A gas leakage monitoring method and system for the environmental protection field of gasholders

Abstract

The application provides a gas leakage monitoring method and system for the environment-friendly field of gas-filled cabinets. The method obtains the current ambient temperature, current load current, and current monitoring data in the gas chamber of the gas-filled cabinet. The current ambient temperature, current load current, and current gas moisture content are input into the physical model of the gas-filled cabinet to obtain the theoretical value of the current decomposition product content. The current ambient temperature, current load current, and current gas moisture content are input into the prediction model of the gas-filled cabinet to reconstruct the estimated value of the current decomposition product content. According to the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content, the gas leakage monitoring result in the gas chamber of the gas-filled cabinet is determined. The application can improve the detection probability of slight leakage, improve the accuracy and reliability of the gas leakage monitoring result, and reduce the safety hazards and environmental pollution caused by gas leakage.

Description

Technical Field

[0001] This application relates to the field of gas leak monitoring technology, and in particular to a gas leak monitoring method and system for the environmental protection of gas-filled cabinets. Background Technology

[0002] In power systems, gas-filled switchgear, as an important electrical device, is widely used in distribution networks of various voltage levels. The gas inside is typically sulfur hexafluoride (SF6), which has excellent insulating properties, effectively improving the equipment's insulation performance and arc-extinguishing capability. However, during actual operation, gas leaks may occur inside the switchgear due to equipment aging, seal failure, or manufacturing defects. Gas leaks not only lead to decreased equipment performance and affect the safe and stable operation of the power grid, but the leaked gas can also cause serious environmental pollution (for example, SF6 gas has an extremely high greenhouse effect potential, and its impact on the ecological environment cannot be ignored). Therefore, timely detection and handling of gas leaks has become a crucial issue for ensuring the safe operation of gas-filled switchgear and protecting the environment.

[0003] Currently, gas leak detection technology in related technologies relies directly on the monitoring results of gas density relays to determine gas leaks. Although this can meet the monitoring needs to a certain extent, it has shortcomings such as false detection and insensitivity to minor leaks because the gas pressure in the gas filling cabinet is affected by factors such as ambient temperature, load current, and gas decomposition. Therefore, it is necessary to improve the gas leak detection methods in related technologies. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, this application provides a gas leakage monitoring method and system for the environmental protection field of gas-filled cabinets to solve the above-mentioned technical problems.

[0005] According to one aspect of the embodiments of this application, a gas leakage monitoring method for the environmental protection field of gas-filled cabinets is provided. The method includes: acquiring the current ambient temperature, current load current, and current monitoring data of the gas chamber of the gas-filled cabinet, wherein the current monitoring data includes: current gas moisture content and current decomposition product content; inputting the current ambient temperature, current load current, and current gas moisture content into a physical model of the gas-filled cabinet to obtain a theoretical value of the current decomposition product content; the physical model of the gas-filled cabinet is based on the historical ambient temperature, historical load current, and historical monitoring data of the gas chamber of the gas-filled cabinet. The data is fitted to obtain the current ambient temperature, current load current, and current gas moisture content. These are then input into the gas chamber prediction model to reconstruct the estimated value of the current decomposition product content. The gas chamber prediction model is trained based on the historical ambient temperature, historical load current, and historical monitoring data within the gas chamber. The gas leakage monitoring result within the gas chamber is determined based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content.

[0006] According to another aspect of the embodiments of this application, a gas leakage monitoring system for the environmental protection field of gas-filled switchgear is also provided, comprising: a data acquisition module for acquiring the current ambient temperature, current load current, and current monitoring data of the gas chamber of the gas-filled switchgear, wherein the current monitoring data includes: current gas moisture content and current decomposition product content; a physical model module for inputting the current ambient temperature, current load current, and current gas moisture content into a physical model of the gas-filled switchgear to obtain a theoretical value of the current decomposition product content; wherein the physical model of the gas-filled switchgear is based on the historical ambient temperature, historical load current, and historical monitoring data of the gas chamber of the gas-filled switchgear. The model is obtained by fitting the data; the estimation model module is used to input the current ambient temperature, the current load current and the current gas moisture content into the gas chamber prediction model to reconstruct the estimated value of the current decomposition product content; the gas chamber prediction model is obtained by training the pre-built prediction model based on the historical ambient temperature, historical load current and historical monitoring data of the gas chamber; the leakage determination model is used to determine the gas leakage monitoring result in the gas chamber based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content and the current gas moisture content.

[0007] The beneficial effects of this application are as follows: This application obtains the current ambient temperature, current load current, and current monitoring data of the gas chamber of the gas-filling cabinet. The current monitoring data includes the current gas moisture content and the current decomposition product content. The current ambient temperature, current load current, and current gas moisture content are input into the physical model of the gas-filling cabinet to obtain the theoretical value of the current decomposition product content. The current ambient temperature, current load current, and current gas moisture content are input into the prediction model of the gas-filling cabinet to reconstruct the estimated value of the current decomposition product content. Based on the theoretical value, the estimated value, and the current decomposition product content, the application obtains the theoretical value, the estimated value, and the current decomposition product content. By combining the content of pollutants, historical gas moisture content, and current gas moisture content, the gas leakage monitoring results in the gas chamber of the gas filling cabinet are determined. The above process, through the integration of theoretical derivation of physical models and data-driven advantages of prediction models, combined with the dynamic change characteristics of gas moisture content, achieves high-precision monitoring of the gas leakage status of the gas filling cabinet chamber. Compared with the method of comparing single parameter data with thresholds, it can effectively reduce the impact of environmental interference, equipment aging and other factors on monitoring results, improve the detection probability of minor leaks, improve the accuracy and reliability of gas leakage monitoring results, and reduce safety hazards and environmental pollution caused by gas leaks.

[0008] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0009] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings:

[0010] Figure 1 This is a flowchart illustrating a gas leakage monitoring method for the environmental protection field of gas-filled cabinets, as shown in an exemplary embodiment of this application;

[0011] Figure 2 This is a flowchart illustrating a gas leakage monitoring method for the environmental protection field of gas-filled cabinets, as shown in another exemplary embodiment of this application;

[0012] Figure 3 This is a flowchart illustrating a gas leakage monitoring method for the environmental protection field of gas-filled cabinets, as shown in another exemplary embodiment of this application.

[0013] Figure 4 This is a block diagram illustrating a gas leak monitoring system for the environmental protection field of gas-filled cabinets, as shown in an exemplary embodiment of this application. Detailed Implementation

[0014] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0015] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0016] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the present application. However, it will be apparent to those skilled in the art that embodiments of the present application may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the present application.

[0017] The implementation details of the technical solutions in the embodiments of this application are described in detail below:

[0018] Figure 1 This is a flowchart illustrating a gas leak monitoring method for the environmental protection field of gas-filled cabinets, as shown in an exemplary embodiment of this application. (Refer to...) Figure 1 As shown, this gas leak monitoring method for the environmental protection field of gas-filled cabinets includes at least steps S110 to S140, which are described in detail below:

[0019] In step S110, the current ambient temperature, current load current, and current monitoring data of the gas chamber of the gas filling cabinet are acquired. In one embodiment of this application, the current monitoring data includes: current gas moisture content, current decomposition product content, current ambient temperature (e.g., thermocouple or resistance temperature detector), current load current (e.g., Hall effect current sensor), current gas moisture content (e.g., high-precision humidity sensor based on capacitive principle), and current decomposition product content (e.g., electrochemical sensor, optical sensor).

[0020] In step S120, the current ambient temperature, current load current, and current gas moisture content are input into the gas-filled cabinet physical model to obtain the theoretical value of the current decomposition product content. In one embodiment of this application, the gas-filled cabinet physical model is obtained by fitting the historical ambient temperature, historical load current, and historical monitoring data inside the gas chamber of the gas-filled cabinet. The historical monitoring data includes: historical gas moisture content, historical decomposition product content, and historical temperature rise data. The historical ambient temperature is collected by a temperature sensor, the historical load current is collected by a current sensor, the historical gas moisture content is collected by a humidity sensor, the historical decomposition product content is collected by a gas sensor, and the historical temperature rise data includes: the temperature rise at the terminals, contacts, or busbars inside the gas-filled cabinet, etc. The historical temperature rise data is measured by attaching fiber optic temperature sensors to the contacts, busbars, or cable joints.

[0021] In step S130, the current ambient temperature, current load current, and current gas moisture content are input into the gas holder prediction model to reconstruct the estimated value of the current decomposition product content. In one embodiment of this application, the gas holder prediction model is trained based on the historical ambient temperature, historical load current, and historical monitoring data of the gas chamber of the gas holder. The pre-built prediction model can employ deep learning networks, such as Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), or a combination of LSTM and CNN. By learning from historical data, the pre-built prediction model captures the dynamic influence of changes in ambient temperature, load current, and gas moisture content on the decomposition product content.

