the data acquisition module, and differential processing and iterative fitting are performed after Kalman filtering to eliminate the switching device action error and reduce uncertainty.
This improved the accuracy and precision of the in-situ calibration of the flow meter, reduced the uncertainty of sensor measurements, and met the accuracy requirements of pump-turbine model tests.
Smart Images
Figure CN121740194A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of data processing technology, and in particular to a rapid in-situ calibration processing method, system and medium for flow meters based on the weighing method. Background Technology
[0002] As flow sensors are increasingly widely used, in-situ calibration is often employed to reduce measurement errors. This process aims to obtain the functional relationship between the flow meter signal and the flow rate, thereby correcting the sensor readings, improving the sensor's measurement accuracy, reducing measurement errors, and ensuring that the flow meter's accuracy meets the requirements of the test bench.
[0003] During the in-situ calibration of the flow sensor, the errors in the switching action and the calibration methods of related technologies were not eliminated, which increased the uncertainty of the in-situ calibration of the flow sensor and reduced the accuracy of the in-situ calibration. Summary of the Invention
[0004] In view of this, the purpose of this application is to propose a rapid in-situ calibration method, system and medium for flow meters based on the weighing method, so as to solve or partially solve the above-mentioned technical problems.
[0005] To achieve the above objectives, this application provides a rapid in-situ calibration processing method for flow meters based on the weighing method, which is applied to an analysis system. The analysis system includes: a data acquisition module, a filter analysis and calculation module, and an in-situ calibration analysis module connected in sequence, wherein the data acquisition module is connected to the flow meter sensor and the weight sensor. The method includes: Using the data acquisition module, the flow detection signal from the flow meter sensor and the weight detection signal from the weight sensor are received, and the flow detection signal and weight detection signal are sent to the filtering analysis and calculation module. The weight detection signal is processed by Kalman filtering using the filtering analysis and calculation module to obtain a predicted weight detection signal, and the predicted weight detection signal is sent to the in-situ calibration analysis module. The flow detection signal is processed by Kalman filtering using the filtering analysis and calculation module to obtain a predicted flow detection signal, and the predicted flow detection signal is sent to the in-situ calibration and analysis module. Using the in-situ calibration analysis module, the predicted flow detection signal and the predicted weight detection signal are subjected to pre-processing differential processing to obtain flow differential data and weight differential data. Based on the flow differential data and weight differential data, iterative fitting processing is performed until the first convergence condition is met to determine the target in-situ calibration data. Based on the target in-situ calibration data, the flow meter sensor is calibrated in-situ.
[0006] Based on the same inventive concept, this application also provides an analysis system, including: a data acquisition module, a filtering analysis and calculation module and an in-situ calibration analysis module connected in sequence, wherein the data acquisition module is connected to a flow meter sensor and a weight sensor; The data acquisition module is configured to receive the flow detection signal from the flow meter sensor and the weight detection signal from the weight sensor, and send the flow detection signal and the weight detection signal to the filtering analysis and calculation module. The filtering analysis and calculation module is configured to perform Kalman filtering on the weight detection signal to obtain a predicted weight detection signal, and send the predicted weight detection signal to the in-situ calibration analysis module. The filtering analysis and calculation module is configured to perform Kalman filtering on the flow detection signal to obtain a predicted flow detection signal, and send the predicted flow detection signal to the in-situ calibration analysis module. The in-situ calibration analysis module is configured to perform pre-processing on the predicted flow detection signal and the predicted weight detection signal to obtain flow difference data and weight difference data, perform iterative fitting processing on the flow difference data and weight difference data until the first convergence condition is met to determine the target in-situ calibration data, and perform in-situ calibration on the flow meter sensor based on the target in-situ calibration data.
[0007] Based on the same inventive concept, this application also provides a non-transitory computer-readable storage medium that stores computer instructions for causing a computer to perform the method described above.
[0008] As can be seen from the above, the rapid in-situ calibration processing method, system, and medium for flow meters based on the weighing method provided in this application can utilize a data acquisition module to receive the flow detection signal from the flow meter sensor and the weight detection signal from the weight sensor. The flow detection signal and weight detection signal are then sent to a filtering analysis and calculation module. The filtering analysis module then processes the weight detection signal and flow detection signal using Kalman filtering to obtain relatively accurate predicted flow detection signal and predicted weight detection signal. Finally, the in-situ calibration analysis module performs pre-difference processing on the predicted flow detection signal and predicted weight detection signal obtained from the filtering analysis and calculation module. Based on the obtained flow difference data and weight difference data, iterative fitting processing is performed until the first convergence condition is met, at which point the iterative fitting process stops. The final result of the iterative fitting process is used as the target in-situ calibration data. This target in-situ calibration data is more accurate, and using this target in-situ calibration data for the flow meter sensor will also be more accurate, reducing the uncertainty of in-situ calibration of the flow meter sensor. Attached Figure Description
[0009] To more clearly illustrate the technical solutions in this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0010] Figure 1 This is a schematic flowchart of the rapid in-situ calibration method for flow meters based on the weighing method according to an embodiment of this application. Figure 2 This is a logical schematic diagram of the rapid in-situ calibration method for flow meters based on the weighing method according to an embodiment of this application; Figure 3 This is a schematic diagram of the analysis system according to an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation
[0011] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.
[0012] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this application should have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," and similar words used in the embodiments of this application do not indicate any order, The terms "include" or "contain" are used to distinguish different components, not to indicate quantity or importance. Words like "include" or "contain" mean that the element or object preceding the word covers the elements or objects listed after it, and their equivalents, without excluding other elements or objects. Words like "connect" or "link" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms like "up," "down," "left," and "right" are used only to indicate relative positional relationships; these relative relationships may change if the absolute position of the described object changes.
[0013] Definitions: In-situ calibration is the correction of a measuring instrument using a higher-precision instrument under normal installation conditions at the test site. The accuracy of the flow meter after in-situ calibration is related to the accuracy of the higher-precision instrument. GB / T 15613-2023 indicates that in-situ calibration can be used as a method to ensure flow accuracy on a pump-turbine model test bench.
