A kind of early warning method and early warning system of power monitoring index

Abstract

The present invention discloses an early warning method and early warning system for power monitoring indicators, and relates to the technical field of data processing; the index data of the power monitoring system is obtained at a first frequency, the index data is preprocessed to obtain a standard index, and a comprehensive score is calculated according to the standard index; the comprehensive score is sampled at a second frequency to obtain a score set, the comprehensive score in the score set with a preset number as a group is used as a target group, and the multiple target groups are sequentially tested for deviation from a characteristic curve; if the deviation between the target group and the characteristic curve is higher than the preset deviation, the target group is determined to be abnormal data, the second frequency is increased to obtain a third frequency, the comprehensive score is sampled at the third frequency, and an early warning is issued if the number of occurrences of abnormal data within a preset period exceeds an early warning threshold. Combined with frequency adjustment and abnormality detection mechanisms, potential risks can be effectively identified and warned, and maintenance costs and risks can be reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing, and in particular relates to an early warning method and an early warning system for electric power monitoring indicators. Background Art

[0002] Transformers are indispensable equipment in power systems, mainly used for AC voltage devices to meet the needs of different power equipment. The background is that the transmission and distribution of voltage in power systems require different levels of voltage, and transformers are key equipment to achieve this conversion, ensuring that electricity can be efficiently and safely transmitted and used.

[0003] When a transformer fails, a fault indicator will send a signal. The indicator will send out an obvious signal, such as a flashing red light or a sound, so that the staff can find and deal with it in time. The staff will regularly inspect and maintain the transformer. If an abnormal situation or failure is found, the staff can immediately repair it. Regular inspections can find some potential problems and deal with them in time to avoid failures. The transformer's sensor detects abnormal data and automatically warns.

[0004] Although the above method can solve the faults of some transformers, the efficiency of real-time monitoring of transformer power indicators is low because of the large amount of power indicator data resulting in large amount of calculation. Summary of the invention

[0005] The purpose of the present invention is to solve the problem of low efficiency in real-time monitoring of transformer power indicators, and to propose an early warning method and early warning system for power monitoring indicators.

[0006] In a first aspect of the present invention, a method for early warning of power monitoring indicators is first proposed, the method comprising:

[0007] Optionally, the method comprises:

[0008] Acquire index data of the power monitoring system at a first frequency, preprocess the index data to obtain standard indexes, and calculate a comprehensive score based on the standard indexes;

[0009] The comprehensive score is sampled at a second frequency to obtain a score set, the comprehensive scores in the score set grouped by a preset number are used as target groups, and the multiple target groups are sequentially tested for deviation from the characteristic curve; the characteristic curve is obtained by obtaining a preset number of normal operating indicator data in a historical database, calculating the historical indicator data by a weight formula to obtain a characteristic score, and fitting the characteristic score to obtain the characteristic curve; the first frequency is greater than the second frequency;

[0010] If the deviation between the target group and the characteristic curve is higher than the preset deviation, the target group is determined to be abnormal data, the second frequency is increased to obtain the third frequency, and the comprehensive score is sampled at the third frequency. If the number of occurrences of abnormal data within the preset period exceeds the warning threshold, a warning is issued.

[0011] Optionally, the index data includes: temperature, current, transformer internal magnetic field and external magnetic field; the standard index includes: standard magnetic field index, standard temperature index and standard current index; the standard index obtained by preprocessing the index data includes:

[0012] The first indicator data is obtained by performing data cleaning on the indicator data to filter out abnormal values, and the first temperature indicator and the first current indicator in the first indicator data are normalized to obtain a standard temperature indicator and a standard current indicator;

[0013] The first internal magnetic field and the first external magnetic field in the first index data are converted from the time domain to the frequency domain, the center frequency of all magnetic fields in the frequency domain is determined, the peak closest to the first external magnetic field in the first internal magnetic field is eliminated to obtain a corrected magnetic field index, the corrected magnetic field index is converted from the time domain back to the frequency domain to obtain a second magnetic field index, the second magnetic field index is normalized to obtain a standard magnetic field index.

[0014] Optionally, the method for calculating a comprehensive score based on the standard index further includes:

[0015] The standard data is calculated through a weight formula to obtain a comprehensive score;

[0016] Weight formula:

[0017]

[0018] Among them, W represents the comprehensive score, P T Indicates standard temperature index, P I Indicates the standard current index, P C Indicates standard magnetic field index.