[0022] In step S140, the gas leakage monitoring result in the gas chamber of the gas filling cabinet is determined based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content. In one embodiment of this application, the gas leakage monitoring result includes: gas leakage, gas not leaking. By combining the theoretical derivation of the physical model with the data-driven advantage of the prediction model, and combining the dynamic change characteristics of the gas moisture content, high-precision monitoring of the gas leakage status of the gas filling cabinet chamber is achieved. Compared with the method of comparing single parameter data with thresholds, it can effectively reduce the impact of environmental interference, equipment aging, and other factors on the monitoring results, improve the detection probability of minor leaks, enhance the accuracy and reliability of gas leakage monitoring results, and reduce safety hazards and environmental pollution caused by gas leaks.

[0023] In one embodiment of this application, if the physical model of the gas-insulated switchgear includes: a physical model of the gas-insulated switchgear in a non-leaking state and a physical model of the gas-insulated switchgear in a leaking state, and the theoretical value of the current decomposition product content includes: the theoretical value of the decomposition products in the non-leaking state and the theoretical value of the decomposition products in the leaking state, then the process of inputting the current ambient temperature, current load current, and current gas moisture content into the physical model of the gas-insulated switchgear to obtain the theoretical value of the current decomposition product content includes:

[0024] The historical ambient temperature, historical load current, and historical monitoring data inside the gas-insulated switchgear are time-aligned to obtain a time-series data combination. In one embodiment of this application, before time-aligning the historical ambient temperature, historical load current, and historical monitoring data inside the gas-insulated switchgear, anomaly detection is performed on the historical ambient temperature, historical load current, and historical monitoring data inside the gas-insulated switchgear. If abnormal data is found, it is removed, and data is filled into the positions of the abnormal data through interpolation to obtain a pre-processed time-series data combination. This pre-processed time-series data combination is then used to replace the original time-series data combination.

[0025] Clustering time-series data combinations yields multiple time-series data combination clusters. Based on the data variance within each time-series data combination cluster, these clusters are filtered to obtain historical operational status data of the gas-insulated switchgear in a leak-free state. In one embodiment of this application, the process of clustering time-series data combinations includes: normalizing the data in different time-series data combinations to obtain normalized time-series data combinations; calculating the DTW (Dynamic Time Warping) distance between different normalized time-series data combinations; constructing a DTW distance matrix based on the distance between different normalized time-series data combinations; and inputting the DTW distance matrix into a hierarchical clustering algorithm to obtain multiple time-series data combination clusters. The process of filtering multiple time-series data clusters based on the data variance in each cluster includes: calculating the variance of each type of historical data (i.e., the variance of historical temperature data, historical load current, historical gas moisture content, historical decomposition product content, and historical temperature rise data) for each cluster; normalizing the variance of the corresponding type of historical data in each cluster based on the variance of each type of historical data to obtain the normalized historical data variance in each cluster (i.e., the normalized historical temperature data variance, normalized historical load current variance, normalized historical gas moisture content variance, normalized historical decomposition product variance, and normalized historical temperature rise data variance in each cluster); and calculating the normalized historical temperature data variance and normalized historical load current variance for each cluster. The weighted average of the variances of current, normalized historical gas moisture content, normalized historical decomposition product content, and normalized historical temperature rise data is used to obtain the weighted average variance value of each time series data cluster. The data in the time series data cluster corresponding to the smallest weighted average variance value is taken as the historical operating status data of the gas-filled cabinet in the non-leaking state. The normalization of the variances of historical temperature data, historical gas moisture content, historical temperature rise data, historical load current, and historical decomposition product content in each time series data cluster is carried out using the same normalization method, such as Min-Max normalization or Z-score standardization.

[0026] The current ambient temperature, current load current, and current gas moisture content are input into the physical model of the gas-filled cabinet in a leak-free state to obtain the theoretical value of the decomposition products in this state. In one embodiment of this application, the physical model of the gas-filled cabinet in a leak-free state is obtained by fitting historical operating data. The physical model of the gas-filled cabinet in a leak-free state includes: a temperature-current-thermal effect model and an energy-moisture-decomposition product physical model in a leak-free state. The temperature-current-thermal effect model describes the temperature rise inside the gas-filled cabinet under the combined effect of ambient temperature and load current. Its inputs are the current ambient temperature and current load current, and its output is the theoretical value of temperature rise. The energy-moisture-decomposition product physical model takes the theoretical value of temperature rise and the current gas moisture content as inputs, and establishes a quantitative relationship between temperature rise, moisture content, and decomposition product generation to output the theoretical value of decomposition products in a leak-free state. These two sub-models are linked together through the physical mechanisms of energy transfer and material transformation to form a complete calculation link for the theoretical value of decomposition products in a leak-free state, providing a physical reference for subsequent leak monitoring.

[0027] The current ambient temperature, current load current, and current gas moisture content are input into the physical model of the gas-filled cabinet under leakage conditions to obtain the theoretical value of decomposition products under leakage conditions. The physical model of the gas-filled cabinet under leakage conditions is obtained by fitting the historical ambient temperature, historical load current, and experimental monitoring data of the gas chamber under leakage conditions. In one embodiment of this application, the experimental monitoring data includes: experimental monitoring gas moisture content, experimental monitoring decomposition product content, and experimental monitoring temperature rise data. The physical model of the gas-filled cabinet under leakage conditions includes: a temperature-current-thermal effect model under leakage conditions and an energy-moisture-decomposition product physical model under leakage conditions. The temperature-current-thermal effect model under leakage conditions describes the correlation between ambient temperature, load current, and internal temperature rise of the gas-filled cabinet under leakage conditions. Its input is the current ambient temperature and current load current, and its output is the theoretical value of temperature rise under leakage conditions. The energy-moisture-decomposition product physical model under leakage conditions describes the dynamic relationship between temperature rise, gas moisture content, and decomposition product generation under leakage conditions. Its input is the theoretical value of temperature rise under leakage conditions and current gas moisture content, and its output is the theoretical value of decomposition products under leakage conditions. The test monitoring data was obtained through simulated leakage tests under different leakage orifice diameters, different gas pressure conditions, and different load current conditions. This ensures that the physical model of the gas-filled cabinet under leakage conditions can cover the characteristics of various leakage levels, providing a comprehensive theoretical data reference for subsequent comparison with actual monitoring data.

[0028] In one embodiment of this application, if the physical model of the gas-filled cabinet in a leak-free state includes: a temperature-current-thermal effect model and an energy-moisture-decomposition product physical model in a leak-free state, then the process of inputting the current ambient temperature, current load current, and current gas moisture content into the physical model of the gas-filled cabinet in a leak-free state to obtain the theoretical values ​​of the decomposition products in a leak-free state includes:

[0029] The current ambient temperature and current load current are input into the temperature-current-thermal effect model under non-leakage conditions to obtain the theoretical value of the temperature rise under non-leakage conditions. In one embodiment of this application, the temperature-current-thermal effect model under non-leakage conditions is obtained by fitting historical ambient temperature, historical load current, and historical temperature rise data from historical operating data. The expression of the temperature-current-thermal effect model under non-leakage conditions is as follows:

[0030] Equation (1)

[0031] in, This indicates historical temperature rise data (e.g., historical temperature rise at terminals, contacts, or busbars inside the gas-insulated cabinet). Indicates the coefficient of thermal effect of electric current. Indicates historical load current. Indicates the influence coefficient of ambient temperature. Indicates historical ambient temperature. This represents a constant term.

[0032] In one embodiment of this application, the process of fitting the historical ambient temperature, historical load current and historical temperature rise data in the historical operating status data to obtain the temperature-current-thermal effect model under non-leakage condition is achieved by linear regression fitting of the historical ambient temperature, historical load current and historical temperature rise data in the historical operating status data using the least squares method (OLS), and solving for the current-thermal effect coefficient, ambient temperature influence coefficient and constant term.

[0033] The theoretical value of temperature rise under non-leakage conditions and the current gas moisture content are input into the energy-moisture-decomposition product physical model under non-leakage conditions to obtain the theoretical value of decomposition products under non-leakage conditions. In one embodiment of this application, the energy-moisture-decomposition product physical model under non-leakage conditions is obtained by fitting historical temperature rise data, historical gas moisture content, and historical decomposition product content from historical operating data. The expression of the energy-moisture-decomposition product physical model under non-leakage conditions is as follows:

[0034] Equation (2)

[0035] in, Indicates the content of historical decomposition products (e.g., sulfur dioxide, hydrogen sulfide, etc.). This represents the pre-exponential factor (a constant characterizing the frequency of gas decomposition reactions). This represents the normalized historical gas moisture content. This indicates the reaction order of water (usually set to 1 or 2, indicating the degree of influence of water on the rate of gas decomposition reaction). The activation energy (unit: J / mol) represents the energy of a gas decomposition reaction. Represents the universal gas constant. This represents historical temperature rise data under non-leaking conditions (e.g., temperature rise at terminals, contacts, or busbars inside a gas-insulated cabinet). Indicates the ambient reference absolute temperature. This indicates the error term.

[0036] In one embodiment of this application, the process of fitting historical temperature rise data, historical gas moisture content and historical decomposition product content in historical operating status data to obtain the energy moisture-decomposition product physical model under non-leakage conditions is obtained by fitting with a nonlinear fitting tool.

