Power distribution network transformer health state on-line monitoring method and device

Abstract

The invention provides a power distribution network transformer health state on-line monitoring method and device, and belongs to the technical field of transformer state monitoring, and the method comprises the steps: building a state monitoring on-line database and a historical database according to a normal operation state, a historical operation state, an aging state and a fault state of a power distribution network transformer; the method comprises the following steps: establishing feature data of different periodic weights based on multi-source time series data in a historical database, integrating the expanded multi-source time series data into a reconstructed feature data set by applying the periodic weights, integrating the reconstructed feature data set into a single fusion index, and establishing a monitoring model of the fusion index and a state result based on a random degradation model; constructing an optimization objective function of parameters and feature weights of the monitoring model based on the deviation value of the prediction state and the actual state; and performing reverse optimization adjustment on the parameters and the feature weights of the monitoring model based on the historical database, and monitoring the health state in the online database according to the optimized monitoring model.

Description

Technical Field

[0001] The present invention belongs to the technical field of transformer status monitoring, and specifically relates to an online monitoring method and device for the health status of a transformer in a distribution network. Background Art

[0002] The distribution network transformer is the core equipment connecting the distribution network and the user's electricity consumption side. The stable operation of the transformer plays a vital role in the safe and stable operation of the entire power network. At present, the two methods of regular maintenance and post-maintenance are generally used for the distribution network transformers in operation. However, these two methods will cause the problems of "over-maintenance" and "disrepair". On the one hand, it will waste a lot of manpower and material resources. On the other hand, it may cause the transformer to be out of operation unplanned, affecting the stable operation of the power system. Therefore, online monitoring of the health status of the transformer is a necessary link for the safe and stable operation of the distribution network transformer. At the same time, accurate and effective status monitoring can effectively extend the life cycle of the distribution network transformer and improve its performance and economic benefits.

[0003] At present, the online monitoring technology of transformer health status is mainly divided into two categories: model-driven and data-driven. One is the model-driven online health status monitoring method, which predicts the health status by establishing a mathematical model of the transformer. The model-driven method has limited application of historical data sets and is highly dependent on the status online monitoring model. The other is the data-driven online health status monitoring method, which predicts the health status through statistical or neural network analysis of historical data. The data-driven method requires a large number of data samples and may overfit. This type of method has been widely used in existing application research, but it requires a large number of data samples to support the research. At the same time, large samples will also cause the model to overfit.

[0004] As the distribution network is undergoing informatization and digital transformation, how to reasonably, efficiently and accurately utilize the existing monitoring data is a major difficulty faced by the transformer status monitoring of the distribution network. A large amount of online monitoring data and offline test data are stored in the transformer status monitoring system of the distribution network, and the time series data of each monitoring module cannot be effectively utilized. In order to solve the above problems, an online monitoring method for distribution network transformers is invented by considering the periodic characteristics of monitoring data, which plays a vital role in ensuring the safe and stable operation of distribution network transformers. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a method for online monitoring of the health status of a transformer in a distribution network, comprising the following steps:

[0006] Step 1: Establish a status monitoring online database and a historical database based on the normal operating status, historical operating status, aging status and fault status of the distribution network transformer;

[0007] Step 2: Based on the multi-source time series data in the historical database, establish feature data with different period weights, and use the period weights to integrate the expanded multi-source time series data into a reconstructed feature data set, integrate the reconstructed feature data set into a single fusion indicator, and establish a monitoring model for the fusion indicator and status results based on the random degradation model;

[0008] Step 3: Construct an optimization objective function of the parameters and feature weights of the monitoring model based on the deviation value between the predicted state and the actual state;

[0009] Step 4: Perform reverse optimization and adjustment on the parameters and feature weights of the monitoring model based on the historical database, and monitor the health status in the online database according to the optimized monitoring model.

