High-voltage electrical appliance SF6 gas measurement verification method and system based on virtual etalon

By using a machine learning method based on a virtual standard, the problem of high workload and cost in periodic calibration of SF6 gas density relays has been solved, achieving high-precision and low-cost online calibration and ensuring the safe operation of high-voltage electrical appliances.

CN118937148APending Publication Date: 2024-11-12XIAN HIGH VOLTAGE APP RES INST CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411033828.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the existing technology, the periodic calibration of SF6 gas density relays is labor-intensive and costly, and cannot accurately assess their actual operating status, posing safety hazards.

Method used

A virtual standard-based approach is adopted, which establishes a virtual standard through machine learning, records the measurement values ​​of SF6 meters or sensors over a long period of time, calculates the error and analyzes the instrument degradation, and provides calibration guidance based on operating status.

Benefits of technology

It reduces the workload of calibration, improves measurement accuracy and reliability, lowers operation and maintenance costs, supports condition-based maintenance, and ensures the safe operation of high-voltage electrical appliances.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118937148A_ABST
    Figure CN118937148A_ABST
Patent Text Reader

Abstract

The invention discloses a high-voltage electrical appliance SF6 gas measurement verification method and system based on a virtual etalon, and belongs to the technical field of high-voltage electrical equipment detection, and the verification method comprises the steps: collecting a measurement value of an SF6 meter or a sensor; a first virtual standard device and a second virtual standard device are constructed, the first virtual standard device takes the SF6 meter or the sensor after being calibrated for a certain time as a base point, and a machine learning method is used for fitting measured values within set time to form a function of a first standard value; the second virtual etalon uses the characteristic relation between the pressure and the temperature of the SF6 gas of the high-voltage electric appliance under different density conditions as a function of a second standard value; calculating an error between the measured value and a first standard value obtained by using the first virtual standard device, and fitting an error between the measured value and a second standard value obtained by using the second virtual standard device; analyzing the two obtained errors and trends thereof, and generating error evaluation and degradation evaluation of the SF6 meter or sensor; according to the method, the verification workload is reduced, and a more reliable guarantee is provided for safe operation of the high-voltage electric appliance.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of high-voltage electrical appliance detection, and in particular relates to a method and system for measuring and verifying SF6 gas of a high-voltage electrical appliance based on a virtual standard device. Background Art

[0002] SF6 gas is widely used as an insulating medium in the field of high-voltage electrical appliances. The state of SF6 gas is related to the safe operation of high-voltage electrical appliances. At present, density relays (meters) and SF6 sensors are used to measure the temperature, pressure, density and other state quantities of the gas. Since density relays often use mechanical temperature compensation, the measurement is inaccurate during long-term operation. Therefore, the density relay needs to be calibrated regularly (in accordance with the provisions of DL / T596-2021 "Preventive Test Procedures for Power Equipment"), which is a very large workload and has a certain impact on the safe operation of the equipment.

[0003] SF6 gas is often used as an insulating medium and arc-extinguishing isolation medium in high-voltage electrical appliances. Its good insulation performance and arc-extinguishing ability make it a major option in the power industry. GB / T 12022 "Industrial Sulfur Hexafluoride" stipulates the characteristic relationship between pressure and temperature under different density conditions of sulfur hexafluoride gas in high-voltage electrical equipment. The Beattie-Bridgman formula (hereinafter referred to as the BB formula) can calculate the relationship between SF6 gas state parameters more accurately.

[0004] P=(56.9×ρ×T×(1+B)-ρ 2 ×A)×10 -6 (1)

[0005] Where: A = 74 × (1-0.727 × 10 -3 ×ρ)

[0006] B = 2.51 × 10 -3 ×ρ×(1-0.846×10 -3 ×ρ)

[0007] Where:

[0008] P—SF6 gas pressure (gauge pressure), MPa;

[0009] ρ—SF6 gas density, kg / m 3 ;

[0010] T—absolute temperature of SF6 gas, K.

