Data calibration method and system for high-precision direct current collector

Abstract

The invention provides a data calibration method and system for a high-precision direct current collector, and relates to the technical field of direct current collectors, and the method comprises the steps: integrating a traceable reference source and a temperature sensor, and building a collector calibration architecture; error calibration is carried out on the target direct current collector to obtain a collection error data set, and periodic calibration is carried out on the operational amplifier bias voltage; environment temperature values are collected in real time, and calibration time is recorded at the same time; and constructing a collector drift prediction network, and introducing the collector drift prediction network to perform data calibration on the DC optimized collector based on the temperature-time joint compensation channel to obtain a reconstruction signal calibration result. The technical problem that in the prior art, due to environment interference or long-term operation, drifting of the direct current collector can be accumulated step by step, and consequently the collection precision of the direct current collector is poor is solved, error calibration is conducted by introducing the traceable reference source and the drifting prediction network, and the precision and reliability of direct current signal collection are improved.

Description

Technical Field

[0001] The present application relates to the technical field of DC collectors, and in particular to a data calibration method and system for a high-precision DC collector. Background Art

[0002] A DC collector is an instrument used for real-time monitoring and acquisition of DC signals. It features high precision, high resolution, and exceptional anti-interference capabilities. Designed for use in high-interference environments in the energy storage industry, it supports bipolar, bidirectional measurement, and features signed output. In practical applications, DC collectors often operate under highly variable environmental conditions, such as high and low temperatures, high humidity, or environments with strong electromagnetic interference. Factors such as temperature and humidity can significantly affect the collector's electronic components, causing drift in measurement accuracy. Especially for high-precision measuring instruments, temperature-induced errors may be minimal, but they can still significantly impact overall stability and accuracy. Furthermore, after prolonged operation, the performance of the DC collector's internal electronic components naturally degrades. Measurement errors may gradually arise over time due to bias voltage, aging, or other factors, leading to a decrease in accuracy over time, thus impacting the DC collector's measurement accuracy and stability.

[0003] In summary, the prior art has a technical problem in which the drift of the DC collector gradually accumulates due to environmental interference or long-term operation, resulting in poor collection accuracy of the DC collector. Summary of the Invention

[0004] The purpose of this application is to provide a data calibration method and system for a high-precision DC collector, so as to solve the technical problem in the prior art that the drift of the DC collector gradually accumulates due to environmental interference or long-term operation, resulting in poor collection accuracy of the DC collector.

[0005] In view of the above problems, the present application provides a data calibration method and system for a high-precision DC collector.

[0006] In the first aspect, the present application provides a data calibration method for a high-precision DC collector, which is implemented by a data calibration system for a high-precision DC collector, wherein the data calibration method for a high-precision DC collector includes: integrating a traceable reference source and a temperature sensor in a target DC collector, connecting the traceable reference source and the collector signal channel in parallel through a multiplexer, and building a collector calibration architecture; using the collector calibration architecture to perform error calibration and record the target DC collector to obtain an acquisition error data set, and periodically calibrating the operational amplifier bias voltage based on the acquisition error data set to obtain a DC optimized collector; real-time acquisition of ambient temperature values through the temperature sensor, and recording calibration time data at the same time, performing regression fitting based on the ambient temperature value, calibration time data and the acquisition error data set to build a temperature-time joint compensation channel; constructing a collector drift prediction network, introducing the collector drift prediction network to perform data calibration on the DC optimized collector based on the temperature-time joint compensation channel, and obtaining a reconstructed signal calibration result.

[0007] Optionally, a calibration strategy analysis is performed on the collector calibration architecture to obtain a collector switching calibration strategy, which includes a signal acquisition strategy and a switching calibration strategy; a collector calibration period is set according to a data calibration target; if the collector is not in the calibration period, the signal acquisition strategy is triggered, and the collector signal channel is activated through the signal acquisition strategy to perform DC signal acquisition; when the collector is in the calibration period, the switching calibration strategy is triggered, and the error calibration record is switched to the traceable reference source through the switching calibration strategy to obtain an acquisition error data set.

[0008] Optionally, the switching calibration strategy is used to switch to the traceable reference source for signal acquisition to obtain a reference source acquisition value set; the theoretical standard value set of the traceable reference source is recorded and obtained; the reference source acquisition value set and the theoretical standard value set are arranged and aligned according to the acquisition time to obtain an acquisition value sequence set and a standard value sequence set; the deviation value set of the acquisition value sequence set and the standard value sequence set is calculated and obtained, and the deviation value set is used as the acquisition error data set.

[0009] Optionally, a PID controller is initialized according to characteristic information of the target DC collector, the PID controller including a proportional coefficient, an integral coefficient, and a differential coefficient; the PID controller is verified and tuned to generate a target PID controller; the operational amplifier bias voltage is used as the control output, and the acquisition error data set is calibrated and calculated based on the target PID controller according to the collector calibration period to obtain the DC optimized collector.

[0010] Optionally, the ambient temperature value, calibration time data and the acquisition error data set are cleaned and normalized to obtain a standard ambient temperature value, a standard calibration time data and a standard acquisition error data set; the standard ambient temperature value and the standard calibration time data are used as independent variables, and the standard acquisition error data set is used as a dependent variable, and multiple regression fitting is performed on the independent variables and the dependent variables to generate a temperature-time error prediction model; the original DC signal data set is obtained through the DC optimization collector, and the original DC signal data set is compensated and analyzed based on the temperature-time error prediction model to build the temperature-time joint compensation channel.

[0011] Optionally, a historical data set of a DC collector is mined and obtained, wherein the historical data set includes historical collection values, temperature values, time data, collector aging data, and corresponding collection error data; an LSTM neural network structure is used to label the drift degree of the historical data set of the DC collector and perform predictive supervision training to generate an initial drift prediction network; and loss assessment and iterative tuning are performed on the initial drift prediction network to obtain a collector drift prediction network.

[0012] Optionally, a cross entropy loss function is used to perform loss assessment feedback on the initial drift prediction network to determine model loss data; based on the model loss data, a model optimizer is selected, and the initial drift prediction network is iteratively tuned based on the model optimizer to obtain the collector drift prediction network.

[0013] Optionally, signal compensation is performed on the DC optimization collector based on the temperature-time joint compensation channel to obtain a DC signal compensation data set; the collector drift prediction network is introduced to perform drift prediction on the DC optimization collector and the current DC signal acquisition data to determine the collector drift prediction parameters; data calibration is performed on the DC signal compensation data set based on the collector drift prediction parameters to obtain a reconstructed signal calibration result.

[0014] Optionally, a drift level evaluation is performed on the collector drift prediction parameters to obtain the collector drift degree level, and a calibration strategy is analyzed based on the collector drift degree level to determine the drift correction coefficient; the DC signal compensation data set is calibrated and corrected based on the drift correction coefficient to obtain the reconstructed signal calibration result.