[0010] In a preferred embodiment: the characteristics in the status online database include gas online characteristics and temperature online characteristics; wherein the gas online characteristics include: hydrogen, methane, acetylene, ethylene, ethane, carbon monoxide, carbon dioxide, total hydrocarbons, and total hydrocarbons and values; the historical database includes gas online characteristics, temperature online characteristics, and historical health status.

[0011] In a preferred embodiment: the characteristic data of different period weights include: for the characteristic data X in the historical database, M}, set the initial sliding window period sequence S = {s1,s2,···,s N}, based on S, the long time series feature data with a time series length of L is divided into n Then, these small one-dimensional data blocks are stacked into a length of s n , width is p n Two-dimensional time series feature data set, where L = s n ×p n .

[0012] In a preferred embodiment: The reconstruction feature is represented as:

[0013]

[0014] where a i is the cycle weight, y i is the feature data divided into small blocks of data, x m ' represents the mth reconstructed feature.

[0015] In a preferred embodiment: a composite index after fusing multiple performance data is established, and errors and feature weights are considered to obtain a composite health index Z(t) after fusing multiple reconstructed feature data:

[0016]

[0017] where x t,m ' represents the reconstructed feature data of the mth reconstructed feature at time t, b m represents the feature weight corresponding to the mth reconstructed feature; c m Represents the error term corresponding to the mth reconstructed feature.

[0018] In a preferred embodiment: a random degradation function of the degradation model is established, and the time-varying degradation process of the composite health index Z(t) of the linear Wiener process of the standard Brownian motion is considered to construct a random degradation function of the composite health index:

[0019] Z(t)=z1+θt+σD(t)

[0020] Among them, Z(t) is the composite health index; z1 is the initial state of the composite health index; D(t) is the standard Brownian motion, θ is the drift coefficient, and σ is the diffusion coefficient.

[0021] In a preferred embodiment: the predicted state value is calculated by a random degradation function model, and the state value T' of the random degradation of the distribution network transformer is defined as follows:

[0022] T′=inf{t:Z(t)≥v|z1<v}

[0023] Among them, inf represents the definition, v is the critical threshold corresponding to the composite health indicator;

[0024] The predicted state of the distribution network transformer is calculated as follows:

[0025]

[0026] in, represents the expected value, M represents the number of reconstructed features, Δt represents the time interval between two monitoring points, T' k is the predicted state at time k, K is the monitoring time, x k,m ' represents the reconstructed feature data of the mth reconstructed feature at time k.

[0027] In a preferred embodiment: in step 3, a prediction mean square error optimization objective function is constructed,

[0028]

[0029] Among them, A is the set of all weights; T' k and T k are the predicted value and actual value of the life of the distribution network transformer corresponding to the kth group of data respectively.

[0030] In a preferred embodiment: the reverse optimization adjustment of model parameters and feature weights based on the historical database includes: the widely used quasi-Newton method is used to solve the evaluation weights and critical thresholds:

[0031]

[0032] Where A' is the optimal weight set, v' is the optimal critical threshold, and v is the critical threshold corresponding to the composite health indicator; the optimal weight determined according to the above formula includes the time period weight a i , feature weight b corresponding to the reconstructed feature m The error term c corresponding to the reconstructed feature m ; Update the state monitoring model according to the optimal weight set and optimal critical threshold of reverse optimization.

[0033] The present invention also proposes an online monitoring device for the health status of a distribution network transformer, which is used to implement the above-mentioned online monitoring method for the health status of a distribution network transformer, comprising:

[0034] An upload module is used to upload the temperature time series data, gas time series data and health status of the target distribution network transformer within the historical time;

[0035] A determination module, for determining, for the first cycle in the initial sliding window cycle sequence and the characteristic data of different cycle weights, the temperature time series data and the gas time series data based on the sliding segmentation of the first cycle from the temperature time series data and the gas time series data in the historical time;

[0036] A generation module is used to generate a first-cycle two-dimensional feature data set according to the temperature time series data and gas time series data of the first-cycle sliding segmentation, and fuse the first-cycle two-dimensional feature data set to form the first-cycle feature data; fuse the first-cycle feature data to form composite health degradation data based on the random degradation model, establish an optimization objective function for minimizing the predicted mean square error, and form a monitoring model for the health status in the online database;

