Insulation on-line monitoring device and method for tower electrical equipment

Abstract

The invention discloses a tower electrical equipment insulation online monitoring device and method, and relates to the technical field of electrical equipment fault monitoring, and the method comprises the steps: collecting environmental parameters and insulation resistance data of a tower, building a nonlinear compensation relation between the environmental parameters and insulation resistance, and dynamically compensating an insulation resistance value; based on the insulation resistance value and the environment parameters after dynamic compensation, an insulation state data set is constructed through a time sequence decomposition and high-dimensional data embedding method; and constructing an insulation aging prediction model, inputting the insulation state data set into the insulation aging prediction model, analyzing the future aging trend of the insulation performance of the tower equipment, and generating a detection report according to the future aging trend of the insulation performance of the tower equipment. According to the method, the insulation aging prediction model constructed by the LSTM is utilized to capture the long-term dependency relationship of the insulation performance change, so that the future insulation resistance value and the aging trend are accurately predicted, the maintenance suggestion report is generated, and powerful support is provided for safe operation of a power system.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical equipment fault monitoring, and particularly to an on-line insulation monitoring device and method for pole and tower electrical equipment. Background Art

[0002] As an important part of the power transmission system, the insulation performance of pole and tower electrical equipment is directly related to the safe operation of the power grid. With the rapid development of modern power systems, the operating environment of pole and tower electrical equipment has become increasingly complex, and its insulation performance is affected by a variety of factors, such as environmental temperature, humidity, atmospheric pressure, and equipment surface contamination. These external factors can cause fluctuations in the insulation resistance value, thereby affecting the insulation state of the equipment. Traditional insulation performance monitoring methods usually adopt regular manual detection or simple static on-line monitoring methods, which cannot reflect the dynamic change characteristics of insulation performance in complex environments in real time. In addition, current insulation monitoring technologies are mostly based on single parameters and lack in-depth analysis of the complex relationship between environmental parameters and equipment insulation performance, so they cannot effectively predict the insulation aging trend and potential failure risks of equipment.

[0003] Although the existing technologies have made certain progress in improving the insulation performance monitoring of poles and towers, there are still significant deficiencies. First, traditional insulation monitoring methods do not fully consider the non-linear impact of dynamic changes in environmental parameters on insulation performance, resulting in a reduction in the accuracy and reliability of insulation resistance measurement results. Second, existing methods lack the ability to accurately predict the aging trend of equipment insulation performance. Insulation aging prediction requires combining a large amount of historical data and real-time monitoring data, analyzing time series characteristics, and capturing the long-term dependence relationship of insulation performance changes. Traditional methods usually rely on linear analysis or empirical formulas and are difficult to adapt to complex non-linear aging processes. These deficiencies limit the application effect of existing technologies in on-line insulation monitoring and prediction of pole and tower electrical equipment. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an on-line insulation monitoring method for pole and tower electrical equipment to solve the problems in the prior art that the non-linear impact of environmental parameters on insulation performance cannot be dynamically compensated and the accurate prediction of the future aging trend of insulation performance is lacking.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides an online monitoring method for insulation of pole tower electrical equipment, which includes collecting environmental parameters and insulation resistance data of the pole tower, establishing a nonlinear compensation relationship between the environmental parameters and the insulation resistance, and dynamically compensating the insulation resistance value; based on the insulation resistance value and environmental parameters after dynamic compensation, constructing an insulation status data set through time series decomposition and high-dimensional data embedding method; constructing an insulation aging prediction model, inputting the insulation status data set into the insulation aging prediction model, analyzing the future aging trend of the insulation performance of the pole tower equipment, and generating a detection report based on the future aging trend of the insulation performance of the pole tower equipment.

[0007] As a preferred solution of the method for online monitoring insulation of tower electrical equipment of the present invention, wherein: the environmental parameters include temperature, humidity, atmospheric pressure and equipment surface contamination; The insulation resistance data includes transient overvoltage data and transient overcurrent data.

