Online calibration method and system for low-voltage current transformer

Abstract

The invention relates to the technical field of electrical measurement calibration, in particular to an online calibration method and system for a low-voltage current transformer, and the method comprises the steps: collecting data through a sensor, and generating an original data set; calculating a dynamic contact resistance value and a temperature rise rate by using the original data set, and generating a state index; performing correlation analysis on the state index and the basic error measurement value, and outputting an error contribution coefficient; adjusting an error calibration algorithm parameter by adopting an error contribution coefficient, and compensating a real-time measured value to obtain a calibrated error value; and comparing the calibrated error value with a standard transformer model to realize verification feedback, outputting a final calibration error value and storing related data. The system comprises a data acquisition module, a state evaluation module, an error source identification module, an adaptive calibration module, a verification feedback module and a data management module. According to the invention, the problems of contact resistance increase caused by aging of the crimping device and error measurement deviation caused by local heating are solved, and the calibration precision and the system reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of electrical metrology and calibration technology, and in particular to an online calibration method and system for low-voltage current transformers. Background Technology

[0002] Low-voltage current transformers are key equipment in the field of power metering, used to convert large primary currents into standard secondary signals. Their metering accuracy directly affects the fairness of electricity trading. Automated verification technology integrates computer control, precision measurement units, and mechanical actuators to automate the entire verification process of current transformers. The system first automatically completes the installation and wiring of the current transformer, then applies stepped current excitation and synchronously collects output data. Next, it uses digital algorithms to compare the measured values ​​with the standard model, thereby evaluating key indicators such as ratio error and phase angle error.

[0003] Existing automated verification technology for low-voltage current transformers suffers from the following technical challenges: Firstly, the primary circuit crimping device may experience increased contact resistance due to aging from long-term use or inadequate initial design. This leads to localized heating in high-current testing scenarios, causing changes in the transformer's temperature characteristics and consequently resulting in measurement deviations that affect calibration accuracy. Secondly, during continuous operation in automated production lines, oxidation or mechanical wear at the crimping points can cause poor contact, increased resistance, and significant localized temperature rise during testing. This results in fluctuations in measurement data and ultimately reduces the reliability of the verification results. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an online calibration method and system for low-voltage current transformers. This invention solves the technical problem that in high-current testing scenarios on automated calibration lines for low-voltage current transformers, aging or inadequate design of the primary circuit crimping device leads to increased contact resistance and localized heating, resulting in measurement deviations and decreased calibration accuracy.

[0005] To solve the above-mentioned technical problems, the specific contents of the present invention are as follows:

[0006] In a first aspect, the present invention provides an online calibration method for a low-voltage current transformer, comprising: Step 1: Collect real-time primary current value, secondary voltage value, and real-time temperature data of the contact point of the primary circuit crimping device of the low-voltage current transformer under the high current test scenario of the primary circuit, and generate the original dataset. Step 2: Using the original dataset, calculate the dynamic contact resistance value and temperature rise rate of the crimping device to generate status indicators; Step 3: Perform correlation analysis between the status indicators and the basic error measurement values ​​of the low-voltage current transformer, and output the error contribution coefficient; Step 4: Adjust the parameters of the error calibration algorithm using the error contribution coefficient to compensate for the basic error value of the real-time measurement and obtain the calibrated error value. Step 5: Compare the calibrated error value with the standard current transformer model. If the comparison result exceeds the allowable range, re-execute the parameter adjustment until the comparison result is within the allowable range. Then output the final calibration error value, save the final calibration error value, status indicators and calibration parameters.

[0007] Furthermore, the online calibration method for low-voltage current transformers of the present invention is characterized in that step 1 includes: The primary current value is acquired using a Hall effect current sensor; The secondary voltage value is acquired using a differential voltage sensor; The temperature of the contact point of the crimping device is collected using an infrared temperature sensor; The collected primary current, secondary voltage, and temperature data are timestamped and filtered to generate the original dataset.

[0008] Furthermore, the online calibration method for low-voltage current transformers of the present invention is characterized in that step 2 includes: Instantaneous current and voltage values ​​are extracted from the original dataset, and Ohm's law is applied to calculate the contact resistance. The temperature rise rate is obtained by performing first-order differential processing on the temperature data in the original dataset. The calculated contact resistance value and temperature rise rate are input into the fuzzy logic system, and the output status index is generated.

[0009] Furthermore, the online calibration method for low-voltage current transformers of the present invention is characterized in that step 3 includes: Using state indicators and basic error measurements, principal component analysis was employed to separate the error components. Call up the historical error patterns of the same type of current transformer and calculate the similarity between the current error pattern and the historical error pattern; The output error contribution coefficient is based on the proportion of the error component caused by the crimping device in the similarity measurement error component.

[0010] Furthermore, the online calibration method for low-voltage current transformers of the present invention is characterized in that step 4 includes: The parameters of the error calibration algorithm are adjusted using the error contribution coefficient as a weighting factor. The basic error measurement value is processed using the error calibration algorithm with adjusted parameters to obtain the calibrated error value.