[0011] The temperature compensation material of the SF6 gas density relay in field operation will be deformed due to aging after a certain period of use, resulting in poor temperature compensation performance and serious errors. When the ambient temperature changes suddenly, the alarm and lockout values ​​will have large errors, leading to major accidents. At present, the standards for SF6 gas detection or verification include DL / T 917-2005 "Determination of sulfur hexafluoride gas density", DL / T 1986-2019 "Technical Specifications for Gas Detection of Sulfur Hexafluoride Mixed Gas Insulation Equipment", DL / T 259-2012 "Sulfur Hexafluoride Gas Density Relay Verification Procedure", etc. GB / T8905-1996 "Guidelines for Gas Management and Detection in Sulfur Hexafluoride Electrical Equipment", etc. DL / T596-2021 "Preventive Test Procedures for Power Equipment" stipulates that each SF6 switch user should regularly calibrate the SF6 gas density relay. From the actual operation situation, it is also very necessary to regularly calibrate the SF6 density relay (meter) or sensor in field operation. However, the number of density relays is huge, and the workload of indiscriminate regular calibration is very large and uneconomical.

[0012] The calibration methods for density relays are divided into offline calibration and online calibration. Offline calibration requires the SF6 density relay to be disassembled from the high-voltage electrical appliance. If the high-voltage electrical appliance continues to operate without the density relay being calibrated, there is a risk that the safety performance of the equipment cannot be guaranteed because the SF6 gas status is not effectively detected. However, if the equipment is turned off to ensure the safety performance during the SF6 density relay calibration, it will cause a power outage, which is very costly. Therefore, the density relay is generally calibrated regularly through planned shutdown and maintenance, which has a large workload, and if the disassembly method is used, it will affect the equipment in operation, and there are technical problems that the safety performance of the high-voltage electrical equipment cannot be guaranteed or the power outage cost is high.

[0013] Online calibration does not require disassembly of the density relay, but requires specific calibration instruments, and works under power, which is not environmentally friendly and poses risks such as gas leakage and power station safety.

[0014] Therefore, the outstanding problem of the prior art is that the real operating status of the instrument is unclear, the specific error is unknown, the actual operating status of the density relay cannot be monitored for a long time, and it is difficult to evaluate the process and degree of degradation evolution. The current regular calibration method is not based on the actual operating conditions (accuracy) of the density relay, but is an indiscriminate calibration, which is labor-intensive and costly. Summary of the invention

[0015] The present invention provides a method and system for measuring and verifying SF6 gas for high-voltage electrical appliances based on a virtual standard. The method records the measured values ​​of SF6 meters or sensors online for a long time, adopts a machine learning method to establish a virtual standard according to the characteristics of SF6 gas, calculates and analyzes the error of the instrument, evaluates the degradation of the instrument, and guides the verification work based on the instrument status, thereby reducing the verification workload, avoiding invalid verification, and providing more reliable protection for the safe operation of high-voltage electrical appliances.

[0016] In order to achieve the above object, the present invention adopts the following technical solution:

[0017] A method for measuring and verifying SF6 gas for high-voltage electrical appliances based on a virtual standard includes:

[0018] Step 1: Collecting the measured values ​​of the SF6 meter or sensor, wherein the measured values ​​include: pressure, temperature and density of SF6 gas;

[0019] Step 2: construct a first virtual standard and / or a second virtual standard, wherein the first virtual standard takes a sensor after a certain calibration as a base point, and uses a machine learning method to fit the measured values ​​within a set time to form a function of the first standard value; the second virtual standard uses the characteristic relationship between pressure and temperature under different density conditions of SF6 gas of high-voltage electrical equipment as a function of the second standard value;

[0020] Step 3, calculating a first error and / or a second error, wherein the first error is an error between a measured value and a first standard value obtained by using a first virtual standard device; the second error is an error between a fitted measured value and a second standard value obtained by using a second virtual standard device; the fitted measured value is obtained by the measured value and a measured value fitting function;

[0021] Step 4: Analyze the first error and / or the second error and their trends to generate an error evaluation and degradation evaluation of the SF6 meter or sensor.

[0022] Furthermore, in step 2, the first virtual standard is obtained through the following training process:

[0023] 1) Take the measured value of the SF6 meter or sensor within a period of time d1 after a certain calibration As the reference value of the virtual standard; where P * is the reference measurement pressure vector, is the pressure of SF6 gas at time i, T * is the reference measurement temperature vector, is the temperature of SF6 gas at time i, ρ * is the reference measurement density vector, is the density of SF6 gas at time i;

[0024] 2) Use the benchmark value and XGBOOST to train a virtual standard S1: ρ = F1 (P, T), where ρ is the SF6 gas density, P is the SF6 gas pressure, and T is the SF6 gas temperature.