[0014] The weighing method: Calibration using the weighing method involves passing a known mass of liquid or gas into the flow meter, then measuring the difference between the displayed value and the actual mass to calculate a calibration regression curve (a linear equation). Accurately calculating the calibration regression curve improves the accuracy and reliability of flow meter measurements, ensuring the flow meter's accuracy in practical applications.
[0015] Overall error: Overall error reflects the overall error of the entire measurement system composed of instruments, including the systematic error of the instruments and the random error introduced during measurement.
[0016] Systematic errors: Systematic errors are measurement errors caused by inherent factors in the analytical process. In this article, they mainly refer to measurement errors caused by the inherent precision limitations of the instruments themselves.
[0017] Random error: refers to the error that is mutually compensating due to the small random fluctuations of a series of related factors during the measurement process.
[0018] In-situ calibration uncertainty: The uncertainty of the calibration regression curve calculated after in-situ calibration of the existing measuring instrument.
[0019] Normal Distribution: Also known as the normal distribution or normal range distribution, it is usually denoted as X ~ N(μ, σ²). Here, μ is the expected value (mean) of the normal distribution, and σ² is the variance of the normal distribution. Random errors in the measurement process of sensors and measuring instruments typically conform to a normal distribution.
[0020] Kalman filtering is a technique used in signal denoising where sensor measurements often follow a normal distribution after repeated measurements. Kalman filtering is an algorithm that uses the state equations of a linear system and Gaussian function fusion to optimally estimate the system state based on the system's input and output observation data. In signal denoising, the observed values can be considered as noisy signals, while the state estimate represents the clean signal we desire. Kalman filtering estimates the clean signal by fusing the observed signal and the state prediction.
[0021] Kernel density estimation is a nonparametric method used to estimate the probability density function of data. Based on a kernel function, it estimates the probability density of each data point by weighting the kernel functions near each data point with a certain bandwidth parameter; that is, it makes inferences about the population based on a finite data sample. The Gaussian kernel function is a recommended function used to describe Gaussian distributed samples.
[0022] When conducting pump-turbine runner model tests, it is typically required that the overall efficiency measurement error be less than ±0.25% near the highest efficiency point of the runner model. For a pump-turbine model test bench, the overall efficiency measurement error originates from random errors in efficiency measurement, systematic errors in flow rate measurement, systematic errors in head measurement, and systematic errors in torque measurement. Among these, the systematic error in flow rate measurement is the primary influencing factor on the overall efficiency error. Therefore, the systematic error in flow rate measurement on the model test bench should not exceed ±0.25%, and the smaller the error, the better. Due to practical limitations, pump-turbine test benches cannot guarantee the configuration of sufficiently high-precision flow sensors. In this case, to reduce the error during flow sensor measurement, in-situ calibration is often used to obtain the functional relationship between the flowmeter signal and the flow rate, thereby correcting the sensor reading, improving the sensor's measurement accuracy, reducing measurement errors, and ensuring that the flowmeter accuracy meets the accuracy requirements of the test bench.
[0023] The gravimetric method is a common method for in-situ calibration of flow sensors. It involves using a gravimetric sensor to measure the weight of water flowing into an open weighing tank over a period of time, passing through the instrument being calibrated. The actual volume of water in the tank is calculated using the mass-volume formula, and then divided by the reading on a timer to obtain the standard flow rate. When calibrating an electromagnetic flowmeter in situ using the gravimetric method, before calibration, a water pump supplies water to the test system. The water flows from a constant-level water tank, through the flowmeter being calibrated, and back to the water tank via a commutator and outlet. After the operating conditions stabilize, the water is allowed to flow through the commutator into the weighing tank. After a period of time, the commutator resets. The mass of water collected in the weighing tank during this period and the water exchange time are recorded, and the time average is calculated to determine the standard volumetric flow rate and the time average of the sensor signal, thus completing the in-situ calibration of the flowmeter under a specific flow condition.
[0024] By using the weighing method, testers can obtain standard flow values corresponding to multiple flow sensor signals. Then, through linear fitting, they can finally obtain a linear functional relationship between the flow sensor signal and the flow rate, correct the flow meter reading, and reduce measurement system errors.
[0025] A typical pipeline system for in-situ calibration of a flow meter using the weighing method is as follows: Figure 1As shown. The liquid is driven by a water supply pump and flows through a water pipe into a pressure stabilizing container, the flow meter under calibration, and a switch, before flowing into the weighing tank. The total injection time is recorded by a timer. After reading the timer reading, the weighing sensor reading located below the weighing container, and the cumulative flow value of the flow meter under test, the liquid is then released into the return water pipeline through the drain valve below the container. Before the flow meter on the test bench is calibrated, water is supplied to the test system by the water supply pump, and the water returns to the water supply tank through the drain port via the switch. After the flow rate stabilizes, the switch activates to guide water into the weighing tank, and resets after a set time interval. The standard flow rate is calculated by taking the water mass collected in the large weighing container (hereinafter referred to as the "large tank") during this period and the injection time, and matched with the flow meter reading to complete the in-situ calibration of a certain flow point.
[0026] If the in-situ calibration of the flow sensor is carried out using the weighing method, the in-situ calibration error mainly comes from factors such as errors caused by potential equipment leakage, water injection errors due to commutator operation, time measurement errors, weighing tank errors, and density measurement errors. Among these, errors caused by potential equipment leakage and water injection errors due to commutator operation are the main influencing factors on the in-situ calibration error of the flow meter. Reducing these errors helps to reduce the flow measurement uncertainty of the test bench.