[0011] Furthermore, the online calibration method for low-voltage current transformers of the present invention is characterized in that step 5 includes: Calculate the residual between the calibrated error value and the expected value of the standard current transformer model; Analyze the residual values ​​and trends of multiple consecutive sampling points; When the residual value exceeds the preset threshold and the trend of change conforms to the predetermined pattern, it is determined that the parameter adjustment in step 4 will be re-executed.

[0012] Furthermore, the online calibration method for low-voltage current transformers of the present invention is characterized in that step 5 includes: The integrity of the final calibration error value, status indicators, and calibration parameters is verified. The verified data is packaged according to a predetermined structure; The encapsulated data is written to the local storage medium, and a corresponding data index is generated; Key information is extracted from the data index and transmitted to the monitoring platform via a secure communication link.

[0013] Furthermore, the online calibration method for low-voltage current transformers of the present invention is characterized by further comprising: Before performing step 1, the sensors used to collect primary current, secondary voltage, and temperature data are calibrated. After performing step 5, a calibration report is generated, which includes the basic error measurements before calibration, the final calibration error values ​​after calibration, and status indicators.

[0014] Furthermore, the online calibration method for low-voltage current transformers of the present invention is characterized in that calling the historical error modes of transformers of the same model includes: Build a historical database to store historical error data and corresponding status indicators of various types of instrument transformers; When it is necessary to identify the error source of the current transformer, retrieve historical data with the same or similar model as the current transformer from the historical database; When the same historical error pattern exists in the historical database, it is directly used as a reference template. When no historical data for the same model exists, the optimal matching pattern is selected from the historical data of similar models based on the similarity of status indicators as a reference template.

[0015] Secondly, the present invention provides an online calibration system for a low-voltage current transformer, applied to the online calibration method for a low-voltage current transformer as described above, comprising: The data acquisition module is configured to acquire real-time primary current value, secondary voltage value, and real-time temperature data of the contact point of the primary circuit crimping device of the low-voltage current transformer under the high current test scenario of the primary circuit, and generate raw dataset. The condition assessment module is configured to receive the raw dataset, calculate the dynamic contact resistance value and temperature rise rate of the crimping device, and generate condition indicators. The error source identification module is configured to receive status indicators and basic error measurement values ​​of low-voltage current transformers, perform correlation analysis, and output error contribution coefficients. The adaptive calibration module is configured as the receiving error contribution coefficient, which adjusts the parameters of the error calibration algorithm accordingly and compensates for the basic error value of the real-time measurement to obtain the calibrated error value. The verification feedback module is configured to compare the calibrated error value with the standard current transformer model. When the comparison result exceeds the allowable range, a recalibration command is triggered until the comparison result is within the allowable range, at which point the final calibration error value is output. The data management module is configured to store the final calibration error value, status indicators, and calibration parameters, and to complete the data upload.

[0016] Beneficial effects of this invention: This invention generates a raw dataset by real-time acquisition of primary current, secondary voltage, and temperature data. Based on this raw dataset, it calculates the dynamic contact resistance and temperature rise rate of the crimping device to generate status indicators. Then, it performs correlation analysis with basic error measurements to output an error contribution coefficient. Subsequently, it uses this error contribution coefficient to adjust the parameters of the error calibration algorithm and compensate for the basic error values. Finally, it achieves closed-loop control by comparing the calibrated error values ​​with a standard model. This effectively solves the problem of error measurement deviation caused by increased contact resistance and localized heating due to aging or insufficient design of the primary circuit crimping device in the automated calibration production line of low-voltage current transformers, improving calibration accuracy and system reliability. Furthermore, the construction of a historical database and a similarity matching mechanism enhance the adaptability of error source identification, while data integrity verification and secure transmission ensure process traceability. Attached Figure Description

[0017] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the drawings without creative effort.

[0018] Figure 1 The flowchart illustrates an online calibration method for low-voltage current transformers provided by this invention. Detailed Implementation

[0019] To make the technical solution of the present invention clearer, the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. The present invention provided by various embodiments will be described in detail below with reference to the accompanying drawings. To better understand the purpose of the present invention, the present invention will be described in further detail below.

[0020] Firstly, please refer to Figure 1 The present invention provides an online calibration method for a low-voltage current transformer, comprising: Step 1: Collect real-time primary current value, secondary voltage value, and real-time temperature data of the contact point of the primary circuit crimping device of the low-voltage current transformer under the high current test scenario of the primary circuit, and generate the original dataset. Step 2: Using the original dataset, calculate the dynamic contact resistance value and temperature rise rate of the crimping device to generate status indicators; Step 3: Perform correlation analysis between the status indicators and the basic error measurement values ​​of the low-voltage current transformer, and output the error contribution coefficient; Step 4: Adjust the parameters of the error calibration algorithm using the error contribution coefficient to compensate for the basic error value of the real-time measurement and obtain the calibrated error value. Step 5: Compare the calibrated error value with the standard current transformer model. If the comparison result exceeds the allowable range, re-execute the parameter adjustment until the comparison result is within the allowable range. Then output the final calibration error value, save the final calibration error value, status indicators and calibration parameters.