[0015] In the second aspect, the present application also provides a data calibration system for a high-precision DC collector, which is used to execute a data calibration method for a high-precision DC collector as described in the first aspect, wherein the data calibration system for a high-precision DC collector includes: a calibration architecture building module, which is used to integrate a traceable reference source and a temperature sensor in a target DC collector, and connect the traceable reference source and the collector signal channel in parallel through a multiplexer to build a collector calibration architecture; a periodic calibration module, which is used to use the collector calibration architecture to perform error calibration and record the target DC collector to obtain an acquisition error data set, and periodically calibrate the op amp bias voltage based on the acquisition error data set to obtain a DC optimized collector; a compensation channel building module, which is used to collect ambient temperature values in real time through the temperature sensor, and record calibration time data at the same time, perform regression fitting based on the ambient temperature value, calibration time data and the acquisition error data set, and build a temperature-time joint compensation channel; a signal calibration module, which is used to construct a collector drift prediction network, introduce the collector drift prediction network to perform data calibration on the DC optimized collector based on the temperature-time joint compensation channel, and obtain a reconstructed signal calibration result.

[0016] One or more technical solutions provided in this application have at least the following beneficial effects: A collector calibration architecture is constructed by integrating a traceable reference source and a temperature sensor in a target DC collector, and connecting the traceable reference source and the collector signal channel in parallel through a multiplexer. The target DC collector is calibrated and recorded using the collector calibration architecture to obtain an acquisition error data set, and the operational amplifier bias voltage is periodically calibrated based on the acquisition error data set to obtain a DC optimized collector. The ambient temperature value is collected in real time by the temperature sensor, and calibration time data is recorded at the same time. Regression fitting is performed based on the ambient temperature value, calibration time data, and the acquisition error data set to construct a temperature-time joint compensation channel. A collector drift prediction network is constructed, and the collector drift prediction network is introduced to perform data calibration on the DC optimized collector based on the temperature-time joint compensation channel to obtain a reconstructed signal calibration result. That is to say, by integrating a traceable reference source and a temperature sensor, the traceable reference source is connected in parallel with the DC collector signal channel through a multiplexer, a calibration architecture is established, the op amp bias voltage is periodically calibrated, the acquisition error is regressed and fitted using the temperature sensor data and calibration time data, and a temperature-time joint compensation channel is built for data calibration to obtain the reconstructed signal calibration result, thereby improving the accuracy and reliability of DC signal acquisition.

[0017] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, which can be implemented in accordance with the contents of the description, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically listed below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and a person of ordinary skill in the art can obtain other drawings based on the provided drawings without creative work.

[0019] Figure 1 This is a flow chart of a data calibration method for a high-precision DC collector in this application.

[0020] Figure 2 This is a structural diagram of a data calibration system for a high-precision DC collector in this application.

[0021] Description of the reference numerals: calibration architecture building module 11 , periodic calibration module 12 , compensation channel building module 13 , signal calibration module 14 . DETAILED DESCRIPTION

[0022] This application provides a data calibration method and system for a high-precision DC collector, addressing the existing technical problem of poor DC collector accuracy due to the gradual accumulation of drift in the DC collector caused by environmental interference or long-term operation. By integrating a traceable reference source and a temperature sensor, connecting the traceable reference source in parallel with the DC collector signal channel via a multiplexer, establishing a calibration architecture, periodically calibrating the op amp bias voltage, performing regression fitting of the acquisition error using temperature sensor data and calibration time data, and establishing a temperature-time joint compensation channel for data calibration, the calibration results of the reconstructed signal are obtained, thereby improving the accuracy and reliability of DC signal acquisition.

[0023] Below, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited to the example embodiments described herein. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. It should also be noted that, for the convenience of description, only the parts related to this application, rather than all of them, are shown in the accompanying drawings.

[0024] For example 1, please refer to the attached Figure 1 The present application provides a data calibration method for a high-precision DC collector, wherein the data calibration method for a high-precision DC collector is performed by a data calibration system for a high-precision DC collector, and the data calibration method for a high-precision DC collector specifically includes the following steps: S100: Integrate a traceable reference source and a temperature sensor in a target DC collector, connect the traceable reference source and the collector signal channel in parallel through a multiplexer, and build a collector calibration architecture.

[0025] Specifically, the target DC data collector is a high-precision DC data collector that requires error calibration and optimization. It features high accuracy (better than 0.1%), high resolution (one hundred thousandth of the range), and exceptionally strong anti-interference capabilities. Designed for use in the high-interference environments of the energy storage industry, it supports bipolar, bidirectional measurement and has a signed output. This device typically achieves high-precision data acquisition through a 24-bit analog-to-digital converter (ADC). Its linear accuracy reaches one ten-thousandth of the range. Four independent ADC channels simultaneously sample data, allowing for complete acquisition of all electrical parameters across all four channels in as little as 20ms. It also features positive and negative bipolar measurement capabilities and provides 24-bit ADC sampling output. The target DC data collector utilizes high-performance isolation components to isolate the power supply, communication, measurement terminals, and individual channels from each other. This not only effectively prevents external electromagnetic interference but also ensures device stability and safety in high-voltage environments. The isolation strength between each signal channel exceeds 2500VDC, enabling it to withstand higher voltage surges without damage. The target DC collector supports various current (e.g., 5mA, 10A) and voltage (e.g., 10V, 500V) measurements, and is capable of measuring extremely low currents (uA level). It also has a very strong overload capability, allowing it to continue measuring without damage even after an overload condition within a certain range.

[0026] A traceable reference source is integrated into the target DC collector. A traceable reference source is a standard power source whose output values are traceable to conventional standards and serves as a reference for calibration. A temperature sensor is also integrated to measure ambient temperature. The accuracy and stability of the target DC collector may be affected by temperature fluctuations. Therefore, the integrated temperature sensor works in conjunction with the traceable reference source to adjust measurement results based on the real-time ambient temperature, ensuring that accuracy is unaffected by ambient temperature. A multiplexer is used to connect the traceable reference source in parallel with the DC collector's signal channel. In other words, a multiplexer is used to connect multiple signal sources to the target DC collector. One signal source is the traceable reference source, providing a stable and traceable electrical signal, while another is the target DC collector's signal channel, receiving actual external current and voltage signals. Through the multiplexer, the outputs of the reference source and signal channel can be alternately fed into the target DC collector for measurement.