[0027] Although current in-situ calibration methods using weighing require testing after the pipeline flow rate has stabilized, in actual measurements, flow pulsations and noise will cause the quasi-transient flow rate readings measured by the flowmeter to fluctuate continuously. Furthermore, persistent hidden leaks in the equipment will also lead to overall measurement inaccuracies. Current methods reduce the impact of noise on the measured values by calculating the time-averaged values of the flowmeter readings and the time-averaged changes in the water volume in the weighing tank. This involves measuring the total flowmeter reading and the time-averaged mass of injected purified water under multiple operating conditions within the same time interval, and establishing a calibration regression curve between the flowmeter readings and standard flow data. However, the time-averaging of data over the entire time interval will incorporate the mismatch between water volume and flowmeter readings caused by switch operations, and it does not guarantee the complete elimination of the effects of noise, flow pulsations, and pipeline leaks, ultimately leading to increased errors.
[0028] In summary, current in-situ calibration techniques for flow meters have the following two problems: In current weighing methods, it is suggested that errors caused by the switching device's operation be addressed by setting the timing point at the highest water volume point. However, this is difficult to implement in practice, and the timing point can only be set at the midpoint of the switching device's operation. There is no reasonable and scientific technology to eliminate the errors caused by this setting.
[0029] In in-situ calibration of flow sensors, errors caused by factors such as load cells, switch actions, fluctuations in flow meter data, and pipeline leaks are reduced rather than eliminated through time-averaging. However, this method cannot determine the total calibration time in advance, nor can it detect potential cumulative errors during the calibration process. Innovative technologies are needed to identify and eliminate these errors.
[0030] The main technical problem is that while the pump-turbine model test requires controllable overall system error and individual sensor error levels, errors from the switch's operation and the inability of the calibration method to eliminate sensor in-situ calibration uncertainty during the flowmeter's in-situ calibration process hinder the model test. Currently, in the model test, the flowmeter is calibrated in-situ using a weighing method. The water volume calculated from the water mass and the flowmeter signal output are time-averaged, and then a least-squares linear fit is performed to obtain the sensor calibration regression curve. The core idea of this strategy is to perform time-averaged filtering on the measurement data under each flow condition and then assign equal weights to all filtered data. However, in actual measurement, if abnormal water loss or switch malfunction occurs under a certain condition, systematic deviations will appear in the data points. In this case, the equal-weighted fitting method cannot eliminate the errors.
[0031] In summary, the methods of the relevant technologies neither consider the measurement uncertainty of the weighing barrel during the in-situ calibration process, nor reasonably eliminate the uncertainty of sensor measurement, the error caused by the switching action, etc., nor have the ability to eliminate cumulative and trend-type errors.
[0032] The embodiments of this application will be described in detail below with reference to the accompanying drawings.
[0033] The embodiments of this application provide a rapid in-situ calibration processing method for flow meters based on the weighing method, which is applied to an analysis system. The analysis system includes a data acquisition module, a filtering analysis and calculation module, and an in-situ calibration analysis module connected in sequence. The data acquisition module is connected to the flow meter sensor and the weight sensor.
[0034] like Figure 1 Combination Figure 2 As shown, the method includes: Step 101: Using the data acquisition module, receive the flow detection signal from the flow meter sensor and the weight detection signal from the weight sensor, and send the flow detection signal and weight detection signal to the filtering analysis and calculation module.
[0035] In practical implementation, the data acquisition module mainly receives the following two signals: real-time weight detection signals from a weight sensor (e.g., a large weighing cylinder sensor) (e.g., the reading of the large weighing cylinder sensor), and real-time signals from a flow sensor (e.g., an electromagnetic flow meter sensor) (e.g., the electromagnetic flow meter sensor is set to output pulse count and therefore outputs a pulse count signal). Both the weight sensor and the flow sensor readings are transmitted to the data acquisition module in digital signal form.
[0036] The data acquisition module calculates and records the average value and measurement deviation of the flow detection signal and weight detection signal. If the obtained digital signals from the weight sensor and flow sensor are stable, the measurement deviation of the digital signal is set to... ( (This corresponds to the smallest scale of the sensor).
[0037] Step 102: Using the filtering analysis and calculation module, perform Kalman filtering on the weight detection signal to obtain a predicted weight detection signal, and send the predicted weight detection signal to the in-situ calibration analysis module.
[0038] In practice, a filtering analysis and calculation module is added after the data acquisition module. The filtering analysis and calculation module performs Kalman filtering on the weight detection signal and outputs an estimated predicted weight detection signal.
[0039] Step 103: Using the filtering analysis and calculation module, perform Kalman filtering on the flow detection signal to obtain a predicted flow detection signal, and send the predicted flow detection signal to the in-situ calibration analysis module.
[0040] In practice, the filtering analysis and calculation module performs Kalman filtering on the flow detection signal and outputs an estimated predicted flow detection signal.
[0041] Step 104: Using the in-situ calibration analysis module, perform pre-processing on the predicted flow detection signal and the predicted weight detection signal to obtain flow difference data and weight difference data. Perform iterative fitting processing on the flow difference data and weight difference data until the first convergence condition is met to determine the target in-situ calibration data. Perform in-situ calibration on the flow meter sensor based on the target in-situ calibration data.
[0042] In practice, by considering the operating conditions of the flow sensor, determining the variable weights corresponding to the residual distribution based on the preceding differential processing, and carrying out iterative least squares fitting, the target in-situ calibration data (e.g., the slope and intercept of the regression curve) for in-situ calibration of the flow sensor is determined by using the flow differential data and weight differential data, thus realizing the in-situ calibration process of the flow sensor.
[0043] The above scheme allows the data acquisition module to receive flow detection signals from the flow meter sensor and weight detection signals from the weight sensor. These signals are then sent to the filtering and analysis module. The filtering module processes these signals using Kalman filtering to obtain relatively accurate predicted flow and weight signals. Finally, the in-situ calibration analysis module performs pre-difference processing on the predicted flow and weight signals obtained from the filtering and analysis module. Iterative fitting is then performed based on the obtained flow and weight difference data until the first convergence condition is met. The final result of this iterative fitting is used as the target in-situ calibration data. This results in more accurate target in-situ calibration data, leading to more accurate in-situ calibration of the flow meter sensor and reducing the uncertainty of in-situ calibration.
