Bus duct temperature and humidity anomaly monitoring system

Abstract

The present application relates to the technical field of power system monitoring, in particular to a bus duct temperature and humidity anomaly monitoring method, comprising the following steps: S1, periodically collecting temperature and humidity data by a temperature and humidity sensor, and performing data preprocessing to obtain processed data; S2, performing mutation detection and trend analysis according to the processed data to obtain mutation data and trend data respectively; S3, respectively self-adaptively adjusting a weighting strategy according to the mutation data and the trend data; when in use, the present application is beneficial to improving the credibility of data, the present method can effectively reduce sensor noise and error, improve system adaptability, and the data after denoising and outlier elimination can be used as high-quality input for mutation detection, trend analysis and time series prediction, thereby improving the accuracy of the whole temperature and humidity anomaly monitoring, improving the accuracy of anomaly monitoring by comprehensively detecting short-time windows and long-time windows, and being beneficial to avoiding false positives caused by pure mutation detection and improving the response speed of the present application.

Description

Technical Field

[0001] This invention relates to the field of power system monitoring technology, specifically to a busbar trunking temperature and humidity anomaly monitoring system. Background Technology

[0002] Busbar trunking plays a crucial role in power systems, primarily for the efficient and safe distribution and transmission of high-current electrical energy. Busbar trunking is typically composed of metal strips (such as copper or aluminum), which are encased in insulating material and installed within a sealed metal casing, forming a power distribution channel. Because busbar trunking usually carries high current loads, any overheating or excessive humidity can lead to electrical faults or even serious accidents such as fires. Therefore, implementing effective temperature and humidity monitoring is essential for preventing potential risks.

[0003] The patent with publication number CN118640975A describes in its specification a method for monitoring abnormal temperature and humidity in busbar trunking, relating to the field of power system monitoring technology. It involves collecting regional values ​​of the busbar trunking, including the main body, internal components, and joints. Based on these regional values, a monitoring layout is established for the busbar trunking, comprising temperature and humidity monitoring, with the two types of values ​​labeled to construct an anomaly matrix. The monitored values ​​obtained within this matrix are compared with preset alarm rules. When the alarm rules are met, an alarm is generated, improving the accuracy and reliability of busbar trunking temperature and humidity monitoring, reducing false alarms or missed alarms, and notifying staff of abnormal situations in a more intuitive and convenient way for rapid location and handling of anomalies. This achieves multiple judgments of temperature and humidity, further improving monitoring accuracy. While this technology improves the stability and accuracy of monitoring data through moving average filtering and multi-index comparison, the use of moving average filtering may lead to system response lag. In cases of rapid temperature increases or sudden humidity changes, the system may not respond immediately, affecting safety.

[0004] In conclusion, developing a busbar temperature and humidity anomaly monitoring system remains a critical issue that urgently needs to be addressed in the field of power system monitoring technology. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies. Although the aforementioned technologies improve the stability and accuracy of monitoring data through moving average filtering and multi-index comparison, the use of moving average filtering may lead to system response lag, and the system may not be able to respond immediately when there is a rapid increase in temperature or a sudden change in humidity.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention provides a method for monitoring abnormal temperature and humidity in busbar trunking, comprising the following steps:

[0008] S1. Collect temperature and humidity data periodically using a temperature and humidity sensor, and perform data preprocessing to obtain processed data;

[0009] S2. Based on the processed data, perform mutation detection and trend analysis to obtain mutation data and trend data respectively;

[0010] S3. Adaptively adjust the weighting strategy based on mutation data and trend data respectively;

[0011] S4. Employ time-series forecasting models to analyze historical and trend data and provide forecast data;

[0012] S5. By monitoring the mutation data, trend data, and prediction data in real time, anomaly detection and anomaly alarm and early warning are performed.

[0013] Further, in step S1, temperature and humidity data are periodically collected using a temperature and humidity sensor, and the data is preprocessed to obtain the processed data.

[0014] The temperature and humidity sensor is installed in key parts of the busbar trunking, including the connection end, the enclosed conductor, and areas prone to water accumulation in humid environments. The sensor collects temperature and humidity data X(t) using a periodic sampling method of T. At time t, the temperature and humidity data are X(t) and X(t), respectively. W (t) and X S (t), expressed by the formula: Q W (t),Q S (t) is the actual temperature and humidity signal, E W (t),E S (t) is random noise, assumed to have a mean of 0 and a variance of σ. 2 Gaussian white noise, expressed by the formula: E W (t),E S (t)~ζ(0,σ 2 The data preprocessing includes using Kalman filtering to denoise the temperature and humidity data, resulting in denoised temperature and humidity data as follows: and The Grubbs test was used to remove outliers from the temperature and humidity data. The temperature data X was defined as follows: W The mean and standard deviation of (t) are expressed by the following formula:

[0015] in This represents the average temperature data within a time window, where N represents the total number of data points within the statistical time window. This represents the temperature value after denoising at time point i. This represents summing over all data points. Indicates the degree of fluctuation in temperature data. This represents the difference between the i-th temperature value and the average temperature. It calculates the square of the temperature deviation from the mean. It is the sum of the squared deviations of all data points. It is an unbiased estimate used when calculating the root mean square error. This involves taking the square root of the result to calculate the Grubbs statistic, expressed as: The threshold for the Grubbs test is: Where t β / (2N),N-2 It is the critical value of the t-distribution with N-2 degrees of freedom, when R > R ctil The data points are then identified as outliers and removed. The method for removing outliers from humidity data is the same.

[0016] Further, in step S2, the method for performing mutation detection and trend analysis based on the processed data to obtain mutation data and trend data respectively is as follows:

[0017] The mutation detection includes using short-time window processing to calculate the temperature and humidity change trend within the most recent 1-3 seconds. At time t, temperature and humidity data within the range of 0 (1-3 seconds) are selected, and the rate of change of the short-time window is calculated, expressed by the formula: Among them U W (t) is the rate of temperature change, U S (t) represents the rate of change of humidity, O represents the short-time window (1 second to 3 seconds), and ψ i This refers to the short-time window weighting coefficient. A dynamic thresholding method is employed, using historical mean and standard deviation to set an adaptive abrupt change threshold. Discrete wavelet transform is then used to perform multi-scale analysis of the temperature and humidity signal, extracting high-frequency components and identifying abrupt change points. The formula is as follows: Where Y W (a,b),Y S (a, b) are wavelet transform coefficients, where a is the scaling factor (used to separate high and low frequencies), b is the translation factor, and χ² is the shift factor. * (·) is the mother wavelet (Daubechies wavelet).

[0018] Further, in step S2, the method for performing mutation detection and trend analysis based on the processed data to obtain mutation data and trend data respectively is as follows:

[0019] The trend analysis includes using long-term window processing to calculate the temperature and humidity change trend over the most recent 10-60 seconds, selecting data from the past A (10-60 seconds), and calculating a weighted moving average to smooth the trend. The formula is as follows: Among them W ted (t) is the temperature trend value, S ted (t) is the humidity trend value, A is the length of the trend analysis window (within the past 10-60 seconds), and δ i These are long-term window weighting coefficients, satisfying ∑δ i =1, weighting coefficient expression formula: Where ε controls the rate of weight decay. It is an exponentially decaying weight. It is the normalization factor.

[0020] Furthermore, in step S3, the method for adaptively adjusting the weighting strategy based on the mutation data and trend data is as follows:

[0021] The weighting strategy includes increasing the weight of the short-time window if either the temperature surge is ≥5℃ or the humidity surge is ≥10%RH in the mutation data, and defining the mutation response factor D. jmp (t), expressed by the formula: Where ΔW = W(t) - W(tO) is the short-term temperature change, and ΔS = S(t) - S(tO) is the short-term humidity change, in D jmp If (t)≥1, a mutation is triggered, and the short-time window weight is updated. The formula is as follows:

[0022] Where φ w The adjustment rate for the short-term window is set to 0.1-0.3, ψ i (t) is the short-time window weight at time t, ψ i (t-1) is the short-time window weight of the previous time step, (1-ψ i (t-1)) is the complement of the short-time window, ζ i (t)=1-ψ i (t) is the long-term window weight used for long-term trend analysis. When the short-term window weight increases, the long-term window weight decreases accordingly, and vice versa.

