Transformer oil chromatogram acquisition and discrimination method

Abstract

The invention discloses a transformer oil chromatogram acquisition and discrimination method, which comprises the following steps: firstly, installing an oil extraction valve and an oil-gas separation device at an oil extraction port of a transformer, extracting gas components in oil, introducing the gas components into a gas sensor, and acquiring concentration data of acetylene, hydrogen and carbon monoxide; then, installing a gas flow and temperature sensor, monitoring the gas flow and temperature in real time, and calculating the gas flow and temperature ratio; a comprehensive evaluation index is formed by associating gas component data with a gas flow temperature ratio, and a fault risk is judged according to a multi-stage threshold value, so that graded alarm is realized. According to the transformer oil chromatography obtaining and judging method, the accuracy and timeliness of transformer fault detection can be effectively improved, and reliable guarantee is provided for stable operation of the transformer.

Description

Technical Field

[0001] The invention belongs to the technical field of power equipment monitoring and fault diagnosis, and in particular relates to a transformer oil chromatogram acquisition and identification method. Background Art

[0002] Existing transformer oil chromatography monitoring systems mainly rely on oil circulation equipment, sampling equipment and gas analysis systems. These systems extract and analyze the components of dissolved gases, such as hydrogen, acetylene and carbon monoxide, from transformer oil to monitor and warn of potential faults inside the transformer. However, due to the following technical and economic limitations, traditional oil chromatography monitoring is difficult to apply to a wide range of transformer types, especially in low-cost, widely distributed transformer equipment.

[0003] The existing oil chromatography monitoring system includes oil circulation loop, circulation equipment, oil extraction and sampling devices. These components are large in size and require a large space around the transformer for assembly. In addition, the installation process is complicated and requires professional technicians to modify and seal the oil circulation system to ensure the normal operation of the gas sampling system. This complex assembly requirement has restrictions on the use scenario and is not suitable for transformers with limited installation space or inconvenient geographical locations.

[0004] The current oil chromatography monitoring system is very expensive. Even if it does not include the basic equipment of the data processing system, the cost of a single system is usually tens of thousands of RMB, and the installation cost is also several thousand RMB. This is a relatively large expense for medium- and low-cost transformers. Many transformers, especially small distribution transformers that cost only a few thousand RMB, are not cost-effective if equipped with expensive oil chromatography monitoring systems, and are difficult to popularize on a large scale.

[0005] During the installation and maintenance of traditional oil chromatography monitoring systems, the transformer usually needs to be shut down to ensure safety, which will cause production or power supply losses during the power outage. For industrial scenarios or residential power supply scenarios with high requirements for continuous operation, the economic losses and user inconvenience caused by power outages are more significant. Therefore, this technology is not easy to promote in many key scenarios and civil occasions.

[0006] Due to the high cost and high installation requirements of traditional systems, oil chromatography monitoring suitable for large and critical equipment is difficult to extend to small and widely distributed transformers. For small transformers costing only a few thousand yuan, such as those widely used in rural power grids and civil power distribution networks, installing traditional oil chromatography monitoring systems is costly and uneconomical. Therefore, the application of existing technologies in low-cost transformers is greatly limited and cannot meet the actual monitoring needs of medium- and low-priced, widely distributed transformers.

[0007] In summary, the limitations of existing oil chromatography monitoring systems in terms of volume, cost and installation complexity have hindered the popularization of early warning technology for transformer faults. With the widespread application of small transformers in power grid systems, there is an urgent need for a lightweight, low-cost and easy-to-install monitoring solution to make the status monitoring of small and medium-sized transformers more economical and effective, and to achieve all-round protection of distributed transformer networks. Summary of the invention

[0008] The technical problem to be solved by the present invention is to provide a transformer oil chromatogram acquisition and identification method which can effectively improve the accuracy and timeliness of transformer fault detection and provide reliable guarantee for the stable operation of the transformer.