[0025] Furthermore, in step 3, calculating the error between the measured value and the first standard value obtained by using the first virtual standard device comprises the following steps:

[0026] 1) Collect the SF6 gas state data (P, T, ρ) during the actual operation of the SF6 meter or sensor, substitute it into the first virtual standard device, and obtain the first standard value ρ is the SF6 gas density, P is the SF6 gas pressure, and T is the SF6 gas temperature;

[0027] 2) Calculate the error e1 between the actual measured value and the standard value of gas density,

[0028] Furthermore, in step 2, the characteristic relationship between pressure and temperature under different density conditions of SF6 gas in high-voltage electrical appliances is used to obtain a second virtual standard S2: ρ=F2(P,T), where ρ is the SF6 gas density, P is the SF6 gas pressure, and T is the SF6 gas temperature.

[0029] Furthermore, in step 3, the process of obtaining the fitted measurement value is:

[0030] Collect the gas state data (P, T, ρ) of the actual operation process of the SF6 meter or sensor in a certain period of time d1, and use the gas state data (P, T, ρ) of the actual operation process of the SF6 meter or sensor in a certain period of time d2 and XGBOOST to train a model that reflects the measurement state of the SF6 meter or sensor in this period of time Where, ρ is the SF6 gas density, P is the SF6 gas pressure, T is the SF6 gas temperature, It is the fitted measurement value, that is, the measured density value of SF6 gas taking into account the actual degradation of the SF6 meter or sensor.

[0031] Further, in step 3, calculating the error between the measured value and the second standard value obtained by the second virtual standard device comprises the following steps:

[0032] 1) Take the set standard SF6 gas state value and bring it into the measurement value fitting function and the second virtual standard to obtain the density value of the SF6 meter or sensor under actual degradation. And the ideal density value

[0033] 2) Calculate the error e2 between the actual density value and the second standard value,

[0034] Furthermore, the measurement value fitting function model is trained by the following method:

[0035] 1) Loading and preprocessing data: Load the newly collected training data into the memory, perform necessary preprocessing, and divide the training data set and the test data set; the training data is the SF6 gas state data (P, T, ρ) during the actual operation of the SF6 meter or sensor within a certain period of time d2. The preprocessing includes data validity judgment, missing value, and abnormal value processing;

[0036] 2) Model training: Use the XGBOOST model and training data set to perform model training to obtain a model M that reflects the measurement status of the SF6 meter or sensor;

[0037] 3) Model evaluation: Use the test data set to evaluate the performance of the model and calculate indicators such as accuracy and F1 score;

[0038] 4) Model update: Determine whether the newly trained model meets the accuracy standard based on the accuracy rate and F1 score. If so, update the model M; if not, retrain.

[0039] Further, step 4 includes the following steps:

[0040] The first error and the second error are continuously updated during the actual use of the SF6 meter or sensor;

[0041] Observe the changing trend of the first error. If the first error continues to expand, it indicates that the temperature compensation capability of the SF6 meter or sensor is deteriorating. If the first error is greater than the first set value, it is determined that the SF6 meter or sensor needs to be calibrated.

[0042] Observe the changing trend of the second error. If the second error continues to expand, it is considered that the temperature compensation capability of the SF6 meter or sensor is continuously deteriorating. If the second error is greater than the second set value, it is determined that the SF6 meter or sensor needs to be calibrated.

[0043] A high-voltage electrical appliance SF6 gas measurement and calibration system based on a virtual standard device, comprising:

[0044] A collection device, used to collect the measured values ​​of the SF6 meter or sensor, the measured values ​​including: pressure, temperature and density of SF6 gas;

[0045] The monitoring and analysis device is used to receive data from the acquisition device, store, calculate and analyze the data, on the one hand to form a virtual standard, and on the other hand to compare the calculated value and the measured value of the virtual standard to obtain the measurement error of the instrument; the error is recorded for a long time to form an error sequence, and the trend analysis of the sequence is performed to obtain the error evolution.