[0023] Furthermore, in step S3, the method for adaptively adjusting the weighting strategy based on the mutation data and trend data is as follows:

[0024] The weighting strategy includes determining the presence of a trend anomaly if either the temperature continuously rises by ≥3℃ / min or the humidity continuously rises by ≥5%RH / min. This strategy reduces the weight of the short-term window, strengthens trend data analysis, and defines a trend anomaly factor D. ted (t), expressed by the formula: in It is the long-term temperature change rate. It is the long-term humidity change rate, in D tedIf (t)≥1, an abnormal trend is triggered, and the weight of the long-term window is adjusted. The formula is as follows:

[0025] Where φ s The trend adjustment coefficient is set to 0.05-0.2, ζ i (t) is the long-term window weight at the current time t, ζ i (t-1) is the long-term window weight of the previous time step, D ted (t) is the trend detection factor, (1-ζ) i (t-1)) is the adjustment factor, and the weights are updated using exponential smoothing, expressed as follows: in The value is set between 0.2 and 0.5 to ensure smooth variation.

[0026] Furthermore, in step S4, the method for using a time-series forecasting model to analyze historical and trend data and provide forecast data is as follows:

[0027] Collect temperature and humidity data X(t) for time series modeling and short-term forecasting, expressed by the formula: Where X(t) is the temperature and humidity data at the current time t, c represents either the average level or the long-term trend of the temperature and humidity data, and ∈(t) represents unpredictable noise at the current time point. γ represents how past forecast errors affect current temperature and humidity forecasts; F represents how many past temperature and humidity data points are used for forecasting; γ represents how many past temperature and humidity data points are used for forecasting. i It measures the impact of past temperature and humidity data on future forecasts. G represents how many past forecast errors are used in the forecast, and η represents the number of past forecast errors used in the forecast. k It measures the impact of past prediction errors on future predictions and uses a Long Short-Term Memory (LSTM) network for long-term prediction. The output hidden states are used to predict temperature and humidity data for the next few seconds. The formula is as follows:

[0028] in This refers to the temperature and humidity data at time t+n, where F represents the number of past temperature and humidity data points used for prediction, and γ... i It measures the impact of past temperature and humidity data on future predictions. X(t+ni) represents the temperature and humidity values ​​at time (t+ni), G represents the number of past prediction errors used in the prediction, and η k It measures the impact of past forecast errors on future forecasts. This represents the error between the predicted value and the actual value at a past point in time.

[0029] Furthermore, in step S4, the method for using a time-series forecasting model to analyze historical and trend data and provide forecast data is as follows:

[0030] A weighted fusion method is used to combine the temperature and humidity data predicted by the Long Short-Term Memory Network and the Temporal Series Prediction Model, expressed as follows:

[0031] Where ι is the weighting factor. This represents the final prediction result at time t+n. It is the prediction result of the time series forecasting model at time t+n. This is the prediction result of the Long Short-Term Memory (LSTM) network at time t+n, which depends on the error between the LTM network and the time series prediction model. The formula is as follows: Where ι is the weighting factor, h is the long short-term memory network, κ is the time series prediction model, e is the base of the natural logarithm, and MSE is the mean squared error.

[0032] Furthermore, in step S5, the method for detecting anomalies and issuing alarms and warnings by real-time monitoring of the mutation data, trend data, and prediction data is as follows:

[0033] The anomaly detection is based on a triple threshold judgment method, which calculates an anomaly score for each data source, expressed by the formula: λ(t)=μ1·λ v (t)+μ2·λ θ (t)+μ3·λ κ (t), where μ1, μ2, μ3 are weighting coefficients satisfying μ1+μ2+μ3=1 to adjust the contribution of different data sources, λ v (t) reflects the degree of abrupt change in temperature and humidity, λ θ (t) reflects the degree of long-term trend anomaly in temperature and humidity, λ κ (t) reflects the degree of anomaly in the predicted temperature and humidity. The degree of abrupt change in temperature and humidity can be calculated by the rate of change within a short time window, expressed by the formula: in This represents the rate of change in temperature and humidity within a short-term window. The standard deviation of historical temperature and humidity changes is used for normalization. The degree of anomaly in the trend data is calculated by the rate of temperature and humidity change over the past A (10 seconds - 60 seconds), expressed by the formula: Where λ θ (t) measures the overall trend of temperature and humidity data at time t, and A is the size of the long-term window. It is the standard deviation of trend data. The trend change at time ti represents the rate of change of temperature and humidity data. The prediction anomaly score measures the deviation between the predicted temperature and humidity and the current data, expressed by the formula: Where λ κ (t) is the predicted anomaly score, which measures the degree of deviation between the predicted and current measurements of temperature and humidity. The standard deviation of the predicted data is used to normalize the predicted score. These are the predicted temperature and humidity values ​​at time t+n′. These are the actual measured values ​​of temperature and humidity at the current time t. It is the absolute error between the predicted value and the current actual measurement value. Based on the comprehensive anomaly score λ(t), an adaptive alarm threshold Θ is set. bj (t), expressed by the formula: Where Θ bj (t) is the adaptive alarm threshold. It is the average of historical anomaly scores. It is the standard deviation of historical anomaly scores, and r′ is the adaptive coefficient that determines alarm sensitivity. The alarm sensitivity includes Level 1 warning, Level 2 warning and Level 3 alarm.

[0034] On the other hand, the present invention also provides a busbar trunking temperature and humidity anomaly monitoring system, comprising:

[0035] The data acquisition module periodically collects temperature and humidity data through temperature and humidity sensors, and performs data preprocessing to obtain processed data.

[0036] The detection and analysis module performs mutation detection and trend analysis based on the processed data to obtain mutation data and trend data, respectively.

[0037] The strategy adjustment module adaptively adjusts the weighting strategy based on both mutation and trend data.

[0038] The time series forecasting module uses a time series forecasting model to analyze historical and trend data and provide forecast data.

[0039] The anomaly alarm module detects anomalies and issues alarms and warnings by monitoring the mutation data, trend data, and prediction data in real time.

[0040] Furthermore, the process of periodically collecting temperature and humidity data through a temperature and humidity sensor and performing data preprocessing to obtain the processed data includes: installing the temperature and humidity sensor at key locations in the busbar trunking, including connection ends, enclosed conductors, and areas prone to water accumulation in humid environments; the sensor collecting temperature and humidity data X(t) at periodic sampling intervals T; and setting the temperature and humidity data at time t as X(t). W (t) and X S (t), expressed by the formula: Q W (t),Q S (t) is the actual temperature and humidity signal, E W (t),E S(t) is random noise, assumed to have a mean of 0 and a variance of σ. 2 Gaussian white noise, expressed by the formula: E W (t),E S (t)~ζ(0,σ 2 The data preprocessing includes using Kalman filtering to denoise the temperature and humidity data, resulting in denoised temperature and humidity data as follows: and The Grubbs test was used to remove outliers from the temperature and humidity data. The temperature data X was defined as follows: W The mean and standard deviation of (t) are expressed by the following formula: in This represents the average temperature data within a time window, where N represents the total number of data points within the statistical time window. This represents the temperature value after denoising at time point i. This represents summing over all data points. Indicates the degree of fluctuation in temperature data. This represents the difference between the i-th temperature value and the average temperature. It calculates the square of the temperature deviation from the mean. It is the sum of the squared deviations of all data points. It is an unbiased estimate used when calculating the root mean square error. This involves taking the square root of the result to calculate the Grubbs statistic, expressed as: The threshold for the Grubbs test is: Where t β / (2N),N-2 It is the critical value of the t-distribution with N-2 degrees of freedom, when R > R ctil The data points are then identified as outliers and removed. The method for removing outliers from humidity data is the same.

[0041] Furthermore, the operational procedures for performing mutation detection and trend analysis based on the processed data to obtain mutation data and trend data respectively include: the mutation detection includes using short-time window processing to calculate the temperature and humidity change trend within the most recent 1-3 seconds; at time t, selecting temperature and humidity data within 0 (1-3 seconds), calculating the rate of change of the short-time window, expressed by the formula: Among them U W (t) is the rate of temperature change, U S (t) represents the rate of change of humidity, O represents the short-time window (1 second to 3 seconds), and ψ i This refers to the short-time window weighting coefficient. A dynamic thresholding method is employed, using historical mean and standard deviation to set an adaptive abrupt change threshold. Discrete wavelet transform is then used to perform multi-scale analysis of the temperature and humidity signal, extracting high-frequency components and identifying abrupt change points. The formula is as follows: Where Y W(a,b),Y S (a, b) are wavelet transform coefficients, where a is the scaling factor (used to separate high and low frequencies), b is the translation factor, and χ² is the shift factor. * (·) is the mother wavelet (Daubechies wavelet).