[0021] In the online calibration method for low-voltage current transformers, the real-time primary current value, secondary voltage value, and real-time temperature data of the primary circuit crimping device contact points are first acquired using sensors under a high-current test scenario. Specifically, a Hall effect current sensor is used to acquire the primary current value, a differential voltage sensor is used to acquire the secondary voltage value, and an infrared temperature sensor is used to acquire the temperature of the crimping device contact points. The acquired primary current value, secondary voltage value, and temperature data are then timestamped and filtered to generate the raw dataset.

[0022] The dynamic contact resistance and temperature rise rate of the crimping device are calculated using the original dataset to generate a status index. Instantaneous current and voltage values ​​are extracted from the original dataset, and Ohm's law is applied to calculate the contact resistance. Simultaneously, the temperature data in the original dataset is processed using the first-order derivative to obtain the temperature rise rate. The calculated contact resistance and temperature rise rate are input into a fuzzy logic system, which outputs the status index.

[0023] The system performs correlation analysis between the status indicators and the basic error measurements of the low-voltage current transformer, outputting the error contribution coefficient. Principal component analysis is used to separate the error components, and historical error patterns of transformers of the same model are retrieved to calculate the similarity between the current error pattern and historical error patterns. Based on the similarity measure, the proportion of error components caused by the crimping device is quantified, and the error contribution coefficient is output.

[0024] The parameters of the error calibration algorithm are adjusted using an error contribution coefficient to compensate for the basic error value of the real-time measurement, thus obtaining the calibrated error value. The error contribution coefficient is then used as a weighting factor to adjust the parameters of the error calibration algorithm. The adjusted error calibration algorithm is then used to process the basic error measurement value to obtain the calibrated error value.

[0025] The calibrated error value is compared with the standard current transformer model. If the comparison result exceeds the allowable range, parameter adjustment is performed again until the comparison result is within the allowable range, at which point the final calibration error value is output. The residual between the calibrated error value and the expected value of the standard current transformer model is calculated, and the residual values ​​and trends of multiple consecutive sampling points are analyzed. When the residual value exceeds a preset threshold and the trend conforms to a predetermined pattern, parameter adjustment is performed again. Finally, the final calibration error value, status indicators, and calibration parameters are saved. The integrity of the final calibration error value, status indicators, and calibration parameters is verified. The verified data is encapsulated according to a predetermined structure, written to local storage media, and a data index is generated. Key information is extracted based on the data index and transmitted to the monitoring platform via a secure communication link.

[0026] In step 1 of this invention, a primary current value is acquired using a Hall effect current sensor, a secondary voltage value is acquired using a differential voltage sensor, and the temperature of the contact point of the crimping device is acquired using an infrared temperature sensor. The acquired primary current value, secondary voltage value, and temperature data are then time-stamped and filtered to generate the original dataset. Time-stamp alignment synchronizes the data from different sensors in time, and filtering removes noise and improves data quality, thus providing accurate input for subsequent calculations.

[0027] In step 2 of this invention, instantaneous current and voltage values ​​are extracted from the original dataset, and Ohm's law is applied to calculate the contact resistance value. The temperature data in the original dataset is processed using first-order differentiation to obtain the temperature rise rate. The calculated contact resistance value and temperature rise rate are input into a fuzzy logic system, which outputs a status index. Ohm's law is used to directly calculate resistance, the first-order derivative reflects the rate of temperature change, and the fuzzy logic system integrates various parameters to generate a status index reflecting the state of the crimping device.

[0028] In step 3 of this invention, principal component analysis (PCA) is used to separate error components using state indicators and basic error measurements. Historical error patterns of current transformers of the same model are retrieved, and the similarity between the current error pattern and the historical error patterns is calculated. Based on the similarity metric, the proportion of error components caused by the crimping device is quantified, and the error contribution coefficient is output. PCA identifies the source of error, historical pattern comparison provides a reference, and similarity metric helps determine the degree of influence of the crimping device.

[0029] In step 4 of this invention, the error contribution coefficient is used as a weighting factor to adjust the parameters of the error calibration algorithm. The adjusted error calibration algorithm is then used to process the basic error measurement values ​​to obtain the calibrated error values. Adjusting the weighting factor parameters makes the calibration algorithm more focused on the errors caused by the crimping device, thereby compensating for the measured values ​​and improving calibration accuracy.

[0030] In step 5 of this invention, the residual between the calibrated error value and the expected value of the standard transformer model is calculated; the residual values ​​and trends of multiple consecutive sampling points are analyzed; when the residual value exceeds a preset threshold and the trend conforms to a predetermined pattern, the parameter adjustment in step 4 is re-executed. Residual analysis monitors the calibration effect, and trend judgment ensures the timeliness of adjustment and avoids the accumulation of errors.

[0031] In step 5 of this invention, the final calibration error value, status indicators, and calibration parameters are verified for integrity; the verified data is encapsulated according to a predetermined structure; the encapsulated data is written to a local storage medium, and a corresponding data index is generated; key information is extracted based on the data index and transmitted to the monitoring platform via a secure communication link. Integrity verification ensures data reliability, encapsulation and indexing facilitate storage and retrieval, and secure transmission ensures data security.