[0009] The technical solution of the present invention is:

[0010] A method for obtaining and distinguishing transformer oil chromatogram comprises the following steps:

[0011] Step 1, oil extraction and oil-gas separation, including the following steps:

[0012] S1. Install the oil valve and oil-gas separation device at the oil inlet of the transformer;

[0013] S2, using an oil-gas separation device to extract the gas component in the oil and introduce it into the gas sensor;

[0014] S3. Perform gas component testing on the gas sensor to obtain concentration data of acetylene, hydrogen and carbon monoxide in the gas;

[0015] Step 2: Install the gas flow sensor and temperature sensor: Install the gas flow sensor and temperature sensor near the gas sensor to monitor the gas flow L and temperature T in real time;

[0016] Step 3: Calculate the air flow-temperature ratio. The calculation formula of the air flow-temperature ratio is:

[0017]

[0018] Among them, k1 represents the ratio of gas flow rate to temperature, L represents gas flow rate, and T represents gas temperature. The larger the k1 value is, the higher the potential danger level of the transformer is.

[0019] Step 4: Joint judgment, including the following steps:

[0020] S4, performing correlation analysis on the gas composition data measured by the gas sensor and the k1 value to form a comprehensive evaluation index;

[0021] S5. Select an insulated exhaust pipe with waterproof function as the exhaust port, and ensure that the height of the exhaust terminal of the exhaust port is greater than or equal to the height of the transformer oil tank;

[0022] Step 5: Alarm threshold setting, including the following steps:

[0023] S6. Monitor the operating environment of multiple transformers, and when the temperature T changes by more than 20 degrees Celsius, calculate the average value k0 of the air flow temperature ratio k1 as a benchmark;

[0024] S7, set alarm conditions: when When ΔT is less than 10 degrees Celsius, it is considered that the airflow temperature exceeds the reference value by 5%, triggering an alarm.

[0025] Furthermore, the insulated exhaust duct in step S5 is oriented vertically upward.

[0026] Furthermore, the concentration data and temperature data of acetylene and hydrogen in the gas are read using LoRa, Wi-Fi, Zigbee, RS-485 bus or Ethernet cable.

[0027] Furthermore, the step S4 of correlating and analyzing the gas composition data measured by the gas sensor with the k1 value to form a comprehensive evaluation index includes the following steps:

[0028] (1) Obtain gas composition data: Use gas sensors to detect the gas components separated from transformer oil in real time, and record the concentrations of these gas components as C 乙炔 , C 氢气 and C 一氧化碳 ;

[0029] (2) Standardized gas composition data: The concentration of each gas is standardized to obtain a dimensionless gas concentration value. The expression of the standardized gas composition data is:

[0030]

[0031] Among them, C′ 气体 Indicates the standardized value of gas concentration, C 气体 Indicates the currently measured gas concentration, μ 气体 represents the average concentration of the gas under historical normal conditions, σ 气体 It represents the standard deviation of the gas under the historical normal state. 气体 When it is close to 0, it means that the current gas concentration is close to the normal level. 气体 When it is much greater than 0 or much less than 0, it means that the current gas concentration deviates significantly from the normal level;

[0032] (3) Correlation analysis and calculation of comprehensive evaluation index: weighted combination of the standardized gas concentration value and the airflow temperature ratio k1 is performed to calculate the comprehensive evaluation index R. The expression of the comprehensive evaluation index R is:

[0033] R=w1×C′ 乙炔 +w2×C′ 氢气 +w3×C′ 一氧化碳 +...+w n ×k1,

[0034] Among them, w1 represents the weight coefficient of acetylene gas, w2 represents the weight coefficient of hydrogen, w3 represents the weight coefficient of carbon monoxide, and w n represents the influence weight of air flow temperature ratio k1 on transformer failure risk, C′ 乙炔 Represents the standardized value of acetylene concentration, C′ 氢气 Represents the standardized value of hydrogen concentration, C′ 一氧化碳 Represents the normalized value of carbon monoxide concentration;

[0035] (4) Determine the alarm threshold: Obtain the safety threshold R of the comprehensive evaluation index R through historical data or simulation experiments 阈值 , if the real-time monitored R value exceeds the safety threshold R 阈值 , that is, it is determined that the transformer has an abnormal risk and triggers an alarm.