[0042] The trend analysis includes using long-term window processing to calculate the temperature and humidity change trend over the most recent 10-60 seconds, selecting data from the past A (10-60 seconds), and calculating a weighted moving average to smooth the trend. The formula is as follows: Among them W ted (t) is the temperature trend value, S ted (t) is the humidity trend value, A is the length of the trend analysis window (within the past 10-60 seconds), and δ i These are long-term window weighting coefficients that satisfy Σδ i =1, weighting coefficient expression formula: Where ε controls the rate of weight decay. It is an exponentially decaying weight. It is a normalization factor;

[0043] Furthermore, the operational procedure for adaptively adjusting the weighting strategy based on mutation data and trend data includes: if either a temperature surge ≥5℃ or a humidity surge ≥10%RH occurs in the mutation data, the short-term window weight is increased, and a mutation response factor D is defined. jmp (t), expressed by the formula: Where ΔW = W(t) - W(tO) is the short-term temperature change, and ΔS = S(t) - S(tO) is the short-term humidity change, in D jmp If (t)≥1, a mutation is triggered, and the short-time window weight is updated. The formula is as follows: Where φ w The adjustment rate for the short-term window is set to 0.1-0.3, ψ i (t) is the short-time window weight at time t, ψ i (t-1) is the short-time window weight of the previous time step, (1-ψ i (t-1)) is the complement of the short-time window, ζ i (t)=1-ψ i (t) is the long-term window weight used for long-term trend analysis. When the short-term window weight increases, the long-term window weight decreases accordingly, and vice versa.

[0044] The weighting strategy includes determining the presence of a trend anomaly if either the temperature continuously rises by ≥3℃ / min or the humidity continuously rises by ≥5%RH / min. This strategy reduces the weight of the short-term window, strengthens trend data analysis, and defines a trend anomaly factor D. ted (t), expressed by the formula: in It is the long-term temperature change rate. It is the long-term humidity change rate, in D ted If (t)≥1, an abnormal trend is triggered, and the weight of the long-term window is adjusted. The formula is as follows:

[0045] Where φ s The trend adjustment coefficient is set to 0.05-0.2, ζ i (t) is the long-term window weight at the current time t, ζ i (t-1) is the long-term window weight of the previous time step, D ted (t) is the trend detection factor, (1-ζ) i (t-1)) is the adjustment factor, and the weights are updated using exponential smoothing, expressed as follows: in The value should be between 0.2 and 0.5 to ensure smooth variation;

[0046] Furthermore, a time-series forecasting model is used to analyze historical and trend data, and the operational process for forecasting data includes: collecting temperature and humidity data X(t), performing time-series modeling for short-term forecasting, expressed by the formula:

[0047] Where X(t) is the temperature and humidity data at the current time t, c represents either the average level or the long-term trend of the temperature and humidity data, and ∈(t) represents unpredictable noise at the current time point. γ represents how past forecast errors affect current temperature and humidity forecasts; F represents how many past temperature and humidity data points are used for forecasting; γ represents how many past temperature and humidity data points are used for forecasting. i It measures the impact of past temperature and humidity data on future forecasts. G represents how many past forecast errors are used in the forecast, and η represents the number of past forecast errors used in the forecast. k It measures the impact of past prediction errors on future predictions and uses a Long Short-Term Memory (LSTM) network for long-term prediction. The output hidden states are used to predict temperature and humidity data for the next few seconds. The formula is as follows:

[0048] in This refers to the temperature and humidity data at time t+n, where F represents the number of past temperature and humidity data points used for prediction, and γ... i It measures the impact of past temperature and humidity data on future predictions. X(t+ni) represents the temperature and humidity values ​​at time (t+ni), G represents the number of past prediction errors used in the prediction, and η k It measures the impact of past forecast errors on future forecasts. This represents the error between the predicted value and the actual value at a past point in time.

[0049] A weighted fusion method is used to combine the temperature and humidity data predicted by the Long Short-Term Memory Network and the Temporal Series Prediction Model, expressed as follows:

[0050] Where ι is the weighting factor. This represents the final prediction result at time t+n. It is the prediction result of the time series forecasting model at time t+n. This is the prediction result of the Long Short-Term Memory (LSTM) network at time t+n, which depends on the error between the LTM network and the time series prediction model. The formula is as follows: Where ι is the weighting factor, h is the long short-term memory network, κ is the time series prediction model, e is the base of the natural logarithm, and MSE is the mean squared error.

[0051] Furthermore, the operational process for anomaly detection and alarm / early warning through real-time monitoring of the mutation data, trend data, and prediction data includes: the anomaly detection is based on a triple threshold judgment method, calculating the anomaly score for each data source, expressed by the formula: λ(t)=μ1·λ v (t)+μ2·λ θ (t)+μ3·λ κ (t), where μ1, μ2, μ3 are weighting coefficients satisfying μ1+μ2+μ3=1 to adjust the contribution of different data sources, λ v (t) reflects the degree of abrupt change in temperature and humidity, λ θ (t) reflects the degree of long-term trend anomaly in temperature and humidity, λ κ (t) reflects the degree of anomaly in the predicted temperature and humidity. The degree of abrupt change in temperature and humidity can be calculated by the rate of change within a short time window, expressed by the formula: in This represents the rate of change in temperature and humidity within a short-term window. The standard deviation of historical temperature and humidity changes is used for normalization. The degree of anomaly in the trend data is calculated by the rate of temperature and humidity change over the past A (10 seconds - 60 seconds), expressed by the formula: Where λ θ (t) measures the overall trend of temperature and humidity data at time t, and A is the size of the long-term window. It is the standard deviation of trend data. The trend change at time ti represents the rate of change of temperature and humidity data. The prediction anomaly score measures the deviation between the predicted temperature and humidity and the current data, expressed by the formula: Where λ κ (t) is the predicted anomaly score, which measures the degree of deviation between the predicted and current measurements of temperature and humidity. The standard deviation of the predicted data is used to normalize the predicted score. These are the predicted temperature and humidity values ​​at time t+n′. These are the actual measured values ​​of temperature and humidity at the current time t. It is the absolute error between the predicted value and the current actual measurement value. Based on the comprehensive anomaly score λ(t), an adaptive alarm threshold Θ is set. bj (t), expressed by the formula: Where Θ bj (t) is the adaptive alarm threshold. It is the average of historical anomaly scores. It is the standard deviation of historical anomaly scores, and r′ is the adaptive coefficient that determines alarm sensitivity. The alarm sensitivity includes Level 1 warning, Level 2 warning and Level 3 alarm.

[0052] Beneficial effects

[0053] Compared with known public technologies, the technical solution provided by this invention has the following advantages:

[0054] Beneficial effects:

[0055] When used, this invention helps to improve the reliability of data. This method can effectively reduce sensor noise and error, improve system adaptability, and the data after noise reduction and outlier removal can be used as high-quality input for mutation detection, trend analysis and time series prediction, thereby improving the accuracy of the entire temperature and humidity anomaly monitoring. By combining short-term and long-term window detection, the accuracy of anomaly monitoring is improved, which helps to avoid false alarms caused by simply relying on mutation detection and improves the response speed of this invention.