[0008] As a preferred solution of the method for online monitoring insulation of electrical equipment in towers of the present invention, the nonlinear compensation relationship between environmental parameters and insulation resistance is established to dynamically compensate insulation resistance value, and the specific steps are as follows: Clean the collected environmental parameters and insulation resistance data, remove abnormal data and noise, and normalize the environmental parameters and insulation resistance data of different dimensions; Identify the impact of environmental parameters on insulation resistance through statistical analysis of historical data; According to the influence of environmental parameters on insulation resistance, the ANN algorithm is used in combination with nonlinear activation function to establish the nonlinear compensation relationship between environmental parameters and insulation resistance. Based on the cleaned and normalized environmental parameters and insulation resistance data, the periodicity, coupling and exponential decay relationship of the environmental parameters are quantified through the nonlinear compensation relationship between the environmental parameters and the insulation resistance, and the dynamic compensation value of the insulation resistance is predicted. The expression is: ; in, Indicates the ambient temperature, Indicates the ambient humidity. Indicates the degree of contamination on the surface of the equipment. represents the ambient atmospheric pressure, is the temperature weight coefficient, is the weight coefficient of humidity, is the weight coefficient of equipment surface contamination, is the relative contribution of temperature and humidity coupling to the compensation value, is the dynamic compensation value of insulation resistance; Add the dynamic compensation value of the insulation resistance to the measured actual insulation resistance value to generate the insulation resistance value after dynamic compensation.

[0009] As a preferred solution of the on-line insulation monitoring method for pole and tower electrical equipment according to the present invention, wherein: based on the insulation resistance value after dynamic compensation and environmental parameters, an insulation state data set is constructed by time series decomposition and high-dimensional data embedding method. The specific steps are as follows. Use STL to decompose the insulation resistance value into a long-term trend term, a periodic term, and a residual term. Use DTW to synchronously align the long-term trend term, periodic term, and residual term of the insulation resistance value with the environmental parameters at the corresponding moments, and establish the embedding relationship between the long-term trend term, periodic term, residual term, and environmental parameters. According to the embedding relationship between the long-term trend term, periodic term, residual term, and environmental parameters, embed the environmental parameters as high-dimensional data into the insulation resistance value after dynamic compensation corresponding to different time points to construct an insulation state data set.

[0010] As a preferred solution of the on-line insulation monitoring method for pole and tower electrical equipment according to the present invention, wherein: the construction of the insulation aging prediction model is as follows. Use LSTM as the framework of the insulation aging prediction model. Based on the framework of the insulation aging prediction model, construct an input layer, a time series processing layer, a fully connected layer, and an output layer. Input layer, receive the insulation state data set, extract time series features through a sliding window method, and generate a time series input feature matrix. Time series processing layer, receive the time series feature matrix of the input layer, use LSTM units to process the time series data, capture the long-term dependence relationship of insulation aging through memory gates and forget gates, and generate a hidden state representation of the time series. Fully connected layer, receive the hidden state representation generated by the time series processing layer, perform a linear transformation on the hidden state using a weight matrix and a bias vector, and perform activation processing through a ReLU activation function to extract the deep feature representation of insulation aging. Output layer, receive the deep feature representation generated by the fully connected layer, perform a linear transformation, and combine with a Sigmoid activation function to map the prediction result to the predicted value of the future insulation resistance. Combine the input layer, time series processing layer, fully connected layer, and output layer to form an insulation aging prediction model.

[0011] As a preferred embodiment of the on-line insulation monitoring method for pole and tower electrical equipment of the present invention, the steps of inputting the insulation status data set into the insulation aging prediction model to analyze the future aging trend of the insulation performance of the pole and tower equipment and based on the future aging trend of the insulation performance of the pole and tower equipment are as follows: Input the insulation status data set into the insulation aging prediction model to predict the future insulation resistance value, and the expression is: ; Wherein, represents the insulation resistance value at the future time point , is the Sigmoid activation function, is the current time point, is the length of the future time period, is the total number of data in the insulation status data set, is the index of the data in the insulation status data set, is the th data in the insulation status data set, is the weight coefficient of the th data in the insulation status data set, is the weight matrix of the time series processing layer, is the bias vector of the time series processing layer, is the smoothing factor, is the bias vector of the fully connected layer, is the weight matrix of the output layer, is the hidden state representation generated by the time series processing layer through the LSTM unit; Define the aging threshold based on the historical insulation resistance data and the equipment operation records; When , it is considered that the future aging trend of the insulation performance of the pole and tower equipment is serious; When , it is considered that the future aging trend of the insulation performance of the pole and tower equipment is slight.

[0012] As a preferred embodiment of the on-line insulation monitoring method for pole and tower electrical equipment of the present invention, the test report includes the equipment operation records, the aging threshold , the insulation resistance value at the future time point , the analysis result of the future aging trend of the insulation performance of the pole and tower equipment, and the maintenance suggestions.