[0037] In one embodiment of this application, if the physical model of the gas-filled cabinet in a leaking state includes: a temperature-current-thermal effect model and an energy-moisture-decomposition product physical model in a leaking state, then the process of inputting the current ambient temperature, current load current, and current gas moisture content into the physical model of the gas-filled cabinet in a leaking state to obtain the theoretical values ​​of the decomposition products in a leaking state includes:

[0038] The current ambient temperature and current load current are input into the temperature-current-thermal effect model under leakage conditions to obtain the theoretical value of the temperature rise under leakage conditions. In one embodiment of this application, the temperature-current-thermal effect model under leakage conditions is obtained by fitting historical ambient temperature, historical load current from historical operating data, and experimental monitoring temperature rise data from experimental monitoring data. The expression of the temperature-current-thermal effect model under leakage conditions is as follows:

[0039] Equation (3)

[0040] in, This indicates the temperature rise data monitored during the test. Indicates the resistance loss coefficient. Indicates historical load current. Indicates the influence coefficient of ambient temperature. Indicates historical ambient temperature. Represents the cross-coupling coefficient. This indicates the first error term. This represents the first constant term.

[0041] A temperature-current-thermal effect model under leakage conditions is obtained by fitting historical ambient temperature, historical load current, and experimental temperature rise data from historical operating status data and experimental monitoring data. This model is implemented through linear regression fitting.

[0042] The theoretical value of the temperature rise under leakage conditions and the current gas moisture content are input into the energy-moisture-decomposition product physical model under leakage conditions to obtain the theoretical value of the decomposition products under leakage conditions. In one embodiment of this application, the energy-moisture-decomposition product physical model under leakage conditions is obtained by fitting the experimentally monitored gas moisture content, experimentally monitored decomposition product content, and experimentally monitored temperature rise data from the experimental monitoring data. The expression of the energy-moisture-decomposition product physical model under leakage conditions is as follows:

[0043] Equation (4)

[0044] in, This indicates that the test monitors the content of decomposition products. This represents the pre-exponential factor in the leaked state. This represents the normalized moisture content of the gas monitored in the experiment. This indicates the order of the water reaction under leak conditions. The activation energy (in J / mol) represents the gas decomposition reaction under leak conditions. This indicates the temperature rise data monitored during the test. Represents the universal gas constant. Indicates the ambient reference absolute temperature. This indicates the second error term.

[0045] In one embodiment of this application, the process of fitting the physical model of energy moisture-decomposition products under leakage state based on the experimental monitoring data of moisture content of experimental monitoring gas, content of experimental monitoring decomposition products and temperature rise data of experimental monitoring data is obtained by fitting using nonlinear least squares method.

[0046] In one embodiment of this application, if the gas-insulated switchgear prediction model includes: a prediction model of the gas-insulated switchgear in a non-leaking state and a prediction model of the gas-insulated switchgear in a leaking state, and the estimated value of the current decomposition product content includes: the estimated value of the decomposition products in the non-leaking state and the estimated value of the decomposition products in the leaking state, then the process of inputting the current ambient temperature, the current load current, and the current gas moisture content into the gas-insulated switchgear prediction model to obtain the estimated value of the current decomposition product content includes:

[0047] The current ambient temperature, current load current, and current gas moisture content are input into the prediction model of the gas-filled cabinet in a leak-free state to obtain an estimate of the decomposition products in this state. In one embodiment of this application, the prediction model of the gas-filled cabinet in a leak-free state is obtained by training a pre-built prediction model based on historical operating state data. The process of training the pre-built prediction model based on historical operating state data includes: using historical ambient temperature, historical load current, and historical gas moisture content from the historical operating state data as input features, and historical decomposition product content as output data, to train the pre-built prediction model. During the training process, a sliding window method is used to divide the training samples, taking a continuous time series of historical operating state data as an input sample and the decomposition product content at the corresponding time as an output sample, in order to capture the temporal dependency of the data. Simultaneously, by using the Adam optimizer to minimize the mean squared error between the estimated decomposition products in the leak-free state output by the pre-built prediction model and the output samples, and after multiple rounds of iterative training, training is stopped when the loss of the trained prediction model on the validation set tends to stabilize and no longer decreases, thus obtaining a converged prediction model of the gas-filled tank in the leak-free state. The estimated decomposition products in the leak-free state are the estimated decomposition products at the current time.

[0048] The current ambient temperature, current load current, and current gas moisture content are input into the prediction model for the gas holder under leakage conditions to obtain estimated values ​​of decomposition products under leakage conditions. In one embodiment of this application, the prediction model for the gas holder under leakage conditions is trained based on the historical ambient temperature, historical load current, and experimental monitoring data of the gas holder chamber at different leakage rates. The experimental monitoring data covers experimental monitoring gas moisture content, experimental monitoring decomposition product content, and experimental monitoring temperature rise data at different leakage rates (characterized by different leakage orifice diameters). The process of training the pre-built prediction model for the gas holder under leakage conditions based on the historical ambient temperature, historical load current, and experimental monitoring gas moisture content at different leakage rates includes: using synchronously collected historical ambient temperature, historical load current, and experimental monitoring gas moisture content at different leakage rates as input features, and the experimental monitoring decomposition product content at the corresponding leakage rate as output data, to train the pre-built prediction model. During training, a sliding window method was used to divide the training samples. A continuous time series of multivariate time-series data (including historical ambient temperature, historical load current, and moisture content of the test monitoring gas under different leakage rates) was used as an input sample, and the content of decomposition products at the corresponding time point was used as the output sample, in order to capture the temporal dependencies of the data. Simultaneously, the Adam optimizer was used to minimize the mean square error between the estimated decomposition products under leakage conditions output by the pre-built prediction model and the output sample. After multiple rounds of iterative training, training was stopped when the loss of the trained prediction model on the validation set tended to stabilize and no longer decreased, resulting in a converged prediction model of the gas-filled tank under leakage conditions. The estimated decomposition products under leakage conditions are the estimated decomposition products at the current time point.

[0049] In one embodiment of this application, if the prediction model for the gas-filled cabinet in a leak-free state includes: a temperature-current-temperature rise prediction model and an energy-moisture-decomposition product prediction model in a leak-free state, then the process of inputting the current ambient temperature, current load current, and current gas moisture content into the prediction model for the gas-filled cabinet in a leak-free state to obtain the estimated values ​​of the decomposition products in a leak-free state includes:

[0050] The current ambient temperature and current load current are input into the temperature-current-temperature rise prediction model under non-leakage conditions to obtain an estimated temperature rise value under non-leakage conditions. In one embodiment of this application, the temperature-current-temperature rise prediction model under non-leakage conditions is trained on a pre-built temperature rise prediction model based on historical ambient temperature, historical load current, and historical temperature rise data from historical operating status data. The pre-built temperature rise prediction model adopts a long short-term memory network model or a bidirectional long short-term memory network model. The temperature-current-temperature rise prediction model under non-leakage conditions is used to achieve accurate mapping of the current ambient temperature and current load current to temperature rise data, providing a data-driven theoretical reference for subsequent comparison with actual monitoring data.

[0051] The estimated temperature rise and current gas moisture content under non-leaking conditions are input into an energy-moisture-decomposition product prediction model under non-leaking conditions to obtain an estimated value of decomposition products under non-leaking conditions. In one embodiment of this application, the energy-moisture-decomposition product prediction model under non-leaking conditions is trained on a pre-constructed decomposition product prediction model based on historical temperature rise data, historical gas moisture content, and historical decomposition product content from historical operating data. The pre-constructed temperature rise prediction model and the pre-constructed decomposition product prediction model constitute a pre-constructed prediction model, which can employ a long short-term memory network model or a bidirectional long short-term memory network model, etc. The energy-moisture-decomposition product prediction model under non-leaking conditions uses the estimated temperature rise and current gas moisture content under non-leaking conditions as input features. By learning the nonlinear mapping relationship between historical temperature rise data, historical gas moisture content, and historical decomposition product content, it achieves accurate prediction of the current decomposition product content.

[0052] In one embodiment of this application, if the prediction model for the gas-filled cabinet under leakage conditions includes: a temperature-current-temperature rise prediction model under leakage conditions and an energy-moisture-decomposition product prediction model under leakage conditions, then the process of inputting the current ambient temperature, current load current, and current gas moisture content into the prediction model for the gas-filled cabinet under leakage conditions to obtain the estimated values ​​of decomposition products under leakage conditions includes:

[0053] The current ambient temperature and current load current are input into the temperature-current-temperature rise prediction model under leakage conditions to obtain the estimated temperature rise data under leakage conditions. The temperature-current-temperature rise prediction model under leakage conditions is trained on a pre-built temperature rise prediction model based on the historical ambient temperature, historical load current, and experimental monitoring temperature rise data of the gas-insulated switchgear. The process of training the pre-built temperature rise prediction model under leakage conditions based on the historical ambient temperature, historical load current, and experimental monitoring temperature rise data of the gas-insulated switchgear to obtain the temperature-current-temperature rise prediction model under non-leakage conditions is the same as the process of training the pre-built temperature rise prediction model under non-leakage conditions based on historical ambient temperature, historical load current, and historical temperature rise data from historical operating status data.