[0032] This invention also includes calibrating the sensor used to collect primary current, secondary voltage, and temperature data before performing step 1; and generating a calibration report after performing step 5. The calibration report includes the basic error measurement values ​​before calibration, the final calibration error values ​​after calibration, and status indicators. Sensor calibration improves data acquisition accuracy, and the calibration report records the process, facilitating traceability and analysis.

[0033] In step 3 of this invention, calling historical error patterns of the same model of current transformer includes constructing a historical database to store historical error data and corresponding status indicators for various models of current transformers. When error source identification of the current current transformer is required, historical data of the same or similar model to the current current transformer is retrieved from the historical database. If a historical error pattern of the same model exists in the historical database, it is directly called as a reference template. If no historical data of the same model exists, the optimal matching pattern is selected from historical data of similar models based on the similarity of the status indicators as a reference template. The historical database provides rich references, and similarity matching ensures that a suitable pattern can be found even if the models are different, enhancing the adaptability of error source identification.

[0034] Secondly, the present invention provides an online calibration system for a low-voltage current transformer, applied to the online calibration method for a low-voltage current transformer as described above, comprising: The data acquisition module is configured to acquire real-time primary current value, secondary voltage value, and real-time temperature data of the contact point of the primary circuit crimping device of the low-voltage current transformer under the high current test scenario of the primary circuit, and generate raw dataset. The condition assessment module is configured to receive the raw dataset, calculate the dynamic contact resistance value and temperature rise rate of the crimping device, and generate condition indicators. The error source identification module is configured to receive status indicators and basic error measurement values ​​of low-voltage current transformers, perform correlation analysis, and output error contribution coefficients. The adaptive calibration module is configured as the receiving error contribution coefficient, which adjusts the parameters of the error calibration algorithm accordingly and compensates for the basic error value of the real-time measurement to obtain the calibrated error value. The verification feedback module is configured to compare the calibrated error value with the standard current transformer model. When the comparison result exceeds the allowable range, a recalibration command is triggered until the comparison result is within the allowable range, at which point the final calibration error value is output. The data management module is configured to store the final calibration error value, status indicators, and calibration parameters, and to complete the data upload.

[0035] This invention addresses the issues of increased contact resistance and localized overheating caused by aging or inadequate design of the primary circuit crimping device in automated calibration lines for low-voltage current transformers through a multi-step collaborative processing approach. First, sensors collect real-time data on the primary current, secondary voltage, and contact point temperature of the crimping device under high-current testing scenarios, generating a raw dataset to provide a foundation for condition monitoring. Next, the raw dataset is used to calculate the dynamic contact resistance and temperature rise rate of the crimping device, generating condition indicators to quantify its real-time status. Then, the condition indicators are correlated with basic error measurements, and principal component analysis is used to separate error components. Based on historical error pattern similarity, the error contribution coefficient is calculated to accurately identify the proportion of error caused by the crimping device. Subsequently, the error contribution coefficient is used as a weighting factor to adjust the parameters of the error calibration algorithm, compensating for the real-time measured basic error value to obtain a calibrated error value that dynamically offsets the deviation introduced by the crimping device. Finally, the calibrated error value is compared with a standard transformer model. Residual analysis and trend judgment are used to verify the calibration effect. When the result exceeds limits, the parameters are readjusted to form a closed-loop control until the final calibration error value is output, and the relevant data is saved. The process effectively suppresses the negative impact of increased contact resistance and localized heating on measurement accuracy and improves calibration reliability through a continuous data stream of real-time monitoring, status assessment, error source identification, adaptive calibration, and verification feedback.

[0036] This invention relates to the online calibration process of low-voltage current transformers in an automated verification production line. Addressing the issues of increased contact resistance and localized overheating in the primary circuit crimping device due to aging or design deficiencies, it achieves error compensation and accuracy improvement through coordinated processing of real-time data acquisition, status assessment, error source identification, adaptive calibration, and verification feedback. The implementation process is as follows: During the data acquisition phase, a Hall effect current sensor was used to collect the primary current value under high-current testing conditions in the primary circuit, a differential voltage sensor was used to collect the secondary voltage value, and an infrared temperature sensor was used to monitor the temperature of the contact point of the crimping device. The collected primary current, secondary voltage, and temperature data were timestamped to synchronize the multi-source data on the timeline. Then, a filtering algorithm was used to remove noise interference, generating a high-quality raw dataset that provides reliable input for subsequent analysis.

[0037] In the condition assessment phase, the dynamic contact resistance and temperature rise rate of the crimping device are calculated using the original dataset. Instantaneous current and voltage values ​​are extracted from the original dataset, and Ohm's law is applied to calculate the contact resistance, reflecting the conductivity characteristics of the crimping point. Simultaneously, the temperature data is processed using first-order differentiation to obtain the temperature rise rate, characterizing the heating trend. The calculated contact resistance and temperature rise rate are input into a fuzzy logic system. Membership functions and inference rules are used to handle parameter uncertainties, outputting a comprehensive condition index to quantify the real-time operating status of the crimping device.