[0036] Furthermore, step S4 also includes dynamic adjustment and multi-level alarm steps, and the dynamic adjustment and multi-level alarm are specifically: setting a multi-level alarm threshold, and dividing the degree of exceeding the limit of the R value into different levels of "early warning", "warning" and "serious alarm"; if the R value is close to the threshold but does not exceed it, the system prompts to take preventive maintenance, if it far exceeds the threshold, the emergency shutdown or maintenance process is initiated.

[0037] Furthermore, the calculation of the average value k0 of the airflow-temperature ratio k1 in step S6 includes the following steps:

[0038] (1) Collect gas flow and temperature data: monitor and record gas flow L and temperature T for a long time under different operating conditions of multiple transformers to ensure that the data covers the typical operating temperature range of the transformer and various working load conditions;

[0039] (2) Calculate the air flow-temperature ratio k1 of each sampling, repeat the sampling and obtain a large number of k1 values. The calculation formula of the air flow-temperature ratio k1 is:

[0040]

[0041] Wherein, k1 represents the ratio of gas flow rate to temperature at each sampling time point, L represents gas flow rate, and T represents gas temperature;

[0042] (3) Data cleaning and screening: Clean the collected k1 value data and delete abnormal values, such as abnormal flow or extreme temperature values ​​during equipment maintenance;

[0043] (4) Calculate the mean value by temperature zone: Group the data according to the range of temperature T, and calculate the average value k0 of k1 in each temperature zone to ensure that the k0 used as the benchmark reflects the characteristics of different temperature zones. Specifically, the temperature is divided into multiple temperature zones with a range of 10 degrees Celsius;

[0044] (5) Calculate the overall average value k0 of k1. The expression for calculating the overall average value k0 of k1 is:

[0045] Where N is the total number of samples, k 1,i Represents the airflow-temperature ratio for each sample.

[0046] Furthermore, the method for deleting outliers in step (3) is a 3-sigma screening method, and the 3-sigma screening method includes the following steps:

[0047] (1) Calculate the mean and standard deviation. The expressions of the mean and standard deviation are:

[0048]

[0049]

[0050] Where N represents the total number of data points, x i represents the value of the i-th data point, μ represents the mean of the data, and σ represents the standard deviation of the data;

[0051] (2) Determine the 3-sigma range: define the normal data range, that is, [μ-3σ, μ+3σ]. Data points within this range are considered normal data, while data points outside this range are considered outliers.

[0052] (3) Screening outliers: traverse each data point to determine whether it is within the range of [μ-3σ,μ+3σ]. For data points outside this range, mark them as outliers and remove them;

[0053] (4) Update the data set: After removing the outliers, the retained data set is used for subsequent analysis steps. If the data set is small and the outliers significantly affect the mean and standard deviation, repeat the above steps until there are no significant outliers.

[0054] Furthermore, the step S6 also includes a step of setting dynamic adjustment rules, and the dynamic adjustment rules are specifically: if the ambient temperature fluctuates significantly during the monitoring period, a weighted average method is used to make k0 reflect the latest airflow temperature ratio characteristics, thereby realizing dynamic updating of k0.

[0055] Furthermore, the weighted average method comprises the following steps:

[0056] (1) Set the weighting coefficient w i : The weight coefficient is set to an arithmetic decreasing weight or an exponential decreasing weight. The expression of the arithmetic decreasing weight is:

[0057]

[0058] Among them, w i Represents the weight coefficient, satisfying The closer the data point is to the current one, the higher the weight is. n represents the window size, and i represents the data point number.

[0059] The expression of the exponentially decreasing weight is:

[0060]

[0061] Among them, w i Represents the weight coefficient, satisfying The closer the data point is to the current one, the higher the weight is. λ represents the attenuation coefficient, n represents the window size, i represents the data point number, and j represents each data point in the sliding window.