[0056] In use, this invention employs an adaptive weighting strategy, dynamically adjusting the weights of short-term and long-term windows based on real-time changes in temperature and humidity. This helps avoid misjudgments that may arise from fixed-weight schemes. The adaptive adjustment of the weights allows for rapid response to anomalies and accurate prediction of potential trends, improving the overall stability and accuracy of monitoring. By combining short-term, long-term, and predictive data with a triple threshold judgment method, false alarms and missed alarms are reduced, improving detection accuracy. The adaptive alarm threshold dynamically adjusts the alarm standard to adapt to different environmental changes, avoiding misjudgments caused by fixed thresholds. Attached Figure Description

[0057] Figure 1 This is a flowchart of a method for monitoring abnormal temperature and humidity in a busbar trunking according to the present invention;

[0058] Figure 2 This is a system diagram of a busbar trunking temperature and humidity anomaly monitoring system according to the present invention. Detailed Implementation

[0059] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0060] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but includes other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0061] The present invention will now be described in further detail with reference to the accompanying drawings:

[0062] Example 1:

[0063] like Figure 1 As shown, the present invention provides a method for monitoring abnormal temperature and humidity in busbar trunking, comprising the following steps: S1, periodically collecting temperature and humidity data through a temperature and humidity sensor, and performing data preprocessing to obtain processed data;

[0064] Further, in step S1, temperature and humidity data are periodically collected using a temperature and humidity sensor, and the data is preprocessed to obtain the processed data.

[0065] The temperature and humidity sensor is installed in key parts of the busbar trunking, including the connection end, the enclosed conductor, and areas prone to water accumulation in humid environments. The sensor collects temperature and humidity data X(t) using a periodic sampling method of T. At time t, the temperature and humidity data are X(t) and X(t), respectively. W (t) and X S (t), expressed by the formula: Q W (t),Q S (t) is the actual temperature and humidity signal, E W (t),E S (t) is random noise, assumed to have a mean of 0 and a variance of σ. 2 Gaussian white noise, expressed by the formula: EW (t),E S (t)~ζ(0,σ 2 The data preprocessing includes using Kalman filtering to denoise the temperature and humidity data, resulting in denoised temperature and humidity data as follows: and The Grubbs test was used to remove outliers from the temperature and humidity data. The temperature data X was defined as follows: W The mean and standard deviation of (t) are expressed by the following formula:

[0066] in This represents the average temperature data within a time window, where N represents the total number of data points within the statistical time window. This represents the temperature value after denoising at time point i. This represents summing over all data points. Indicates the degree of fluctuation in temperature data. This represents the difference between the i-th temperature value and the average temperature. It calculates the square of the temperature deviation from the mean. It is the sum of the squared deviations of all data points. It is the unbiased estimate used when calculating the root mean square error. This involves taking the square root of the result to calculate the Grubbs statistic, expressed as: The threshold for the Grubbs test is: Where t β / (2N),N-2 It is the critical value of the t-distribution with N-2 degrees of freedom, when R > R ctil The data points are then identified as outliers and removed. The method for removing outliers from humidity data is the same.

[0067] In this embodiment, Kalman filtering effectively removes Gaussian noise, making the temperature and humidity data smoother, reducing errors, and improving the reliability of monitoring. Grubbs criterion can identify and remove outliers, avoiding interference from extreme data on the overall trend judgment, which helps to improve the credibility of the data. This method can effectively reduce sensor noise and errors, improve system adaptability, and the data after denoising and outlier removal can be used as high-quality input for mutation detection, trend analysis, and time series prediction, thereby improving the accuracy of the entire temperature and humidity anomaly monitoring.

[0068] S2. Based on the processed data, perform mutation detection and trend analysis to obtain mutation data and trend data respectively;

[0069] Further, in step S2, the method for performing mutation detection and trend analysis based on the processed data to obtain mutation data and trend data respectively is as follows:

[0070] The mutation detection includes using short-time window processing to calculate the temperature and humidity change trend within the most recent 1-3 seconds. At time t, temperature and humidity data within the range of 0 (1-3 seconds) are selected, and the rate of change of the short-time window is calculated, expressed by the formula: Among them U W (t) is the rate of temperature change, U S (t) represents the rate of change of humidity, O represents the short-time window (1 second to 3 seconds), and ψ i This refers to the short-time window weighting coefficient. A dynamic thresholding method is employed, using historical mean and standard deviation to set an adaptive abrupt change threshold. Discrete wavelet transform is then used to perform multi-scale analysis of the temperature and humidity signal, extracting high-frequency components and identifying abrupt change points. The formula is as follows: Where Y W (a,b),Y S (a, b) are wavelet transform coefficients, where a is the scaling factor (used to separate high and low frequencies), b is the translation factor, and χ² is the shift factor. * (·) is the mother wavelet (Daubechies wavelet).

[0071] Further, in step S2, the method for performing mutation detection and trend analysis based on the processed data to obtain mutation data and trend data respectively is as follows:

[0072] The trend analysis includes using long-term window processing to calculate the temperature and humidity change trend over the most recent 10-60 seconds, selecting data from the past A (10-60 seconds), and calculating a weighted moving average to smooth the trend. The formula is as follows: Among them W ted (t) is the temperature trend value, S ted (t) is the humidity trend value, A is the length of the trend analysis window (within the past 10-60 seconds), and δ i These are long-term window weighting coefficients, satisfying ∑δ i =1, weighting coefficient expression formula: Where ε controls the rate of weight decay. It is an exponentially decaying weight. It is the normalization factor.

[0073] In this embodiment, by combining short-term window analysis with dynamic thresholds and wavelet transform, anomalies such as local overheating of busbars and sudden increases in humidity can be effectively identified in a short period of time. Long-term window analysis uses weighted moving average and smooth decay weights to avoid interference from short-term fluctuations on long-term trend judgment and improve prediction accuracy. By combining short-term and long-term window detection, the accuracy of anomaly monitoring is improved, which helps to avoid false alarms caused by simply relying on mutation detection and improves the response speed of the present invention.

[0074] S3. Adaptively adjust the weighting strategy based on mutation data and trend data respectively;

[0075] Furthermore, in step S3, the method for adaptively adjusting the weighting strategy based on the mutation data and trend data is as follows:

[0076] The weighting strategy includes increasing the weight of the short-time window if either the temperature surge is ≥5℃ or the humidity surge is ≥10%RH in the mutation data, and defining the mutation response factor D. jmp (t), expressed by the formula: Where ΔW = W(t) - W(tO) is the short-term temperature change, and ΔS = S(t) - S(tO) is the short-term humidity change, in D jmp If (t)≥1, a mutation is triggered, and the short-time window weight is updated. The formula is as follows:

[0077] Where φ w The adjustment rate for the short-term window is set to 0.1-0.3, ψ i (t) is the short-time window weight at time t, ψ i (t-1) is the short-time window weight of the previous time step, (1-ψ i (t-1)) is the complement of the short-time window, ζ i (t)=1-ψ i (t) is the long-term window weight used for long-term trend analysis. When the short-term window weight increases, the long-term window weight decreases accordingly, and vice versa.

[0078] Furthermore, in step S3, the method for adaptively adjusting the weighting strategy based on the mutation data and trend data is as follows:

[0079] The weighting strategy includes determining the presence of a trend anomaly if either the temperature continuously rises by ≥3℃ / min or the humidity continuously rises by ≥5%RH / min. This strategy reduces the weight of the short-term window, strengthens trend data analysis, and defines a trend anomaly factor D. ted (t), expressed by the formula: in It is the long-term temperature change rate. It is the long-term humidity change rate, in D ted If (t)≥1, an abnormal trend is triggered, and the weight of the long-term window is adjusted. The formula is as follows:

[0080] Where φ s The trend adjustment coefficient is set to 0.05-0.2, ζ i (t) is the long-term window weight at the current time t, ζ i (t-1) is the long-term window weight of the previous time step, D ted (t) is the trend detection factor, (1-ζ) i(t-1)) is the adjustment factor, and the weights are updated using exponential smoothing, expressed as follows: in The value is set between 0.2 and 0.5 to ensure smooth variation.

[0081] In this embodiment, an adaptive weighting strategy is adopted, which dynamically adjusts the weights of short-term and long-term windows based on real-time changes in temperature and humidity. This helps avoid misjudgment problems that may arise from fixed-weight schemes. When sudden changes occur, the weight of the short-term window is rapidly increased, improving the detection capability for sudden anomalies, such as the rapid identification of sudden temperature rises at busbar connection points or water accumulation in humid environments. These are only short-term fluctuations and will not be misjudged as anomalies. Instead, a more accurate judgment is made with the support of trend data, thereby reducing false alarms and missed alarms. The short-term window is used to detect sudden changes, while the long-term window is used to analyze the long-term operating trend of the equipment. Adaptively adjusting the weights of both can not only respond quickly to anomalies but also accurately predict trend-based hidden dangers, improving the overall stability and accuracy of monitoring.