[0027] A multiplexer is an electronic switching device that combines multiple signal sources into a single signal stream for transmission. It is used to combine multiple signal channels and switch them to the input of the target DC collector, ensuring that the target DC collector can receive signals from both the reference source and the signal channels for processing. A traceable reference source is connected in parallel with the target DC collector's signal channel. The target DC collector receives both the standard signal output by the reference source and the actual measured electrical signal in real time. This parallel connection ensures that both signal sources can be transmitted in parallel to the target DC collector's input for multi-channel signal processing and calibration. By combining multiplexing and parallel connections, a collector calibration architecture is established that continuously corrects measurement errors in the target DC collector. Whenever the target DC collector receives the standard signal from the reference source, it compares it with the actual signal being collected and adjusts its internal parameters based on the error, ensuring that the target DC collector maintains high accuracy over extended use.

[0028] By establishing a data collector calibration framework, the long-term stability and measurement accuracy of the target DC data collector are significantly improved, especially in environments with large temperature fluctuations. A traceable reference source ensures the accuracy of the measurement standard, and the temperature compensation function further eliminates the influence of ambient temperature on measurement errors, thus achieving high-precision data collection in a variety of complex application scenarios.

[0029] S200: Utilizing the collector calibration architecture to perform error calibration on the target DC collector to obtain a collection error data set, and periodically calibrating the operational amplifier bias voltage based on the collection error data set to obtain a DC optimized collector.

[0030] Furthermore, the present application S200 includes: A calibration strategy analysis is performed on the collector calibration architecture to obtain a collector switching calibration strategy, which includes a signal acquisition strategy and a switching calibration strategy. According to the data calibration target, a collector calibration period is set. If the collector is not in the collector calibration period, the signal acquisition strategy is triggered, and the collector signal channel is activated through the signal acquisition strategy to perform DC signal acquisition. When the collector is in the collector calibration period, the switching calibration strategy is triggered, and the error calibration record is switched to the traceable reference source through the switching calibration strategy to obtain an acquisition error data set.

[0031] Through the switching calibration strategy, the signal is acquired by switching to the traceable reference source to obtain a reference source acquisition value set; the theoretical standard value set of the traceable reference source is recorded and obtained; the reference source acquisition value set and the theoretical standard value set are arranged and aligned according to the acquisition time to obtain an acquisition value sequence set and a standard value sequence set; the deviation value set of the acquisition value sequence set and the standard value sequence set is calculated and obtained, and the deviation value set is used as the acquisition error data set.

[0032] Specifically, a calibration strategy analysis is conducted on the collector calibration architecture. By analyzing the target DC collector's operating environment, performance, and usage frequency, a reasonable calibration strategy is determined. The purpose of the calibration strategy analysis is to determine when and how to perform calibration to minimize errors and ensure high accuracy. The collector switching calibration strategy, in actual application, adjusts the collector's signal source and calibration method to enable the collector to switch between different operating cycles (such as normal operating cycle and calibration cycle). This strategy includes a signal acquisition strategy and a switching calibration strategy. The signal acquisition strategy is used during non-calibration periods to acquire data by activating the collector's signal channels. The switching calibration strategy is used during calibration periods to perform error calibration by switching to a traceable reference source.

[0033] During normal operating cycles (i.e., non-calibration cycles), the signal acquisition strategy is triggered, activating the collector's signal channels and performing real-time data acquisition. For example, this involves collecting battery voltage and current data. When the set calibration cycle begins, the calibration strategy is triggered and compared with a standard reference source, recording error data for use in calibrating the target DC collector.

[0034] Based on the calibration target and the target DC collector's usage requirements, set the collector calibration cycle, specifying the calibration interval. This interval is typically adjusted dynamically based on the target DC collector's operating environment, frequency of use, and accuracy requirements. During non-calibration periods, the trigger signal acquisition strategy continues normal signal acquisition and activates the collector's signal channels for DC signal acquisition. By triggering the signal acquisition strategy, the collector's signal channels are activated, allowing the collector to begin acquiring DC signals, including voltage and current.

[0035] When the collector is in the calibration cycle, the switching calibration strategy is triggered, and the multiplexer switches the target DC collector from the actual signal channel to the traceable reference source to perform error calibration and record the error. When connected to the standard traceable reference source, the standard voltage or current value can be obtained, compared with the actual measurement result of the target DC collector, and the error data is recorded. By switching the calibration strategy, you can switch to the traceable reference source. The reference source acquisition value refers to the voltage or current value actually measured and recorded by the DC collector when calibrating with a traceable reference source. It is the signal obtained by the collector from the reference source at a specific point in time, reflecting the measurement result of the collector at that moment. It is crucial for evaluating the accuracy and stability of the collector because it can be directly compared with the theoretical standard value of the reference source to determine the error of the collector.

[0036] Obtain and record the theoretical standard value set of the traceable reference source. This set of ideal voltage or current values that the traceable reference source should output is typically provided by the manufacturer or determined according to international standards. For example, for a 10V reference source, the theoretical standard value set consists of multiple 10V values. Align the reference source acquisition value set and the theoretical standard value set based on the acquisition time, specifically by the acquisition timestamp. Align the sorted reference source acquisition value set and the theoretical standard value set to form two sequences: the acquisition value sequence set and the standard value sequence set. At this point, each acquisition value in the reference source acquisition value set corresponds to the theoretical standard value in the standard value sequence set at the corresponding time point. For example, the reference source acquisition value at 0s is 10.002V, and the theoretical standard value is 10.000V; the reference source acquisition value at 11s is 10.010V, and the theoretical standard value is 10.000V; and the reference source acquisition value at 110s is 30.009V, and the theoretical standard value is 30.000V.

[0037] The deviation value set is calculated from the collection value sequence set and the standard value sequence set. Specifically, the difference between each collection value and the corresponding standard value is calculated to obtain the deviation value. All these deviation values are combined to form a deviation value set, the collection error data set, which records the error between the target DC collector and the reference source. For example, the reference source collection value at 0s is 10.002V, the theoretical standard value is 10.000V, and the deviation value is 0.002V; the reference source collection value at 11s is 10.010V, the theoretical standard value is 10.000V, and the deviation value is 0.010V; the reference source collection value at 110s is 30.009V, the theoretical standard value is 30.000V, and the deviation value is 0.009V.

[0038] Based on the calibration cycle settings, the target DC data collector automatically calibrates and records errors during each calibration cycle, generating a set of deviation values in real time. This improves the accuracy and stability of the data collector over the long term. Even in high-interference or extreme environments, the data collector maintains high accuracy and reduces errors caused by device aging or environmental factors.

[0039] Furthermore, the present application further comprises the following steps: According to the characteristic information of the target DC collector, a PID controller is initialized, the PID controller including a proportional coefficient, an integral coefficient, and a differential coefficient; the PID controller is verified and tuned to generate a target PID controller; and the operational amplifier bias voltage is used as the control output. The acquisition error data set is calibrated and calculated based on the target PID controller according to the collector calibration period to obtain the DC optimized collector.