[0013] In a second aspect, the present invention provides an on-line insulation monitoring device for pole and tower electrical equipment, including a data acquisition module, a data set construction module, and an aging trend prediction module; the data acquisition module is used to collect the environmental parameters and insulation resistance data of the pole and tower, establish a non-linear compensation relationship between the environmental parameters and the insulation resistance, and dynamically compensate the insulation resistance value; the data set construction module is used to construct an insulation state data set based on the dynamically compensated insulation resistance value and environmental parameters through time series decomposition and high-dimensional data embedding method; the aging trend prediction module is used to construct an insulation aging prediction model, input the insulation state data set into the insulation aging prediction model, analyze the future aging trend of the insulation performance of the pole and tower equipment, and generate a detection report according to the future aging trend of the insulation performance of the pole and tower equipment.

[0014] In a third aspect, the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and: when the computer program is executed by the processor, any step of the on-line insulation monitoring method for pole and tower electrical equipment as described in the first aspect of the present invention is implemented.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and: when the computer program is executed by the processor, any step of the on-line insulation monitoring method for pole and tower electrical equipment as described in the first aspect of the present invention is implemented.

[0016] The beneficial effects of the present invention are as follows: By establishing a non-linear compensation relationship between the environmental parameters and the insulation resistance, the dynamic adjustment of the insulation resistance value is realized, ensuring the accuracy and reliability of the measurement results in a complex environment. Further, the insulation aging prediction model constructed by using LSTM captures the long-term dependence relationship of the insulation performance change, so as to accurately predict the future insulation resistance value and aging trend, and generate a maintenance suggestion report, providing strong support for the safe operation of the power system, while optimizing the maintenance cost and efficiency. Description of the Drawings

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings according to these drawings without creative efforts.

[0018] Figure 1 It is a flowchart of the on-line insulation monitoring method for pole and tower electrical equipment in Embodiment 1.

[0019] Figure 2 It is a schematic diagram of the on-line insulation monitoring device for pole and tower electrical equipment in Embodiment 1. Detailed Embodiments

[0020] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following provides a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings of the specification.

[0021] In the following description, many specific details are set forth to facilitate a thorough understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art can make similar generalizations without departing from the spirit of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0022] Secondly, the so-called "one embodiment" or "embodiment" herein refers to specific features, structures, or characteristics that may be included in at least one implementation manner of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments.

[0023] Example 1, referring to Figure 1 and Figure 2 , is the first embodiment of the present invention. This embodiment provides a method for on-line monitoring of the insulation of pole and tower electrical equipment, including the following steps: S1. Collect the environmental parameters and insulation resistance data of the pole and tower, establish a non-linear compensation relationship between the environmental parameters and the insulation resistance, and dynamically compensate the insulation resistance value.

[0024] The environmental parameters include temperature, humidity, atmospheric pressure, and the pollution degree of the equipment surface; The insulation resistance data includes transient overvoltage data and transient overcurrent data.

[0025] It should be noted that for the environmental parameters, corresponding sensors are deployed around the pole and tower and on the equipment surface to continuously monitor the changes in temperature, humidity, atmospheric pressure, and the pollution degree of the equipment surface. For the insulation resistance data, detection instruments installed at key parts of the electrical equipment accurately record the transient overvoltage data and transient overcurrent data without affecting the normal operation of the equipment. These data are then transmitted to the data analysis platform to provide a basis for subsequent non-linear compensation and aging prediction.

[0026] Clean the collected environmental parameters and insulation resistance data, remove abnormal data and noise, and normalize the environmental parameters and insulation resistance data with different dimensions; Furthermore, during the data cleaning process, the original data obtained from temperature, humidity, atmospheric pressure, and equipment surface contamination sensors, as well as transient overvoltage data and transient overcurrent data detection instruments, is first reviewed to identify and remove outliers and noise. For example, data points that significantly exceed the reasonable range or do not conform to the time series continuity will be marked as outliers and corrected or removed through interpolation or other appropriate algorithms. In addition, the consistency and integrity of the data will be checked to ensure that no missing values affect subsequent analysis. This process ensures that the dataset used for analysis is of high quality and reliable.

[0027] Furthermore, the normalization process aims to transform temperature, humidity, atmospheric pressure, and equipment surface contamination with different dimensions, as well as transient overvoltage data and transient overcurrent data, to a common scale for easy comparison and modeling. For example, through the min-max scaling method, all environmental parameters and insulation resistance data will be linearly transformed into the range between 0 and 1; or the Z-score normalization is adopted to make the average value of each data type become 0 and the standard deviation become 1.