[0054] The estimated temperature rise and current gas moisture content under leakage conditions are input into the energy moisture-decomposition product prediction model under leakage conditions to obtain the estimated decomposition products under leakage conditions. In one embodiment of this application, the energy moisture-decomposition product prediction model under leakage conditions is trained on a pre-built decomposition product prediction model based on the experimental monitoring data of gas moisture content, decomposition product content, and temperature rise. The process of training the pre-built decomposition product prediction model under leakage conditions based on the experimental monitoring data of gas moisture content, decomposition product content, and temperature rise is the same as the process of training the pre-built decomposition product prediction model under non-leakage conditions based on historical temperature rise data, historical gas moisture content, and historical decomposition product content from historical operating data.

[0055] In one embodiment of this application, if the theoretical value of the current decomposition product content is the theoretical value of the decomposition product content under a non-leaking state, and the estimated value of the current decomposition product content is the estimated value of the decomposition product content under a non-leaking state, then the process of determining the gas leakage monitoring result in the gas chamber of the gas holder based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content includes:

[0056] The difference between the current content of decomposition products and the theoretical value of decomposition products in the non-leaking state is taken as the theoretical residual value of decomposition products in the non-leaking state; the difference between the current content of decomposition products and the estimated value of decomposition products in the non-leaking state is taken as the estimated residual value of decomposition products in the non-leaking state. In one embodiment of this application, the formula for calculating the theoretical residual value of decomposition products in the non-leaking state is as follows:

[0057] Equation (5)

[0058] in, This represents the theoretical residual value of the decomposition products in the unleashed state. Indicates the current content of decomposition products. This represents the theoretical value of the decomposition products in the unleashed state.

[0059] In one embodiment of this application, the formula for calculating the estimated residual value of the decomposition products in the unleashed state is as follows:

[0060] Equation (6)

[0061] in, This represents the estimated residual value of the decomposition products under the unleashed state. Indicates the current content of decomposition products. This represents the estimated value of the decomposition products in the unleashed state.

[0062] The variation range of gas moisture content is determined based on historical and current gas moisture content. In one embodiment of this application, the variation range of gas moisture content is characterized by the slope of moisture content data within a preset sliding window; the moisture content data within the preset sliding window consists of moisture content data from the n most recent time points preceding the current gas moisture content, and the expression for the slope of the moisture content data within the preset sliding window is as follows:

[0063] Equation (7)

[0064] in, This represents the slope of the moisture content data within the preset sliding window. This indicates the number of moisture content data points within the preset sliding window. Indicates the first in the preset sliding window One moisture content data, Indicates the first in the preset sliding window The time sequence number corresponding to each moisture content data point.

[0065] The gas leakage monitoring results within the gas chamber of the gas holder are determined based on the theoretical residual value of the decomposition products under the non-leaking state, the estimated residual value of the decomposition products under the non-leaking state, and the variation range of gas moisture content. In one embodiment of this application, a comprehensive judgment mechanism is constructed based on the theoretical residual value of the decomposition products under the non-leaking state, the estimated residual value of the decomposition products under the non-leaking state, and the variation range of gas moisture content. This multi-dimensional fusion judgment mechanism can effectively reduce the probability of misjudgment based on a single indicator, improve the accuracy and reliability of gas leakage monitoring, and provide strong protection for the safe and stable operation of the gas holder.

[0066] In one embodiment of this application, the process of determining the gas leakage monitoring result in the gas chamber of the gas holder based on the theoretical residual value of the decomposition products under the non-leaking state, the estimated residual value of the decomposition products under the non-leaking state, and the variation range of the gas moisture content includes:

[0067] A basic probability assignment function is constructed for the theoretical residual value of decomposition products under a non-leaking state; a basic probability assignment function is constructed for the estimated residual value of decomposition products under a non-leaking state; and a basic probability assignment function is constructed for the variation range of gas moisture content. In one embodiment of this application, the expression of the basic probability assignment function for the theoretical residual value of decomposition products under a non-leaking state includes:

[0068] Equation (8)

[0069] in, This indicates the confidence level for determining that the decomposition products under the unleashed state have not been leaked. This represents the theoretical residual value of the decomposition products in the unleashed state. The standard deviation represents the theoretical residual value of the decomposition products in the unleashed state (calculated based on the theoretical residual value of the decomposition products at historical moments in the unleashed state).

[0070] Equation (9)

[0071] in, This represents the uncertainty in determining the undisclosed state based on the theoretical residual value of the decomposition products under the undisclosed state. This indicates the confidence level of determining that the decomposition products under the unleashed state are unleashed.

[0072] In one embodiment of this application, the expression for the basic probability assignment function of the estimated residual value of the decomposition product in the undisclosed state includes:

[0073] Equation (10)

[0074] in, This indicates the confidence level for determining that the decomposition products under the undisclosed state have not been disclosed based on the estimated residual values. This represents the estimated residual value of the decomposition products under the unleashed state. This represents the standard deviation of the estimated residuals of the decomposition products under the unleashed state (calculated based on the estimated residuals of the decomposition products at historical times under the unleashed state).

[0075] Equation (11)

[0076] in, This indicates the uncertainty in determining the undisclosed state based on the estimated residual value of the decomposition products. This indicates the confidence level of determining that the residual value of the decomposition products under the undisclosed state is undisclosed.

[0077] In one embodiment of this application, the expression for the basic probability assignment function of the variation range of gas moisture content includes:

[0078] Equation (12)

[0079] in, This indicates that the moisture content is considered to be in a stable state based on the magnitude of the change in gas moisture content. confidence level This represents the absolute value of the slope of the moisture content data within the preset sliding window. Represents the steady-state threshold. This indicates the threshold for change.

[0080] Equation (13)

[0081] in, This indicates that the change in gas moisture content is determined based on the magnitude of the change. confidence level This represents the absolute value of the slope of the moisture content data within the preset sliding window. Represents the steady-state threshold. This indicates the threshold for change.

[0082] Equation (14)

[0083] in, This indicates that the determination of the moisture content trend is uncertain based on the magnitude of the change in gas moisture content. confidence level This indicates that the moisture content is considered to be in a stable state based on the magnitude of the change in gas moisture content. confidence level This indicates that the change in gas moisture content is determined based on the magnitude of the change. The confidence level.

[0084] According to the Dempster-Shafer evidence combination rule, the basic probability assignment function of the theoretical residual value of the decomposition product in the undisclosed state and the basic probability assignment function of the estimated residual value of the decomposition product in the undisclosed state are fused to obtain an intermediate fusion result. In one embodiment of this application, the process of fusing the basic probability assignment function of the theoretical residual value of the decomposition product in the undisclosed state and the basic probability assignment function of the estimated residual value of the decomposition product in the undisclosed state to obtain an intermediate fusion result according to the Dempster-Shafer evidence combination rule includes: (1) calculating the first conflict based on the confidence level of determining that the product is undisclosed based on the theoretical residual value of the decomposition product in the undisclosed state, the uncertainty level of determining that the product is undisclosed based on the theoretical residual value of the decomposition product in the undisclosed state, the confidence level of determining that the product is undisclosed based on the estimated residual value of the decomposition product in the undisclosed state, and the uncertainty level of determining that the product is undisclosed based on the estimated residual value of the decomposition product in the undisclosed state. (2) Calculate the confidence level of the undisclosed state based on the first conflict coefficient, the confidence level of the theoretical residual value of the decomposition product under the undisclosed state as undisclosed, and the confidence level of the estimated residual value of the decomposition product under the undisclosed state as undisclosed; (3) Calculate the confidence level of the uncertain state based on the conflict coefficient, the uncertainty level of the theoretical residual value of the decomposition product under the undisclosed state as undisclosed, and the uncertainty level of the estimated residual value of the decomposition product under the undisclosed state as undisclosed; (4) Use the confidence level of the undisclosed state and the confidence level of the uncertain state as intermediate fusion results.

[0085] In one embodiment of this application, the formula for calculating the first conflict coefficient is as follows:

[0086] Equation (15)

[0087] in, Indicates the first conflict coefficient. This indicates the confidence level for determining that the decomposition products under the unleashed state have not been leaked. This represents the uncertainty in determining the undisclosed state based on the theoretical residual value of the decomposition products under the undisclosed state. This indicates the confidence level for determining that the decomposition products under the undisclosed state have not been disclosed based on the estimated residual values. This represents the uncertainty in determining whether the residual value of the decomposition products under the undisclosed state is undisclosed.

[0088] In one embodiment of this application, the formula for calculating the degree of trust in the undisclosed state is as follows:

[0089] Equation (16)

[0090] in, This indicates the level of trust in the undisclosed state. This indicates the confidence level for determining that the decomposition products under the unleashed state have not been leaked. This indicates the confidence level for determining that the decomposition products under the undisclosed state have not been disclosed based on the estimated residual values. This represents the first conflict coefficient.