[0062] (2) Calculate the weighted average value, the calculation formula of the weighted average value is:

[0063]

[0064] Among them, k0 represents the average value of the air flow temperature ratio after update, w i Represents the weight coefficient, k 1,i Represents the i-th airflow temperature ratio within the sliding window.

[0065] Beneficial effects of the present invention:

[0066] 1. Reduce costs and simplify installation:

[0067] The present invention realizes the oil chromatography detection function by installing an oil extraction valve and an oil-gas separation device at the oil extraction port of the transformer and directly detecting the gas composition in combination with a gas sensor, thereby eliminating the need for traditional oil circulation equipment and complex sampling devices. This design greatly simplifies the installation process, reduces the installation and maintenance costs, and enables the monitoring system to be more economically and efficiently applied to small and medium-sized transformers, especially low-cost equipment;

[0068] 2. Improved real-time performance and accuracy:

[0069] The present invention can quickly identify the potential danger level of the transformer by real-time monitoring of gas flow and temperature and calculating the gas flow-temperature ratio. Real-time data collection and analysis helps to detect abnormal conditions inside the transformer early, avoids the delay risk of traditional systems, and improves the timeliness and accuracy of fault detection.

[0070] 3. Comprehensive judgment of multiple parameters to enhance fault identification capability:

[0071] The present invention correlates and analyzes the concentration data of the gas sensor with the airflow temperature ratio to form a comprehensive evaluation index. By standardizing and weighting the concentrations of acetylene, hydrogen and carbon monoxide, the reliability and accuracy of the judgment are improved. Compared with the detection method of a single indicator, the comprehensive multi-parameter monitoring method can more comprehensively reflect the actual operating status of the transformer.

[0072] 4. Dynamic adjustment and multi-level alarm to improve monitoring sensitivity:

[0073] The present invention adopts a dynamic adjustment rule in setting the alarm threshold, updates the reference value by a weighted average method, and can reflect the latest airflow temperature ratio characteristics in real time. At the same time, setting multi-level alarm thresholds (such as early warning, warning, and severe alarm) can trigger graded alarms according to the degree of over-limit of the comprehensive evaluation index, effectively improving the sensitivity and adaptability of the monitoring system and avoiding the occurrence of false alarms.

[0074] 5. 3-sigma screening method improves data quality:

[0075] The present invention incorporates a 3-sigma screening method for data cleaning, which ensures the validity and accuracy of the data by eliminating outliers. The data cleaning step filters out abnormal data caused by equipment maintenance, extreme environments, etc., improves the quality and accuracy of subsequent analysis, and thus improves the reliability of fault judgment.

[0076] 6. Signal transmission methods that can flexibly adapt to different scenarios:

[0077] The present invention selects LoRa, Wi-Fi, Zigbee, RS-485 bus or Ethernet cable signal transmission mode according to the application scenario, making data acquisition and transmission more flexible. The wireless or wired data reading mode can adapt to different installation scenarios, has strong adaptability, is convenient for remote data monitoring and real-time transmission, and is conducive to promotion and application in complex scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 The present invention is a schematic flow chart of a transformer oil chromatogram acquisition and identification method. DETAILED DESCRIPTION

[0079] like Figure 1As shown, a method for obtaining and distinguishing transformer oil chromatogram comprises the following steps:

[0080] Step 1, oil extraction and oil-gas separation, including the following steps:

[0081] S1. Install the oil valve and oil-gas separation device at the oil inlet of the transformer;

[0082] S2, using an oil-gas separation device to extract the gas component in the oil and introduce it into the gas sensor;

[0083] S3. Perform gas component testing on the gas sensor to obtain concentration data of acetylene, hydrogen and carbon monoxide in the gas;

[0084] Step 2: Install the gas flow sensor and temperature sensor: Install the gas flow sensor and temperature sensor near the gas sensor to monitor the gas flow L and temperature T in real time;

[0085] Step 3: Calculate the air flow-temperature ratio. The calculation formula of the air flow-temperature ratio is:

[0086]

[0087] Among them, k1 represents the ratio of gas flow rate to temperature, L represents gas flow rate, and T represents gas temperature. The larger the k1 value is, the higher the potential danger level of the transformer is.