[0037] A calculation module, which calculates a first minimum mean square error between the first cycle characteristic time series data and the first cycle health state of the detection point from the first cycle characteristic time series data and the first cycle health state;

[0038] The update module is used to update the model parameters and cycle weights of the optimized monitoring model according to the characteristic data of each different cycle and the health status of different cycles, and use the optimized monitoring model for subsequent online monitoring;

[0039] The acquisition module is used to collect the temperature time series data and gas time series data of the target distribution network transformer in the current time period; the determination module uses the online data of the acquisition module as input and substitutes it into the optimized monitoring model to determine the health status of the target distribution network transformer.

[0040] Compared with the prior art, the present invention has the following beneficial technical effects:

[0041] According to the method and device for online monitoring of the health status of a distribution network transformer according to the embodiments of the present invention, in order to address the shortcoming that a large amount of monitoring data cannot be effectively utilized in the online monitoring of the status of a distribution network transformer, the monitoring time series data is periodically decomposed, and the weights of the periodic characteristics and the parameters of the detection model are updated and optimized through the combination of the monitoring model and historical data, so that the online monitoring method for the health status of the transformer is more reasonable and accurate, while reducing the impact of status monitoring, thereby greatly improving the utilization rate of the monitoring data. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of an embodiment of a method for online monitoring of the health status of a transformer in a distribution network provided by the present invention;

[0043] Figure 2 It is a schematic diagram of periodic decomposition of the online monitoring method for the health status of a transformer in a distribution network provided by the present invention;

[0044] Figure 3 It is a structural schematic diagram of a distribution network transformer health status monitoring device provided by the present invention. DETAILED DESCRIPTION

[0045] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings. The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings. ② Wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0046] In the description of the present invention, it is necessary to understand that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0047] In order to make the purpose, technical solution and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0048] Embodiment 1:

[0049] like Figure 1 FIG. 1 is a flow chart of an online monitoring method for the health status of a transformer in a distribution network according to the present invention. The online monitoring method for the health status of a transformer in a distribution network comprises the following steps:

[0050] Step 1: Establish a condition monitoring online database and a historical database based on the normal operating status, historical operating status, aging status and fault status of the distribution network transformer.

[0051] The features in the condition monitoring online database include gas online features and temperature online features. The gas online features include: hydrogen, methane, acetylene, ethylene, ethane, carbon monoxide, carbon dioxide, total hydrocarbons, and total hydrocarbons and values.

[0052] The historical database includes gas online characteristics, temperature online characteristics, and historical health status.

[0053] The health status includes dividing the operating status of the distribution network transformer into five levels according to the transformer status inspection and maintenance regulations of the State Grid Corporation of China, namely excellent, good, general, fault, and serious fault.

[0054] Step 2: Establish feature data with different period weights based on multi-source time series data in the historical database, and establish a monitoring model;

[0055] The characteristic data of different period weights include: for the multi-source time series data X={x1,x2,···,x M}, set the initial sliding window period sequence S = {s1,s2,···,s N}, based on S, the long time series feature data with a time series length of L is divided into n Then, these small one-dimensional data blocks are stacked into a length of s n , width is p n Two-dimensional time series feature data set, where L = s n ×p n .

[0056] The expanded multi-source time series data are integrated into a reconstructed feature data set by applying periodic weights, and the reconstructed feature data set is integrated into a single fusion indicator. A monitoring model of the fusion indicator and status results is established based on the random degradation model.

[0057] Figure 2 A schematic diagram of periodic decomposition provided in an embodiment of the present application is shown, such as Figure 2 As shown,

[0058] The reconstructed feature is expressed as:

[0059]

[0060] where a i is the cycle weight, y i Divide the feature data into small chunks. m ' represents the mth reconstructed feature.