[0019] Optionally, the early warning threshold is greater than the alert threshold, and the method of increasing the second frequency to obtain the third frequency includes:

[0020] If the number of occurrences of abnormal data within the preset period does not exceed the warning threshold, the third frequency returns to the second frequency sampling, and the corrector is started to correct the sensor parameters.

[0021] Optionally, if the number of occurrences of abnormal data within a preset period exceeds a warning threshold, the method may further include:

[0022] Acquire a transformer image taken by an infrared thermal imager, perform noise reduction on the transformer image by using a median filtering algorithm to obtain a first image, and perform a contrast-limited adaptive histogram equalization algorithm on the first image to enhance detail information of the image to obtain a target image;

[0023] The pixel value f at the pixel point (x, y) at time t on the target image t (x, y) is matched with the mean of a preset number of Gaussian models through a matching formula, and if the pixel value of the target image fails to match the background model, it is determined to be a foreground point;

[0024] Identify the image area containing the foreground point, calculate the edge complexity of the image area, and if the edge complexity exceeds a preset value, determine that the image area is a fire point;

[0025] If there is a fire point in the target image and the number of occurrences of abnormal data within a preset period exceeds the warning threshold, it is determined that there is a fire risk and a warning is issued;

[0026] Edge complexity calculation formula:

[0027] Among them, A i is the edge pixel f t The curvature at (x, y), B is the mean curvature, N is the total number of edge pixels, and K represents the complexity.

[0028] In a second aspect of the present invention, an early warning system for power monitoring indicators is provided, comprising: the device comprising:

[0029] Optionally, the device includes a preprocessing module, a deviation detection module, and an early warning module:

[0030] The preprocessing module is used to obtain index data of the power monitoring system at a first frequency, preprocess the index data to obtain standard indexes, and calculate a comprehensive score based on the standard indexes;

[0031] The deviation detection module is used to sample the comprehensive score at a second frequency to obtain a score set, take the comprehensive scores in the score set in groups of a preset number as target groups, and sequentially perform deviation detection on the multiple target groups with a characteristic curve; the characteristic curve is obtained by obtaining a preset number of normal operation indicator data in a historical database, calculating the historical indicator data by a weight formula to obtain a characteristic score, and fitting the characteristic score to obtain a characteristic curve; the first frequency is greater than the second frequency;

[0032] The early warning module is used to determine that the target group is abnormal data if the deviation between the target group and the characteristic curve is higher than a preset deviation, increase the second frequency to obtain a third frequency, sample the comprehensive score with the third frequency, and issue an early warning if the number of occurrences of abnormal data within a preset period exceeds a warning threshold.

[0033] Optionally, the preprocessing module includes an index processing module and a magnetic field module:

[0034] The indicator processing module is used to obtain first indicator data by performing data cleaning on the indicator data to filter out abnormal values, and to perform normalization processing on the first temperature indicator and the first current indicator in the first indicator data to obtain a standard temperature indicator and a standard current indicator;

[0035] The magnetic field module is used to convert the first internal magnetic field and the first external magnetic field in the first indicator data from the time domain to the frequency domain, determine the center frequency of all magnetic fields in the frequency domain, eliminate the peak closest to the first external magnetic field in the first internal magnetic field to obtain a corrected magnetic field index, convert the corrected magnetic field index from the time domain back to the frequency domain to obtain a second magnetic field index, and normalize the second magnetic field index to obtain a standard magnetic field index.

[0036] Optionally, the preprocessing module includes a weight calculation module:

[0037] The weight calculation module is used to calculate the standard data through the weight formula to obtain a comprehensive score;

[0038] Weight formula:

[0039]

[0040] Among them, W represents the comprehensive score, P T Indicates standard temperature index, P I Indicates the standard current index, P C Indicates standard magnetic field index.

[0041] Optionally, the early warning module includes a warning module:

[0042] The warning module is used to return the third frequency to the second frequency for sampling if the number of occurrences of abnormal data within a preset period does not exceed the warning threshold, and start the corrector to correct the sensor parameters.