[0082] S4. Employ time-series forecasting models to analyze historical and trend data and provide forecast data;

[0083] Furthermore, in step S4, the method for using a time-series forecasting model to analyze historical and trend data and provide forecast data is as follows:

[0084] Collect temperature and humidity data X(t) for time series modeling and short-term forecasting, expressed by the formula: Where X(t) is the temperature and humidity data at the current time t, c represents either the average level or the long-term trend of the temperature and humidity data, and ∈(t) represents unpredictable noise at the current time point. γ represents how past forecast errors affect current temperature and humidity forecasts; F represents how many past temperature and humidity data points are used for forecasting; γ represents how many past temperature and humidity data points are used for forecasting. i It measures the impact of past temperature and humidity data on future forecasts. G represents how many past forecast errors are used in the forecast, and η represents the number of past forecast errors used in the forecast. k It measures the impact of past prediction errors on future predictions and uses a Long Short-Term Memory (LSTM) network for long-term prediction. The output hidden states are used to predict temperature and humidity data for the next few seconds. The formula is as follows:

[0085] in This refers to the temperature and humidity data at time t+n, where F represents the number of past temperature and humidity data points used for prediction, and γ... i It measures the impact of past temperature and humidity data on future predictions. X(t+ni) represents the temperature and humidity values ​​at time (t+ni), G represents the number of past prediction errors used in the prediction, and η kIt measures the impact of past forecast errors on future forecasts. This represents the error between the predicted value and the actual value at a past point in time.

[0086] Furthermore, in step S4, the method for using a time-series forecasting model to analyze historical and trend data and provide forecast data is as follows:

[0087] A weighted fusion method is used to combine the temperature and humidity data predicted by the Long Short-Term Memory Network and the Temporal Series Prediction Model, expressed as follows:

[0088] Where ι is the weighting factor. This represents the final prediction result at time t+n. It is the prediction result of the time series forecasting model at time t+n. This is the prediction result of the Long Short-Term Memory (LSTM) network at time t+n, which depends on the error between the LTM network and the time series prediction model. The formula is as follows: Where ι is the weighting factor, h is the long short-term memory network, κ is the time series prediction model, e is the base of the natural logarithm, and MSE is the mean squared error.

[0089] In this embodiment, short-term prediction can detect sudden trends a few seconds before temperature and humidity anomalies occur. Long Short-Term Memory Network (LSTM) can predict the overall trend of temperature and humidity in the busbar trunking through long-term time-dependent learning, which helps to plan equipment maintenance in advance, avoid unnecessary alarms, and improve the accuracy of early warning. Adaptive adjustment of weights automatically selects the most suitable prediction method based on real-time data changes, which is beneficial to improving prediction accuracy.

[0090] S5. By monitoring the mutation data, trend data and prediction data in real time, anomaly detection and anomaly alarm and early warning are performed;

[0091] Furthermore, in step S5, the method for detecting anomalies and issuing alarms and warnings by real-time monitoring of the mutation data, trend data, and prediction data is as follows:

[0092] The anomaly detection is based on a triple threshold judgment method, which calculates an anomaly score for each data source, expressed by the formula: λ(t)=μ1·λ v (t)+μ2·λ θ (t)+μ3·λ κ (t), where μ1, μ2, μ3 are weighting coefficients satisfying μ1+μ2+μ3=1 to adjust the contribution of different data sources, λ v (t) reflects the degree of abrupt change in temperature and humidity, λ θ (t) reflects the degree of long-term trend anomaly in temperature and humidity, λ κ(t) reflects the degree of anomaly in the predicted temperature and humidity. The degree of abrupt change in temperature and humidity can be calculated by the rate of change within a short time window, expressed by the formula: in This represents the rate of change in temperature and humidity within a short-term window. The standard deviation of historical temperature and humidity changes is used for normalization. The degree of anomaly in the trend data is calculated by the rate of temperature and humidity change over the past A (10 seconds - 60 seconds), expressed by the formula: Where λ θ (t) measures the overall trend of temperature and humidity data at time t, and A is the size of the long-term window. It is the standard deviation of trend data. The trend change at time ti represents the rate of change of temperature and humidity data. The prediction anomaly score measures the deviation between the predicted temperature and humidity and the current data, expressed by the formula: Where λ κ (t) is the predicted anomaly score, which measures the degree of deviation between the predicted and current measurements of temperature and humidity. The standard deviation of the predicted data is used to normalize the predicted score. These are the predicted temperature and humidity values ​​at time t+n′. These are the actual measured values ​​of temperature and humidity at the current time t. It is the absolute error between the predicted value and the current actual measurement value. Based on the comprehensive anomaly score λ(t), an adaptive alarm threshold Θ is set. bj (t), expressed by the formula: Where Θ bj (t) is the adaptive alarm threshold. It is the average of historical anomaly scores. It is the standard deviation of historical anomaly scores, and r′ is the adaptive coefficient that determines alarm sensitivity. The alarm sensitivity includes Level 1 warning, Level 2 warning and Level 3 alarm.

[0093] In this embodiment, the first-level warning includes the system recording anomalies but not triggering external alarms; it is only used for internal monitoring. The second-level warning includes a further increase in the anomaly score, triggering a notification to maintenance personnel for inspection. The third-level alarm includes a serious exceedance of the threshold, immediately triggering an audible and visual alarm and linking with equipment protection measures. By combining short-term, long-term, and predictive data with the triple threshold judgment method, false alarms and missed alarms can be reduced, and detection accuracy can be improved. An adaptive alarm threshold is adopted to dynamically adjust the alarm standard to adapt to different environmental changes and avoid misjudgments caused by fixed thresholds.

[0094] Example 2:

[0095] like Figure 2 As shown, Embodiment 2 provides a busbar trunking temperature and humidity anomaly monitoring system, including:

[0096] The data acquisition module periodically collects temperature and humidity data through temperature and humidity sensors, and performs data preprocessing to obtain processed data.

[0097] The detection and analysis module performs mutation detection and trend analysis based on the processed data to obtain mutation data and trend data, respectively.

[0098] The strategy adjustment module adaptively adjusts the weighting strategy based on both mutation and trend data.

[0099] The time series forecasting module uses a time series forecasting model to analyze historical and trend data and provide forecast data.

[0100] The anomaly alarm module detects anomalies and issues alarms and warnings by monitoring the mutation data, trend data, and prediction data in real time.

[0101] Furthermore, the process of periodically collecting temperature and humidity data through a temperature and humidity sensor and performing data preprocessing to obtain the processed data includes: installing the temperature and humidity sensor at key locations in the busbar trunking, including connection ends, enclosed conductors, and areas prone to water accumulation in humid environments; the sensor collecting temperature and humidity data X(t) at periodic sampling intervals T; and setting the temperature and humidity data at time t as X(t). W (t) and X S (t), expressed by the formula: Q W (t),Q S (t) is the actual temperature and humidity signal, E W (t),E S (t) is random noise, assumed to have a mean of 0 and a variance of σ. 2 Gaussian white noise, expressed by the formula: E W (t),E S (t)~ζ(0,σ 2 The data preprocessing includes using Kalman filtering to denoise the temperature and humidity data, resulting in denoised temperature and humidity data as follows: and The Grubbs test was used to remove outliers from the temperature and humidity data. The temperature data X was defined as follows: W The mean and standard deviation of (t) are expressed by the following formula: in This represents the average temperature data within a time window, where N represents the total number of data points within the statistical time window. This represents the temperature value after denoising at time point i. This represents summing over all data points. Indicates the degree of fluctuation in temperature data. This represents the difference between the i-th temperature value and the average temperature. It calculates the square of the temperature deviation from the mean. It is the sum of the squared deviations of all data points. It is an unbiased estimate used when calculating the root mean square error. This involves taking the square root of the result to calculate the Grubbs statistic, expressed as: The threshold for the Grubbs test is: Where t β / (2N),N-2 It is the critical value of the t-distribution with N-2 degrees of freedom, when R > R ctil The data points are then identified as outliers and removed. The method for removing outliers from humidity data is the same.