[0091] In one embodiment of this application, the formula for calculating the degree of confidence in an uncertain state is as follows:

[0092] Equation (17)

[0093] in, It indicates the degree of trust in an uncertain state. This represents the uncertainty in determining the undisclosed state based on the theoretical residual value of the decomposition products under the undisclosed state. This indicates the uncertainty in determining the undisclosed state based on the estimated residual value of the decomposition products. This represents the first conflict coefficient.

[0094] According to the Dempster-Shafer evidence combination rule, the intermediate fusion result is fused with the basic probability assignment function of the change range of gas moisture content to obtain the leakage confidence, non-leakage confidence, and uncertainty confidence. In one embodiment of this application, the process of fusing the intermediate fusion result with the basic probability assignment function of the change range of gas moisture content according to the Dempster-Shafer evidence combination rule to obtain the leakage confidence, non-leakage confidence, and uncertainty confidence includes: (1) calculating the degree of confidence in the leakage state based on the intermediate fusion result; (2) determining the moisture content as stable based on the degree of confidence in the non-leakage state, the degree of confidence in the leakage state, and the change range of gas moisture content. The confidence level is determined based on the magnitude of the change in gas moisture content to define the state of moisture content change. (3) Calculate the second conflict coefficient based on the confidence level of the water content in a stable state; As a non-leaking state, the moisture content is stable. As a leakage state; (4) Based on the second conflict coefficient, the degree of confidence in the non-leakage state, the degree of confidence in the uncertain state, and the change range of gas moisture content, it is determined to be a stable moisture content state. The confidence level, based on the magnitude of change in gas moisture content, is determined to be in a state of uncertainty regarding the trend of moisture content. (5) Calculate the confidence level of no leakage based on the second conflict coefficient, the degree of confidence in the leakage state, the degree of confidence in the uncertain state, and the change in gas moisture content. The confidence level, based on the magnitude of change in gas moisture content, is determined to be in a state of uncertainty regarding the trend of moisture content. (6) Calculate the leakage confidence level based on the second conflict coefficient and the change range of gas moisture content, and determine the state of uncertainty of moisture content trend. The degree of confidence in the uncertain state is used to calculate the uncertainty confidence level.

[0095] In one embodiment of this application, the formula for calculating the degree of trust in the leakage status is as follows:

[0096] Equation (18)

[0097] in, This indicates the level of trust in the leaked status. This indicates the level of trust in the undisclosed state. It indicates the degree of trust in an uncertain state.

[0098] In one embodiment of this application, the formula for calculating the second conflict coefficient is as follows:

[0099] Equation (19)

[0100] in, Indicates the second conflict coefficient. This indicates the level of trust in the undisclosed state. This indicates that the change in gas moisture content is determined based on the magnitude of the change. confidence level This indicates the level of trust in the leaked status. This indicates that the moisture content is considered to be in a stable state based on the magnitude of the change in gas moisture content. The confidence level.

[0101] In one embodiment of this application, the formula for calculating the undisclosed confidence level is as follows:

[0102] Equation (20)

[0103] in, This indicates that the confidence level was not disclosed. This indicates the level of trust in the undisclosed state. This indicates that the moisture content is considered to be in a stable state based on the magnitude of the change in gas moisture content. confidence level This indicates that the determination of the moisture content trend is uncertain based on the magnitude of the change in gas moisture content. confidence level It indicates the degree of trust in an uncertain state. This represents the second conflict coefficient.

[0104] In one embodiment of this application, the formula for calculating the disclosure confidence level is as follows: Equation (21)

[0105] in, Indicates the confidence level of the leak. This indicates the level of trust in the leaked status. This indicates that the change in gas moisture content is determined based on the magnitude of the change. confidence level This indicates that the determination of the moisture content trend is uncertain based on the magnitude of the change in gas moisture content. confidence level It indicates the degree of trust in an uncertain state. This represents the second conflict coefficient.

[0106] In one embodiment of this application, the formula for calculating the uncertainty confidence level is as follows:

[0107] Equation (22)

[0108] in, Indicates the confidence level of uncertainty. This indicates that the determination of the moisture content trend is uncertain based on the magnitude of the change in gas moisture content. confidence level It indicates the degree of trust in an uncertain state. This represents the second conflict coefficient.

[0109] If the confidence level of leakage is greater than the first preset confidence threshold, and the confidence level of leakage is greater than the confidence level of no leakage, then the gas leakage monitoring result is determined to be a gas leakage. In one embodiment of this application, the first preset confidence threshold can be set according to actual conditions, for example, 0.5.

[0110] If the confidence level of no leakage is greater than the first preset confidence threshold, and the confidence level of no leakage is greater than the confidence level of leakage, then the gas leakage monitoring result is determined to be no gas leakage. In one embodiment of this application, the first preset confidence threshold can be set according to actual conditions, for example, 0.5.

[0111] If the uncertainty confidence level is greater than the second preset confidence threshold, the gas leak monitoring result is determined to be uncertain. In one embodiment of this application, the second preset confidence threshold is less than the first preset confidence threshold. The second preset confidence threshold is set according to the actual situation, for example, 0.3.

[0112] In one embodiment of this application, when the monitoring result is determined to be uncertain, the ambient temperature, load current, and monitoring data of the gas chamber in the gas holder can be collected at the next moment. The ambient temperature, load current, and monitoring data of the next moment are then input into the physical model of the gas holder to obtain the theoretical value of the decomposition product content at the next moment. The ambient temperature, load current, and monitoring data of the next moment are then input into the prediction model of the gas holder to reconstruct the estimated value of the decomposition product content at the next moment. Based on the theoretical value, estimated value, and content of the decomposition products at the next moment, the historical gas moisture content, and the gas moisture content at the next moment, the gas leakage monitoring result in the gas chamber of the gas holder is determined. If the monitoring results for consecutive preset time periods are all uncertain, an early warning is issued, recommending that technicians conduct on-site manual inspection to ensure the safe and stable operation of the gas holder.

[0113] In one embodiment of this application, following the Dempster-Shafer evidence combination rule, the basic probability assignment functions of the theoretical residual values ​​of decomposition products in the non-leaking state, the estimated residual values ​​of decomposition products in the non-leaking state, and the basic probability assignment functions of the variation range of gas moisture content are fused twice. This facilitates the full integration of the complementarity and correlation of multi-dimensional monitoring data, improving the accuracy and reliability of gas leakage status determination. Through phased fusion, the basic probability assignment functions of the theoretical residual values ​​of decomposition products in the non-leaking state and the estimated residual values ​​of decomposition products in the non-leaking state are first fused to initially screen out the core characteristics of the non-leaking state. Then, the key indicator of the variation range of gas moisture content is introduced for secondary fusion. This effectively reduces the interference of single data fluctuations on monitoring results, enhances the sensitivity to capturing gas leakage trends under complex operating conditions, and provides a more scientific and comprehensive decision-making basis for gas leakage monitoring in the field of gas-filled cabinet environmental protection.

[0114] In one embodiment of this application, the process of determining the gas leakage monitoring result in the gas chamber of the gas holder based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content includes:

[0115] The difference between the current content of decomposition products and the theoretical value of decomposition products in the non-leaking state is taken as the theoretical residual value of decomposition products in the non-leaking state; the difference between the current content of decomposition products and the theoretical value of decomposition products in the leaking state is taken as the theoretical residual value of decomposition products in the leaking state. In one embodiment of this application, the theoretical residual value of decomposition products in the leaking state is the difference between the current content of decomposition products and the theoretical value of decomposition products in the leaking state.

[0116] The difference between the current decomposition product content and the estimated decomposition product content in the non-leaking state is used as the residual value of the decomposition product estimation in the non-leaking state; the difference between the current decomposition product content and the estimated decomposition product content in the leaking state is used as the residual value of the decomposition product estimation in the leaking state. In one embodiment of this application, the residual value of the decomposition product estimation in the leaking state is the difference between the current decomposition product content and the estimated decomposition product content in the leaking state.

[0117] The theoretical residual value of the final decomposition product is determined based on the theoretical residual values ​​of the decomposition products under the non-leaking state and the leaking state; the estimated residual value of the final decomposition product is determined based on the estimated residual values ​​of the decomposition products under the non-leaking state and the leaking state. In one embodiment of this application, the process of determining the final theoretical residual value of the decomposition products based on the theoretical residual values ​​of the decomposition products in the non-leaking state and the decomposition products in the leaking state includes: calculating the absolute value of the theoretical residual value of the decomposition products in the non-leaking state to obtain the theoretical absolute value of the decomposition products in the non-leaking state; calculating the absolute value of the theoretical residual value of the decomposition products in the leaking state to obtain the theoretical absolute value of the decomposition products in the leaking state; calculating a normalized feature value based on the theoretical absolute value of the decomposition products in the non-leaking state and the theoretical absolute value of the decomposition products in the leaking state; if the normalized feature value is less than a preset feature threshold, then the theoretical residual value of the decomposition products corresponding to the smallest absolute value among the theoretical absolute values ​​of the decomposition products in the non-leaking state and the theoretical absolute values ​​of the decomposition products in the leaking state is taken as the final theoretical residual value of the decomposition products; if the normalized feature value is greater than or equal to the preset feature threshold, then the weighted average of the theoretical absolute values ​​of the decomposition products in the non-leaking state and the theoretical absolute values ​​of the decomposition products in the leaking state is taken as the final theoretical residual value of the decomposition products.