[0040] Specifically, the PID controller is initialized based on the target DC collector's characteristic information. The PID controller's primary function is to calculate the output control signal based on the measured error, thereby optimizing the collector's measured values. During initialization, the three parameters of the PID controller are first determined: the proportional coefficient, the integral coefficient, and the differential coefficient. The proportional coefficient is set to an initial value, such as 1.0, based on experience or preliminary experimental data. The proportional coefficient primarily affects the speed and magnitude of the response to the current error. A larger proportional coefficient makes the PID controller more sensitive to error changes. The integral coefficient is typically set to 0 initially and then adjusted based on calibration data. The integral effect eliminates steady-state errors, but increases the target DC collector's response time. It is suitable for eliminating long-term drift after the target DC collector stabilizes. The differential coefficient is typically set to 0 initially. The differential effect reduces overshoot and suppresses oscillations in the target DC collector. By predicting the error trend, it accelerates the target DC collector's response and is suitable for reducing oscillations in the target DC collector. The PID controller is initialized based on the target DC collector's response speed and error characteristics, ensuring that the initial PID parameters do not lead to excessive oscillation or slow response.

[0041] Verify and tune the initialized PID controller to ensure it effectively reduces acquisition error and optimizes the target DC collector performance. PID (Proportional-Integral-Derivative) controller tuning is a process aimed at selecting appropriate proportional coefficients (Kp), integral coefficients (Ki), and differential coefficients (Kd) to effectively reduce error and maintain the target DC collector temperature. Set a control target, for example, reducing acquisition error to less than 0.001V. Experimentally obtain acquisition error data for the collector, such as 0.005V, 0.003V, 0.004V, and 0.002V. Adjust the collector signal based on the PID controller output. After each adjustment, evaluate the change in acquisition error. If the error decreases, the PID parameter combination is effective. For example, if the error decreases from 0.005V to 0.002V, this indicates improvement in the current PID parameter settings.

[0042] Adjusting PID parameters is key to PID controller tuning. For example, gradually increasing the proportional coefficient makes the controller more sensitive to errors and responds more quickly. PID parameter tuning can be accomplished by gradually increasing the proportional coefficient until rapid convergence of the target DC power supply output is observed. For example, adjust the proportional coefficient from 1.0 to 1.5. Alternatively, adjust the integral coefficient. Increasing the integral coefficient can help eliminate long-term accumulated errors. Experiment with the results to see if increasing the integral coefficient reduces steady-state error. If the integral coefficient is too large, it may cause over-response and oscillation. Alternatively, adjust the differential coefficient. Increasing the differential coefficient can help reduce overshoot and oscillation in the target DC power supply. When adjusting the differential coefficient, ensure that it effectively reduces the rate of error change and avoids system oscillation. For example, suppose that after gradual tuning, the target PID controller's PID parameters are a proportional coefficient of 1.5, an integral coefficient of 0.8, and a differential coefficient of 0.3.

[0043] When the adjusted PID parameters can meet the target error requirements, the final parameter combination is the target PID controller. At this point, the output of the PID controller can stably and accurately control the error of the collector. For example, the final error data may be reduced to: [0.001V, 0.0008V, 0.001V, 0.0006V], indicating that through PID tuning, the error of the target DC collector has met the accuracy requirements.

[0044] Op amp bias voltage is a voltage deviation at the input of an operational amplifier (OPA) due to design or manufacturing errors, often affecting signal accuracy. Operational amplifiers (OPA) are commonly used electronic components. Bias voltage refers to the DC voltage applied to the input of an OP amp. It regulates the circuit's operating state and ensures the amplifier operates correctly within its required operating range. The OP amp bias voltage is used as the output signal of a PID controller to adjust the error of a DC data collector to reduce errors. The OP amp bias voltage can affect the input signal and, in turn, the output of the data collector. Precisely adjusting the bias voltage can correct for measurement errors caused by device imperfections. Assuming the initial OP amp bias voltage value, Vbias, is 1.0V, this voltage can be adjusted to achieve the calibration target.

[0045] During the data collector calibration cycle, a PID controller is activated based on the error data set to adjust the target DC data collector's bias. Based on the current error data set, the adjusted output value is calculated using the principles of proportional, integral, and differential. The PID controller's output affects the bias voltage of the operational amplifier. By calculating the PID controller's output, the bias voltage is adjusted to correct the data collector's output. For example, assume the target PID controller's parameters have been tuned to: proportional coefficient 1.5, integral coefficient 0.8, and differential coefficient 0.3. Based on the error data, the PID controller calculates a new control value to adjust the bias voltage. At a certain point, the PID controller calculates a bias voltage adjustment value of ΔVbias = 0.02V, thus adjusting the original bias voltage (1.0V) to 1.02V.

[0046] By calibrating and calculating the acquisition error data set, a DC-optimized data collector was ultimately developed, capable of providing more accurate measurements. Bias voltage adjustment ensures that the data collector's output is closer to the actual value, thereby reducing errors. For example, after a complete PID controller calibration cycle, the acquisition errors were reduced from 0.003V and 0.004V to 0.0005V and 0.0006V, demonstrating a significant performance improvement and effective error correction for the DC-optimized data collector. The DC-optimized data collector reduces acquisition errors through PID controller calibration, providing more accurate measurements. Through PID controller adjustment, the data collector's error was reduced from the initial millivolts to the microvolt level, significantly improving accuracy. Through a real-time feedback mechanism, the PID controller automatically adjusts based on the current error, ensuring the data collector maintains high accuracy during real-time operation.

[0047] S300: Collecting the ambient temperature value in real time through the temperature sensor, and recording the calibration time data at the same time, performing regression fitting based on the ambient temperature value, the calibration time data and the acquisition error data set, and building a temperature-time joint compensation channel.

[0048] Furthermore, the present application S300 includes: The ambient temperature value, calibration time data and acquisition error data set are cleaned and normalized to obtain a standard ambient temperature value, a standard calibration time data and a standard acquisition error data set; the standard ambient temperature value and the standard calibration time data are used as independent variables, and the standard acquisition error data set is used as a dependent variable, and multiple regression fitting is performed on the independent variables and the dependent variables to generate a temperature-time error prediction model; the original DC signal data set is obtained through the DC optimization collector, and compensation analysis is performed on the original DC signal data set based on the temperature-time error prediction model to build the temperature-time joint compensation channel.

[0049] Specifically, an integrated temperature sensor collects real-time temperature data from the target high-precision DC harvester's current environment. Temperature affects its operating state, so real-time monitoring of ambient temperature is crucial for accurate measurement and calibration. The temperature sensor converts the ambient temperature into an electrical signal. Assuming a digital temperature sensor is used, an analog-to-digital converter (ADC) converts the electrical signal into a processable digital value. The sensor might output a temperature value such as 25°C. The specific accuracy and range of the data collected depend on the sensor used. The temperature values collected by the temperature sensor can be collected in real time by a microcontroller. For example, the sensor collects temperature data every 1 second and sends it to the main controller, which records the current temperature value. The values obtained at 1 second are 24.5°C, 25.0°C at 2 seconds, 25.2°C at 3 seconds, 25.1°C at 4 seconds, and 24.8°C at 5 seconds, representing the ambient temperature at each acquisition moment.