[0046] Furthermore, the monitoring and analysis system includes:

[0047] A virtual standard construction module, used to construct a first virtual standard and a second virtual standard;

[0048] A first error calculation module, used for calculating an error between a measured value and a first standard value obtained by using a first virtual standard device, recorded as a first error;

[0049] A parameter fitting module is used for training a model reflecting the real measurement state of the SF6 meter or sensor based on the measured value when the second virtual standard is used, so as to obtain a model fitting the measured value;

[0050] A second error calculation module is used to calculate the error between the fitted measurement value and the second standard value obtained by the second virtual standard device, which is recorded as the second error;

[0051] The trend analysis module is used to analyze the first error and the second error and their trends, and generate an error evaluation and a degradation evaluation of the SF6 meter or sensor.

[0052] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0053] The present invention combines a virtual standard with online monitoring of SF6 gas to provide an online calibration method for SF6 measuring instruments and sensors, which has the following advantages:

[0054] 1) Strong traceability, can evaluate service status, and predict future trends: Using virtual standards, it is possible to record historical evaluation results for a long time and send them back immediately, predict the future operating trends of sensors, provide an evaluation of the entire life cycle of sensors after they are put into service, and ensure the consistency between measurement results and expected standards based on known standard data or models. This provides traceability of measurement results, allowing them to be compared and verified with existing standards.

[0055] 2) High precision: The virtual standard can eliminate the errors introduced by measuring equipment and environmental factors through parameter estimation and data processing during the calibration process, thereby improving the accuracy of the measurement results.

[0056] 3) Low cost, flexibility, and strong scalability: Compared with the use of real traceable standard equipment, virtual standards reduce the need for expensive real standard equipment, save calibration time, and improve efficiency. When more measurement parameters are involved or more complex calibration is required, the virtual standard can be adjusted and customized to adapt to different models and specifications of high-voltage electrical SF6 measurement equipment, expand to more measurement parameters or characteristics, and meet different measurement conditions and requirements.

[0057] 4) It can provide operation and inspection judgment based on the operating status, support "condition-based maintenance", reduce operation and maintenance costs, and save time. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic diagram of the system structure of the present invention;

[0059] Figure 2 is a flow chart of the method of the present invention;

[0060] Figure 3 This is a schematic diagram of the monitoring and analysis device module of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose and technical solution of the present invention clearer and easier to understand, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. The specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0062] In the description of the present invention, it should be understood that the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, the meaning of "multiple" is two or more. In the description of the present invention, it should be noted that, unless otherwise clearly specified and defined, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or it can be the internal connection of two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0063] Explanation of relevant terms

[0064] 1) Virtual Standardizer: A model-based or algorithm-based method for converting actual measurement data into data that conforms to a predefined standard or theoretical model. Virtual Standardizers improve the accuracy and comparability of measurement data by correcting, amending or converting actual measurement data to make it consistent with the expected standard or model.

[0065] The main purpose of a virtual standard is to standardize measurement data by simulation or estimation in the absence of a true traceable standard or theoretical model. It uses existing standard data or models as a reference and adjusts them according to the characteristics of the actual measurement data, so that the measurement results are closer to the expected standard.

[0066] Virtual standards are widely used in analytical chemistry, biomedicine, environmental science, industrial manufacturing, etc. They can be used for calibration of experimental data, instrument calibration, data normalization, etc., which helps to improve the accuracy, reliability and traceability of measurements.

[0067] Virtual standards are an approximate processing method based on models or algorithms, and there are still some differences between them and real standards or models. Therefore, in specific applications, it is necessary to reasonably select and verify the applicability of virtual standards based on specific needs and actual conditions.

[0068] 2) Measurement Verification refers to the process of verifying and confirming measurement results to ensure the accuracy and reliability of the measurement. It is very important in the fields of science, engineering and technology to confirm whether the measurement data meets the expected standards, specifications or requirements.