[0043] Optionally, the early warning module further includes an image processing module, an image judgment module, a risk judgment module and a fire risk module:

[0044] The image processing module is used to obtain a transformer image taken by an infrared thermal imager, perform noise reduction on the transformer image by a median filtering algorithm to obtain a first image, and perform a contrast-limited adaptive histogram equalization algorithm on the first image to enhance detail information of the image to obtain a target image;

[0045] The image judgment module is used to convert the pixel value f at the pixel point (x, y) at time t on the target image t (x, y) is matched with the mean of a preset number of Gaussian models through a matching formula, and if the pixel value of the target image fails to match the background model, it is determined to be a foreground point;

[0046] The risk judgment module is used to identify the image area containing the foreground point, calculate the edge complexity of the image area, and if the edge complexity exceeds a preset value, judge the image area as a fire point;

[0047] The fire risk module is used to determine that there is a fire risk and issue an early warning if there is a fire point in the target image and the number of occurrences of abnormal data within a preset period exceeds the early warning threshold;

[0048] Edge complexity calculation formula:

[0049] Among them, A i is the edge pixel f t The curvature at (x, y), B is the mean curvature, N is the total number of edge pixels, and K represents the complexity.

[0050] Beneficial effects of the present invention:

[0051] The present invention proposes an early warning method for power monitoring indicators, which obtains the indicator data of the power monitoring system at a first frequency, pre-processes the indicator data to obtain a standard indicator, calculates a comprehensive score according to the standard indicator, and enters the comprehensive score into a scoring queue in chronological order; the comprehensive score in the scoring queue is dequeued at a first frequency, and the comprehensive score dequeued in the scoring queue is sampled at a second frequency and entered into a target queue, and the comprehensive scores of the target queue with a preset number of dequeued are used as target groups, and the multiple target groups are sequentially tested for deviation from the characteristic curve; if the deviation between the target group and the characteristic curve is higher than the preset deviation, the target group is determined to be abnormal data, the second frequency is increased to obtain a third frequency, and the comprehensive score dequeued in the scoring queue is sampled at the third frequency and entered into the target queue, and an early warning is issued if the number of occurrences of abnormal data within a preset period exceeds the early warning threshold. Real-time monitoring of the power system, combined with frequency adjustment and abnormality detection mechanisms, effectively identifies and warns of potential risks, improves the stability and reliability of the transformer, and reduces maintenance costs and risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The present invention will be further described below in conjunction with the accompanying drawings.

[0053] Figure 1 A flowchart of an early warning method for power monitoring indicators is provided for an embodiment of the present invention;

[0054] Figure 2 A flowchart of another early warning method for power monitoring indicators is provided for an embodiment of the present invention;

[0055] Figure 3 A schematic diagram of an early warning method for power monitoring indicators is provided for an embodiment of the present invention;

[0056] Figure 4 A schematic structural diagram of an early warning system for power monitoring indicators is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the present invention, the description of "first", "second", etc. is only used for descriptive purposes, and cannot be understood as indicating or implying its relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" can explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0058] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of the present invention.

[0059] The embodiment of the present invention provides an early warning method for power monitoring indicators. Figure 1 , Figure 1 A flowchart of an early warning method for power monitoring indicators provided by an embodiment of the present invention. The method comprises the following steps:

[0060] S101, acquiring index data of a power monitoring system at a first frequency, preprocessing the index data to obtain standard indexes, and calculating a comprehensive score based on the standard indexes.

[0061] S102, sampling the comprehensive scores at the second frequency to obtain a score set, taking the comprehensive scores in the score set in groups of a preset number as target groups, and performing deviation detection on the multiple target groups and the characteristic curve in turn.

[0062] S103, if the deviation between the target group and the characteristic curve is higher than the preset deviation, the target group is determined to be abnormal data, the second frequency is increased to obtain the third frequency, and the comprehensive score is sampled at the third frequency. If the number of occurrences of abnormal data within the preset period exceeds the warning threshold, a warning is issued.

[0063] The characteristic curve is obtained by obtaining a preset number of normal operating indicator data in the historical database, calculating the historical indicator data through a weight formula to obtain a characteristic score, and fitting the characteristic score to obtain the characteristic curve; the first frequency is greater than the second frequency.

[0064] An early warning method for power monitoring indicators provided in an embodiment of the present invention pre-processes indicator data, accurately calculates a comprehensive score, and combines frequency adjustment and anomaly detection mechanisms to effectively identify and warn of potential risks, thereby improving the stability and reliability of the transformer and reducing maintenance costs and risks.

[0065] In one implementation, the comprehensive scores of a preset number of out-of-queue targets in the target queue are taken as target groups, and multiple target groups are tested for deviation from the characteristic curve in turn. For example, three comprehensive scores in the target queue are tested for deviation from the characteristic curve. The characteristic curve obtains the characteristic score by calculation through the indicator data of normal operation in three historical databases, and fits it into a characteristic curve. Each target group can be tested according to the position on the corresponding characteristic curve, and each target group is tested in turn, which is conducive to ensuring data retrieval accuracy and improving operation efficiency.