[0061] A composite index is established after integrating multiple performance data. Considering the error and feature weight, the composite health index Z(t) after integrating multiple reconstructed feature data is obtained as follows:

[0062]

[0063] where x t,m ' represents the reconstructed feature data of the mth reconstructed feature at time t, b m represents the feature weight corresponding to the mth reconstructed feature; c m Represents the error term corresponding to the mth reconstructed feature.

[0064] The random degradation function of the degradation model is established. Considering the time-varying degradation process of the composite health index Z(t) of the linear Wiener process of standard Brownian motion, the random degradation function of the composite health index is constructed:

[0065] Z(t)=z1+θt+σD(t)

[0066] Among them, Z(t) is the composite health index; z1 is the initial state of the composite health index; D(t) is the standard Brownian motion, which reflects the randomness of time changes in the degradation process; θ is the drift coefficient, which reflects the trend of the composite health index over time; σ is the diffusion coefficient, which reflects the random uncertainty of the degradation process;

[0067] Calculating the predicted state value through the random degradation function model further includes:

[0068] The state value T' of the random degradation of the distribution network transformer is defined as follows:

[0069] T′=inf{t:Z(t)≥v|z1<v}

[0070] Among them, inf represents the definition, v is the critical threshold corresponding to the composite health indicator;

[0071] The predicted status of the distribution network transformers is derived by calculation, including:

[0072]

[0073] in, represents the expected value, M represents the number of reconstructed features, Δt represents the time interval between two monitoring points, T' k is the predicted state at time k, K is the monitoring time, x k,m ' represents the reconstructed feature data of the mth reconstructed feature at time k.

[0074] Step 3: Construct the optimization objective function of model parameters and feature weights based on the deviation value between the predicted state and the actual state;

[0075] The optimization objective function of model parameters and feature weights based on the deviation between the predicted state and the actual state includes constructing the optimization objective function of minimizing the predicted mean square error.

[0076]

[0077] Where A is the set of all weights; T' k and T k are the predicted value and actual value of the life of the distribution network transformer corresponding to the kth group of data respectively.

[0078] Step 4: Perform reverse optimization adjustment on the model parameters and feature weights based on the historical database, and monitor the health status in the online database according to the adjusted monitoring model;

[0079] Based on the historical database, the reverse optimization adjustment of model parameters and feature weights includes the widely used quasi-Newton method to solve the evaluation weights and critical thresholds. The calculation formula is:

[0080]

[0081] Where A' is the optimal weight set, v' is the optimal critical threshold; the optimal weight determined according to the above formula includes the time period weight a i , the feature weight b corresponding to the reconstructed feature m The error term c corresponding to the reconstructed feature m The condition monitoring model is updated according to the optimal weight set and optimal critical threshold of reverse optimization to improve the accuracy and effectiveness of health status monitoring of distribution network transformers.

[0082] Embodiment 2:

[0083] Based on the same technical concept, the present application also provides an online monitoring device for the health status of a distribution network transformer. Figure 3 The structure diagram of a device for online monitoring the health status of a transformer in a power distribution network provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, the device comprises:

[0084] An upload module is used to upload the temperature time series data, gas time series data and health status of the target distribution network transformer within a historical period; the temperature time series data includes the temperature time series data corresponding to each detection point on the target distribution network transformer; the gas time series data includes the gas time series data of hydrogen, methane, acetylene, ethylene, ethane, carbon monoxide, carbon dioxide, total hydrocarbons, and total hydrocarbons and values ​​in the target distribution network transformer; the health status includes the maintenance operation status of the transformer.

[0085] A determination module is used to determine, for the first period in the initial sliding window period sequence, the temperature time series data and the gas time series data based on the sliding segmentation of the first period from the historical temperature time series data and the historical gas time series data, with respect to the different period weight characteristic data.