[0050] Each time you calibrate your device, record the start and end times of the calibration operation. This time data is recorded along with the temperature data, typically automatically via the sensor's timestamp. This calibration time data records the entire calibration process from start to finish and can be compared with the temperature data to analyze how ambient temperature affected the device's performance during the calibration period. Each temperature acquisition value and the corresponding calibration time data should be arranged in chronological order to ensure accurate data matching for further analysis. For example, the temperature at 10:00 a.m. might be 24.5°C.

[0051] Data cleaning and normalization are performed on the ambient temperature values, calibration time data, and acquisition error datasets to obtain standard ambient temperature values, standard calibration time data, and standard acquisition error datasets. Data cleaning refers to removing incomplete, inconsistent, or noisy data. For example, since the operating temperature range of a high-precision DC data collector is -40°C to +70°C, outliers such as 100°C or -50°C in the temperature data must be removed as they do not conform to normal ambient temperature values. Normalization converts the data to a standard range (e.g., [0, 1]) for easier comparison and analysis. For example, assume the following cleaned data: ambient temperature: 24.5°C, 25.0°C, 25.5°C, 24.8°C; calibration time: 10:00:00, 10:10:00, 10:20:00; acquisition error: 0.02V, 0.03V, 0.01V. The normalized data are: ambient temperature: 0, 0.5, 1.0, 0.3, calibration time: 0, 0.5, 1.0, acquisition error: 0.5, 1.0, 0.

[0052] A multivariate regression fit is performed using the normalized standard ambient temperature and standard calibration time data as independent variables and the standard acquisition error dataset as the dependent variable. The regression coefficients are solved using the least squares method to obtain the error prediction model. For example, the model might be: Error prediction value = a·temperature + b·time + c, where a, b, and c are regression coefficients. The fitting results in a of 0.03, b of 0.01, and c of 0.02, resulting in a regression fit result of error prediction value = 0.03·temperature + 0.01·time + 0.02. The temperature-time error prediction model is generated based on the results of the multivariate regression fit. It is used to predict the error value of the measurement device under a specific ambient temperature and calibration time. It expresses the relationship between the temperature and time factors and the acquisition error.

[0053] The raw DC signal dataset acquired by the DC-optimized collector is combined with the temperature-time error prediction model, and compensation analysis is performed on the raw signal data based on the output of the prediction model. The raw DC signal dataset, obtained by signal acquisition using the DC-optimized collector (which undergoes PID regulation and environmental calibration), is the uncompensated signal dataset collected by the DC-optimized collector. This dataset may contain errors due to environmental factors, equipment deviations, and other factors. For example, some examples of raw DC signal datasets are shown in Table 1: Table 1. Sample data of some original DC signal datasets ; For each piece of original signal data, a compensation analysis is performed using the temperature-time error prediction model to calculate its error prediction value, and then the error is subtracted from the original signal to obtain the compensated signal value. For example, for the first data: temperature = 25.0°C, calibration time = 10:00:00 (time = 0 minutes), prediction error = 0.03×25+0.01×0+0.02=0.77, compensated signal = 5.00−0.77=4.23V; the second data: temperature = 25.5°C, calibration time = 10:05:00 (time 5 minutes), prediction error = 0.03×25.5+0.01×5+0.02=0.835, compensated signal = 5.02−0.835=4.185V; the third data: temperature = 25.2°C, calibration time = 10:10:00 (time 10 minutes), prediction error = 0.03×25.2+0.01×1 0 + 0.02 = 0.756 + 0.10 + 0.02 = 0.876, compensated signal = 5.01 − 0.876 = 4.134 V; fourth data: temperature = 25.3°C, calibration time = 10:15:00 (time = 15 minutes), prediction error = 0.03 × 25.3 + 0.01 × 15 + 0.02 = 0.929, compensated signal = 5.02 − 0.929 = 4.091 V; fifth data: temperature = 25.1°C, calibration time = 10:20:00 (time = 20 minutes), prediction error = 0.03 × 25.1 + 0.01 × 20 + 0.02 = 0.973, compensated signal = 5.02 − 0.973 = 4.047 V.

[0054] Based on the compensation analysis results, a temperature-time joint compensation channel is created. Each time a raw signal is acquired, the temperature-time error prediction model is automatically applied to compensate the acquired signal in real time. When a new DC signal is acquired, the current ambient temperature and calibration time are automatically detected, the error value is calculated using the error prediction model, and real-time correction is performed on the acquired value. The temperature-time joint compensation channel compensates the raw DC signal data based on the temperature-time error prediction model to correct for measurement errors caused by temperature changes and time factors. By combining real-time temperature monitoring and calibration time data, the device's measurement accuracy in various environments is improved, reducing errors caused by environmental fluctuations. Through the temperature-time error prediction model and compensation analysis, signal acquisition errors are automatically adjusted to ensure that the device consistently outputs high-precision measurement results, effectively improving the stability and reliability of the DC collector under varying ambient temperatures and operating times.

[0055] S400: Constructing a collector drift prediction network, introducing the collector drift prediction network to perform data calibration on the DC optimized collector based on the temperature-time joint compensation channel, and obtaining a reconstructed signal calibration result.

[0056] Furthermore, the present application S400 includes: A historical data set of a DC collector is mined and obtained, wherein the historical data set includes historical collection values, temperature values, time data, collector aging data, and corresponding collection error data. An LSTM neural network structure is used to perform drift degree annotation and predictive supervision training on the historical data set of the DC collector to generate an initial drift prediction network. The initial drift prediction network is subjected to loss assessment and iterative tuning to obtain a collector drift prediction network.

[0057] A cross entropy loss function is used to perform loss assessment feedback on the initial drift prediction network to determine model loss data; based on the model loss data, a model optimizer is selected, and the initial drift prediction network is iteratively tuned based on the model optimizer to obtain the collector drift prediction network.

[0058] Specifically, the historical data set of the DC collector is mined. This data is the historical data collected over a period of time. This typically includes the collector's collection values, ambient temperature, operating hours, and other information at different time points, reflecting the performance and changes of the target DC collector during actual operation. Historical collection values are the DC signal values collected by the target DC collector over a period of time; temperature values are the ambient temperature values during which the target DC collector operated over a period of time. Time data is the timestamp recorded by the collector, indicating the time of each data point. Collector aging data is data reflecting the aging process of the target DC collector, which may include equipment performance degradation, accuracy changes, and other information. Collection error data is the error between the actual measurement value of the DC collector and the true standard value.