[0038] In the error source identification stage, the status indicators are correlated with the basic error measurements of the low-voltage current transformer. Principal component analysis is used to separate error components and identify the main error sources. Historical error patterns of transformers of the same model are retrieved from the historical database, and similar data are calculated between the current error pattern and the historical error pattern. Based on the similarity results, the proportion of error components caused by the crimping device is quantified, and the error contribution coefficient is output to accurately distinguish between device-related errors and inherent transformer errors.

[0039] In the adaptive calibration phase, the error contribution coefficient is used as a weighting factor to adjust the parameters of the error calibration algorithm. The weighting factor modifies the gain or bias settings in the algorithm, prioritizing compensation for deviations caused by the crimping device during the calibration process. The adjusted error calibration algorithm processes the basic error values ​​measured in real time, and a weighted compensation strategy is used to generate calibrated error values, dynamically offsetting the effects of increased contact resistance and localized heating.

[0040] The verification feedback phase compares the calibrated error value with the standard current transformer model, calculates the residual to evaluate the calibration effect, and analyzes the residual values ​​and trends of multiple consecutive sampling points. When the residual exceeds a preset threshold and the trend conforms to a predetermined pattern, the parameter adjustment command is re-triggered, forming a closed-loop control. The iterative process continues until the residual is within the allowable range, and the final calibration error value is output.

[0041] The data management phase verifies the integrity of the final calibration error value, status indicators, and calibration parameters, checking for missing data or format errors. Verified data is packaged according to a predetermined structure, written to local storage media, and a data index is generated. Key information is extracted from the index and transmitted to the monitoring platform via a secure communication link, supporting long-term traceability and auditing. Furthermore, sensors are calibrated before data acquisition to ensure accuracy; after calibration, a calibration report is generated, recording the basic error measurements before and after calibration, the final calibration error value, and status indicators, facilitating operational analysis.

[0042] The implementation process utilizes a multi-step closed-loop data flow, forming a complete adaptive calibration chain from physical parameter acquisition and status indicator generation to error compensation and verification. The construction of a historical database and a similarity matching mechanism enhance the adaptability of error source identification, while real-time monitoring and feedback control ensure the calibration reliability of the system under aging high-voltage connection devices. This method effectively suppresses measurement deviations introduced by increased contact resistance and localized heating, improving the calibration accuracy and stability of low-voltage current transformers in automated verification lines.

[0043] In step 2, the fuzzy logic system includes input variables, output variables, and fuzzy rules. The input variables are contact resistance value and temperature rise rate, and the output variable is a state index. A fuzzy set is defined for the contact resistance value, including low, medium, and high, and a fuzzy set is defined for the temperature rise rate, including slow, medium, and fast. The input values ​​are mapped to the fuzzy sets using a triangular membership function, and then inference is performed according to preset fuzzy rules. The fuzzy rules include that if the contact resistance value is high and the temperature rise rate is fast, the state index is poor; if the contact resistance value is low and the temperature rise rate is slow, the state index is good. The inference result is defuzzified using the centroid method to obtain the accurate state index value.

[0044] In step 3, the principal component analysis method extracts eigenvalues ​​and eigenvectors by calculating the covariance matrix of the state index and the basic error measurement values, and selects the principal components with eigenvalues ​​greater than 1 as error components. The historical database stores historical error data and corresponding state indexes for various types of instrument transformers. The historical error data includes basic error measurements and calibrated error values. When a historical error mode is invoked, records of the same model as the current instrument transformer are retrieved from the historical database. The Euclidean distance between the current error mode and the historical error mode is calculated as the similarity. Based on the similarity-weighted average, the proportion of error components caused by the crimping device is quantified.

[0045] In step 4, the error calibration algorithm employs a linear compensation model, with parameters including a gain coefficient and an offset. The error contribution coefficient is used as a weighting factor to adjust the gain coefficient; the adjusted gain coefficient is equal to the original gain coefficient multiplied by the error contribution coefficient. The basic error measurement value is processed using the adjusted error calibration algorithm, and the calibrated error value is equal to the basic error measurement value minus the product of the adjusted gain coefficient and the basic error measurement value.

[0046] In step 5, the preset threshold is set according to the accuracy requirements of the standard current transformer model, and the preset threshold for the absolute value of the residual is 5% of the basic error measurement value. The predetermined pattern is that the residual increases or decreases monotonically for three consecutive sampling points. When the residual value exceeds the preset threshold and the trend conforms to the predetermined pattern, a recalibration command is triggered. Integrity verification includes checking whether there are missing values ​​in the data fields and whether the values ​​exceed the physical range. The verified data is encapsulated in JSON format, including a timestamp, the final calibration error value, status indicators, and calibration parameters. The secure communication link uses the TLS protocol for encrypted data transmission.

[0047] In step 1, the timestamp alignment process uses GPS synchronization signals to ensure time consistency of data from all sensors, and the filtering process uses a Butterworth low-pass filter to remove high-frequency noise. The Hall effect current sensor's measurement range is set to 0-1000A with an accuracy of ±0.5%; the differential voltage sensor's measurement range is set to 0-5V with an accuracy of ±0.1%; and the infrared temperature sensor's measurement range is set to 0-150°C with an accuracy of ±1°C. The original dataset is stored in tabular form, containing timestamp, primary current value, secondary voltage value, and temperature data fields.