[0086] A generation module is used to generate a first-cycle two-dimensional feature data set according to the first-cycle sliding segmentation data, and fuse the first-cycle two-dimensional feature data set to form the first-cycle feature data; based on the random degradation model, the first-cycle feature data is fused to form composite health degradation data, and the prediction mean square error optimization objective function is established to form an online state monitoring model to be calculated.

[0087] A calculation module, which calculates a first minimum mean square error between the first cycle characteristic time series data and the first cycle health state of the detection point from the first cycle characteristic time series data and the first cycle health state;

[0088] The updating module is used to update and optimize the model parameters and cycle weights of the online monitoring model according to each of the different cycle characteristic data and different cycle health states, and use the optimized online monitoring model for subsequent online monitoring.

[0089] The acquisition module is used to collect the temperature time series data and gas time series data of the target distribution network transformer in the current time period; the temperature time series data includes the temperature time series data corresponding to each detection point on the target distribution network transformer; the gas time series data includes the gas time series data of hydrogen, methane, acetylene, ethylene, ethane, carbon monoxide, carbon dioxide, total hydrocarbons, and total hydrocarbons and values ​​in the target distribution network transformer;

[0090] The determination module uses the online data of the acquisition module as input and substitutes it into the optimized online monitoring model to determine the health status of the target distribution network transformer.

[0091] According to the online monitoring method for the health status of a distribution network transformer in an embodiment of the present invention, in order to address the shortcoming that a large amount of monitoring data cannot be effectively utilized in the online monitoring of the status of a distribution network transformer, the monitoring time series data is periodically decomposed, and the weights of the periodic characteristics and the parameters of the detection model are updated and optimized through the combination of the monitoring model and historical data. This makes the online monitoring method for the health status of the transformer more reasonable and accurate, while reducing the impact of status monitoring and greatly improving the utilization rate of the monitoring data.

[0092] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0093] In the several embodiments provided in the present application, it should be understood that the disclosed methods and devices can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some communication interface, device or unit, which can be electrical, mechanical or other forms.

[0094] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to an embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any verified manner in one or more embodiments or examples.

[0095] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A method for online monitoring of the health status of a distribution network transformer, characterized in that: The following steps are involved: Step 1: Establish a status monitoring online database and a historical database based on the normal operating status, historical operating status, aging status and fault status of the distribution network transformer; Step 2: Based on the multi-source time series data in the historical database, establish feature data with different period weights, and use the period weights to integrate the expanded multi-source time series data into a reconstructed feature data set, integrate the reconstructed feature data set into a single fusion indicator, and establish a monitoring model for the fusion indicator and status results based on the random degradation model; Step 3: Construct an optimization objective function of the parameters and feature weights of the monitoring model based on the deviation value between the predicted state and the actual state; Step 4: Perform reverse optimization and adjustment on the parameters and feature weights of the monitoring model based on the historical database, and monitor the health status in the online database according to the optimized monitoring model.

2. The method for online monitoring of the health status of a distribution network transformer according to claim 1, characterized in that: The features in the state online database include: Gas online characteristics, temperature online characteristics; The gas online characteristics include: hydrogen, methane, acetylene, ethylene, ethane, carbon monoxide, carbon dioxide, total hydrocarbons, and total hydrocarbons and values; the historical database includes gas online characteristics, temperature online characteristics, and historical health status.

3. The method for online monitoring of the health status of a distribution network transformer according to claim 1, characterized in that: The characteristic data of different period weights include: for the characteristic data X in the historical database, M }, set the initial sliding window period sequence S = {s1,s2,···,s N }, based on S, the long time series feature data with a time series length of L is divided into n Then, these small one-dimensional data blocks are stacked into a length of s n , width is p n Two-dimensional time series feature data set, where L = s n ×p n .

4. The method for online monitoring of the health status of a distribution network transformer according to claim 3 is characterized in that: The reconstructed feature is expressed as: where a i is the cycle weight, y i is the feature data divided into small blocks of data, x m ' represents the mth reconstructed feature.