[0059] The LSTM neural network architecture, a recurrent neural network (RNN) structure, is effective at processing sequential data and is particularly well-suited for tasks with long-term dependencies. LSTMs can memorize long-term information in sequential data and are used for time series prediction and modeling. An LSTM neural network is used to annotate drift levels and conduct predictive supervised training on historical datasets of DC data collectors. Drift in the data is labeled based on acquisition error and aging data. Drift annotations can be categorized into different levels, such as mild drift, moderate drift, and severe drift. Annotations can be based on whether acquisition error exceeds a certain threshold or quantified by the difference between actual measurements and standard values. The LSTM neural network is trained using annotated historical datasets (including input features such as ambient temperature, time, and aging data, as well as drift annotations of acquisition error). The LSTM network learns the relationship between input data and drift and builds a drift prediction model. Based on the training data, the LSTM network learns the correlation between temperature, time, aging, and acquisition error, and predicts future error trends.

[0060] During training, model performance is evaluated by calculating the error between the model's predicted values and the true values. If the prediction results deviate significantly, model parameters need to be adjusted and iterative tuning performed to optimize hyperparameters such as the network structure and learning rate. The cross-entropy loss function is a commonly used loss function in classification tasks, particularly suitable for neural networks whose output is a probability distribution (such as multi-classification problems). It calculates the difference between the actual category and the predicted probability distribution. Loss evaluation is performed on the initial drift prediction network to calculate the model's prediction error on the given data. For example, assuming the actual value of the collector drift is 1 (indicating drift has occurred), and the initial prediction network outputs a predicted probability of 0.7, the loss calculated using the cross-entropy loss function is 0.155.

[0061] The model loss data calculated by the loss function will determine the performance of the current model. If the loss value is large, it means that the gap between the prediction and the actual situation is large, and further optimization is needed. To perform optimization, select a suitable optimizer, such as the Adam optimizer. The Adam optimizer can automatically adjust the learning rate based on the gradient of the loss and the historical gradient, thereby making training more efficient. After selecting the optimizer, the next step is to iteratively tune the initial drift prediction network. Each iteration calculates the current loss and adjusts the network parameters through the optimizer. Iterative tuning is a continuous optimization process until the model converges to the optimal solution. Specifically, the training data is input and forward propagated to obtain the predicted value; the loss value is calculated using the cross-entropy loss function; the gradient of the loss function with respect to the network weights is calculated using the backpropagation algorithm; the Adam optimizer updates the weights based on the gradient to reduce the loss value; and the above process is repeated until the loss value stabilizes and approaches zero.

[0062] After multiple iterations of tuning, the resulting network is optimized for collector drift prediction, capable of accurately predicting collector drift and responding promptly to various environmental changes. This optimized network provides more stable drift predictions, significantly improving equipment operational accuracy. By mining historical data from DC collectors and employing an LSTM neural network for drift prediction, and iteratively tuning using a cross-entropy loss function and optimizer, the accuracy of collector drift prediction is improved, errors caused by drift in DC optimized collectors are reduced, and operational efficiency and stability are enhanced.

[0063] Furthermore, the present application further comprises the following steps: Based on the temperature-time joint compensation channel, signal compensation is performed on the DC optimized collector to obtain a DC signal compensation data set; the collector drift prediction network is introduced to perform drift prediction on the DC optimized collector and the current DC signal acquisition data to determine the collector drift prediction parameters; based on the collector drift prediction parameters, data calibration is performed on the DC signal compensation data set to obtain a reconstructed signal calibration result.

[0064] A drift level assessment is performed on the collector drift prediction parameters to obtain a collector drift degree level, and a calibration strategy is analyzed based on the collector drift degree level to determine a drift correction coefficient; and a calibration correction is performed on the DC signal compensation data set based on the drift correction coefficient to obtain the reconstructed signal calibration result.

[0065] Specifically, the DC-optimized data collector is compensated for its signal using a combined temperature-time compensation channel. This channel utilizes a previously generated temperature-time error prediction model to compensate for the output signal of the DC-optimized data collector. Temperature and time are important factors affecting data collector performance. By establishing a compensation channel, the errors caused by these factors can be corrected. This compensation strategy brings the DC signal closer to its true value, reducing errors caused by factors such as temperature and time. Specifically, temperature and time information are input into the compensation channel, and the data collector's output signal is corrected using a pre-established temperature-time error prediction model to generate a compensated signal dataset. For example, suppose a DC data collector outputs 5.0V at 25°C, but at 30°C, the output signal is higher, reaching 5.05V. Using the temperature compensation channel, the 5.05V output is adjusted to 5.0V, eliminating the temperature error.

[0066] After signal compensation, the current DC signal acquisition data is drift predicted using the collector drift prediction network. The DC optimized collector and the current DC signal acquisition data are then input into the collector drift prediction network for drift prediction, yielding the collector drift prediction parameters. The collector drift prediction parameters are generated by the collector drift prediction network and represent the extent and trend of the DC optimized collector drift. The compensated DC signal compensation dataset is calibrated based on the collector drift prediction parameters. The goal of calibration is to eliminate or reduce the error caused by collector drift, thereby obtaining a more accurate signal. The collector drift prediction parameters are applied to the DC signal compensation dataset, and the signal is corrected to obtain the reconstructed signal calibration result, which is the final signal after drift prediction calibration.

[0067] Specifically, the drift prediction parameters of the collector are evaluated for drift level to assess the drift severity of the target DC collector. Drift levels are typically expressed in different levels (such as mild, moderate, and severe) to determine appropriate compensation measures. The drift amplitude, drift rate, and drift trend are derived from the collector drift prediction parameters. Based on the drift amplitude and rate, the drift level of the collector is assessed. For example, a small drift amplitude and a slow drift rate may be considered mild drift, while a large drift amplitude and a fast drift rate may be considered severe drift. A threshold model is typically designed to set the standards for each level. For example, a drift amplitude greater than 0.05V and a drift rate exceeding 0.01V / hour is considered severe drift, while an amplitude less than 0.02V and a rate less than 0.005V / hour is considered mild drift.

[0068] The drift correction factor is determined based on the drift level and the actual operating status of the data collector. When the device's drift level is low, the correction factor may be small, while when the drift level is high, a larger correction factor is required to achieve a stronger correction effect. Based on the drift prediction parameters and the drift level, a calibration strategy analysis is performed to calculate a specific drift correction factor to effectively correct the signal. Calibration strategy analysis involves developing specific correction measures based on the data collector's drift level. The calibration strategy includes how to select the drift correction factor and how to apply it to each signal value in the DC signal dataset.