[0118] In one embodiment of this application, the process of determining the final estimated residual value of the decomposition products based on the estimated residual value of the decomposition products in the unleashed state and the estimated residual value of the decomposition products in the leaked state is the same as the process of determining the final theoretical residual value of the decomposition products based on the theoretical residual value of the decomposition products in the unleashed state and the theoretical residual value of the decomposition products in the leaked state.

[0119] The variation range of gas moisture content is determined based on historical and current gas moisture content. In one embodiment of this application, the variation range of gas moisture content is shown in formula (7).

[0120] The gas leakage monitoring results in the gas chamber of the gas holder are determined based on the theoretical residual value of the final decomposition products, the estimated residual value of the final decomposition products, and the variation range of gas moisture content. In one embodiment of this application, the process of determining the gas leakage monitoring results in the gas chamber of the gas holder based on the theoretical residual value of the final decomposition products, the estimated residual value of the final decomposition products, and the variation range of gas moisture content is the same as the process of determining the gas leakage monitoring results in the gas chamber of the gas holder based on the theoretical residual value of the decomposition products in the non-leaking state, the estimated residual value of the decomposition products in the non-leaking state, and the variation range of gas moisture content.

[0121] In one embodiment of this application, by introducing the theoretical residual value and the estimated residual value of the final decomposition products, the actual deviation of the decomposition product content under the theoretical model and the prediction model can be comprehensively reflected. Combined with the variation range of gas moisture content, multi-dimensional cross-validation of gas leakage status can be achieved. This not only effectively filters out misjudgments caused by abnormal fluctuations of a single parameter, but also more comprehensively captures the complex changing trends inside the gas chamber of the gas filling cabinet, thereby providing a more reliable basis for gas leakage monitoring, ensuring the safety and stability of the gas filling cabinet in the field of environmental protection, and further improving the robustness and accuracy of monitoring results.

[0122] In one embodiment of this application, the process of determining the final theoretical residual value of the decomposition products based on the theoretical residual values ​​of the decomposition products in the unleashed state and the theoretical residual values ​​of the decomposition products in the leaked state includes:

[0123] The absolute value of the theoretical residual value of the decomposition products in the non-leaking state is calculated to obtain the theoretical absolute value of the decomposition products in the non-leaking state. In one embodiment of this application, the formula for calculating the theoretical absolute value of the decomposition products in the non-leaking state is as follows:

[0124] Equation (23)

[0125] in, This represents the theoretical absolute value of the decomposition products under the condition of no leakage. Indicates the current content of decomposition products. This represents the theoretical value of the decomposition products in the unleashed state.

[0126] The absolute value of the theoretical residual value of the decomposition products under leakage conditions is calculated to obtain the theoretical absolute value of the decomposition products under leakage conditions. In one embodiment of this application, the formula for calculating the theoretical absolute value of the decomposition products under leakage conditions is as follows:

[0127] Equation (24)

[0128] in, This represents the theoretical absolute value of the decomposition products under leak conditions. Indicates the current content of decomposition products. This represents the theoretical value of the decomposition products under leakage conditions.

[0129] Based on the theoretical absolute values ​​of the decomposition products under the non-leaking state and the theoretical absolute values ​​of the decomposition products under the leaking state, normalized eigenvalues ​​are calculated. In one embodiment of this application, the formula for calculating the normalized eigenvalues ​​is as follows:

[0130] Equation (25)

[0131] in, Represents the normalized eigenvalues. This represents the theoretical absolute value of the decomposition products under the condition of no leakage. This represents the theoretical absolute value of the decomposition products under leak conditions. Indicates the smoothing term (e.g., 10). -6 ).

[0132] If the normalized eigenvalue is less than a preset eigenvalue threshold, the theoretical residual value of the decomposition product corresponding to the smallest absolute value between the theoretical absolute value of the decomposition product in the unleashed state and the theoretical absolute value of the decomposition product in the leaked state is taken as the final theoretical residual value of the decomposition product. In one embodiment of this application, the preset eigenvalue threshold is set according to the actual situation, for example, 0.6.

[0133] If the normalized eigenvalue is greater than or equal to a preset eigenvalue threshold, the weighted average of the theoretical absolute values ​​of the decomposition products under the non-leaking state and the leaking state is used as the final theoretical residual value of the decomposition products. In one embodiment of this application, a normalized eigenvalue is calculated based on the theoretical absolute values ​​of the decomposition products under the non-leaking state and the leaking state, and the final theoretical residual value of the decomposition products is determined based on the comparison between the normalized eigenvalue and the preset eigenvalue threshold. This process comprehensively considers the distribution characteristics of the theoretical residuals of the decomposition products under both non-leaking and leaking states, avoiding misjudgments caused by abnormal residuals under a single state. When the normalized eigenvalue is small, it indicates that the absolute values ​​of the residuals of the two states are significantly different, and selecting the residual corresponding to the smaller absolute value can highlight the characteristics of the dominant state. When the normalized eigenvalue is large, the information of the two states can be integrated through weighted averaging (arithmetic average or reciprocal weighting, etc.), reducing the impact of extreme values ​​on the results, thereby providing a more robust and reliable residual data foundation for subsequent gas leak status determination, and further improving the adaptability of the monitoring model to complex working conditions.

[0134] Figure 2 This is a flowchart illustrating another exemplary embodiment of a gas leak monitoring method for the environmental protection field of gas-filled cabinets, in which... Figure 2 In this paper, the gas leakage monitoring method for the environmental protection field of gas-filled switchgear includes: (1) obtaining the current ambient temperature, current load current, and current monitoring data of the gas chamber of the gas-filled switchgear; (2) inputting the current ambient temperature, current load current, and current gas moisture content into the physical model of the gas-filled switchgear in the non-leakage state to obtain the theoretical value of the decomposition products in the non-leakage state; (3) inputting the current ambient temperature, current load current, and current gas moisture content into the prediction model of the gas-filled switchgear in the non-leakage state to obtain the estimated value of the decomposition products in the non-leakage state; (4) inputting the current decomposition product content and the theoretical value of the decomposition products in the non-leakage state into the prediction model of the gas-filled switchgear in the non-leakage state. The difference between the two is used as the theoretical residual value of the decomposition products under the non-leakage state; the difference between the current decomposition product content and the estimated value of the decomposition products under the non-leakage state is used as the estimated residual value of the decomposition products under the non-leakage state; (5) the change range of gas moisture content is determined based on the historical gas moisture content and the current gas moisture content; the change range of gas moisture content is characterized by the slope of the moisture content data within the preset sliding window; (6) the gas leakage monitoring results in the gas chamber of the gas filling cabinet are determined based on the theoretical residual value of the decomposition products under the non-leakage state, the estimated residual value of the decomposition products under the non-leakage state, and the change range of gas moisture content.

[0135] Figure 3 This is a flowchart illustrating a gas leakage monitoring method for the environmental protection field of gas-filled cabinets, as shown in another exemplary embodiment of this application. Figure 3The process of the gas leakage monitoring method for the environmental protection field of gas-filled switchgear includes: (1) obtaining the current ambient temperature, current load current, and current monitoring data of the gas chamber of the gas-filled switchgear; (2) inputting the current ambient temperature, current load current, and current gas moisture content into the physical model of the gas-filled switchgear in the non-leakage state to obtain the theoretical value of the decomposition products in the non-leakage state; (3) inputting the current ambient temperature, current load current, and current gas moisture content into the physical model of the gas-filled switchgear in the leakage state to obtain the theoretical value of the decomposition products in the leakage state; (4) inputting the current ambient temperature, current load current, and current gas moisture content into the prediction model of the gas-filled switchgear in the non-leakage state to obtain the estimated value of the decomposition products in the non-leakage state; (5) inputting the current ambient temperature, current load current, and current gas moisture content into the prediction model of the gas-filled switchgear in the leakage state to obtain the estimated value of the decomposition products in the leakage state; (6) taking the difference between the current decomposition product content and the theoretical value of the decomposition products in the non-leakage state as the decomposition product content in the non-leakage state. Solve the theoretical residual value of the decomposition product; take the difference between the current content of the decomposition product and the theoretical value of the decomposition product under the leakage state as the theoretical residual value of the decomposition product under the leakage state; (7) take the difference between the current content of the decomposition product and the estimated value of the decomposition product under the non-leakage state as the estimated residual value of the decomposition product under the non-leakage state; take the difference between the current content of the decomposition product and the estimated value of the decomposition product under the leakage state as the estimated residual value of the decomposition product under the leakage state; (8) determine the final theoretical residual value of the decomposition product based on the theoretical residual value of the decomposition product under the non-leakage state and the theoretical residual value of the decomposition product under the leakage state; determine the final estimated residual value of the decomposition product based on the estimated residual value of the decomposition product under the non-leakage state and the estimated residual value of the decomposition product under the leakage state; (9) determine the change range of the gas moisture content based on the historical gas moisture content and the current gas moisture content; (10) determine the gas leakage monitoring result in the gas chamber of the gas filling cabinet based on the final theoretical residual value of the decomposition product, the final estimated residual value of the decomposition product and the change range of the gas moisture content.