[0028] Identify the impact of environmental parameters on insulation resistance through statistical analysis of historical data; The specific process is that identifying the impact of environmental parameters on insulation resistance through statistical analysis of historical data involves in-depth research on past temperature, humidity, atmospheric pressure, and equipment surface contamination, as well as transient overvoltage data and transient overcurrent data. For example, comparing the temperature records accumulated over the years with the corresponding changes in insulation resistance to find the correlation pattern between the two. At the same time, analyze how the fluctuations of humidity, atmospheric pressure, and equipment surface contamination in different seasons or weather conditions are mapped to the changes in insulation resistance. For transient overvoltage data and transient overcurrent data, through regression analysis, the specific impact degree of various environmental factors on insulation resistance can be quantified.

[0029] According to the impact of environmental parameters on insulation resistance, adopt the ANN algorithm and combine it with a non-linear activation function to establish a non-linear compensation relationship between environmental parameters and insulation resistance; Furthermore, first, based on the statistical analysis results of historical data, identify the specific influence patterns of temperature, humidity, atmospheric pressure, equipment surface contamination, as well as transient overvoltage data and transient overcurrent data on insulation resistance. Next, using the ANN algorithm, take these influence patterns as input features and process them through a multi-layer neural network structure, where each layer uses a non-linear activation function to simulate complex non-linear relationships. For example, for the input environmental conditions such as temperature and humidity, the ANN algorithm will go through a series of hidden layer processes, and each node applies a non-linear activation function to capture the subtle differences in the insulation resistance response when environmental factors change. The final output is a dynamic compensation value, which reflects the expected change in insulation resistance under the current environmental conditions, thereby achieving precise adjustment of the measured insulation resistance value. This process ensures that the measurement of insulation resistance can be appropriately non-linearly compensated under different environments.

[0030] Based on the cleaned and normalized environmental parameters and insulation resistance data, through the non-linear compensation relationship between environmental parameters and insulation resistance, quantify the periodicity, coupling, and exponential decay relationships of environmental parameters, and predict the dynamic compensation value of insulation resistance. The expression is: ; Where, represents the environmental temperature, represents the environmental humidity, represents the contamination degree of the equipment surface, represents the environmental atmospheric pressure, is the weight coefficient of temperature, is the weight coefficient of humidity, is the weight coefficient of the contamination degree of the equipment surface, is the relative contribution degree of the coupling of temperature and humidity to the compensation value, is the dynamic compensation value of insulation resistance; It should be noted that after obtaining and processing temperature, humidity, atmospheric pressure, equipment surface contamination, as well as transient overvoltage data and transient overcurrent data, consider the periodic change of temperature, the square effect of humidity, the logarithmic growth of equipment surface contamination, and the coupling effect between temperature and humidity, and combine the exponential decay characteristics of atmospheric pressure to comprehensively evaluate the influence of these factors on insulation resistance. For example, when the environmental temperature shows seasonal periodic fluctuations, or the equipment surface contamination accumulates gradually, this expression can accurately reflect the specific influence of these changes on insulation resistance, thereby providing an accurate dynamic compensation value ΔR for adjusting the actually measured insulation resistance value to ensure that the monitoring results are closer to the actual situation.

[0031] Add the dynamic compensation value of insulation resistance to the actually measured insulation resistance value to generate the dynamically compensated insulation resistance value.

[0032] S2. Based on the dynamically compensated insulation resistance value and environmental parameters, construct an insulation state dataset through time series decomposition and high-dimensional data embedding method.

[0033] Use STL to decompose the insulation resistance value into a long-term trend term, a periodic term, and a residual term; Further, first, collect the time series data of the insulation resistance value, which is a set of insulation resistance values recorded in chronological order. Then, select appropriate time window parameters for separating the long-term trend term and the periodic term. Next, through STL decomposition, decompose the insulation resistance value sequence into three parts: the long-term trend term, the periodic term, and the residual term. The long-term trend term represents the overall trend of the insulation resistance value changing over time, the periodic term represents the periodic fluctuation law of the insulation resistance value, and the residual term reflects the random fluctuations or noises in the insulation resistance value that cannot be explained by the trend and the period. For example, for a sequence of insulation resistance values recorded in hours, the long-term trend term can be extracted by setting a window containing the data of the whole year, and at the same time, the daily or weekly periodic changes can be extracted using the periodic parameters, and finally, the decomposition results of the long-term trend term, the periodic term, and the residual term are obtained.