[0102] Furthermore, the operational procedures for performing mutation detection and trend analysis based on the processed data to obtain mutation data and trend data respectively include: the mutation detection includes using short-time window processing to calculate the temperature and humidity change trend within the most recent 1-3 seconds; at time t, selecting temperature and humidity data within 0 (1-3 seconds), calculating the rate of change of the short-time window, expressed by the formula: Among them U W (t) is the rate of temperature change, U S (t) represents the rate of change of humidity, O represents the short-time window (1 second to 3 seconds), and ψ i This refers to the short-time window weighting coefficient. A dynamic thresholding method is employed, using historical mean and standard deviation to set an adaptive abrupt change threshold. Discrete wavelet transform is then used to perform multi-scale analysis of the temperature and humidity signal, extracting high-frequency components and identifying abrupt change points. The formula is as follows: Where Y W (a,b),Y S (a, b) are wavelet transform coefficients, where a is the scaling factor (used to separate high and low frequencies), b is the translation factor, and χ² is the shift factor. * (·) is the mother wavelet (Daubechies wavelet).

[0103] The trend analysis includes using long-term window processing to calculate the temperature and humidity change trend over the most recent 10-60 seconds, selecting data from the past A (10-60 seconds), and calculating a weighted moving average to smooth the trend. The formula is as follows: Among them W ted (t) is the temperature trend value, S ted (t) is the humidity trend value, A is the length of the trend analysis window (within the past 10-60 seconds), and δ i These are long-term window weighting coefficients, satisfying ∑δ i =1, weighting coefficient expression formula: Where ε controls the rate of weight decay. It is an exponentially decaying weight. It is a normalization factor;

[0104] Furthermore, the operational procedure for adaptively adjusting the weighting strategy based on mutation data and trend data includes: if either a temperature surge ≥5℃ or a humidity surge ≥10%RH occurs in the mutation data, the short-term window weight is increased, and a mutation response factor D is defined. jmp (t), expressed by the formula: Where ΔW = W(t) - W(tO) is the short-term temperature change, and ΔS = S(t) - S(tO) is the short-term humidity change, in D jmp If (t)≥1, a mutation is triggered, and the short-time window weight is updated. The formula is as follows: Where φ w The adjustment rate for the short-term window is set to 0.1-0.3, ψ i (t) is the short-time window weight at time t, ψ i (t-1) is the short-time window weight of the previous time step, (1-ψ i (t-1)) is the complement of the short-time window, ζ i (t)=1-ψ i (t) is the long-term window weight used for long-term trend analysis. When the short-term window weight increases, the long-term window weight decreases accordingly, and vice versa.

[0105] The weighting strategy includes determining the presence of a trend anomaly if either the temperature continuously rises by ≥3℃ / min or the humidity continuously rises by ≥5%RH / min. This strategy reduces the weight of the short-term window, strengthens trend data analysis, and defines a trend anomaly factor D. ted (t), expressed by the formula: in It is the long-term temperature change rate. It is the long-term humidity change rate, in D ted If (t)≥1, an abnormal trend is triggered, and the weight of the long-term window is adjusted. The formula is as follows:

[0106] Where φ s The trend adjustment coefficient is set to 0.05-0.2, ζ i (t) is the long-term window weight at the current time t, ζ i (t-1) is the long-term window weight of the previous time step, D ted (t) is the trend detection factor, (1-ζ) i (t-1)) is the adjustment factor, and the weights are updated using exponential smoothing, expressed as follows: in The value should be between 0.2 and 0.5 to ensure smooth variation;

[0107] Furthermore, a time-series forecasting model is used to analyze historical and trend data, and the operational process for forecasting data includes: collecting temperature and humidity data X(t), performing time-series modeling for short-term forecasting, expressed by the formula:

[0108] Where X(t) is the temperature and humidity data at the current time t, c represents either the average level or the long-term trend of the temperature and humidity data, and ∈(t) represents unpredictable noise at the current time point. γ represents how past forecast errors affect current temperature and humidity forecasts; F represents how many past temperature and humidity data points are used for forecasting; γ represents how many past temperature and humidity data points are used for forecasting. i It measures the impact of past temperature and humidity data on future forecasts. G represents how many past forecast errors are used in the forecast, and η represents the number of past forecast errors used in the forecast. k It measures the impact of past prediction errors on future predictions and uses a Long Short-Term Memory (LSTM) network for long-term prediction. The output hidden states are used to predict temperature and humidity data for the next few seconds. The formula is as follows:

[0109] in This refers to the temperature and humidity data at time t+n, where F represents the number of past temperature and humidity data points used for prediction, and γ... i It measures the impact of past temperature and humidity data on future predictions. X(t+ni) represents the temperature and humidity values ​​at time (t+ni), G represents the number of past prediction errors used in the prediction, and η k It measures the impact of past forecast errors on future forecasts. This represents the error between the predicted value and the actual value at a past point in time.

[0110] A weighted fusion method is used to combine the temperature and humidity data predicted by the Long Short-Term Memory Network and the Temporal Series Prediction Model, expressed as follows:

[0111] Where ι is the weighting factor. This represents the final prediction result at time t+n. It is the prediction result of the time series forecasting model at time t+n. This is the prediction result of the Long Short-Term Memory (LSTM) network at time t+n, which depends on the error between the LTM network and the time series prediction model. The formula is as follows: Where ι is the weighting factor, h is the long short-term memory network, κ is the time series prediction model, e is the base of the natural logarithm, and MSE is the mean squared error.

[0112] Furthermore, the operational process for anomaly detection and alarm / early warning through real-time monitoring of the mutation data, trend data, and prediction data includes: the anomaly detection is based on a triple threshold judgment method, calculating the anomaly score for each data source, expressed by the formula: λ(t)=μ1·λ v (t)+μ2·λ θ (t)+μ3·λ κ (t), where μ1, μ2, μ3 are weighting coefficients satisfying μ1+μ2+μ3=1 to adjust the contribution of different data sources, λ v (t) reflects the degree of abrupt change in temperature and humidity, λ θ (t) reflects the degree of long-term trend anomaly in temperature and humidity, λ κ (t) reflects the degree of anomaly in the predicted temperature and humidity. The degree of abrupt change in temperature and humidity can be calculated by the rate of change within a short time window, expressed by the formula: in This represents the rate of change in temperature and humidity within a short-term window. The standard deviation of historical temperature and humidity changes is used for normalization. The degree of anomaly in the trend data is calculated by the rate of temperature and humidity change over the past A (10 seconds - 60 seconds), expressed by the formula: Where λ θ (t) measures the overall trend of temperature and humidity data at time t, and A is the size of the long-term window. It is the standard deviation of trend data. The trend change at time ti represents the rate of change of temperature and humidity data. The prediction anomaly score measures the deviation between the predicted temperature and humidity and the current data, expressed by the formula: Where λ κ (t) is the predicted anomaly score, which measures the degree of deviation between the predicted and current measurements of temperature and humidity. The standard deviation of the predicted data is used to normalize the predicted score. These are the predicted temperature and humidity values ​​at time t+n′. These are the actual measured values ​​of temperature and humidity at the current time t. It is the absolute error between the predicted value and the current actual measurement value. Based on the comprehensive anomaly score λ(t), an adaptive alarm threshold Θ is set. bj (t), expressed by the formula: Where Θ bj (t) is the adaptive alarm threshold. It is the average of historical anomaly scores. It is the standard deviation of historical anomaly scores, and r′ is the adaptive coefficient that determines alarm sensitivity. The alarm sensitivity includes Level 1 warning, Level 2 warning and Level 3 alarm.