[0136] The following describes an embodiment of the apparatus described in this application, which can be used to execute the gas leak monitoring system for the environmental protection field of gas-filled cabinets as described in the above embodiments of this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the gas leak monitoring method for the environmental protection field of gas-filled cabinets described above in this application.

[0137] Figure 4 This is a block diagram illustrating a gas leak monitoring system for the environmental protection field of gas-filled cabinets, as shown in an exemplary embodiment of this application.

[0138] like Figure 4 As shown, this exemplary gas leak monitoring system 400 for the environmental protection field of gas-filled cabinets includes:

[0139] The data acquisition module 401 is used to acquire the current ambient temperature, current load current, and current monitoring data of the gas chamber of the gas chamber. The current monitoring data includes the current gas moisture content and the current decomposition product content.

[0140] The physical model module 402 is used to input the current ambient temperature, current load current and current gas moisture content into the gas chamber physical model to obtain the theoretical value of the current decomposition product content; the gas chamber physical model is obtained by fitting the historical ambient temperature, historical load current and historical monitoring data in the gas chamber of the gas chamber.

[0141] The estimation model module 403 is used to input the current ambient temperature, current load current and current gas moisture content into the gas holder prediction model to reconstruct the estimated value of the current decomposition product content. The gas holder prediction model is obtained by training a pre-built prediction model based on the historical ambient temperature, historical load current and historical monitoring data in the gas chamber of the gas holder.

[0142] Leakage determination model 404 is used to determine the gas leakage monitoring results in the gas chamber of the gas holder based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content.

[0143] In one embodiment of this application, the current monitoring data includes: current gas moisture content, current decomposition product content, current ambient temperature (obtained by a temperature sensor, such as a thermocouple or a resistance temperature detector), current load current (obtained by a current sensor, such as a Hall effect current sensor), current gas moisture content (obtained by a humidity sensor, such as a high-precision humidity sensor based on the capacitive principle), and current decomposition product content (obtained by a gas sensor, such as an electrochemical sensor or an optical sensor).

[0144] In one embodiment of this application, the physical model of the gas-insulated switchgear is obtained by fitting the historical ambient temperature, historical load current, and historical monitoring data inside the gas chamber of the gas-insulated switchgear. The historical monitoring data includes: historical gas moisture content, historical decomposition product content, and historical temperature rise data. The historical ambient temperature is collected by a temperature sensor, the historical load current is collected by a current sensor, the historical gas moisture content is collected by a humidity sensor, the historical decomposition product content is collected by a gas sensor, and the historical temperature rise data includes: the temperature rise at the terminals, contacts, or busbars inside the gas-insulated switchgear. The historical temperature rise data is measured by attaching fiber optic temperature sensors to the contacts, busbars, or cable joints.

[0145] In one embodiment of this application, the gas-insulated switchgear prediction model is trained based on historical ambient temperature, historical load current, and historical monitoring data within the gas chamber of the gas-insulated switchgear. The pre-built prediction model can employ a deep learning network, such as a long short-term memory network, a convolutional neural network, or a combination of long short-term memory and convolutional neural networks. By learning from historical data, the pre-built prediction model captures the dynamic influence of changes in ambient temperature, load current, and gas moisture content on the content of decomposition products. During model training, root mean square error or mean absolute error can be used as the loss function, and the model parameters are continuously adjusted using a backpropagation algorithm until the prediction accuracy of the pre-built prediction model meets the preset requirements, thus obtaining the gas-insulated switchgear prediction model.

[0146] In one embodiment of this application, the gas leak monitoring results include: gas leak, gas not leaking, and by leveraging the theoretical derivation of the fusion physical model and the data-driven advantages of the prediction model, combined with the dynamic change characteristics of gas moisture content, high-precision monitoring of the gas leak status of the gas chamber of the gas filling cabinet is achieved. Compared with the method of comparing single parameter data with thresholds, it can effectively reduce the impact of environmental interference, equipment aging and other factors on the monitoring results, improve the detection probability of minor leaks, enhance the accuracy and reliability of gas leak monitoring results, and reduce safety hazards and environmental pollution caused by gas leaks.

[0147] It should be noted that the gas leakage monitoring system for the environmental protection field of gas-filled cabinets provided in the above embodiments and the gas leakage monitoring method for the environmental protection field of gas-filled cabinets provided in the above embodiments belong to the same concept. The specific operation methods of each module and unit have been described in detail in the method embodiments and will not be repeated here. In practical applications, the gas leakage monitoring system for the environmental protection field of gas-filled cabinets provided in the above embodiments can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. This is not a limitation here.

[0148] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A gas leakage monitoring method for gas-filled switchgear in the environmental protection field, characterized in that, The method includes: The current ambient temperature, current load current, and current monitoring data of the gas chamber of the gas filling cabinet are obtained. The current monitoring data includes the current gas moisture content and the current decomposition product content. The current ambient temperature, current load current, and current gas moisture content are input into the gas chamber physical model to obtain the theoretical value of the current decomposition product content; the gas chamber physical model is obtained by fitting the historical ambient temperature, historical load current, and historical monitoring data of the gas chamber. The current ambient temperature, the current load current, and the current gas moisture content are input into the gas chamber prediction model to reconstruct the estimated value of the current decomposition product content. The gas chamber prediction model is obtained by training a pre-built prediction model based on the historical ambient temperature, historical load current, and historical monitoring data of the gas chamber. Based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content, the gas leakage monitoring results in the gas chamber of the gas-filling cabinet are determined.

2. The gas leakage monitoring method for gas-filled switchgear in the environmental protection field according to claim 1, characterized in that, If the physical model of the gas-insulated cabinet includes: a physical model of the gas-insulated cabinet in a non-leaking state and a physical model of the gas-insulated cabinet in a leaking state, and the theoretical value of the current decomposition product content includes: the theoretical value of the decomposition products in a non-leaking state and the theoretical value of the decomposition products in a leaking state, then the process of inputting the current ambient temperature, the current load current, and the current gas moisture content into the physical model of the gas-insulated cabinet to obtain the theoretical value of the current decomposition product content includes: The historical ambient temperature, historical load current, and historical monitoring data of the gas-filled cabinet are time-aligned to obtain a time-series data combination; the historical monitoring data includes: historical gas moisture content, historical decomposition product content, and historical temperature rise data; the historical temperature rise data includes: temperature rise at terminals, contacts, or busbars inside the gas-filled cabinet; The time-series data combinations are clustered to obtain multiple time-series data combination clusters; based on the data variance in each time-series data combination cluster, the multiple time-series data combination clusters are filtered to obtain the historical operating status data of the gas-insulated cabinet in a leak-free state; The current ambient temperature, the current load current, and the current gas moisture content are input into the physical model of the gas filling cabinet in a leak-free state to obtain the theoretical value of the decomposition products in the leak-free state; the physical model of the gas filling cabinet in a leak-free state is obtained by fitting the historical operating state data. The current ambient temperature, the current load current, and the current gas moisture content are input into the physical model of the gas-filled cabinet under leakage conditions to obtain the theoretical values ​​of the decomposition products under leakage conditions. The physical model of the gas-filled cabinet under leakage conditions is obtained by fitting the historical ambient temperature, historical load current, and experimental monitoring data of the gas chamber of the gas-filled cabinet under leakage conditions.

3. The gas leakage monitoring method for gas-filled switchgear in the environmental protection field according to claim 2, characterized in that, If the physical model of the gas-filled cabinet in a leak-free state includes: a temperature-current-thermal effect model and an energy-moisture-decomposition product physical model in a leak-free state, then the process of inputting the current ambient temperature, the current load current, and the current gas moisture content into the physical model of the gas-filled cabinet in a leak-free state to obtain the theoretical values ​​of the decomposition products in the leak-free state includes: The current ambient temperature and the current load current are input into the temperature-current-thermal effect model under non-leakage conditions to obtain the theoretical value of the temperature rise data under non-leakage conditions; the temperature-current-thermal effect model under non-leakage conditions is obtained by fitting the historical ambient temperature, historical load current and historical temperature rise data in the historical operating state data; The theoretical value of the temperature rise data under the non-leakage state and the current gas moisture content are input into the energy moisture-decomposition product physical model under the non-leakage state to obtain the theoretical value of the decomposition products under the non-leakage state; the energy moisture-decomposition product physical model under the non-leakage state is obtained by fitting the historical temperature rise data, historical gas moisture content and historical decomposition product content in the historical operating state data.