[0048] In step 2, the specific process of calculating the contact resistance value using Ohm's law is as follows: Instantaneous current and voltage values ​​at the same timestamp are extracted from the original dataset. The contact resistance value is equal to the instantaneous voltage value divided by the instantaneous current value. The first-order differential processing of the temperature rise rate uses the central difference method, calculating the difference between the current temperature value and the temperature value at the previous sampling point, divided by the sampling time interval. The fuzzy sets of the input variable, contact resistance value, with fuzzy domains of low, medium, and high are set to 0-10mΩ, 10-50mΩ, and 50-100mΩ, respectively. The fuzzy sets of the temperature rise rate with fuzzy domains of slow, medium, and fast are set to 0-5°C / s, 5-10°C / s, and 10-20°C / s, respectively. The output variable, state index, has a domain of 0-100, representing the state of the crimping device from excellent to poor. The fuzzy rule base contains 9 rules; for example, if the contact resistance value is low and the temperature rise rate is slow, the state index is excellent; if the contact resistance value is high and the temperature rise rate is fast, the state index is poor. The defuzzification process uses the centroid method to calculate the precise value of the state index.

[0049] In step 3, the specific process of separating error components using principal component analysis is as follows: Construct a sample matrix of state indicators and basic error measurements; calculate the covariance matrix of the sample matrix; solve for the eigenvalues ​​and eigenvectors of the covariance matrix; select the first k principal components with eigenvalues ​​greater than 1 as error components, where k is determined by the cumulative contribution rate, and the cumulative contribution rate threshold is set to 85%. The similarity calculation of historical error patterns uses the Euclidean distance formula; the smaller the distance, the higher the similarity. The error contribution coefficient quantification process is as follows: calculate the reciprocal of the Euclidean distance between the current error pattern and the historical error pattern, normalize it, use it as a weight, and weighted average the proportion of error caused by the crimping device in the historical error pattern to output the error contribution coefficient.

[0050] In step 4, the linear compensation model of the error calibration algorithm is expressed as: Calibrated error value = Basic error measurement value - (Gain coefficient × Basic error measurement value + Offset). The initial value of the gain coefficient is set to 0.1, and the initial value of the offset is set to 0. When adjusting the parameters, the gain coefficient is adjusted to the original gain coefficient multiplied by the error contribution coefficient, while the offset remains unchanged. The calibrated error value is calculated using the adjusted gain coefficient and offset.

[0051] In step 5, the preset threshold is determined based on the accuracy level of the standard current transformer model. For a 0.5 class current transformer, the preset threshold for the absolute value of the residual is set to 5% of the basic error measurement value. The predetermined mode is defined as the residual monotonically increasing or monotonically decreasing for three consecutive sampling points. The condition for monotonically increasing is residual(n) > residual(n-1) > residual(n-2), and the condition for monotonically decreasing is residual(n) < residual(n-1) < residual(n-2). After the recalibration command is triggered, the system returns to step 4 to readjust the parameters. Integrity verification includes checking whether data fields are missing and whether the values ​​are within a reasonable range. The reasonable range is 0-1000A for primary current, 0-5V for secondary voltage, and 0-150°C for temperature. The predetermined structure is encapsulated in JSON format, containing key-value pairs: timestamp, final_error, status_index, and calibration_params. The local storage medium uses a solid-state drive, and the data index is organized using a B-tree structure. Secure communication links use the TLS 1.2 protocol and the AES-256 encryption algorithm.

[0052] In step 3, the historical database is constructed using a relational database, and the storage table structure includes transformer model, historical error data, status indicators, and timestamps. The optimal matching mode selection process is as follows: calculate the Euclidean distance between the current status indicator and the status indicators of similar transformer models in the historical database, and select the historical data with the smallest distance as the reference template.

[0053] In step 2, the formula for calculating the contact resistance value using Ohm's law is: ; in, This indicates the contact resistance value, measured in ohms. It represents the instantaneous value of voltage, and the unit is volts; This represents the instantaneous current value, measured in amperes. Instantaneous voltage and current values ​​at the same timestamp are extracted from the original dataset and substituted into the formula to calculate the dynamic contact resistance value.

[0054] The rate of temperature rise is calculated using a first-order differential approximation: ; in, This indicates the rate of temperature rise, measured in degrees Celsius per second. This indicates the temperature value at the current sampling point, in degrees Celsius. This indicates the temperature value of the previous sampling point; This represents the sampling time interval, in seconds. The temperature sequence in the original dataset is differentially processed to obtain the temperature rise rate.

[0055] In step 3, the specific process of principal component analysis includes: constructing a sample matrix of state indices and basic error measurements. Each row represents a sample, and each column represents a variable. Calculate the covariance matrix of the sample matrix. : ; in, This represents the sample size. Solve for the covariance matrix. eigenvalues and eigenvectors Principal components with eigenvalues ​​greater than 1 were selected, and the cumulative contribution rate threshold was set to 85%. Principal component scores were calculated through projection. ,in It is the eigenvector matrix.