[0066] In one implementation, the comprehensive score is sampled at the second frequency to obtain a score set and a target group is obtained by grouping according to a preset number. If the target group is determined to be abnormal data, the second frequency is increased to obtain a third frequency, and the comprehensive score is sampled at the third frequency to obtain a score set and a target group is obtained by grouping according to a preset number, and deviation detection is continued. This method can balance the efficiency of real-time data monitoring and the rate of subsequent analysis and processing.

[0067] In one embodiment, step S101 includes:

[0068] The first indicator data is obtained by cleaning the indicator data to filter out abnormal values, and the first temperature indicator and the first current indicator in the first indicator data are normalized to obtain a standard temperature indicator and a standard current indicator.

[0069] The first internal magnetic field and the first external magnetic field in the first indicator data are converted from the time domain to the frequency domain, the center frequency of all magnetic fields in the frequency domain is determined, the peak closest to the first external magnetic field in the first internal magnetic field is eliminated to obtain a corrected magnetic field index, the corrected magnetic field index is converted from the time domain back to the frequency domain to obtain a second magnetic field index, the second magnetic field index is normalized to obtain a standard magnetic field index.

[0070] In one implementation, data cleaning is performed to ensure the accuracy and reliability of indicator data, avoid adverse effects of outliers on subsequent analysis and evaluation, and thus improve the credibility of standard indicators.

[0071] In one implementation, indicator data of different dimensions are converted into a unified standard through normalization processing to facilitate subsequent comprehensive score calculation.

[0072] In one embodiment, step S101 further includes:

[0073] The standard data is calculated using a weighted formula to obtain a comprehensive score.

[0074] Weight formula:

[0075]

[0076] Among them, W represents the comprehensive score, P T Indicates standard temperature index, P I Indicates the standard current index, P C Indicates standard magnetic field index.

[0077] In one implementation, the impact of multiple indicators on transformer performance is comprehensively considered to more comprehensively evaluate the transformer status.

[0078] In one embodiment, see Figure 2 ,exist Figure 1 On the basis of this, step S103 further includes:

[0079] S104: If the number of occurrences of abnormal data within the preset period does not exceed the warning threshold, the third frequency returns to the second frequency sampling, and the corrector is started to correct the sensor parameters.

[0080] In one implementation, see Figure 3 , Figure 3A schematic diagram of an early warning method for power monitoring indicators is provided for an embodiment of the present invention, wherein power indicator data is obtained at a first frequency, and the first frequency is obtained once per second, for example; the indicator data is preprocessed and calculated to obtain a comprehensive score, and the comprehensive score is sampled at a second frequency to obtain a score set, and the second frequency is obtained, for example, once every five seconds; the score set is grouped according to a preset number and a deviation detection is performed, and if the deviation is higher than the preset deviation, the second frequency is increased to obtain a third frequency, and the third frequency is obtained, for example, once every three seconds; if after the frequency is increased, abnormal data within a preset period exceeds the early warning threshold, an early warning is issued, and if it does not exceed the preset period, for example within ten seconds, the third frequency returns to the second frequency and continues to cycle; if the deviation is not higher than the preset deviation, the cycle continues at the second frequency; the first frequency is greater than or equal to the third frequency and greater than the second frequency, the data processing amount is reduced, the detection efficiency is improved, the obtained indicator data is time-sensitive, and can reflect the real-time changes of transformer indicators.

[0081] In one implementation, the sampling frequency is changed to achieve accurate prediction of the transformer. If the number of occurrences of abnormal data within a preset period does not exceed the warning threshold, the third frequency returns to the second frequency sampling, which can effectively reduce the amount of calculation.

[0082] In one implementation, while ensuring the stability of the early warning system, sensor errors are corrected in a timely manner to ensure the accuracy of collected data.

[0083] In one embodiment, step S103 further includes:

[0084] The transformer image taken by the infrared thermal imager is obtained, and the transformer image is denoised by a median filtering algorithm to obtain a first image. The first image is subjected to a contrast-limited adaptive histogram equalization algorithm to enhance the detail information of the image to obtain a target image.