[0034] Use DTW to synchronously align the long-term trend term, the periodic term, and the residual term of the insulation resistance value with the environmental parameters at the corresponding moments, and establish the embedding relationship between the long-term trend term, the periodic term, and the residual term and the environmental parameters; Further, first, extract the time series of the long-term trend term, the periodic term, and the residual term of the insulation resistance value, and the time series of the environmental parameters at the corresponding moments of the insulation resistance value, such as temperature, humidity, etc. Since there may be asynchrony in the time scale or dynamic changes between the different time series of the insulation resistance value and the time series of the environmental parameters, adopt the dynamic time warping algorithm to achieve the alignment in time between each decomposition term of the insulation resistance value and the environmental parameters by calculating the optimal matching path between the time series. The specific steps include: calculating the distance matrix for each decomposition term of the insulation resistance value and the time series of the environmental parameters respectively, using the DTW algorithm to find the optimal matching path between the two to minimize the cumulative distance error of the corresponding relationship between the two time series; then, through the alignment result, embed the environmental parameters into the long-term trend term, the periodic term, and the residual term of the insulation resistance value, and establish the embedding relationship between each decomposition term of the insulation resistance value and the environmental parameters. For example, for a long-term trend term of a section of insulation resistance value and a humidity time series, after aligning the time steps of the two through the DTW algorithm, the relationship between the long-term trend of the insulation resistance value and the change of humidity can be analyzed synchronously, thus providing support for further dynamic compensation.

[0035] According to the embedding relationship between the long-term trend term, the periodic term, the residual term and the environmental parameters, the environmental parameters are embedded as high-dimensional data into the dynamically compensated insulation resistance values corresponding to different time points, and an insulation state data set is constructed.

[0036] Further, first, obtain the time series data of the insulation resistance value, and decompose it into a long-term trend term, a periodic term and a residual term by a decomposition method. At the same time, extract the time series of environmental parameters corresponding to each time point, such as temperature, humidity and pressure. Then, using the embedding method, take the environmental parameters as high-dimensional features and establish a corresponding relationship with the long-term trend term, the periodic term and the residual term. Through correlation analysis and dynamic compensation algorithms, the influence of the change of environmental parameters is reflected in the dynamic compensation of the insulation resistance value, so as to generate the dynamically compensated insulation resistance value. Then, construct a high-dimensional data record by combining the dynamically compensated insulation resistance value at each time point with the corresponding time label, long-term trend term, periodic term, residual term and the embedded environmental parameters. Finally, integrate these high-dimensional data records in chronological order to form an insulation state data set. For example, for the insulation resistance value recorded in a certain hour, extract its dynamically compensated value, and jointly form a data record with the environmental parameters such as temperature, humidity and pressure at this moment, as well as the long-term trend term, periodic term and residual term. After accumulating all time point records in turn, a complete insulation state data set including time, dynamically compensated insulation resistance value, environmental parameters and decomposition terms is generated.

[0037] S3. Construct an insulation aging prediction model, input the insulation state data set into the insulation aging prediction model, analyze the future aging trend of the insulation performance of the pole and tower equipment, and generate a detection report according to the future aging trend of the insulation performance of the pole and tower equipment.

[0038] Use LSTM as the framework of the insulation aging prediction model. Based on the framework of the insulation aging prediction model, construct an input layer, a time series processing layer, a fully connected layer and an output layer; The input layer receives the insulation state data set, extracts time series features by the sliding window method, and generates a time series input feature matrix; The specific process is as follows. After receiving the insulation state data set containing temperature, humidity, atmospheric pressure, equipment surface contamination degree, transient overvoltage data and transient overcurrent data, use the sliding window technology to process these data. For example, set a time window with a fixed length to slide along the time axis. Each time it slides, a continuous data sequence within this window will be captured. For each window position, extract the time series features therein and arrange these features in matrix form to form a time series input feature matrix.