[0088] Step 4: Joint judgment, including the following steps:

[0089] S4, performing correlation analysis on the gas composition data measured by the gas sensor and the k1 value to form a comprehensive evaluation index;

[0090] S5. Select an insulated exhaust pipe with waterproof function as the exhaust port, and ensure that the height of the exhaust terminal of the exhaust port is greater than or equal to the height of the transformer oil tank;

[0091] Step 5: Alarm threshold setting, including the following steps:

[0092] S6. Monitor the operating environment of multiple transformers, and when the temperature T changes by more than 20 degrees Celsius, calculate the average value k0 of the air flow temperature ratio k1 as a benchmark;

[0093] S7, set alarm conditions: when When ΔT is less than 10 degrees Celsius, it is considered that the airflow temperature exceeds the reference value by 5%, triggering an alarm.

[0094] Preferably, the insulated exhaust duct in step S5 is oriented vertically upward.

[0095] Preferably, the concentration data and temperature data of acetylene and hydrogen in the gas are read using LoRa, Wi-Fi, Zigbee, RS-485 bus or Ethernet cable.

[0096] Preferably, the step S4 of correlating and analyzing the gas composition data measured by the gas sensor with the k1 value to form a comprehensive evaluation index comprises the following steps:

[0097] (1) Obtain gas composition data: Use gas sensors to detect the gas components separated from transformer oil in real time, and record the concentrations of these gas components as C 乙炔 , C 氢气 and C 一氧化碳 ;

[0098] (2) Standardized gas composition data: The concentration of each gas is standardized to obtain a dimensionless gas concentration value. The expression of the standardized gas composition data is:

[0099]

[0100] Among them, C′ 气体 Indicates the standardized value of gas concentration, C 气体 Indicates the currently measured gas concentration, μ 气体 represents the average concentration of the gas under historical normal conditions, σ 气体 It represents the standard deviation of the gas under the historical normal state. 气体 When it is close to 0, it means that the current gas concentration is close to the normal level. 气体 When it is much greater than 0 or much less than 0, it means that the current gas concentration deviates significantly from the normal level;

[0101] (3) Correlation analysis and calculation of comprehensive evaluation index: weighted combination of the standardized gas concentration value and the airflow temperature ratio k1 is performed to calculate the comprehensive evaluation index R. The expression of the comprehensive evaluation index R is:

[0102] R=w1×C′ 乙炔 +w2×C′ 氢气 +w3×C′ 一氧化碳 +...+w n ×k1,

[0103] Among them, w1 represents the weight coefficient of acetylene gas, w2 represents the weight coefficient of hydrogen, w3 represents the weight coefficient of carbon monoxide, and w n represents the influence weight of air flow temperature ratio k1 on transformer failure risk, C′ 乙炔 Represents the standardized value of acetylene concentration, C′ 氢气 Represents the standardized value of hydrogen concentration, C′ 一氧化碳Represents the normalized value of carbon monoxide concentration;

[0104] (4) Determine the alarm threshold: Obtain the safety threshold R of the comprehensive evaluation index R through historical data or simulation experiments 阈值 , if the real-time monitored R value exceeds the safety threshold R 阈值 , that is, it is determined that the transformer has an abnormal risk and triggers an alarm.

[0105] More preferably, step S4 also includes a dynamic adjustment and multi-level alarm step, and the dynamic adjustment and multi-level alarm are specifically: setting a multi-level alarm threshold, and dividing the degree of exceeding the limit of the R value into different levels of "early warning", "warning" and "serious alarm"; if the R value is close to the threshold but does not exceed it, the system prompts to take preventive maintenance, if it far exceeds the threshold, the emergency shutdown or maintenance process is initiated.