[0069] Example 1

[0070] Reference Figure 2 , a high-voltage electrical appliance sulfur hexafluoride measurement and calibration method based on a virtual standard device comprises the following steps:

[0071] Step 1, collecting the measured values ​​of the SF6 meter (hereinafter referred to as SF6 meter) or sensor, including: SF6 gas pressure P, SF6 gas temperature T and SF6 gas density ρ and other state parameters;

[0072] Step 2, construct a virtual standard, including the following two standards: the first virtual standard S1: taking the sensor after a certain calibration as the base point, using the machine learning method to fit the measured value within a certain period of time after the base point to form a function of the first standard value; the second virtual standard S2: using the SF6 gas empirical formula BB empirical formula as the function of the second standard value;

[0073] Step 3: Calculate the error between the measured value and the standard value generated by the virtual standard device, including:

[0074] 1) Bringing the measured value within a certain time window d1 into the first virtual standard device to generate the first standard value of the time window, and calculating the first error based on the measured value and the standard value, where the first error = measured value - first standard value;

[0075] 2) Within a certain time window d2, the measured value data set is subjected to machine learning fitting to form a measured value fitting function. According to the provisions of the GB / T 12022 Industrial Sulfur Hexafluoride standard, the set standard SF6 gas state value, that is, several groups of pressure, temperature, and density parameters that meet the above standard are respectively substituted into the measured value fitting function and the second virtual standard device to generate the fitted measured value and the second standard value respectively, and the second error is calculated based on the fitted measured value and the second standard value, and the second error = fitted measured value - second standard value.

[0076] The first error can better reflect the error caused by the degradation of the instrument itself, while the second error can better reflect the actual error of the instrument.

[0077] Step 4: Record the calculated first error and second error, conduct a combined trend analysis, and generate an error evaluation and degradation evaluation for the SF6 meter or sensor.

[0078] It should be noted that since the BB empirical formula is a multivariate cubic polynomial, generally speaking, multivariate polynomial regression can effectively fit the relationship between variables. However, the actual working conditions are more complicated, and the gas state may be coupled with other disturbances such as trace leakage and component changes, which further increases the nonlinearity of the variable relationship. Therefore, a more complex model is needed here to fit the relationship between temperature, pressure and density. Since the tree model is simple and easy to train, intuitive and easy to understand, and can be effectively integrated and learned, the advanced model XGBOOST of the tree model is selected as the fitting model here. It should be noted that there may be more than one model selection, and models such as neural networks and high-order nonlinear regression can also achieve the fitting goal.

[0079] In step 2, when the first virtual standard is used for calculation, the training and use of the model are as follows:

[0080] 1) Take the measured value of the meter within a period of time d1 after a certain calibration As the reference value of the virtual standard; where P * is the reference measurement pressure vector, is the pressure of SF6 gas at time i, T * is the reference measurement temperature vector, is the temperature of SF6 gas at time i, ρ * is the reference measurement density vector, is the density of SF6 gas at time i;

[0081] 2) Using the reference value and XGBOOST to train the virtual standard S1: ρ = F1(P, T), it is considered that the first virtual standard reflects the basic state of the meter; where ρ is the SF6 gas density, P is the SF6 gas pressure, and T is the SF6 gas temperature; F1(·) is the fitting function of the virtual standard S1;

[0082] 3) Collect the SF6 gas state data (P, T, ρ) during the actual operation of the SF6 meter or sensor, substitute the SF6 gas pressure P and SF6 gas temperature T into the virtual standard S1, and obtain the first standard value Among them, the first standard value is the ideal value of SF6 gas density when the meter is in the basic state;

[0083] 4) Calculate the error e1 between the actual measured value of SF6 gas density and the first standard value,

[0084] 5) During the actual use of the SF6 meter or sensor, repeat steps 3) and 4) continuously to observe the changing trend of the error e1. If the error e1 continues to expand, it is considered that the temperature compensation capability of the SF6 meter or sensor is continuously deteriorating. If the error e1 is greater than a certain set value, it is determined that the SF6 meter or sensor needs to be calibrated.