[0113] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring abnormal temperature and humidity in busbar trunking, characterized in that, Includes the following steps: S1. Collect temperature and humidity data periodically using a temperature and humidity sensor, and perform data preprocessing to obtain processed data; S2. Based on the processed data, perform mutation detection and trend analysis to obtain mutation data and long-term window data, respectively. S3. Adaptively adjust the weighting strategy based on the mutation data and long-term window data respectively; S4. Employ a time-series forecasting model to analyze historical data and long-term window data, and provide forecast data; S5. By monitoring the mutation data, long-term window data and prediction data in real time, anomaly detection and anomaly alarm and early warning are performed; In step S1, temperature and humidity data are periodically collected using a temperature and humidity sensor, and the data is preprocessed to obtain the processed data. The temperature and humidity sensor is installed in key parts of the busbar trunking, including the connection end, the enclosed conductor, and areas prone to water accumulation in humid environments. The sensor is used periodically. Temperature and humidity data are collected through sampling. Set at time At what time, the temperature and humidity data are respectively and Formula: ,in It is a real temperature and humidity signal. It is random noise, assumed to have a mean of 0 and a variance of . Gaussian white noise, expressed by the following formula: The data preprocessing includes using Kalman filtering to denoise the temperature and humidity data, resulting in denoised temperature and humidity data as follows: and Grubbs statistic was used to remove outliers from the temperature and humidity data. Temperature data was defined. The mean and standard deviation are expressed by the following formula: ,in This represents the average value of temperature data within a time window. This represents the total number of data points within the statistical time window. Indicates the first Temperature values ​​after noise reduction at each time point This represents summing over all data points. This indicates the degree of fluctuation in temperature data. Indicates the first The difference between a given temperature value and the average temperature. It calculates the square of the temperature deviation from the mean. It is the sum of the squared deviations of all data points. It is the unbiased estimate used when calculating the root mean square error. This involves taking the square root of the result to calculate the Grubbs statistic, expressed as: The threshold for the Grubbs test is: ,in It has degrees of freedom. of Distribution critical value, in The data points are then identified as outliers and removed. The method for removing outliers from humidity data is the same. In step S2, the method for performing mutation detection and trend analysis based on the processed data to obtain mutation data and long-term window data is as follows: The mutation detection includes calculating the temperature and humidity change trend within the most recent 1-3 seconds using short-time window processing, within a specific time frame. Time, Select The temperature and humidity data within the area are used to calculate the rate of change over a short time window, expressed by the following formula: ,in For the rate of temperature change, The rate of change of humidity. A short time window of 1 to 3 seconds. This refers to the short-time window weighting coefficient. A dynamic thresholding method is employed, using historical mean and standard deviation to set an adaptive abrupt change threshold. Discrete wavelet transform is then used to perform multi-scale analysis of the temperature and humidity signal, extracting high-frequency components and identifying abrupt change points. The formula is as follows: ,in These are wavelet transform coefficients. It is a scale factor. It is the translation factor. It is the mother wavelet; In step S2, the method for performing mutation detection and trend analysis based on the processed data to obtain mutation data and long-term window data is as follows: The trend analysis includes using long-term window processing to calculate the temperature and humidity change trends over the most recent 10-60 seconds, selecting past... Calculate the weighted moving average to smooth the trend, expressed by the formula: ,in It is a temperature trend value. This is the humidity trend value. It is the length of the trend analysis window over the past 10-60 seconds. These are long-term window weighting coefficients, satisfying... The formula for expressing the weighting coefficient is: ,in It controls the rate of weight decay. It is an exponentially decaying weight. It is a normalization factor; In step S3, the method for adaptively adjusting the weighting strategy based on the mutation data and the long-term window data is as follows: The weighting strategy includes increasing the weight of the short-time window if either the temperature surge is ≥5℃ or the humidity surge is ≥10%RH in the mutation data, thus defining a mutation response factor. Formula: ,in It is a short-term temperature change. It is a short-term humidity change, in Then the mutation is triggered, and the short-term window weight is updated, expressed by the formula: ,in The adjustment rate for the short-term window is set to a value between 0.1 and 0.

3. It is a moment The short-term window weight, It is the short-term window weight of the previous time step. It is the complement of the short-time window. The long-term window weight is used for long-term trend analysis. When the short-term window weight increases, the long-term window weight decreases accordingly, and vice versa. In step S3, the method for adaptively adjusting the weighting strategy based on the mutation data and the long-term window data is as follows: The weighting strategy includes determining the presence of a trend anomaly in long-term window data if either the temperature continuously rises by ≥3℃ / min or the humidity continuously rises by ≥5%RH / min. This strategy reduces the weight of short-term window data and strengthens long-term window data analysis, defining a trend anomaly factor. Formula: ,in It is the long-term temperature change rate. It is the long-term humidity change rate, in If an abnormal trend is triggered, the weight of the long-term window will be adjusted, expressed by the formula: ,in The trend adjustment factor is set between 0.05 and 0.

2. It is the current moment. Long-term window weights, It is the weight of the long-term window from the previous time step. It is a trend detection factor. It is an adjustment factor, and the weights are updated using an exponential smoothing method. The formula is as follows: ,in The value should be between 0.2 and 0.5 to ensure smooth variation; In step S4, the method for using a time-series forecasting model to analyze historical data and long-term window data and provide forecast data is as follows: Collect temperature and humidity data For short-term forecasting using time series modeling, the formula is as follows: ,in It is the current time. Temperature and humidity data, It represents either the average level or the long-term trend of temperature and humidity data. This represents unpredictable noise at the current time point. This represents how past forecast errors affect current temperature and humidity forecasts. This represents how many past temperature and humidity data points are used for the prediction. It measures the impact of past temperature and humidity data on future forecasts. This represents how many past prediction errors are used in the prediction. It measures the impact of past prediction errors on future predictions and uses a Long Short-Term Memory (LSTM) network for long-term prediction. The output hidden states are used to predict temperature and humidity data for the next few seconds. The formula is as follows: ,in It is the predicted time. Temperature and humidity data at any given time. This represents how many past temperature and humidity data points are used for the prediction. It measures the impact of past temperature and humidity data on future forecasts. Representing time Temperature and humidity values ​​at any given time. This represents how many past prediction errors are used in the prediction. It measures the impact of past forecast errors on future forecasts. This represents the error between the predicted value and the actual value at a past point in time. In step S4, the method for using a time-series forecasting model to analyze historical data and long-term window data and provide forecast data is as follows: A weighted fusion method is used to combine the temperature and humidity data predicted by the Long Short-Term Memory Network and the Temporal Series Prediction Model, expressed as follows: ,in It is a weighting factor. Indicates time The final prediction result of the point It is a time series prediction model in time The prediction results of the points It is the Long Short-Term Memory Network in time The prediction result for a point depends on the error of the Long Short-Term Memory network and the temporal prediction model, expressed by the formula: ,in It is a weighting factor. It is a Long Short-Term Memory network. It is a time series prediction model. It is the base of the natural logarithm, where It is the mean square error; In step S5, the method for detecting anomalies and issuing alarms and warnings by real-time monitoring of the mutation data, long-term window data, and prediction data is as follows: The anomaly detection is based on a triple threshold judgment method, which calculates an anomaly score for each data source, expressed as follows: ,in The weighting coefficients satisfy Used to adjust the contribution of different data sources. It reflects the degree of sudden changes in temperature and humidity. It reflects the degree of long-term trend anomalies in temperature and humidity. This reflects the degree of anomaly in predicted temperature and humidity. The degree of abrupt change in temperature and humidity can be calculated by the rate of change within a short time window, expressed by the formula: ,in This represents the rate of change in temperature and humidity within a short-term window. The standard deviation of historical temperature and humidity variations was used for normalization, and the anomaly of long-term window data was reduced from the past. The rate of change of temperature and humidity is calculated using the following formula: ,in It measures temperature and humidity data over time. The overall trend of change in the area. It is the standard deviation of the long-time window data. It is in time The trend change value represents the rate of change of temperature and humidity data. The prediction anomaly score measures the deviation between the predicted temperature and humidity and the current data. The formula is as follows: ,in The anomaly score measures the degree of deviation between predicted and current temperature and humidity values. The standard deviation of the predicted data is used to normalize the predicted score. It is in time Predicted temperature and humidity values ​​at the location It is the current time. The actual measured values ​​of temperature and humidity at the location. It is the absolute error between the predicted value and the current actual measurement value, based on the comprehensive anomaly score. Set adaptive alarm threshold Formula: ,in It is an adaptive alarm threshold. It is the average of historical anomaly scores. It is the standard deviation of historical outlier scores. The adaptive coefficient determines the alarm sensitivity. The alarm sensitivity includes level one warning, level two warning and level three alarm.