[0106] Preferably, the calculation of the average value k0 of the airflow-temperature ratio k1 in step S6 comprises the following steps:

[0107] (1) Collect gas flow and temperature data: monitor and record gas flow L and temperature T for a long time under different operating conditions of multiple transformers to ensure that the data covers the typical operating temperature range of the transformer and various working load conditions;

[0108] (2) Calculate the air flow-temperature ratio k1 of each sampling, repeat the sampling and obtain a large number of k1 values. The calculation formula of the air flow-temperature ratio k1 is:

[0109]

[0110] Wherein, k1 represents the ratio of gas flow rate to temperature at each sampling time point, L represents gas flow rate, and T represents gas temperature;

[0111] (3) Data cleaning and screening: Clean the collected k1 value data and delete abnormal values, such as abnormal flow or extreme temperature values ​​during equipment maintenance;

[0112] (4) Calculate the mean value by temperature zone: Group the data according to the range of temperature T, and calculate the average value k0 of k1 in each temperature zone to ensure that the k0 used as the benchmark reflects the characteristics of different temperature zones. Specifically, the temperature is divided into multiple temperature zones with a range of 10 degrees Celsius;

[0113] (5) Calculate the overall average value k0 of k1. The expression for calculating the overall average value k0 of k1 is:

[0114] Where N is the total number of samples, k 1,i Represents the airflow-temperature ratio for each sample.

[0115] More preferably, the method for deleting outliers in step (3) is a 3-sigma screening method, which comprises the following steps:

[0116] (1) Calculate the mean and standard deviation. The expressions of the mean and standard deviation are:

[0117]

[0118]

[0119] Where N represents the total number of data points, x i represents the value of the i-th data point, μ represents the mean of the data, and σ represents the standard deviation of the data;

[0120] (2) Determine the 3-sigma range: define the normal data range, that is, [μ-3σ, μ+3σ]. Data points within this range are considered normal data, while data points outside this range are considered outliers.

[0121] (3) Screening outliers: traverse each data point to determine whether it is within the range of [μ-3σ,μ+3σ]. For data points outside this range, mark them as outliers and remove them;

[0122] (4) Update the data set: After removing the outliers, the retained data set is used for subsequent analysis steps. If the data set is small and the outliers significantly affect the mean and standard deviation, repeat the above steps until there are no significant outliers.

[0123] More preferably, step S6 also includes the step of setting dynamic adjustment rules, and the step of setting dynamic adjustment rules is specifically: if the ambient temperature fluctuates significantly during the monitoring period, a weighted average method is used to make k0 reflect the latest airflow temperature ratio characteristics, thereby realizing dynamic updating of k0.

[0124] More preferably, the weighted average method comprises the following steps:

[0125] (1) Set the weighting coefficient w i : The weight coefficient is set to an arithmetic decreasing weight or an exponential decreasing weight. The expression of the arithmetic decreasing weight is:

[0126]

[0127] Among them, w i Represents the weight coefficient, satisfying The closer the data point is to the current one, the higher the weight is. n represents the window size, and i represents the data point number.

[0128] The expression of the exponentially decreasing weight is:

[0129]

[0130] Among them, w i Represents the weight coefficient, satisfying The closer the data point is to the current one, the higher the weight is. λ represents the attenuation coefficient, n represents the window size, i represents the data point number, and j represents each data point in the sliding window.

[0131] (2) Calculate the weighted average value, the calculation formula of the weighted average value is:

[0132]

[0133] Among them, k0 represents the average value of the air flow temperature ratio after update, w i Represents the weight coefficient, k 1,i Represents the i-th airflow temperature ratio within the sliding window.