[0085] When the second virtual standard is used for calculation, the model is trained and used as follows:

[0086] 1) Using the BB formula to obtain the virtual standard S2: ρ = F2(P, T); F2(·) is the fitting function of the virtual standard S2 measurement value;

[0087] 2) Collect the SF6 gas state data (P, T, ρ) during the actual operation of the SF6 meter or sensor within a period of time d2, and use the SF6 gas state data (P, T, ρ) during the actual operation of the meter within this period of time and XGBOOST to train a model that reflects the meter measurement status within this period of time in, is the fitted measurement value, i.e., the measured density value of SF6 gas considering the actual degradation of the meter; G(·) is the fitting function of model M;

[0088] 3) Take the set standard SF6 gas state value and bring it into the model M and virtual standard S2 to obtain the density value of the SF6 meter or sensor under actual deterioration. and the second standard value The second standard value is the density value under ideal conditions;

[0089] 4) Calculate the error between the SF6 gas measured density value and the second standard value considering the actual degradation of the SF6 meter or sensor

[0090] 5) During the actual use of the SF6 meter or sensor, steps 2) to 4) are continuously repeated, the sliding window continuously trains the new model M, calculates the error e2, and observes the changing trend of the error e2. If the error e2 continues to expand, it is considered that the temperature compensation ability of the SF6 meter or sensor is continuously deteriorating. If the error e2 is greater than a certain set value, it is determined that the meter needs to be calibrated.

[0091] When the second virtual standard is used for calculation, the model M is obtained by continuously training the sliding window on the time series, and the idea of ​​online training model is adopted, as follows:

[0092] 1) Load and preprocess data: Load the newly collected training data into the memory, perform necessary preprocessing, and divide the training data set into the test data set. The training data is the SF6 gas state data (P, T, ρ) during the actual operation of the SF6 meter or sensor within a certain period of time d2. The preprocessing includes data validity judgment, missing value, and abnormal value processing.

[0093] 2) Model training: Use the XGBOOST model and training dataset for model training.

[0094] 3) Model evaluation: Use the test dataset to evaluate the performance of the model and calculate indicators such as accuracy and F1 score.

[0095] 4) Model update: Determine whether the newly trained model meets the accuracy standard based on the accuracy rate and F1 score. If so, update the model M; if not, retrain.

[0096] The present invention selects an applicable virtual standard model to simulate the characteristics and behavior of SF6 measurement. The model should take into account the physical properties, measurement parameters and instrument characteristics of SF6 gas to ensure that the virtual standard can accurately simulate the actual measurement process.

[0097] By estimating and adjusting the parameters of the virtual standard model, it is ensured that it can produce accurate standardized results based on the actual measurement data. Parameter calibration can be performed based on known standard data or calibration equipment to make the virtual standard model closer to the actual measurement results.

[0098] Compare and analyze the actual calibrated measurement data with the known standard data to evaluate the accuracy and reliability of the calibration results. Verify the consistency between the measurement data and the expected standard by calculating indicators such as deviation and correlation.

[0099] Ensure that the calibration process is traceable, that is, the calibration results can be traced back to credible standard data or models. At the same time, make necessary records, including calibration parameters, calibration equipment information, calibration results, etc., for subsequent review and traceability.

[0100] Example 2

[0101] The present embodiment provides a method for measuring and verifying SF6 gas of high-voltage electrical appliances based on a virtual standard. The difference between this method and Embodiment 1 is that the present embodiment does not construct a second virtual standard S2, does not form a measurement value fitting function, and accordingly, does not calculate the second error. Only based on the first error and its trend analysis, an error evaluation and degradation evaluation of the instrument are generated.

[0102] Example 3

[0103] The present embodiment provides a method for measuring and verifying SF6 gas of high-voltage electrical appliances based on a virtual standard. The difference between this method and Embodiment 1 is that the present embodiment does not construct a first virtual standard S1, and accordingly, does not calculate the first error. It only generates an error evaluation and degradation evaluation of the instrument based on the second error and its trend analysis.

[0104] Example 4

[0105] Reference Figure 1 and Figure 3 The general system of the present invention includes the following parts: SF6 density relay (meter) or sensor, collection device and monitoring and analysis device.

[0106] SF6 density relay (meter) or sensor refers to the density relay, sensor and other instruments used to measure the SF6 gas state of high-voltage electrical appliances. It is installed in the SF6 gas chamber of high-voltage electrical appliances to measure the SF6 gas state in real time, including temperature, pressure, density and other state values. The instrument has a measurement value transmission interface and supports RS485, Ethernet or wireless communication methods.