[0069] A correction factor is applied to the compensated DC signal dataset, adjusting each signal value to eliminate drift errors. In other words, the reconstructed signal calibration result is equal to the compensated DC signal dataset minus the drift correction factor. For example, assuming the compensated DC signal dataset is [5.03V, 5.04V, 5.05V, 5.07V], the corresponding drift correction factor is 0.03V (mild drift), and the reconstructed signal calibration result after calibration is [5.00V, 5.01V, 5.02V, 5.04V]. By introducing a temperature-time joint compensation channel and a drift prediction network, signal errors caused by temperature changes, time factors, and device drift can be effectively reduced, thereby improving signal accuracy. By predicting and correcting drift, signal drift issues that may occur after long-term use of high-precision DC data loggers can be effectively prevented, thereby maintaining the long-term stability and reliability of high-precision DC data loggers.

[0070] In summary, the data calibration method for a high-precision DC collector provided in the present application has the following beneficial effects: by integrating a traceable reference source and a temperature sensor in the target DC collector, the traceable reference source and the collector signal channel are connected in parallel through a multiplexer to build a collector calibration architecture; the target DC collector is calibrated and recorded using the collector calibration architecture to obtain an acquisition error data set, and the op amp bias voltage is periodically calibrated based on the acquisition error data set to obtain a DC optimized collector; the ambient temperature value is collected in real time through the temperature sensor, and the calibration time data is recorded at the same time, and regression fitting is performed based on the ambient temperature value, calibration time data and the acquisition error data set to build a temperature-time joint compensation channel; a collector drift prediction network is constructed, and the collector drift prediction network is introduced to perform data calibration on the DC optimized collector based on the temperature-time joint compensation channel to obtain a reconstructed signal calibration result. That is to say, by integrating a traceable reference source and a temperature sensor, the traceable reference source is connected in parallel with the DC collector signal channel through a multiplexer, a calibration architecture is established, the op amp bias voltage is periodically calibrated, the acquisition error is regressed and fitted using the temperature sensor data and calibration time data, and a temperature-time joint compensation channel is built for data calibration to obtain the reconstructed signal calibration result, thereby improving the accuracy and reliability of DC signal acquisition.

[0071] Example 2: Based on the same inventive concept as the data calibration method for a high-precision DC collector in the above-mentioned Example 1, this application also provides a data calibration system for a high-precision DC collector, see the attached Figure 2 , the data calibration system of a high-precision DC collector includes: A calibration architecture building module 11 is used to integrate a traceable reference source and a temperature sensor in a target DC collector, connect the traceable reference source and the collector signal channel in parallel through a multiplexer, and build a collector calibration architecture; a periodic calibration module 12 is used to use the collector calibration architecture to perform error calibration and record the target DC collector to obtain an acquisition error data set, and periodically calibrate the operational amplifier bias voltage based on the acquisition error data set to obtain a DC optimized collector; a compensation channel building module 13 is used to collect ambient temperature values in real time through the temperature sensor, and record calibration time data at the same time, perform regression fitting based on the ambient temperature value, calibration time data and the acquisition error data set, and build a temperature-time joint compensation channel; a signal calibration module 14 is used to construct a collector drift prediction network, introduce the collector drift prediction network to perform data calibration on the DC optimized collector based on the temperature-time joint compensation channel, and obtain a reconstructed signal calibration result.

[0072] Furthermore, the periodic calibration module 12 in the data calibration system of a high-precision DC collector is also used to: perform calibration strategy analysis on the collector calibration architecture to obtain a collector switching calibration strategy, wherein the collector switching calibration strategy includes a signal acquisition strategy and a switching calibration strategy; set a collector calibration period according to a data calibration target; if the collector is not in the calibration period, trigger the signal acquisition strategy, and activate the collector signal channel for DC signal acquisition through the signal acquisition strategy; when the collector is in the calibration period, trigger the switching calibration strategy, and switch to the traceable reference source through the switching calibration strategy to perform error calibration recording and obtain an acquisition error data set.

[0073] Furthermore, the periodic calibration module 12 in the data calibration system of the high-precision DC collector is also used to: switch to the traceable reference source through the switching calibration strategy to perform signal acquisition to obtain a reference source acquisition value set; record and obtain the theoretical standard value set of the traceable reference source; arrange and align the reference source acquisition value set and the theoretical standard value set according to the acquisition time to obtain an acquisition value sequence set and a standard value sequence set; calculate and obtain a deviation value set of the acquisition value sequence set and the standard value sequence set, and use the deviation value set as the acquisition error data set.

[0074] Furthermore, the periodic calibration module 12 in the data calibration system for a high-precision DC collector is also used to: initialize a PID controller according to characteristic information of the target DC collector, the PID controller including a proportional coefficient, an integral coefficient, and a differential coefficient; verify and tune the PID controller to generate a target PID controller; use the op amp bias voltage as the control output, and perform calibration calculations on the acquisition error data set based on the target PID controller according to the collector calibration period to obtain the DC optimized collector.

[0075] Furthermore, the compensation channel building module 13 in the data calibration system of the high-precision DC collector is also used to: perform data cleaning and normalization processing on the ambient temperature value, calibration time data and the acquisition error data set to obtain a standard ambient temperature value, a standard calibration time data and a standard acquisition error data set; use the standard ambient temperature value and the standard calibration time data as independent variables and the standard acquisition error data set as a dependent variable, perform multivariate regression fitting on the independent variables and the dependent variables, and generate a temperature-time error prediction model; obtain the original DC signal data set through the DC optimization collector, perform compensation analysis on the original DC signal data set based on the temperature-time error prediction model, and build the temperature-time joint compensation channel.

[0076] Furthermore, the signal calibration module 14 in the data calibration system of the high-precision DC collector is also used to: mine and obtain a historical data set of the DC collector, wherein the historical data set of the DC collector includes historical collection values, temperature values, time data, collector aging data, and corresponding collection error data; use an LSTM neural network structure to perform drift degree annotation and predictive supervision training on the historical data set of the DC collector to generate an initial drift prediction network; perform loss assessment and iterative optimization on the initial drift prediction network to obtain a collector drift prediction network.

[0077] Furthermore, the signal calibration module 14 in the data calibration system of the high-precision DC collector is also used to: use the cross-entropy loss function to perform loss assessment feedback on the initial drift prediction network to determine the model loss data; select a model optimizer based on the model loss data, and iteratively tune the initial drift prediction network based on the model optimizer to obtain the collector drift prediction network.

[0078] Furthermore, the signal calibration module 14 in the data calibration system of the high-precision DC collector is also used to: perform signal compensation on the DC optimized collector based on the temperature-time joint compensation channel to obtain a DC signal compensation data set; introduce the collector drift prediction network to perform drift prediction on the DC optimized collector and the current DC signal acquisition data to determine the collector drift prediction parameters; perform data calibration on the DC signal compensation data set based on the collector drift prediction parameters to obtain a reconstructed signal calibration result.