[0134] The basic principles, main features and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which shall fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for obtaining and distinguishing transformer oil chromatogram, characterized in that: The following steps are involved: Step 1, oil extraction and oil-gas separation, including the following steps: S1. Install the oil valve and oil-gas separation device at the oil inlet of the transformer; S2, using an oil-gas separation device to extract the gas component in the oil and introduce it into the gas sensor; S3. Perform gas component testing on the gas sensor to obtain concentration data of acetylene, hydrogen and carbon monoxide in the gas; Step 2: Install the gas flow sensor and temperature sensor: Install the gas flow sensor and temperature sensor near the gas sensor to monitor the gas flow L and temperature T in real time; Step 3: Calculate the air flow-temperature ratio. The calculation formula of the air flow-temperature ratio is: Among them, k1 represents the ratio of gas flow rate to temperature, L represents gas flow rate, and T represents gas temperature. The larger the k1 value is, the higher the potential danger level of the transformer is. Step 4: Joint judgment, including the following steps: S4, performing correlation analysis on the gas composition data measured by the gas sensor and the k1 value to form a comprehensive evaluation index; S5. Select an insulated exhaust pipe with waterproof function as the exhaust port, and ensure that the height of the exhaust terminal of the exhaust port is greater than or equal to the height of the transformer oil tank; Step 5: Setting the alarm threshold, including the following steps: S6. Monitor the operating environment of multiple transformers, and when the temperature T changes by more than 20 degrees Celsius, calculate the average value k0 of the air flow temperature ratio k1 as a benchmark; S7, set alarm conditions: when When ΔT is less than 10 degrees Celsius, it is considered that the airflow temperature exceeds the reference value by 5%, triggering an alarm.

2. A transformer oil chromatogram acquisition and identification method according to claim 1, characterized in that: The insulated exhaust duct in step S5 is oriented vertically upward.

3. The transformer oil chromatogram acquisition and identification method according to claim 1 is characterized in that: The concentration data and temperature data of acetylene and hydrogen in the gas are read using LoRa, Wi-Fi, Zigbee, RS-485 bus or Ethernet cable.

4. A transformer oil chromatogram acquisition and discrimination method according to claim 1, characterized in that: The step S4 of correlating and analyzing the gas composition data measured by the gas sensor with the k1 value to form a comprehensive evaluation index includes the following steps: (1) Obtain gas composition data: Use gas sensors to detect the gas components separated from transformer oil in real time, and record the concentrations of these gas components as C 乙炔 , C 氢气 and C 一氧化碳 ; (2) Standardized gas composition data: The concentration of each gas is standardized to obtain a dimensionless gas concentration value. The expression of the standardized gas composition data is: Among them, C 气体 Indicates the standardized value of gas concentration, C 气体 Indicates the currently measured gas concentration, μ 气体 represents the average concentration of the gas under historical normal conditions, σ 气体 It represents the standard deviation of the gas under the historical normal state. 气体 When it is close to 0, it means that the current gas concentration is close to the normal level. 气体 When it is much greater than 0 or much less than 0, it means that the current gas concentration deviates significantly from the normal level; (3) Correlation analysis and calculation of comprehensive evaluation index: weighted combination of the standardized gas concentration value and the airflow temperature ratio k1 is performed to calculate the comprehensive evaluation index R. The expression of the comprehensive evaluation index R is: R=w1×C′ 乙炔 +w2×C′ 氢气 +w3×C′ 一氧化碳 +...+w n ×k1, Among them, w1 represents the weight coefficient of acetylene gas, w2 represents the weight coefficient of hydrogen, w3 represents the weight coefficient of carbon monoxide, and w n represents the influence weight of air flow temperature ratio k1 on transformer failure risk, C′ 乙炔 Represents the standardized value of acetylene concentration, C′ 氢气 Represents the standardized value of hydrogen concentration, C′ 一氧化碳 Represents the normalized value of carbon monoxide concentration; (4) Determine the alarm threshold: Obtain the safety threshold R of the comprehensive evaluation index R through historical data or simulation experiments 阈值 , if the real-time monitored R value exceeds the safety threshold R 阈值 , that is, it is determined that the transformer has an abnormal risk and triggers an alarm.