[0107] The acquisition device collects the measurement data of multiple instruments in a periodic manner and preprocesses the data. The preprocessing includes data validity judgment, abnormal value filtering, etc. The collected data is packaged according to the instrument number and timestamped. The sorted data can be stored in the acquisition device. The acquisition device has an Ethernet or fiber optic communication interface and supports Modbus, IEC61850 or custom protocols. It supports active upload or query upload.

[0108] The monitoring and analysis device receives the data packets from the acquisition device, and stores, calculates and analyzes the data packets. On the one hand, a virtual standard is formed, and on the other hand, the measurement error of the instrument is obtained by comparing the calculated value and the measured value of the virtual standard. The error is recorded for a long time to form an error data set, and the trend analysis of the data set is performed to obtain the error evolution.

[0109] Reference Figure 3 , the monitoring and analysis device includes:

[0110] A virtual standard construction module, used to construct a first virtual standard and a second virtual standard;

[0111] A first error calculation module, used for calculating an error between a measured value and a first standard value obtained by using a first virtual standard device, recorded as a first error;

[0112] A parameter fitting module is used for training a model reflecting the real measurement state of the SF6 meter or sensor based on the measured value when the second virtual standard is used, so as to obtain a model fitting the measured value;

[0113] A second error calculation module is used to calculate the error between the fitted measurement value and the second standard value obtained by the second virtual standard device, which is recorded as the second error;

[0114] The trend analysis module is used to analyze the first error and the second error and their trends, and generate an error evaluation and a degradation evaluation of the SF6 meter or sensor.

[0115] The above contents are only for explaining the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A method for measuring and verifying SF6 gas of high-voltage electrical appliances based on a virtual standard, characterized in that: include: Step 1: Collecting the measured values ​​of the SF6 meter or sensor, wherein the measured values ​​include: pressure, temperature and density of SF6 gas; Step 2: construct a first virtual standard and / or a second virtual standard, wherein the first virtual standard takes a sensor after a certain calibration as a base point, and uses a machine learning method to fit the measured values ​​within a set time to form a function of the first standard value; the second virtual standard uses the characteristic relationship between pressure and temperature under different density conditions of SF6 gas of high-voltage electrical equipment as a function of the second standard value; Step 3, calculating a first error and / or a second error, wherein the first error is an error between a measured value and a first standard value obtained by using a first virtual standard device; the second error is an error between a fitted measured value and a second standard value obtained by using a second virtual standard device; the fitted measured value is obtained by the measured value and a measured value fitting function; Step 4: Analyze the first error and / or the second error and their trends to generate an error evaluation and degradation evaluation of the SF6 meter or sensor.

2. The method for measuring and verifying SF6 gas of high-voltage electrical appliances based on a virtual standard according to claim 1, characterized in that: In step 2, the first virtual standard is obtained through the following training process: 1) Take the measured value of the SF6 meter or sensor within a period of time d1 after a certain calibration As the reference value of the virtual standard; where P * is the reference measurement pressure vector, is the pressure of SF6 gas at time i, T * is the reference measurement temperature vector, is the temperature of SF6 gas at time i, ρ * is the reference measurement density vector, is the density of SF6 gas at time i; 2) Use the benchmark value and XGBOOST to train a virtual standard S1: ρ = F1 (P, T), where ρ is the SF6 gas density, P is the SF6 gas pressure, and T is the SF6 gas temperature.

3. The method for measuring and verifying SF6 gas of high-voltage electrical appliances based on a virtual standard according to claim 1, characterized in that: In step 3, calculating the error between the measured value and the first standard value obtained by using the first virtual standard device comprises the following steps: 1) Collect the SF6 gas state data (P, T, ρ) during the actual operation of the SF6 meter or sensor, substitute it into the first virtual standard device, and obtain the first standard value ρ is the SF6 gas density, P is the SF6 gas pressure, and T is the SF6 gas temperature; 2) Calculate the error e1 between the actual measured value and the standard value of gas density, 4. The method for measuring and verifying SF6 gas of high-voltage electrical appliances based on a virtual standard according to claim 1, characterized in that: In step 2, the characteristic relationship between pressure and temperature under different density conditions of SF6 gas of high-voltage electrical appliances is used to obtain the second virtual standard S2: ρ=F2(P,T), where ρ is the SF6 gas density, P is the SF6 gas pressure, and T is the SF6 gas temperature.