[0044] In some embodiments, step 101 includes: Step 1011: Use the data acquisition module to perform noise reduction processing on the initial flow detection signal sent by the flow meter sensor to obtain the flow detection signal.
[0045] Step 1012: Use the data acquisition module to perform noise reduction processing on the initial weight detection signal sent by the weight sensor to obtain the weight detection signal.
[0046] Step 1013: The data acquisition module sends the flow detection signal and the weight detection signal to the filtering analysis and calculation module.
[0047] Through the above scheme, the data acquisition module can filter and denoise the initial flow detection signal to remove some noise interference, thereby obtaining a clearer flow detection signal. It can also filter and denoise the initial weight detection signal to remove some noise interference, thereby obtaining a clearer weight detection signal. Then, the clear and accurate flow detection signal and weight detection signal are sent to the filtering analysis and calculation module for processing, so that the subsequent filtering analysis and calculation module will not be affected by noise interference and can be more accurate.
[0048] In some embodiments, step 102 is performed using the filter analysis calculation module: Step 1021: Determine the predicted weight value at the previous moment, determine the first-order difference mean of all predicted weight values at the current moment, determine the liquid drop at the current moment and the corresponding impact force correction term, and predict the actual weight value at the current moment based on the liquid drop and the impact force correction term.
[0049] In practice, the weight detection signal of the weight sensor is determined to be M(k).
[0050] Initialize the state covariance P=1 (since the data is a scalar, P is a first-order matrix, i.e., a scalar). Based on the accuracy of the weight sensor and the expected observations, initialize the system noise Q and the observation noise R.
[0051] Determine the predicted weight value of the previous time k-1 at the current time k. , where M_kfe(1) = M(1) at the first moment.
[0052] The formula for predicting the true weight reading Mpre at the current moment is: ; Where avg_diff is the first-order difference mean of the weight detection signal, G is the impact force correction term, and w is the noise term.
[0053] ; in, Let q be the liquid density, q = [Z(k) - Z(k-1)] / [t(k) - t(k-1)], t(k) be the current time of the timer, t(k-1) be the previous time of the timer, A1 be the outlet area of the load cell switch, A2 be the cross-sectional area of the load cell weighing barrel, Z(k) be the liquid level observation value of the load cell at the current time k, Z(k-1) be the liquid level observation value of the load cell at the previous time k-1, and Z(k) - Z(k-1) be the liquid drop at the current time.
[0054] Step 1022: Update the state covariance at the current time and determine the Kalman gain value based on the updated state covariance.
[0055] In practice, the state covariance P is updated based on the current state covariance P, and the updated state covariance PP is determined using the following formula: PP = F1 × P × F1' + Q; where, since the values processed by the Kalman filter are scalars, the state transition matrix F1 is a first-order matrix (i.e., F1 = 1), F1' is the first derivative of F1, and Q is the system noise.
[0056] The Kalman gain K1 is determined by the following formula: K1 = PP × H' / (H × PP × H' + R); where H is the observation matrix, which is a first-order matrix, H = 1, H' is the first derivative of H, and R is the observation noise.
[0057] Step 1023: Combine the actual weight value at the current moment, the Kalman gain value, and the weight detection signal to determine the predicted weight value at the current moment.
[0058] In practical implementation, the formula for the predicted weight indication M_kfe(k) at the current moment is: M_kfe(k)=Mpre+K1×(M(k)-Mpre).
[0059] Then, the updated state covariance PP is updated again to obtain the state covariance P to be used in the next time step, with the formula P = PP - K1 × H × PP.
[0060] Step 1024: Determine the first-order difference mean of the predicted weight value at the current moment, obtain the first difference of the first-order difference mean of two consecutive adjacent moments, and determine that the absolute value of the first difference of the two groups is less than the first difference threshold, then control the first value of the counter to be accumulated.
[0061] In practice, the first difference between the two groups is: avg_diff k - avg_diff k-1 ; avg_diff k+1 - avg_diff k .
[0062] Confirm | avg_diff k - avg_diff k-1 |<ε1 and|avg_diff k+1 - avg_diff k If |<ε1, where ε1 is the first difference threshold, then the first value S of the corresponding counter increases by 1, i.e., S=S+1.
[0063] Step 1025: In response to the first value accumulated by the counter being less than the first counting threshold, the state covariance at the next moment is adjusted, and Kalman filtering is iteratively performed based on the state covariance at the next moment to predict the predicted weight value at the next moment. During the iteration process, the first value of the counter is controlled to accumulate.
[0064] In practice, if S < the first counting threshold (e.g., 10), then the process of steps 1021 to 1024 above is repeated based on the state covariance of the next time step determined in step 1023.
[0065] Step 1026: In response to the first value accumulated by the counter being greater than or equal to the first counting threshold, it is determined that the second convergence condition is met. The last determined predicted weight values are sequence-integrated to obtain the predicted weight detection signal, and the predicted weight detection signal is sent to the in-situ calibration analysis module.
[0066] In practice, if S ≥ the first counting threshold (e.g., 10), the iteration stops, and the final determined predicted weight values are sequence-integrated to obtain the predicted weight detection signal. And the predicted weight detection signal Send to the in-situ calibration analysis module.
[0067] The above method enables accurate Kalman filtering of the weight detection signal, thereby improving the accuracy of the predicted weight detection signal.
[0068] In some embodiments, step 103 is performed using the filter analysis calculation module: Step 1031: Determine the predicted flow rate at the previous moment and predict the actual flow rate at the current moment.
[0069] In practice, the flow detection signal of the flow sensor is determined to be P(k).
[0070] Initialize state covariance P (1) =1 (because the data is a scalar, P) (1) (The system noise Q is a first-order matrix, i.e., a scalar). Based on the accuracy of the weight sensor and the expected observations, the system noise Q is initialized. (1) and observation noise R (1) .