[0079] Furthermore, the signal calibration module 14 in the data calibration system of the high-precision DC collector is also used to: perform drift level evaluation on the collector drift prediction parameters to obtain the collector drift degree level, and perform calibration strategy analysis based on the collector drift degree level to determine the drift correction coefficient; calibrate and correct the DC signal compensation data set based on the drift correction coefficient to obtain the reconstructed signal calibration result.

[0080] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The data calibration method and specific examples of a high-precision DC collector in Example 1 are also applicable to a data calibration system of a high-precision DC collector in this embodiment. Through the above detailed description of the data calibration method of a high-precision DC collector, those skilled in the art can clearly understand the data calibration system of a high-precision DC collector in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here.

[0081] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0082] Obviously, for those skilled in the art, several improvements and modifications can be made to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the scope of protection of the present application.

Claims

1. A data calibration method for a high-precision DC collector, characterized in that: include: Integrate a traceable reference source and a temperature sensor in the target DC collector, connect the traceable reference source and the collector signal channel in parallel through a multiplexer, and build a collector calibration architecture; Performing error calibration and recording on the target DC collector using the collector calibration architecture to obtain a collection error data set, and periodically calibrating the operational amplifier bias voltage based on the collection error data set to obtain a DC optimized collector; The temperature sensor collects the ambient temperature value in real time, and records the calibration time data at the same time, and performs regression fitting based on the ambient temperature value, the calibration time data and the acquisition error data set to build a temperature-time joint compensation channel; A collector drift prediction network is constructed, and the collector drift prediction network is introduced to perform data calibration on the DC optimized collector based on the temperature-time joint compensation channel to obtain a reconstructed signal calibration result.

2. The data calibration method for a high-precision DC collector according to claim 1, characterized in that: The obtaining of the acquisition error data set includes: Performing calibration strategy analysis on the collector calibration architecture to obtain a collector switching calibration strategy, wherein the collector switching calibration strategy includes a signal acquisition strategy and a switching calibration strategy; Set the collector calibration cycle according to the data calibration target; If the collector is not in the calibration period, triggering the signal acquisition strategy, activating the collector signal channel to perform DC signal acquisition through the signal acquisition strategy; When the collector is in a calibration cycle, the switching calibration strategy is triggered, and the error calibration record is switched to the traceable reference source through the switching calibration strategy to obtain a collection error data set.

3. The data calibration method of a high-precision DC collector according to claim 2, characterized in that: The obtaining of the acquisition error data set includes: Switching to the traceable reference source through the switching calibration strategy to perform signal acquisition, and obtaining a reference source acquisition value set; Recording and obtaining a set of theoretical standard values of the traceable reference source; Aligning the reference source collected value set and the theoretical standard value set according to collection time to obtain a collection value sequence set and a standard value sequence set; A deviation value set between the collection value sequence set and the standard value sequence set is calculated and obtained, and the deviation value set is used as the collection error data set.

4. The data calibration method for a high-precision DC collector according to claim 2, characterized in that: The DC optimized collector is obtained, comprising: Initializing a PID controller according to characteristic information of the target DC collector, wherein the PID controller includes a proportional coefficient, an integral coefficient, and a differential coefficient; Verifying and tuning the PID controller to generate a target PID controller; The operational amplifier bias voltage is used as the control output, and the acquisition error data set is calibrated and calculated based on the target PID controller according to the collector calibration period to obtain the DC optimized collector.

5. The data calibration method for a high-precision DC collector according to claim 1, wherein: The construction of the temperature-time joint compensation channel includes: Performing data cleaning and normalization processing on the ambient temperature value, calibration time data, and acquisition error data set to obtain a standard ambient temperature value, a standard calibration time data, and a standard acquisition error data set; Taking the standard ambient temperature value and the standard calibration time data as independent variables and the standard acquisition error data set as dependent variables, performing multiple regression fitting on the independent variables and the dependent variables to generate a temperature-time error prediction model; The original DC signal data set is acquired through the DC optimization collector, and compensation analysis is performed on the original DC signal data set based on the temperature-time error prediction model to build the temperature-time joint compensation channel.

6. The data calibration method for a high-precision DC collector according to claim 1, characterized in that: The construction of the collector drift prediction network includes: Mining and obtaining a DC collector historical data set, wherein the DC collector historical data set includes historical collection values, temperature values, time data, collector aging data, and corresponding collection error data; An LSTM neural network structure is used to perform drift degree annotation and prediction supervision training on the DC collector historical data set to generate an initial drift prediction network; The initial drift prediction network is subjected to loss assessment and iterative tuning to obtain a collector drift prediction network.

7. The data calibration method for a high-precision DC collector according to claim 6, characterized in that: The step of obtaining a collector drift prediction network includes: Using a cross entropy loss function to perform loss assessment feedback on the initial drift prediction network to determine model loss data; A model optimizer is selected according to the model loss data, and the initial drift prediction network is iteratively tuned based on the model optimizer to obtain the collector drift prediction network.

8. The data calibration method for a high-precision DC collector according to claim 1, characterized in that: Obtaining the reconstructed signal calibration result includes: Performing signal compensation on the DC optimization collector based on the temperature-time joint compensation channel to obtain a DC signal compensation data set; Introducing the collector drift prediction network to perform drift prediction on the DC optimization collector and the current DC signal collection data, and determining the collector drift prediction parameters; The DC signal compensation data set is calibrated based on the collector drift prediction parameter to obtain a reconstructed signal calibration result.

9. The data calibration method for a high-precision DC collector according to claim 8, characterized in that: The obtaining of the reconstructed signal calibration result includes: Performing drift level evaluation on the collector drift prediction parameter to obtain a collector drift degree level, and performing calibration strategy analysis based on the collector drift degree level to determine a drift correction coefficient; The DC signal compensation data set is calibrated and corrected based on the drift correction coefficient to obtain the reconstructed signal calibration result.

10. A data calibration system for a high-precision DC collector, characterized in that: The method for calibrating the data of a high-precision DC collector according to any one of claims 1 to 9 is implemented, wherein the data calibration system for the high-precision DC collector comprises: A calibration architecture building module is used to integrate a traceable reference source and a temperature sensor in the target DC collector, and connect the traceable reference source and the collector signal channel in parallel through a multiplexer to build a collector calibration architecture; A periodic calibration module, configured to perform error calibration and record on the target DC collector using the collector calibration architecture to obtain a collection error data set, and to periodically calibrate the operational amplifier bias voltage based on the collection error data set to obtain a DC optimized collector; A compensation channel building module is used to collect the ambient temperature value in real time through the temperature sensor, record the calibration time data at the same time, perform regression fitting based on the ambient temperature value, the calibration time data and the acquisition error data set, and build a temperature-time joint compensation channel; The signal calibration module is used to construct a collector drift prediction network, introduce the collector drift prediction network to perform data calibration on the DC optimization collector based on the temperature-time joint compensation channel, and obtain a reconstructed signal calibration result.

Images