5. The method for measuring and verifying SF6 gas of high-voltage electrical appliances based on a virtual standard according to claim 1, characterized in that: In step 3, the process of obtaining the fitting measurement value is: Collect the gas state data (P, T, ρ) of the actual operation process of the SF6 meter or sensor in a certain period of time d1, and use the gas state data (P, T, ρ) of the actual operation process of the SF6 meter or sensor in a certain period of time d2 and XGBOOST to train a model that reflects the measurement state of the SF6 meter or sensor in this period of time Where, ρ is the SF6 gas density, P is the SF6 gas pressure, T is the SF6 gas temperature, It is the fitted measurement value, that is, the measured density value of SF6 gas taking into account the actual degradation of the SF6 meter or sensor.

6. The high-voltage electrical appliance SF6 gas calibration method based on a virtual standard according to claim 5 is characterized in that: In step 3, calculating the error between the measured value and the second standard value obtained by the second virtual standard device comprises the following steps: 1) Take the set standard SF6 gas state value and bring it into the measurement value fitting function and the second virtual standard to obtain the density value of the SF6 meter or sensor under actual degradation. And the ideal density value 2) Calculate the error e2 between the actual density value and the second standard value, 7. The method for measuring and verifying SF6 gas of high-voltage electrical appliances based on a virtual standard according to claim 6, characterized in that: The measured value fitting function model is trained by the following method: 1) Loading and preprocessing data: Load the newly collected training data into the memory, perform necessary preprocessing, and divide the training data set and the test data set; the training data is the SF6 gas state data (P, T, ρ) during the actual operation of the SF6 meter or sensor within a certain period of time d2. The preprocessing includes data validity judgment, missing value, and abnormal value processing; 2) Model training: Use the XGBOOST model and training data set to perform model training to obtain a model M that reflects the measurement status of the SF6 meter or sensor; 3) Model evaluation: Use the test data set to evaluate the performance of the model and calculate indicators such as accuracy and F1 score; 4) Model update: Determine whether the newly trained model meets the accuracy standard based on the accuracy rate and F1 score. If so, update the model M; if not, retrain.

8. The method for measuring and verifying SF6 gas of high-voltage electrical appliances based on a virtual standard according to claim 1, characterized in that: The step 4 comprises the following steps: The first error and the second error are continuously updated during the actual use of the SF6 meter or sensor; Observe the changing trend of the first error. If the first error continues to expand, it indicates that the temperature compensation capability of the SF6 meter or sensor is deteriorating. If the first error is greater than the first set value, it is determined that the SF6 meter or sensor needs to be calibrated. Observe the changing trend of the second error. If the second error continues to expand, it is considered that the temperature compensation capability of the SF6 meter or sensor is continuously deteriorating. If the second error is greater than the second set value, it is determined that the SF6 meter or sensor needs to be calibrated.

9. A high-voltage electrical appliance SF6 gas measurement and calibration system based on a virtual standard, characterized in that: include: A collection device, used to collect the measured values ​​of the SF6 meter or sensor, the measured values ​​including: pressure, temperature and density of SF6 gas; The monitoring and analysis device is used to receive data from the acquisition device, store, calculate and analyze the data, on the one hand to form a virtual standard, and on the other hand to compare the calculated value and the measured value of the virtual standard to obtain the measurement error of the instrument; the error is recorded for a long time to form an error sequence, and the trend analysis of the sequence is performed to obtain the error evolution.

10. A high-voltage electrical appliance SF6 gas measurement and calibration system based on a virtual standard according to claim 9, characterized in that: The monitoring and analysis device comprises: A virtual standard construction module, used to construct a first virtual standard and a second virtual standard; A first error calculation module, used for calculating an error between a measured value and a first standard value obtained by using a first virtual standard device, recorded as a first error; A parameter fitting module is used for training a model reflecting the real measurement state of the SF6 meter or sensor based on the measured value when the second virtual standard is used, so as to obtain a model fitting the measured value; A second error calculation module is used to calculate the error between the fitted measurement value and the second standard value obtained by the second virtual standard device, which is recorded as the second error; The trend analysis module is used to analyze the first error and the second error and their trends, and generate an error evaluation and a degradation evaluation of the SF6 meter or sensor.