[0071] Determine the predicted flow rate P_kfe(k-1) of the previous time k-1 of the current time k, where P_kfe(1) = P(1) of the first time.
[0072] The formula for predicting the true weight reading Ppre at the current moment is: Ppre = P_kfe(k-1) + avg_diff (1) -G (1) +w (1) ; Among them, avg_diff (1) G is the first-order difference mean of the weight detection signal. (1) For impact force correction, w (1) This is the noise term.
[0073] Step 1032: Update the state covariance at the current time and determine the Kalman gain value based on the updated state covariance.
[0074] In practice, the state covariance P at the current moment is used as the basis for implementation. (1) Perform an update and determine the updated state covariance PP. (1) The formula is: PP (1) = F1 × P (1) × F1' + Q (1) Since the values processed by the Kalman filter are scalars, the state transition matrix F1 is a first-order matrix (i.e., F1=1), F1' is the first derivative of F1, and Q... (1) This represents system noise.
[0075] Determine the Kalman gain K1 (1) The formula is: K1 (1) =PP (1) × H' / (H × PP (1) × H' + R (1) ); where H is the observation matrix, which is a first-order matrix, H=1, H' is the first derivative of H, and R (1) To observe noise.
[0076] Step 1033: Combine the actual flow rate at the current moment, the Kalman gain value, and the flow rate detection signal to determine the predicted flow rate at the current moment.
[0077] In practical implementation, the formula for the predicted weight indication P_kfe(k) at the current moment is: P_kfe(k) = Ppre + K1 (1) × (P(k) -Ppre).
[0078] Then the updated state covariance PP is calculated. (1) Continue updating to obtain the state covariance P needed for the next time step. (1) The formula is P (1) =PP (1) - K1 (1) × H × PP (1) .
[0079] Step 1034: Determine the first-order difference mean of the predicted flow rate at the current time, obtain the second difference of the first-order difference mean of two consecutive adjacent times, and determine that the absolute value of the second difference of the two groups is less than the second difference threshold, then control the second value of the counter to accumulate.
[0080] In practice, the first difference between the two groups is: avg_diff k - avg_diff k-1 ; avg_diff k+1 - avg_diff k .
[0081] Confirm | avg_diff k - avg_diff k-1 |<ε2 and|avg_diff k+1 - avg_diff k If |<ε2, where ε2 is the second difference threshold, then the corresponding second value S of the counter. (1) Increase by 1, i.e., S (1) =S (1) +1.
[0082] Step 1035: In response to the second value accumulated by the counter being less than the second counting threshold, the state covariance at the next moment is adjusted, and Kalman filtering is iteratively performed based on the state covariance at the next moment to predict the predicted flow rate at the next moment. During the iteration process, the second value of the counter is controlled to accumulate.
[0083] In specific implementation, if S (1) If the count threshold is less than 10, then the process of steps 1031 to 1034 above is repeated based on the state covariance of the next time step determined in step 1033.
[0084] Step 1036: In response to the second value accumulated by the counter being greater than or equal to the second counting threshold, it is determined that the third convergence condition is met. The final determined predicted flow rate values are sequence-integrated to obtain the predicted flow rate detection signal, and the predicted flow rate detection signal is sent to the in-situ calibration analysis module.
[0085] In specific implementation, if S (1) If the predicted weight value is greater than or equal to the second counting threshold (e.g., 10), the iteration stops, and the final determined predicted weight value is sequence-integrated to obtain the predicted flow detection signal P_kfe(k). The predicted flow detection signal P_kfe(k) is then sent to the in-situ calibration analysis module.
[0086] The above scheme enables accurate Kalman filtering of the flow detection signal, thereby improving the accuracy of the predicted flow detection signal.
[0087] In some embodiments, step 104 is performed using the in-situ calibration analysis module: Step 1041: Perform pre-processing on the predicted flow rate detection signal and the predicted weight detection signal to obtain flow rate difference data △P_kfe(k) and weight difference data △M_kfe(k).
[0088] Step 1042: Mark the flow difference data and weight difference data under different flow conditions, and determine the first-order difference mean of the predicted flow detection signal to obtain a difference array. The difference array includes: flow difference data for each flow condition, volume difference data corresponding to the weight difference data for each flow condition, and the first-order difference mean of the predicted flow detection signal.
[0089] In practice, the difference array is (△P_kfe(k), △M_kfe(k) / ρ, avg_diff). k ), avg_diff k This is the first-order difference mean of the predicted flow detection signal P_kfe(k).
[0090] Step 1043, iterative execution: determine the current residual, perform least squares fitting on the difference array based on the current residual, and obtain the in-situ calibration data corresponding to the fitting result.
[0091] In some embodiments, step 1043 includes: Step 10431, determine the difference array as follows ,in For traffic differential data, For weight difference data, To determine the density of a liquid, This is volume difference data. The first-order difference mean of the flow detection signal is used to predict the flow rate. k is the total number of parameters in the difference array.
[0092] Step 10432, determine the current residual. , where i is the index of each parameter in the difference array.
[0093] In practice, The residuals between the parameters and the fitted function at each time step.
[0094] Step 10433: Perform least squares fitting on the difference array to obtain the fitting result: .
[0095] Where, for the initial slope a1 (0) With the initial intercept b1 (0) It needs to meet the following requirements. .
[0096] Initial slope a1 (0) With the initial intercept b1 (0) The corresponding formula is: ; .
[0097] in, For data points under different flow conditions, , where m is the total number of all parameters; ; ; ; ; ; .
[0098] The least squares fitting method can be viewed as assigning weights to each parameter based on its flow conditions. Subsequent weight assignments are then based on this weighting.
[0099] In the fitting results obtained by the least squares method, The slope The intercept is... and The following conditions must be met: , The weights of each parameter determined in the previous n-1 iteration.
[0100] In some embodiments, the weights of each parameter The process of determining is as follows: Determine weights using kernel density estimation algorithm The formula is: ; Where h is the bandwidth, , For the residual with current index i, Here, j is the reference residual, and j is the index of the reference residual.