[0056] Similarity is calculated using Euclidean distance: ; in, This represents the Euclidean distance; the smaller the value, the higher the similarity. Indicates the current error pattern's... One eigenvalue; The first representing the historical error pattern One eigenvalue; This indicates the number of features. Similarity weights are calculated based on distance and used to quantify the error contribution coefficient.

[0057] In step 4, the linear compensation model of the error calibration algorithm is expressed as: ; in, This indicates the error value after calibration; Indicates the basic error measurement value; Indicates the gain coefficient; Indicates the offset. Gain coefficient. The initial value is set to 0.1, and the offset is... The initial value is set to 0. When adjusting the parameters, the gain coefficient is adjusted to... ,in Indicates the error contribution coefficient; offset. It remains unchanged.

[0058] In step 5, the residual calculation formula is: ; in, This represents the expected value of the standard current transformer model. The preset threshold is set according to the accuracy class of the current transformer; for a 0.5 class current transformer, the threshold is set to 5% of the basic error measurement value. Trend analysis checks the residual sequence of three consecutive sampling points to determine whether it is monotonically increasing or monotonically decreasing.

[0059] Embodiment 1 of this invention: In an automated calibration line, the primary circuit crimping device of a low-voltage current transformer is subjected to high-current testing for extended periods. Due to oxidation, the contact resistance at the contact points increases from an initial 10mΩ to 50mΩ. Operators deploy Hall effect current sensors to collect primary current values, differential voltage sensors to monitor secondary voltage values, and infrared temperature sensors to track the temperature of the crimping points. Sensor data is time-stamped via GPS and noise is removed using a Butterworth filter to generate a raw dataset containing time series data. The system extracts the instantaneous current value of 500A and the instantaneous voltage value of 25mV from the dataset and calculates the dynamic contact resistance value as 50mΩ using Ohm's law. Temperature data is processed using the center differential method, with a sampling interval of 1 second, and the temperature rise rate is calculated to be 8°C / s. The contact resistance value and temperature rise rate are input into the fuzzy logic system. The membership function for the contact resistance value is set to a low range of 0-10mΩ, a medium range of 10-50mΩ, and a high range of 50-100mΩ. The membership function for the temperature rise rate is set to a slow range of 0-5°C / s, a medium range of 5-10°C / s, and a fast range of 10-20°C / s. The fuzzy rule base contains nine rules. For example, a high contact resistance value and a fast temperature rise rate correspond to a poor state index, and the output state index is 75. Principal component analysis is performed on the state index and the basic error measurement value at 0.5%. A sample matrix is ​​constructed, and the covariance matrix is ​​calculated. The cumulative contribution rate of the principal components with eigenvalues ​​greater than 1 is 85%, thus separating the error components. Historical error patterns of the same model of current transformer are retrieved from the historical database. The Euclidean distance is used to calculate the similarity to 0.9, and the error contribution coefficient is quantified to 0.6. The error calibration algorithm adopts a linear compensation model with an initial gain coefficient of 0.1 and an initial offset value of 0. After adjustment, the gain coefficient is 0.06. The calculated calibration error was 0.48%, with a residual of 0.03% compared to the expected value of 0.45% for the standard current transformer model. The residual sequence for three consecutive sampling points (0.03%, 0.032%, 0.035%) monotonically increased and exceeded the threshold of 0.025%, triggering a recalibration command. The final calibration error value of 0.46% was saved to the solid-state drive, with the data encapsulated in JSON format including a timestamp, status indicator 75, and calibration parameter gain coefficient 0.06, and transmitted to the monitoring platform via TLS 1.2 protocol.

[0060] Embodiment 2 of the present invention: In a long-term operating scenario, mechanical wear of the crimping device causes contact resistance fluctuations within the range of 30-70 mΩ. An infrared temperature sensor detects that the contact point temperature rises from 25°C to 60°C at a rate of 12°C / s. The fuzzy logic system inputs a contact resistance value of 70 mΩ and a temperature rise rate of 12°C / s, and outputs a status index of 85. After principal component analysis of the error components, data from similar current transformers of similar models are matched against a historical database, with an Euclidean distance similarity of 0.7 and an error contribution coefficient of 0.8. The gain coefficient is adjusted to 0.08, and the error value is compensated from the initial 0.6% to 0.52% after calibration. The residual sequence of 0.04%, 0.038%, and 0.036% shows a monotonically decreasing trend, but the absolute value of the residual of 0.04% exceeds the threshold of 0.03%, prompting the system to perform a secondary parameter adjustment. After three iterations, the residual error of the verification feedback module stabilized within 0.02%, and the final calibration error value of 0.49% was recorded in the calibration report. The report included the error value of 0.6% before calibration, the status index 85, and the calibration timestamp. The data management module verified the integrity of the numerical ranges; the primary current value of 800A, the secondary voltage value of 4V, and the temperature of 60°C were all within reasonable ranges. The data index B-tree structure storage supported fast retrieval.