5. A transformer oil chromatogram acquisition and identification method according to claim 4, characterized in that: The step S4 also includes a dynamic adjustment and multi-level alarm step, which specifically includes: setting a multi-level alarm threshold, dividing the degree of exceeding the limit of the R value into different levels of "early warning", "warning" and "serious alarm"; if the R value is close to the threshold but does not exceed it, the system prompts to take preventive maintenance, and if it far exceeds the threshold, the emergency shutdown or maintenance process is initiated.

6. A transformer oil chromatogram acquisition and identification method according to claim 1, characterized in that: The calculation of the average value k0 of the air flow-temperature ratio k1 in step S6 comprises the following steps: (1) Collect gas flow and temperature data: monitor and record gas flow L and temperature T for a long time under different operating conditions of multiple transformers to ensure that the data covers the typical operating temperature range of the transformer and various working load conditions; (2) Calculate the air flow-temperature ratio k1 of each sampling, repeat the sampling and obtain a large number of k1 values. The calculation formula of the air flow-temperature ratio k1 is: Wherein, k1 represents the ratio of gas flow rate to temperature at each sampling time point, L represents gas flow rate, and T represents gas temperature; (3) Data cleaning and screening: Clean the collected k1 value data and delete abnormal values, such as abnormal flow or extreme temperature values ​​during equipment maintenance; (4) Calculate the mean value by temperature zone: Group the data according to the temperature range T, and calculate the average value k0 of k1 in each temperature zone to ensure that the benchmark k0 reflects the characteristics of different temperature zones. Specifically, the grouping includes dividing the temperature into multiple temperature zones with a range of 10 degrees Celsius; (5) Calculate the overall average value k0 of k1. The expression for calculating the overall average value k0 of k1 is: Where N is the total number of samples, k 1,i Represents the airflow-temperature ratio for each sample.

7. A transformer oil chromatogram acquisition and identification method according to claim 6, characterized in that: The method for deleting outliers in step (3) is a 3-sigma screening method, which includes the following steps: (1) Calculate the mean and standard deviation. The expressions of the mean and standard deviation are: Where N represents the total number of data points, x i represents the value of the i-th data point, μ represents the mean of the data, and σ represents the standard deviation of the data; (2) Determine the 3-sigma range: define the normal data range, that is, [μ-3σ, μ+3σ]. Data points within this range are considered normal data, while data points outside this range are considered outliers. (3) Screening outliers: traverse each data point to determine whether it is within the range of [μ-3σ,μ+3σ]. For data points outside this range, mark them as outliers and remove them; (4) Update the data set: After removing the outliers, the retained data set is used for subsequent analysis steps. If the data set is small and the outliers significantly affect the mean and standard deviation, repeat the above steps until there are no significant outliers.

8. The transformer oil chromatogram acquisition and identification method according to claim 6 is characterized in that: The step S6 also includes a step of setting dynamic adjustment rules, wherein the step of setting dynamic adjustment rules is as follows: if the ambient temperature fluctuates significantly during the monitoring period, a weighted average method is used to make k0 reflect the latest airflow temperature ratio characteristics, thereby realizing dynamic updating of k0.

9. A transformer oil chromatogram acquisition and identification method according to claim 8, characterized in that: The weighted average method comprises the following steps: (1) Set the weighting coefficient w i : The weight coefficient is set to an arithmetic decreasing weight or an exponential decreasing weight. The expression of the arithmetic decreasing weight is: Among them, w i Represents the weight coefficient, satisfying The closer the data point is to the current one, the higher the weight. n represents the window size, i represents the data point number; The expression of the exponentially decreasing weight is: Among them, w i Represents the weight coefficient, satisfying The closer the data point is to the current one, the higher the weight is. λ represents the attenuation coefficient, n represents the window size, i represents the data point number, and j represents each data point in the sliding window. (2) Calculate the weighted average value, the calculation formula of the weighted average value is: Among them, k0 represents the average value of the air flow temperature ratio after update, w i Represents the weight coefficient, k 1,i Represents the i-th airflow temperature ratio within the sliding window.

Images