[0101] The above method can yield accurate weight values.
[0102] Step 10434: Determine based on the formula of the fitting result. and This serves as the in-situ calibration data.
[0103] Through the above solution Step 1044: In response to determining that the final in-situ calibration data satisfies the first convergence condition, stop the iteration and use the final in-situ calibration data as the target in-situ calibration data.
[0104] In some embodiments, the first convergence condition includes: The absolute value of the slope difference after the last two iterations. satisfy And the absolute value of the intercept difference after the last two iterations satisfy Where n is the number of iterations. The slope difference threshold. This is the threshold value for the intercept difference.
[0105] Step 1045: Perform in-situ calibration of the flow meter sensor based on the target in-situ calibration data.
[0106] In practice, the final slope a1 will be obtained. (n) and intercept b1 (n) Output, so that the slope a1 can be utilized (n) and intercept b1 (n) The calibration function is used to perform in-situ calibration of the flow meter sensor.
[0107] The effect of this application is: The calibration regression curves of the flow sensor at various operating points and times during the in-situ calibration process were estimated and calculated. The error caused by the switch action can be gradually eliminated through the iterative nested weighted fitting method, which can effectively reduce the uncertainty caused by the system design itself during the in-situ calibration process. At the same time, the correct output of the switch cut-off command is achieved by the convergence judgment in the filtering analysis calculation module, which shortens the in-situ calibration time of the flow sensor.
[0108] The solution proposed in this application has the ability to eliminate interference factors such as switch operation and water flow loss, effectively reducing the uncertainty caused by the system's own design during the in-situ calibration process of the weighing method, and enhancing the anti-interference capability of the analysis system.
[0109] The uncertainty of the calibration function of the flow sensor obtained by the solution in this application is significantly smaller than that of the prior art solution, and the calibration function of the flow sensor obtained is more reliable.
[0110] It should be noted that the method in this embodiment can be executed by a single device, such as a computer or server. The method can also be applied in a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method in this embodiment, and the multiple devices will interact with each other to complete the method described.
[0111] It should be noted that the above description describes some embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0112] Based on the same inventive concept, and corresponding to the methods of any of the above embodiments, this application also provides an analysis system.
[0113] refer to Figure 3 The analysis system includes: a data acquisition module 301, a filter analysis and calculation module 302, and an in-situ calibration analysis module 303 connected in sequence, wherein the data acquisition module 301 is connected to a flow meter sensor and a weight sensor; The data acquisition module 301 is configured to receive the flow detection signal from the flow meter sensor and the weight detection signal from the weight sensor, and send the flow detection signal and the weight detection signal to the filtering analysis and calculation module 302. The filtering analysis and calculation module 302 is configured to perform Kalman filtering on the weight detection signal to obtain a predicted weight detection signal, and send the predicted weight detection signal to the in-situ calibration analysis module 303. The filtering analysis and calculation module 302 is configured to perform Kalman filtering on the flow detection signal to obtain a predicted flow detection signal, and send the predicted flow detection signal to the in-situ calibration analysis module 303. The in-situ calibration analysis module 303 is configured to perform pre-processing on the predicted flow detection signal and the predicted weight detection signal to obtain flow difference data and weight difference data, perform iterative fitting processing on the flow difference data and weight difference data until the first convergence condition is met to determine the target in-situ calibration data, and perform in-situ calibration on the flow meter sensor based on the target in-situ calibration data.
[0114] In some embodiments, the filter analysis calculation module 302 is specifically configured as follows: Determine the predicted weight value at the previous moment, determine the first difference mean of all predicted weight values at the current moment, determine the liquid drop at the current moment and the corresponding impact force correction term, and predict the actual weight value at the current moment based on the liquid drop and the impact force correction term. Update the state covariance at the current moment, and determine the Kalman gain value based on the updated state covariance; The predicted weight value at the current moment is determined by combining the actual weight value at the current moment, the Kalman gain value, and the weight detection signal. Determine the first-order difference mean of the predicted weight value at the current moment, obtain the first difference of the first-order difference mean of two consecutive adjacent moments, and determine that the absolute value of the first difference of the two groups is less than the first difference threshold, then control the first value of the counter to be accumulated. In response to the first value accumulated by the counter being less than the first counting threshold, the state covariance at the next moment is adjusted, and Kalman filtering is iteratively performed based on the state covariance at the next moment to predict the predicted weight value at the next moment. During the iteration process, the first value of the counter is controlled to accumulate. In response to the first value accumulated by the counter being greater than or equal to the first counting threshold, it is determined that the second convergence condition is met. The final determined predicted weight values are then sequence-integrated to obtain the predicted weight detection signal, which is then sent to the in-situ calibration analysis module 303.
[0115] In some embodiments, the filter analysis calculation module 302 is further configured as follows: Determine the predicted flow rate at the previous moment, and predict the actual flow rate at the current moment; Update the state covariance at the current moment, and determine the Kalman gain value based on the updated state covariance; The predicted flow rate at the current moment is determined by combining the actual flow rate at the current moment, the Kalman gain value, and the flow rate detection signal. Determine the first-order difference mean of the predicted flow rate at the current moment, obtain the second difference of the first-order difference mean of two consecutive adjacent moments, and determine that the absolute value of the second difference of the two groups is less than the second difference threshold, then control the second value of the counter to accumulate. In response to the second value accumulated by the counter being less than the second counting threshold, the state covariance at the next moment is adjusted, and Kalman filtering is iteratively performed based on the state covariance at the next moment to predict the predicted flow rate at the next moment. During the iteration process, the second value of the counter is controlled to accumulate. In response to the second value accumulated by the counter being greater than or equal to the second counting threshold, it is determined that the third convergence condition is met. The final determined predicted flow rate values are then sequence-integrated to obtain the predicted flow rate detection signal, which is then sent to the in-situ calibration analysis module 303.