[0039] The time series processing layer receives the time series feature matrix from the input layer, processes the time series data using LSTM cells, captures the long-term dependencies of insulation aging through the memory gate and forget gate, and generates the hidden state representation of the time series. Specifically, the time series processing layer receives the time series feature matrix from the input layer and deeply processes the time series data using LSTM cells. In this process, first, each row in the time series feature matrix represents the insulation state data set at a time point, including temperature, humidity, atmospheric pressure, device surface contamination degree, as well as transient overvoltage data and transient overcurrent data. These data are fed into the LSTM cells in sequence. At each time step, the LSTM cell determines which information should be retained or forgotten through its internal memory gate and forget gate mechanisms. For example, when processing the data of a new time point, the forget gate evaluates the previously saved state information and decides which historical information is no longer relevant, thus allowing the insulation aging prediction model to focus on capturing long-term dependencies. At the same time, the memory gate is responsible for updating the cell state, selectively adding new information to the existing state to ensure that important long-term trends are not overlooked. Subsequently, the LSTM cell generates the hidden state representation of the current time point based on the updated cell state. This hidden state not only contains the insulation state features of the current time point but also incorporates important patterns and trends inherited from past time points, especially those that can reflect the insulation aging process. After the hidden state representation is further processed by a linear transformation and an activation function, it is passed to the subsequent fully connected layer.

[0040] The fully connected layer receives the hidden state representation generated by the time series processing layer, performs a linear transformation on the hidden state using a weight matrix and a bias vector, and performs an activation process through the ReLU activation function to extract the deep feature representation of insulation aging. Specifically, the fully connected layer receives the hidden state representation generated by the time series processing layer and first performs a linear transformation on these hidden states using a weight matrix and a bias vector. Specifically, each hidden state representation is regarded as a vector, multiplied by the weight matrix, and added with the bias vector to achieve the mapping from the hidden space to the new feature space. For example, if the hidden state representation is a high-dimensional vector, then by multiplying it with the weight matrix, this vector can be converted into a new vector with a different dimension, and this process can capture the complex relationships between hidden states. Next, the output after the linear transformation is non-linearly processed through the ReLU (Rectified Linear Unit) activation function. The ReLU function sets all negative values to zero while keeping positive values unchanged, and this step helps to introduce non-linearity, enabling the insulation aging prediction model to learn more complex patterns. After being activated by ReLU, the resulting output is the deep feature representation of insulation aging.

[0041] The output layer receives the deep feature representations generated by the fully connected layer, and through linear transformation and in combination with the Sigmoid activation function, maps the prediction result to the predicted value of the future insulation resistance. Specifically, the deep feature representation is multiplied by the weight matrix of the output layer and added with the bias vector to achieve the mapping from the deep feature space to the predicted value space. For example, if the deep feature representation is a high-dimensional vector, then this vector will be transformed into an output of a single value or a small number of values through linear transformation. Next, to ensure that the prediction result is within a reasonable range of insulation resistance values, the output layer uses the Sigmoid activation function to process the result of the linear transformation. The Sigmoid function compresses the input in any range into an interval between 0 and 1, and then through appropriate scaling and translation operations, maps this interval to the expected range of the actual insulation resistance value. Finally, the output after being processed by the Sigmoid activation function is the predicted value of the future insulation resistance.

[0042] Combining the input layer, the time series processing layer, the fully connected layer and the output layer constitutes an insulation aging prediction model.

[0043] Input the insulation state data set into the insulation aging prediction model to predict the future insulation resistance value. The expression is: ; where represents the insulation resistance value at the future time point , is the Sigmoid activation function, is the current time point, is the length of the future time period, is the total number of data in the insulation state data set, is the index of the data in the insulation state data set, is the th data in the insulation state data set, is the th weight coefficient of the data in the insulation state data set, is the weight matrix of the time series processing layer, is the bias vector of the time series processing layer, is the smoothing factor, is the bias vector of the fully connected layer, is the weight matrix of the output layer, is the hidden state representation generated by the time series processing layer through the LSTM unit; It should be noted that each record in the insulation status dataset, including temperature, humidity, atmospheric pressure, surface contamination degree of the equipment, as well as transient overvoltage data and transient overcurrent data, is processed one by one. For the data at each time point , according to its index in the insulation status dataset and the corresponding weight coefficient , the weight matrix of the time series processing layer and the bias vector are applied for linear transformation, and are normalized through the activation function. Then, the result of this transformation is divided by the norm of this data point plus a smoothing factor to ensure numerical stability. All these processed data points are then weighted and summed, and added to the bias vector of the fully connected layer. Next, this sum is passed into a denominator, which consists of 1 plus an exponentially decaying term, where the exponentially decaying term is based on the weight matrix of the output layer and the time series hidden state representation processed by the ReLU activation function . Finally, the entire expression is mapped to a reasonable range of insulation resistance values through the Sigmoid activation function , thereby generating the predicted value of the insulation resistance at the future time point ; Based on the historical insulation resistance data and the equipment operation records, the aging threshold is defined; When , it is considered that the future aging trend of the insulation performance of the pole tower equipment is serious; When , it is considered that the future aging trend of the insulation performance of the pole tower equipment is slight.