4. The gas leakage monitoring method for gas-filled switchgear in the environmental protection field according to claim 3, characterized in that, If the gas-insulated switchgear prediction model includes: a prediction model of the gas-insulated switchgear in a non-leaking state and a prediction model of the gas-insulated switchgear in a leaking state, and the estimated value of the current decomposition product content includes: the estimated value of the decomposition products in a non-leaking state and the estimated value of the decomposition products in a leaking state, then the process of inputting the current ambient temperature, the current load current, and the current gas moisture content into the gas-insulated switchgear prediction model to obtain the estimated value of the current decomposition product content includes: The current ambient temperature, the current load current, and the current gas moisture content are input into the prediction model of the gas-filled cabinet in the non-leaking state to obtain the estimated value of the decomposition products in the non-leaking state; the prediction model of the gas-filled cabinet in the non-leaking state is obtained by training the pre-built prediction model based on the historical operating state data; the pre-built prediction model includes: a long short-term memory network model. The current ambient temperature, the current load current, and the current gas moisture content are input into the prediction model of the gas-filled cabinet under leakage conditions to obtain the estimated value of the decomposition products under leakage conditions. The prediction model of the gas-filled cabinet under leakage conditions is obtained by training the pre-built prediction model based on the historical ambient temperature, historical load current of the gas-filled cabinet, and experimental monitoring data of the gas chamber of the gas-filled cabinet under different leakage rates.

5. The gas leakage monitoring method for gas-filled switchgear in the environmental protection field according to claim 4, characterized in that, If the prediction model for the gas-filled cabinet in a leak-free state includes: a temperature-current-temperature rise prediction model and an energy-moisture-decomposition product prediction model in a leak-free state, then the process of inputting the current ambient temperature, the current load current, and the current gas moisture content into the prediction model for the gas-filled cabinet in a leak-free state to obtain the estimated value of the decomposition products in the leak-free state includes: The current ambient temperature and the current load current are input into the temperature-current-temperature rise prediction model under the non-leakage state to obtain the temperature rise data estimate under the non-leakage state; the temperature-current-temperature rise prediction model under the non-leakage state is obtained by training a pre-built temperature rise prediction model based on the historical ambient temperature, historical load current and historical temperature rise data in the historical operating state data. The estimated temperature rise data under the non-leakage state and the current gas moisture content are input into the energy moisture-decomposition product prediction model under the non-leakage state to obtain the estimated decomposition product value under the non-leakage state; the energy moisture-decomposition product prediction model under the non-leakage state is trained on a pre-built decomposition product prediction model based on the historical temperature rise data, historical gas moisture content and historical decomposition product content in the historical operating state data; the pre-built temperature rise prediction model and the pre-built decomposition product prediction model constitute the pre-built prediction model.

6. The gas leakage monitoring method for gas-filled switchgear in the environmental protection field according to any one of claims 1-5, characterized in that, If the theoretical value of the current decomposition product content is the theoretical value of the decomposition product under a non-leaking state, and the estimated value of the current decomposition product content is the estimated value of the decomposition product under a non-leaking state, then the process of determining the gas leakage monitoring result in the gas chamber of the gas-filled cabinet based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content includes: The difference between the current content of decomposition products and the theoretical value of decomposition products in the non-leaking state is taken as the theoretical residual value of decomposition products in the non-leaking state; the difference between the current content of decomposition products and the estimated value of decomposition products in the non-leaking state is taken as the estimated residual value of decomposition products in the non-leaking state. The variation range of the gas moisture content is determined based on the historical gas moisture content and the current gas moisture content; the variation range of the gas moisture content is characterized by the slope of the moisture content data within a preset sliding window. Based on the theoretical residual value of the decomposition products under the non-leaking state, the estimated residual value of the decomposition products under the non-leaking state, and the variation range of the gas moisture content, the gas leakage monitoring results in the gas chamber of the gas-filling cabinet are determined.

7. The gas leakage monitoring method for gas-filled switchgear in the environmental protection field according to claim 6, characterized in that, The process of determining the gas leakage monitoring results in the gas chamber of the gas holder based on the theoretical residual value of the decomposition products under the non-leakage state, the estimated residual value of the decomposition products under the non-leakage state, and the variation range of the gas moisture content includes: A basic probability assignment function is constructed for the theoretical residual value of the decomposition products under the non-leaking state; a basic probability assignment function is constructed for the estimated residual value of the decomposition products under the non-leaking state; a basic probability assignment function is constructed for the variation range of the gas moisture content. According to the Dempster-Shafer evidence combination rule, the basic probability assignment function of the theoretical residual value of the decomposition product in the undisclosed state and the basic probability assignment function of the estimated residual value of the decomposition product in the undisclosed state are fused to obtain an intermediate fusion result. According to the Dempster-Shafer evidence combination rule, the intermediate fusion result is fused with the basic probability assignment function of the change range of gas moisture content to obtain the leakage confidence, non-leakage confidence, and uncertainty confidence. If the leakage confidence level is greater than the first preset confidence threshold, and the leakage confidence level is greater than the no-leakage confidence level, then the gas leakage monitoring result is determined to be a gas leakage. If the confidence level of no leakage is greater than the first preset confidence threshold, and the confidence level of no leakage is greater than the confidence level of leakage, then the gas leakage monitoring result is determined to be no gas leakage. If the uncertainty confidence level is greater than the second preset confidence threshold, then the gas leak monitoring result is determined to be uncertain; the second preset confidence threshold is less than the first preset confidence threshold.

8. The gas leakage monitoring method for gas-filled switchgear in the environmental protection field according to any one of claims 4 or 5, characterized in that, The process of determining the gas leakage monitoring result in the gas chamber of the gas-filled cabinet based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content includes: The difference between the current content of decomposition products and the theoretical value of decomposition products in the non-leaking state is taken as the theoretical residual value of decomposition products in the non-leaking state; the difference between the current content of decomposition products and the theoretical value of decomposition products in the leaking state is taken as the theoretical residual value of decomposition products in the leaking state. The difference between the current content of decomposition products and the estimated value of decomposition products in the non-leaking state is taken as the residual value of decomposition product estimation in the non-leaking state; the difference between the current content of decomposition products and the estimated value of decomposition products in the leaking state is taken as the residual value of decomposition product estimation in the leaking state. The theoretical residual value of the final decomposition product is determined based on the theoretical residual value of the decomposition product in the non-leaking state and the theoretical residual value of the decomposition product in the leaking state; the estimated residual value of the final decomposition product is determined based on the estimated residual value of the decomposition product in the non-leaking state and the estimated residual value of the decomposition product in the leaking state. The variation range of the gas moisture content is determined based on the historical gas moisture content and the current gas moisture content; Based on the theoretical residual value of the final decomposition products, the estimated residual value of the final decomposition products, and the variation range of the gas moisture content, the gas leakage monitoring results in the gas chamber of the gas-filling cabinet are determined.

9. The gas leakage monitoring method for gas-filled switchgear in the environmental protection field according to claim 8, characterized in that, The process of determining the final theoretical residual value of the decomposition products based on the theoretical residual values ​​of the decomposition products in the non-leaking state and the theoretical residual values ​​of the decomposition products in the leaking state includes: Calculate the absolute value of the theoretical residual value of the decomposition products in the non-leaking state to obtain the theoretical absolute value of the decomposition products in the non-leaking state; Calculate the absolute value of the theoretical residual value of the decomposition products under the leakage state to obtain the theoretical absolute value of the decomposition products under the leakage state; Calculate the normalized eigenvalues ​​based on the theoretical absolute values ​​of the decomposition products under the non-leaking state and the theoretical absolute values ​​of the decomposition products under the leaking state. If the normalized feature value is less than the preset feature threshold, then the theoretical residual value of the decomposition product corresponding to the smallest absolute value of the decomposition product in the unleashed state and the theoretical absolute value of the decomposition product in the leaked state shall be used as the final theoretical residual value of the decomposition product. If the normalized feature value is greater than or equal to the preset feature threshold, then the weighted average of the theoretical absolute value of the decomposition product in the unleashed state and the theoretical absolute value of the decomposition product in the leaked state is taken as the theoretical residual value of the final decomposition product.

10. A gas leak monitoring system for the environmental protection field of gas-filled switchgear, characterized in that, include: The data acquisition module is used to acquire the current ambient temperature, current load current, and current monitoring data of the gas chamber of the gas chamber. The current monitoring data includes the current gas moisture content and the current decomposition product content. The physical model module is used to input the current ambient temperature, the current load current, and the current gas moisture content into the gas chamber physical model to obtain the theoretical value of the current decomposition product content; the gas chamber physical model is obtained by fitting the historical ambient temperature, historical load current, and historical monitoring data of the gas chamber. The estimation model module is used to input the current ambient temperature, the current load current, and the current gas moisture content into the gas holder prediction model to reconstruct the estimated value of the current decomposition product content; the gas holder prediction model is obtained by training a pre-built prediction model based on the historical ambient temperature, historical load current, and historical monitoring data of the gas holder's gas chamber; The leakage determination model is used to determine the gas leakage monitoring results in the gas chamber of the gas-filling cabinet based on the theoretical value of the current decomposition product content, the estimated value of the current decomposition product content, the current decomposition product content, the historical gas moisture content, and the current gas moisture content.

Images