5. The method for online monitoring of the health status of a distribution network transformer according to claim 4, characterized in that: A composite index is established after integrating multiple performance data, and considering the error and feature weight, a composite health index Z(t) is obtained after integrating multiple reconstructed feature data: where x t,m ' represents the reconstructed feature data of the mth reconstructed feature at time t, b m Represents the feature weight corresponding to the mth reconstructed feature; c m Represents the error term corresponding to the mth reconstructed feature.

6. The method for online monitoring of the health status of a distribution network transformer according to claim 5, characterized in that: The random degradation function of the degradation model is established. Considering the time-varying degradation process of the composite health index Z(t) of the linear Wiener process of standard Brownian motion, the random degradation function of the composite health index is constructed: Z(t)=z1+θt+σD(t) Among them, Z(t) is the composite health index; z1 is the initial state of the composite health index; D(t) is the standard Brownian motion, θ is the drift coefficient, and σ is the diffusion coefficient.

7. The method for online monitoring of the health status of a distribution network transformer according to claim 6, characterized in that: The predicted state value is calculated by the random degradation function model, and the state value T' of the random degradation of the distribution network transformer is defined as follows: T′=inf{t:Z(t)≥v|z1<v} Among them, inf represents the definition, v is the critical threshold corresponding to the composite health indicator; The predicted state of the distribution network transformer is calculated as follows: in, represents the expected value, M represents the number of reconstructed features, Δt represents the time interval between two monitoring points, T' k is the predicted state at time k, K is the monitoring time, x k,m ' represents the reconstructed feature data of the mth reconstructed feature at time k.

8. The method for online monitoring of the health status of a distribution network transformer according to claim 7, characterized in that: In step 3, the optimization objective function of minimizing the predicted mean square error is constructed. Among them, A is the set of all weights; T' k and T k are the predicted value and actual value of the life of the distribution network transformer corresponding to the kth group of data respectively.

9. The method for online monitoring of the health status of a transformer in a distribution network according to claim 1, characterized in that: Reverse optimization adjustment of model parameters and feature weights based on historical databases includes: the widely used quasi-Newton method to solve the evaluation weights and critical thresholds: Where A' is the optimal weight set, v' is the optimal critical threshold, and v is the critical threshold corresponding to the composite health indicator; the optimal weight determined according to the above formula includes the time period weight a i , feature weight b corresponding to the reconstructed feature m The error term c corresponding to the reconstructed feature m ; Update the state monitoring model according to the optimal weight set and optimal critical threshold of reverse optimization.

10. An online monitoring device for the health status of a transformer in a distribution network, characterized in that: A method for online monitoring of the health status of a transformer in a distribution network according to any one of claims 1 to 9, comprising: An upload module is used to upload the temperature time series data, gas time series data and health status of the target distribution network transformer within the historical time; A determination module, for determining, for the first cycle in the initial sliding window cycle sequence and the characteristic data of different cycle weights, the temperature time series data and the gas time series data based on the sliding segmentation of the first cycle from the temperature time series data and the gas time series data in the historical time; A generation module is used to generate a first-cycle two-dimensional feature data set according to the temperature time series data and gas time series data of the first-cycle sliding segmentation, and fuse the first-cycle two-dimensional feature data set to form the first-cycle feature data; fuse the first-cycle feature data to form composite health degradation data based on the random degradation model, establish an optimization objective function for minimizing the predicted mean square error, and form a monitoring model for the health status in the online database; A calculation module, which calculates a first minimum mean square error between the first cycle characteristic time series data and the first cycle health state of the detection point from the first cycle characteristic time series data and the first cycle health state; The update module is used to update the model parameters and cycle weights of the optimized monitoring model according to the characteristic data of each different cycle and the health status of different cycles, and use the optimized monitoring model for subsequent online monitoring; The acquisition module is used to collect the temperature time series data and gas time series data of the target distribution network transformer in the current time period; the determination module uses the online data of the acquisition module as input and substitutes it into the optimized monitoring model to determine the health status of the target distribution network transformer.

Images