[0044] The detection report includes the equipment operation records, the aging threshold , the insulation resistance value at the future time point , the analysis result of the future aging trend of the insulation performance of the pole tower equipment, and the maintenance suggestions.

[0045] This embodiment also provides an on-line insulation monitoring device for pole tower electrical equipment, including: a data acquisition module, a dataset construction module, and an aging trend prediction module; The data acquisition module is used to collect the environmental parameters and insulation resistance data of the pole tower, establish a non-linear compensation relationship between the environmental parameters and the insulation resistance, and dynamically compensate the insulation resistance value; The dataset construction module is used to construct an insulation status dataset based on the dynamically compensated insulation resistance value and environmental parameters through time series decomposition and high-dimensional data embedding method An aging trend prediction module, which is used to construct an insulation aging prediction model, input an insulation state data set into the insulation aging prediction model, analyze the future aging trend of the insulation performance of pole and tower equipment, and generate a detection report according to the future aging trend of the insulation performance of the pole and tower equipment.

[0046] This embodiment also provides a computer device, which is applicable to the case of the on-line insulation monitoring method for pole and tower electrical equipment, and includes: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the on-line insulation monitoring method for pole and tower electrical equipment as proposed in the above embodiment.

[0047] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a carrier network, NFC (Near Field Communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a button, a trackball or a touchpad provided on the shell of the computer device, or an external keyboard, a touchpad or a mouse, etc.

[0048] This embodiment also provides a storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the on-line insulation monitoring method for pole and tower electrical equipment as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM for short), Electrically Erasable Programmable Read-Only Memory (EEPROM for short), Erasable Programmable Read Only Memory (EPROM for short), Programmable Red-Only Memory (PROM for short), Read-Only Memory (ROM for short), magnetic memory, flash memory, a magnetic disk or an optical disc.

[0049] In summary, the present invention realizes the dynamic adjustment of the insulation resistance value by establishing a non-linear compensation relationship between the environmental parameters and the insulation resistance, ensuring the accuracy and reliability of the measurement results in complex environments. Further, the insulation aging prediction model constructed by using LSTM captures the long-term dependence relationship of the insulation performance change, thereby accurately predicting the future insulation resistance value and aging trend, and generating a maintenance recommendation report, which provides strong support for the safe operation of the power system, while optimizing the maintenance cost and efficiency.

[0050] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for online monitoring insulation of a tower electrical equipment, characterized in that: include, Collect the environmental parameters and insulation resistance data of the tower, establish the nonlinear compensation relationship between the environmental parameters and the insulation resistance, and dynamically compensate the insulation resistance value; Based on the insulation resistance value and environmental parameters after dynamic compensation, the insulation status dataset is constructed through time series decomposition and high-dimensional data embedding method; Construct an insulation aging prediction model, input the insulation status data set into the insulation aging prediction model, analyze the future aging trend of the insulation performance of the tower equipment, and generate a test report based on the future aging trend of the insulation performance of the tower equipment.

2. The method for online monitoring insulation of a tower electrical equipment according to claim 1, characterized in that: The environmental parameters include temperature, humidity, atmospheric pressure and equipment surface contamination; The insulation resistance data includes transient overvoltage data and transient overcurrent data.

3. The method for online monitoring insulation of a tower electrical equipment according to claim 2, characterized in that: The nonlinear compensation relationship between the environmental parameters and the insulation resistance is established to dynamically compensate the insulation resistance value. The specific steps are as follows: Clean the collected environmental parameters and insulation resistance data, remove abnormal data and noise, and normalize the environmental parameters and insulation resistance data of different dimensions; Identify the impact of environmental parameters on insulation resistance through statistical analysis of historical data; According to the influence of environmental parameters on insulation resistance, the ANN algorithm is used in combination with nonlinear activation function to establish the nonlinear compensation relationship between environmental parameters and insulation resistance. Based on the cleaned and normalized environmental parameters and insulation resistance data, the periodicity, coupling and exponential decay relationship of the environmental parameters are quantified through the nonlinear compensation relationship between the environmental parameters and the insulation resistance, and the dynamic compensation value of the insulation resistance is predicted. The expression is: ； in, Indicates the ambient temperature, Indicates the ambient humidity. Indicates the degree of contamination on the surface of the equipment. represents the ambient atmospheric pressure, is the temperature weight coefficient, is the weight coefficient of humidity, is the weight coefficient of equipment surface contamination, is the relative contribution of temperature and humidity coupling to the compensation value, is the dynamic compensation value of insulation resistance; The dynamic compensation value of the insulation resistance is added to the actual insulation resistance value measured to generate the insulation resistance value after dynamic compensation.