2. A busbar trunking temperature and humidity anomaly monitoring system, based on the busbar trunking temperature and humidity anomaly monitoring method according to any one of claims 1, characterized in that, include: The data acquisition module periodically collects temperature and humidity data through temperature and humidity sensors, and performs data preprocessing to obtain processed data. The detection and analysis module performs mutation detection and trend analysis based on the processed data to obtain mutation data and long-term window data, respectively. The strategy adjustment module adaptively adjusts the weighting strategy based on both mutation data and long-term window data. The time series forecasting module uses a time series forecasting model to analyze historical data and long-term window data, and provides forecast data. The anomaly alarm module detects anomalies and issues alarms and warnings by monitoring the mutation data, long-term window data and prediction data in real time. The process of periodically collecting temperature and humidity data using a temperature and humidity sensor and preprocessing the data to obtain the processed data includes: installing the temperature and humidity sensor at key locations in the busbar trunking, including connection ends, enclosed conductors, and areas prone to water accumulation in humid environments; and periodically collecting temperature and humidity data using a temperature and humidity sensor. Temperature and humidity data are collected through sampling. Set at time At what time, the temperature and humidity data are respectively and Formula: ,in It is a real temperature and humidity signal. It is random noise, assumed to have a mean of 0 and a variance of . Gaussian white noise, expressed by the following formula: The data preprocessing includes using Kalman filtering to denoise the temperature and humidity data, resulting in denoised temperature and humidity data as follows: and Grubbs statistic was used to remove outliers from the temperature and humidity data. Temperature data was defined. The mean and standard deviation are expressed by the following formula: ,in This represents the average value of temperature data within a time window. This represents the total number of data points within the statistical time window. Indicates the first Temperature values ​​after noise reduction at each time point This represents summing over all data points. This indicates the degree of fluctuation in temperature data. Indicates the first The difference between a given temperature value and the average temperature. It calculates the square of the temperature deviation from the mean. It is the sum of the squared deviations of all data points. It is the unbiased estimate used when calculating the root mean square error. This involves taking the square root of the result to calculate the Grubbs statistic, expressed as: The threshold for the Grubbs test is: ,in It has degrees of freedom. of Distribution critical value, in The data points are then identified as outliers and removed. The method for removing outliers from humidity data is the same. The operational process for performing mutation detection and trend analysis on the processed data to obtain mutation data and long-time window data includes: mutation detection involves calculating the temperature and humidity change trend within the most recent 1-3 seconds using short-time window processing, and analyzing the time frame... Time, Select The temperature and humidity data within the area are used to calculate the rate of change over a short time window, expressed by the following formula: ,in For the rate of temperature change, The rate of change of humidity. A short time window of 1 to 3 seconds. This refers to the short-time window weighting coefficient. A dynamic thresholding method is employed, using historical mean and standard deviation to set an adaptive abrupt change threshold. Discrete wavelet transform is then used to perform multi-scale analysis of the temperature and humidity signal, extracting high-frequency components and identifying abrupt change points. The formula is as follows: ,in These are wavelet transform coefficients. It is a scale factor. It is the translation factor. It is the mother wavelet; The trend analysis includes using long-term window processing to calculate the temperature and humidity change trends over the most recent 10-60 seconds, selecting past... Calculate the weighted moving average to smooth the trend, expressed by the formula: ,in It is a temperature trend value. This is the humidity trend value. It is the length of the trend analysis window over the past 10-60 seconds. These are long-term window weighting coefficients, satisfying... The formula for expressing the weighting coefficient is: ,in It controls the rate of weight decay. It is an exponentially decaying weight. It is a normalization factor; The operational procedure for adaptively adjusting the weighting strategy based on mutation data and long-term window data includes: if either a temperature surge ≥5℃ or a humidity surge ≥10%RH occurs in the mutation data, the short-term window weight is increased, and a mutation response factor is defined. Formula: ,in It is a short-term temperature change. It is a short-term humidity change, in Then the mutation is triggered, and the short-term window weight is updated, expressed by the formula: ,in The adjustment rate for the short-term window is set to a value between 0.1 and 0.

3. It is a moment The short-term window weight, It is the short-term window weight of the previous time step. It is the complement of the short-time window. The long-term window weight is used for long-term trend analysis. When the short-term window weight increases, the long-term window weight decreases accordingly, and vice versa. The weighting strategy includes determining the presence of a trend anomaly in long-term window data if either the temperature continuously rises by ≥3℃ / min or the humidity continuously rises by ≥5%RH / min. This strategy reduces the weight of short-term window data and strengthens long-term window data analysis, defining a trend anomaly factor. Formula: ,in It is the long-term temperature change rate. It is the long-term humidity change rate, in If an abnormal trend is triggered, the weight of the long-term window will be adjusted, expressed by the formula: ,in The trend adjustment factor is set between 0.05 and 0.

2. It is the current moment. Long-term window weights, It is the weight of the long-term window from the previous time step. It is a trend detection factor. It is an adjustment factor, and the weights are updated using an exponential smoothing method. The formula is as follows: ,in The value should be between 0.2 and 0.5 to ensure smooth variation; The time-series forecasting model analyzes historical and long-term window data, and provides an operational workflow for forecasting data, including: collecting temperature and humidity data. For short-term forecasting using time series modeling, the formula is as follows: ,in It is the current time. Temperature and humidity data, It represents either the average level or the long-term trend of temperature and humidity data. This represents unpredictable noise at the current time point. This represents how past forecast errors affect current temperature and humidity forecasts. This represents how many past temperature and humidity data points are used for the prediction. It measures the impact of past temperature and humidity data on future forecasts. This represents how many past prediction errors are used in the prediction. It measures the impact of past prediction errors on future predictions and uses a Long Short-Term Memory (LSTM) network for long-term prediction. The output hidden states are used to predict temperature and humidity data for the next few seconds. The formula is as follows: ,in It is the predicted time. Temperature and humidity data at any given time. This represents how many past temperature and humidity data points are used for the prediction. It measures the impact of past temperature and humidity data on future forecasts. Representing time Temperature and humidity values ​​at any given time. This represents how many past prediction errors are used in the prediction. It measures the impact of past forecast errors on future forecasts. This represents the error between the predicted value and the actual value at a past point in time. A weighted fusion method is used to combine the temperature and humidity data predicted by the Long Short-Term Memory Network and the Temporal Series Prediction Model, expressed as follows: ,in It is a weighting factor. Indicates time The final prediction result of the point It is a time series prediction model in time The prediction results of the points It is the Long Short-Term Memory Network in time The prediction result for a point depends on the error of the Long Short-Term Memory network and the temporal prediction model, expressed by the formula: ,in It is a weighting factor. It is a Long Short-Term Memory network. It is a time series prediction model. It is the base of the natural logarithm, where It is the mean square error; The operational process for anomaly detection and alarm / early warning by real-time monitoring of mutation data, long-term window data, and predicted data includes: the anomaly detection is based on a triple threshold judgment method, calculating an anomaly score for each data source, expressed as: ,in The weighting coefficients satisfy Used to adjust the contribution of different data sources. It reflects the degree of sudden changes in temperature and humidity. It reflects the degree of long-term trend anomalies in temperature and humidity. This reflects the degree of anomaly in predicted temperature and humidity. The degree of abrupt change in temperature and humidity can be calculated by the rate of change within a short time window, expressed by the formula: ,in This represents the rate of change in temperature and humidity within a short-term window. The standard deviation of historical temperature and humidity variations was used for normalization, and the anomaly of long-term window data was reduced from the past. The rate of change of temperature and humidity is calculated using the following formula: ,in It measures temperature and humidity data over time. The overall trend of change in the area. It is the standard deviation of the long-time window data. It is in time The trend change value represents the rate of change of temperature and humidity data. The prediction anomaly score measures the deviation between the predicted temperature and humidity and the current data. The formula is as follows: ,in The anomaly score measures the degree of deviation between predicted and current temperature and humidity values. The standard deviation of the predicted data is used to normalize the predicted score. It is in time Predicted temperature and humidity values ​​at the location It is the current time. The actual measured values ​​of temperature and humidity at the location. It is the absolute error between the predicted value and the current actual measurement value, based on the comprehensive anomaly score. Set adaptive alarm threshold Formula: ,in It is an adaptive alarm threshold. It is the average of historical anomaly scores. It is the standard deviation of historical outlier scores. The adaptive coefficient determines the alarm sensitivity. The alarm sensitivity includes level one warning, level two warning and level three alarm.

Images