[0116] In some embodiments, the in-situ calibration analysis module 303 is specifically configured as follows: The predicted flow rate detection signal and the predicted weight detection signal are subjected to pre-sequence differential processing to obtain flow rate differential data and weight differential data, respectively. The flow differential data and weight differential data under different flow conditions are labeled, and the first-order differential mean of the predicted flow detection signal is determined to obtain a differential array. The differential array includes: flow differential data for each flow condition, volume differential data corresponding to the weight differential data for each flow condition, and the first-order differential mean of the predicted flow detection signal. Iterative execution: Determine the current residual, and perform least squares fitting on the difference array based on the current residual to obtain the in-situ calibration data corresponding to the fitting result; In response to the determination that the final in-situ calibration data satisfies the first convergence condition, the iteration stops, and the final in-situ calibration data is used as the target in-situ calibration data; The flow meter sensor is calibrated in situ based on the target in-situ calibration data.
[0117] In some embodiments, the in-situ calibration analysis module 303 is further configured as follows: Determine the difference array as ,in For traffic differential data, For weight difference data, To determine the density of a liquid, This is volume difference data. The first-order difference mean of the flow detection signal is used to predict the flow rate detection signal, where k is the total number of parameters in the difference array. Determine the current residual , where i is the index of each parameter in the difference array; The least squares method is used to fit the difference array to obtain the fitting result: ; in, The slope The intercept is... and The following conditions must be met: , The weights of each parameter determined in the previous n-1 iteration; Based on the formula of the fitting result, determine and This serves as the in-situ calibration data.
[0118] In some embodiments, the weights of the various parameters The process of determining is as follows: Determine weights using kernel density estimation algorithm The formula is: ; Where h is the bandwidth, , For the residual with current index i, Here, j is the reference residual, and j is the index of the reference residual.
[0119] In some embodiments, the first convergence condition includes: The absolute value of the slope difference after the last two iterations. satisfy And the absolute value of the intercept difference after the last two iterations satisfy Where n is the number of iterations. The slope difference threshold. This is the threshold value for the intercept difference.
[0120] In some embodiments, the data acquisition module 301 is specifically configured as follows: The initial flow detection signal from the flow meter sensor is denoised to obtain the flow detection signal. The initial weight detection signal from the weight sensor is denoised to obtain the weight detection signal. The flow detection signal and the weight detection signal are sent to the filter analysis and calculation module 302.
[0121] For ease of description, the above system is described by dividing it into various modules based on their functions. Of course, in implementing this application, the functions of each module can be implemented in one or more software and / or hardware.
[0122] The system described in the above embodiments is used to implement the corresponding method in any of the foregoing embodiments and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0123] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the methods described in any of the above embodiments.
[0124] Figure 4 This embodiment illustrates a more specific hardware structure of an electronic device. The device may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.
[0125] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.
[0126] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.
[0127] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.
[0128] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).
[0129] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.
[0130] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.
[0131] The electronic devices described above are used to implement the corresponding methods in any of the foregoing embodiments and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0132] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a non-transitory computer-readable storage medium that stores computer instructions for causing the computer to perform the methods described in any of the above embodiments.
[0133] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.
[0134] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to perform the methods described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0135] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this application is limited to these examples; under the concept of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this application as described above, which are not provided in detail for the sake of brevity.
[0136] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this application, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this application, and this also takes into account the fact that the details of the implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this application will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this application, it will be apparent to those skilled in the art that the embodiments of this application can be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.
[0137] Although this application has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.
[0138] The embodiments of this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the claims of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this application should be included within the protection scope of this application.
Claims
1. A rapid in-situ calibration method for flow meters based on the weighing method, characterized in that, The system is applied to an analysis system, which includes: a data acquisition module, a filter analysis and calculation module, and an in-situ calibration analysis module connected in sequence, wherein the data acquisition module is connected to a flow meter sensor and a weight sensor. The method includes: Using the data acquisition module, the flow detection signal from the flow meter sensor and the weight detection signal from the weight sensor are received, and the flow detection signal and weight detection signal are sent to the filtering analysis and calculation module. The weight detection signal is processed by Kalman filtering using the filtering analysis and calculation module to obtain a predicted weight detection signal, and the predicted weight detection signal is sent to the in-situ calibration analysis module. The flow detection signal is processed by Kalman filtering using the filtering analysis and calculation module to obtain a predicted flow detection signal, and the predicted flow detection signal is sent to the in-situ calibration and analysis module. Using the in-situ calibration analysis module, the predicted flow detection signal and the predicted weight detection signal are subjected to pre-processing differential processing to obtain flow differential data and weight differential data. Based on the flow differential data and weight differential data, iterative fitting processing is performed until the first convergence condition is met to determine the target in-situ calibration data. Based on the target in-situ calibration data, the flow meter sensor is calibrated in-situ.
2. The method according to claim 1, characterized in that, The step of using the filtering analysis and calculation module to perform Kalman filtering on the weight detection signal to obtain a predicted weight detection signal, and then sending the predicted weight detection signal to the in-situ calibration analysis module, includes: Execute using the filter analysis calculation module: Determine the predicted weight value at the previous moment, determine the first difference mean of all predicted weight values at the current moment, determine the liquid drop at the current moment and the corresponding impact force correction term, and predict the actual weight value at the current moment based on the liquid drop and the impact force correction term. Update the state covariance at the current moment, and determine the Kalman gain value based on the updated state covariance; The predicted weight value at the current moment is determined by combining the actual weight value at the current moment, the Kalman gain value, and the weight detection signal. Determine the first-order difference mean of the predicted weight value at the current moment, obtain the first difference of the first-order difference mean of two consecutive adjacent moments, and determine that the absolute value of the first difference of the two groups is less than the first difference threshold, then control the first value of the counter to be accumulated. In response to the first value accumulated by the counter being less than the first counting threshold, the state covariance at the next moment is adjusted, and Kalman filtering is iteratively performed based on the state covariance at the next moment to predict the predicted weight value at the next moment. During the iteration process, the first value of the counter is controlled to accumulate. In response to the first value accumulated by the counter being greater than or equal to the first counting threshold, it is determined that the second convergence condition is met. The final determined predicted weight values are then sequence-integrated to obtain the predicted weight detection signal, which is then sent to the in-situ calibration analysis module.