4. The method for online monitoring insulation of a tower electrical equipment according to claim 3, characterized in that: The insulation resistance value and environmental parameters after dynamic compensation are used to construct an insulation status data set through time series decomposition and high-dimensional data embedding. The specific steps are as follows: Use STL to decompose the insulation resistance value into long-term trend term, periodic term and residual term; Use DTW to synchronize the long-term trend term, periodic term, and residual term of the insulation resistance value with the environmental parameters at the corresponding time, and establish an embedded relationship between the long-term trend term, periodic term, residual term, and environmental parameters; According to the embedding relationship between long-term trend terms, periodic terms, residual terms and environmental parameters, the environmental parameters are embedded as high-dimensional data into the insulation resistance values ​​after dynamic compensation corresponding to different time points to construct an insulation status dataset.

5. The method for online monitoring insulation of a tower electrical equipment according to claim 4, characterized in that: The specific steps of constructing the insulation aging prediction model are as follows: Use LSTM as the framework of the insulation aging prediction model. Based on the framework of the insulation aging prediction model, construct the input layer, time series processing layer, fully connected layer, and output layer. The input layer receives the insulation status dataset, extracts the time series features through the sliding window method, and generates the time series input feature matrix; The time series processing layer receives the time series feature matrix of the input layer, processes the time series data using LSTM units, captures the long-term dependency of insulation aging through memory gates and forget gates, and generates hidden state representations of the time series; The fully connected layer receives the hidden state representation generated by the time series processing layer, uses the weight matrix and bias vector to perform linear transformation on the hidden state, and activates it through the ReLU activation function to extract the deep feature representation of insulation aging; The output layer receives the deep feature representation generated by the fully connected layer, and maps the prediction results to the predicted values ​​of future insulation resistance through linear transformation and sigmoid activation function; The input layer, time series processing layer, fully connected layer and output layer are combined to form the insulation aging prediction model.

6. The method for online monitoring insulation of a tower electrical equipment according to claim 5, characterized in that: The insulation state data set is input into the insulation aging prediction model to analyze the future aging trend of the insulation performance of the tower equipment, and according to the future aging trend of the insulation performance of the tower equipment, the specific steps are as follows: The insulation status data set is input into the insulation aging prediction model to predict the future insulation resistance value. The expression is: ； in, Indicates a future time point The insulation resistance value, is the Sigmoid activation function, is the current time point, is the length of the future time period, is the total number of data in the insulation status data set, is the index of the data in the insulation status dataset, is the first data, It is the insulation status data set. The weight coefficient of each data, is the weight matrix of the time series processing layer, is the bias vector of the time series processing layer, is the smoothing factor, is the bias vector of the fully connected layer, is the weight matrix of the output layer, It is the hidden state representation generated by the time series processing layer through the LSTM unit; Define aging thresholds based on historical insulation resistance data and equipment operation records ; when When , it is considered that the insulation performance of the tower equipment will have a serious aging trend in the future; when When , it is considered that the future aging trend of the insulation performance of the tower equipment is slight.

7. The method for online monitoring insulation of a tower electrical equipment according to claim 6, characterized in that: The test report includes equipment operation records, aging thresholds , Future time point Insulation resistance value , future aging trend analysis results of the insulation performance of tower equipment and maintenance recommendations.

8. An online monitoring device for insulation of a tower electrical equipment, based on the online monitoring method for insulation of a tower electrical equipment according to any one of claims 1 to 7, characterized in that: Including, data acquisition module, data set construction module and aging trend prediction module; The data acquisition module is used to collect the environmental parameters and insulation resistance data of the tower, establish the nonlinear compensation relationship between the environmental parameters and the insulation resistance, and dynamically compensate the insulation resistance value; Dataset construction module, used to construct insulation status dataset based on insulation resistance value and environmental parameters after dynamic compensation through time series decomposition and high-dimensional data embedding method The aging trend prediction module is used to build an insulation aging prediction model, input the insulation status data set into the insulation aging prediction model, analyze the future aging trend of the insulation performance of the tower equipment, and generate a test report based on the future aging trend of the insulation performance of the tower equipment.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for online monitoring insulation of tower electrical equipment according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for online monitoring insulation of tower electrical equipment according to any one of claims 1 to 7 are implemented.

Images