Claims

1. A method for online calibration of a low-voltage current transformer, characterized in that, include; Step 1: Collect real-time primary current value, secondary voltage value, and real-time temperature data of the contact point of the primary circuit crimping device of the low-voltage current transformer under the high current test scenario of the primary circuit, and generate the raw dataset. Step 2: Using the original dataset, calculate the dynamic contact resistance value and temperature rise rate of the crimping device to generate status indicators; Step 3: Perform correlation analysis between the status indicators and the basic error measurement values ​​of the low-voltage current transformer, and output the error contribution coefficient; Step 4: Adjust the parameters of the error calibration algorithm using the error contribution coefficient to compensate for the basic error value of the real-time measurement and obtain the calibrated error value. Step 5: Compare the calibrated error value with the standard current transformer model. If the comparison result exceeds the allowable range, re-execute the parameter adjustment until the comparison result is within the allowable range. Then output the final calibration error value, save the final calibration error value, status indicators and calibration parameters.

2. The online calibration method for low-voltage current transformers according to claim 1, characterized in that, Step 1 includes: The primary current value is acquired using a Hall effect current sensor; The secondary voltage value is acquired using a differential voltage sensor; The temperature of the contact point of the crimping device is collected using an infrared temperature sensor; The collected primary current, secondary voltage, and temperature data are timestamped and filtered to generate the original dataset.

3. The online calibration method for low-voltage current transformers according to claim 2, characterized in that, Step 2 includes: Instantaneous current and voltage values ​​are extracted from the original dataset, and Ohm's law is applied to calculate the contact resistance. The temperature rise rate is obtained by performing first-order differential processing on the temperature data in the original dataset. The calculated contact resistance value and temperature rise rate are input into the fuzzy logic system, and the output status index is generated.

4. The online calibration method for low-voltage current transformers according to claim 3, characterized in that, Step 3 includes: Using state indicators and basic error measurements, principal component analysis was employed to separate the error components. Call up the historical error patterns of the same type of current transformer and calculate the similarity between the current error pattern and the historical error pattern; The output error contribution coefficient is based on the proportion of the error component caused by the crimping device in the similarity measurement error component.

5. The online calibration method for low-voltage current transformers according to claim 4, characterized in that, Step 4 includes: The parameters of the error calibration algorithm are adjusted using the error contribution coefficient as a weighting factor. The basic error measurement value is processed using the error calibration algorithm with adjusted parameters to obtain the calibrated error value.

6. The online calibration method for low-voltage current transformers according to claim 5, characterized in that, Step 5 includes: Calculate the residual between the calibrated error value and the expected value of the standard current transformer model; Analyze the residual values ​​and trends of multiple consecutive sampling points; When the residual value exceeds the preset threshold and the trend of change conforms to the predetermined pattern, it is determined that the parameter adjustment in step 4 will be re-executed.

7. The online calibration method for low-voltage current transformers according to claim 6, characterized in that, Step 5 includes: The integrity of the final calibration error value, status indicators, and calibration parameters is verified. The verified data is packaged according to a predetermined structure; The encapsulated data is written to the local storage medium, and a corresponding data index is generated; Key information is extracted from the data index and transmitted to the monitoring platform via a secure communication link.

8. The online calibration method for low-voltage current transformers according to claim 1, characterized in that, Also includes: Before performing step 1, the sensors used to collect primary current, secondary voltage, and temperature data are calibrated. After performing step 5, a calibration report is generated, which includes the basic error measurements before calibration, the final calibration error values ​​after calibration, and status indicators.

9. The online calibration method for low-voltage current transformers according to claim 8, characterized in that, In step 3, calling up the historical error modes of the same model of current transformer includes: Build a historical database to store historical error data and corresponding status indicators of various types of instrument transformers; When it is necessary to identify the error source of the current transformer, retrieve historical data with the same or similar model as the current transformer from the historical database; When the same historical error pattern exists in the historical database, it is directly used as a reference template. When no historical data for the same model exists, the optimal matching pattern is selected from the historical data of similar models based on the similarity of status indicators as a reference template.

10. An online calibration system for a low-voltage current transformer, applied to the online calibration method for a low-voltage current transformer as described in any one of claims 1 to 9, characterized in that, include: The data acquisition module is configured to acquire real-time primary current value, secondary voltage value, and real-time temperature data of the contact point of the primary circuit crimping device of the low-voltage current transformer under the high current test scenario of the primary circuit, and generate raw dataset. The condition assessment module is configured to receive the raw dataset, calculate the dynamic contact resistance value and temperature rise rate of the crimping device, and generate condition indicators. The error source identification module is configured to receive status indicators and basic error measurement values ​​of low-voltage current transformers, perform correlation analysis, and output error contribution coefficients. The adaptive calibration module is configured as the receiving error contribution coefficient, which adjusts the parameters of the error calibration algorithm accordingly and compensates for the basic error value of the real-time measurement to obtain the calibrated error value. The verification feedback module is configured to compare the calibrated error value with the standard current transformer model. When the comparison result exceeds the allowable range, a recalibration command is triggered until the comparison result is within the allowable range, at which point the final calibration error value is output. The data management module is configured to store the final calibration error value, status indicators, and calibration parameters, and to complete the data upload.

Images