3. The method according to claim 1, characterized in that, The step of using the filtering analysis and calculation module to perform Kalman filtering on the flow detection signal to obtain a predicted flow detection signal, and then sending the predicted flow detection signal to the in-situ calibration analysis module, includes: Execute using the filter analysis calculation module: Determine the predicted flow rate at the previous moment, and predict the actual flow rate at the current moment; Update the state covariance at the current moment, and determine the Kalman gain value based on the updated state covariance; The predicted flow rate at the current moment is determined by combining the actual flow rate at the current moment, the Kalman gain value, and the flow rate detection signal. Determine the first-order difference mean of the predicted flow rate at the current moment, obtain the second difference of the first-order difference mean of two consecutive adjacent moments, and determine that the absolute value of the second difference of the two groups is less than the second difference threshold, then control the second value of the counter to accumulate. In response to the second value accumulated by the counter being less than the second counting threshold, the state covariance at the next moment is adjusted, and Kalman filtering is iteratively performed based on the state covariance at the next moment to predict the predicted flow rate at the next moment. During the iteration process, the second value of the counter is controlled to accumulate. In response to the second value accumulated by the counter being greater than or equal to the second counting threshold, it is determined that the third convergence condition is met. The final determined predicted flow rate values are then sequence-integrated to obtain the predicted flow rate detection signal, which is then sent to the in-situ calibration and analysis module.
4. The method according to claim 1, characterized in that, The process involves using the in-situ calibration analysis module to perform pre-processing differential analysis on the predicted flow rate detection signal and the predicted weight detection signal to obtain flow rate differential data and weight differential data. Iterative fitting is then performed based on the flow rate differential data and weight differential data until a first convergence condition is met to determine the target in-situ calibration data. In-situ calibration of the flow meter sensor is then performed based on the target in-situ calibration data, including: Perform the following using the in-situ calibration analysis module: The predicted flow rate detection signal and the predicted weight detection signal are subjected to pre-sequence differential processing to obtain flow rate differential data and weight differential data, respectively. The flow differential data and weight differential data under different flow conditions are labeled, and the first-order differential mean of the predicted flow detection signal is determined to obtain a differential array. The differential array includes: flow differential data for each flow condition, volume differential data corresponding to the weight differential data for each flow condition, and the first-order differential mean of the predicted flow detection signal. Iterative execution: Determine the current residual, and perform least squares fitting on the difference array based on the current residual to obtain the in-situ calibration data corresponding to the fitting result; In response to the determination that the final in-situ calibration data satisfies the first convergence condition, the iteration stops, and the final in-situ calibration data is used as the target in-situ calibration data; The flow meter sensor is calibrated in situ based on the target in-situ calibration data.
5. The method according to claim 4, characterized in that, The process of determining the current residual, performing least-squares fitting on the difference array based on the current residual, and obtaining the in-situ calibration data corresponding to the fitting result includes: Determine the difference array as ,in For traffic differential data, For weight difference data, To determine the density of a liquid, This is volume difference data. The first-order difference mean of the flow detection signal is used to predict the flow rate detection signal, where k is the total number of parameters in the difference array. Determine the current residual , where i is the index of each parameter in the difference array; The least squares method is used to fit the difference array to obtain the fitting result: ； in, The slope The intercept is... and The following conditions must be met: , The weights of each parameter determined in the previous n-1 iteration; Based on the formula of the fitting result, determine and This serves as the in-situ calibration data.
6. The method according to claim 5, characterized in that, The weights of each parameter The process of determining is as follows: Determine weights using kernel density estimation algorithm The formula is: ； Where h is the bandwidth, , For the residual with current index i, Here, j is the reference residual, and j is the index of the reference residual.
7. The method according to claim 5, characterized in that, The first convergence condition includes: The absolute value of the slope difference after the last two iterations. satisfy And the absolute value of the intercept difference after the last two iterations satisfy Where n is the number of iterations. The slope difference threshold. This is the threshold value for the intercept difference.
8. The method according to claim 1, characterized in that, The step of using the data acquisition module to receive the flow detection signal from the flow meter sensor and the weight detection signal from the weight sensor, and sending the flow detection signal and weight detection signal to the filtering analysis and calculation module, includes: The data acquisition module is used to denoise the initial flow detection signal sent by the flow meter sensor to obtain the flow detection signal. The data acquisition module is used to denoise the initial weight detection signal sent by the weight sensor to obtain the weight detection signal. The data acquisition module sends the flow detection signal and the weight detection signal to the filtering analysis and calculation module.
9. An analysis system, characterized in that, include: The system comprises a data acquisition module, a filtering analysis and calculation module, and an in-situ calibration and analysis module connected in sequence, wherein the data acquisition module is connected to the flow meter sensor and the weight sensor. The data acquisition module is configured to receive the flow detection signal from the flow meter sensor and the weight detection signal from the weight sensor, and send the flow detection signal and the weight detection signal to the filtering analysis and calculation module. The filtering analysis and calculation module is configured to perform Kalman filtering on the weight detection signal to obtain a predicted weight detection signal, and send the predicted weight detection signal to the in-situ calibration analysis module. The filtering analysis and calculation module is configured to perform Kalman filtering on the flow detection signal to obtain a predicted flow detection signal, and send the predicted flow detection signal to the in-situ calibration analysis module. The in-situ calibration analysis module is configured to perform pre-processing on the predicted flow detection signal and the predicted weight detection signal to obtain flow difference data and weight difference data, perform iterative fitting processing on the flow difference data and weight difference data until the first convergence condition is met to determine the target in-situ calibration data, and perform in-situ calibration on the flow meter sensor based on the target in-situ calibration data.
10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1 to 8.