[0085] The pixel value f at the pixel point (x, y) at time t on the target image t (x, y) is matched with the mean of a preset number of Gaussian models through a matching formula. If the pixel value of the target image fails to match the background model, it is determined to be a foreground point.

[0086] Identify the image area containing the foreground point, calculate the edge complexity of the image area, and if the edge complexity exceeds a preset value, determine that the image area is a fire point.

[0087] If there are fire spots in the target image and the number of occurrences of abnormal data within a preset period exceeds the warning threshold, it is judged that there is a fire risk and an early warning is issued.

[0088] Edge complexity calculation formula:

[0089] Among them, A i is the edge pixel ft The curvature at (x, y), B is the mean curvature, N is the total number of edge pixels, and K represents the complexity.

[0090] In one implementation, the irregular changes of the edge of the flame image are quantified by edge complexity. The greater the complexity, the more complex the image edge is beyond a preset value, and it is easy to determine whether the image area is a fire point.

[0091] In one implementation, the foreground points in the target image are accurately identified, which improves the accuracy of detecting abnormal areas in the image and provides an important basis for further fault analysis and early warning.

[0092] In one implementation, there is a fire point and the number of occurrences of abnormal data within a preset period exceeds the warning threshold. Through the joint analysis of the two types of data, the early warning system can reduce the misjudgment of people or other interfering objects.

[0093] Based on the same inventive concept, the embodiment of the present invention also provides an early warning system for power monitoring indicators. Figure 4 , Figure 4 A schematic diagram of the structure of an early warning system for power monitoring indicators provided by an embodiment of the present invention includes:

[0094] The preprocessing module is used to obtain index data of the power monitoring system at a first frequency, preprocess the index data to obtain standard indexes, and calculate a comprehensive score based on the standard indexes.

[0095] The deviation detection module is used to sample the comprehensive score at the second frequency to obtain a score set, take the comprehensive score grouped by a preset number in the score set as the target group, and perform deviation detection on multiple target groups in turn with the characteristic curve; the characteristic curve is to obtain the indicator data of the normal operation of a preset number in the historical database, calculate the historical indicator data through a weight formula to obtain the characteristic score, and fit the characteristic score to obtain the characteristic curve; the first frequency is greater than the second frequency.

[0096] The early warning module is used to determine that the target group is abnormal data if the deviation between the target group and the characteristic curve is higher than the preset deviation, increase the second frequency to obtain the third frequency, sample the comprehensive score with the third frequency, and issue an early warning if the number of occurrences of abnormal data within the preset period exceeds the early warning threshold.

[0097] An early warning system for power monitoring indicators provided in an embodiment of the present invention can effectively identify and warn of potential risks by preprocessing indicator data, accurately calculating comprehensive scores, and combining frequency adjustment and anomaly detection mechanisms, thereby improving the stability and reliability of transformers and reducing maintenance costs and risks.

[0098] In one embodiment, the preprocessing module includes an index processing module and a magnetic field module:

[0099] The indicator processing module is used to obtain first indicator data by performing data cleaning on the indicator data and filtering out abnormal values, and to normalize the first temperature indicator and the first current indicator in the first indicator data to obtain a standard temperature indicator and a standard current indicator.

[0100] The magnetic field module is used to convert the first internal magnetic field and the first external magnetic field in the first indicator data from the time domain to the frequency domain, determine the center frequency of all magnetic fields in the frequency domain, eliminate the peak closest to the first external magnetic field in the first internal magnetic field to obtain a corrected magnetic field index, convert the corrected magnetic field index from the time domain back to the frequency domain to obtain a second magnetic field index, and normalize the second magnetic field index to obtain a standard magnetic field index.

[0101] In one embodiment, the preprocessing module includes a weight calculation module:

[0102] The weight calculation module is used to calculate the risk parameters through the weight formula to obtain the comprehensive score of the indicator data.

[0103] Weight formula:

[0104]

[0105] Among them, W represents the comprehensive score, P T Indicates standard temperature index, P I Indicates the standard current index, P C Indicates standard magnetic field index.

[0106] In one embodiment, the early warning module includes an alert module:

[0107] The warning module is used to return the third frequency to the second frequency sampling if the number of occurrences of abnormal data within a preset period does not exceed the warning threshold, and start the corrector to correct the sensor parameters.

[0108] In one embodiment, the early warning module further includes an image processing module, an image judgment module, a risk judgment module and a fire risk module:

[0109] The image processing module is used to obtain the transformer image taken by the infrared thermal imager, reduce the noise of the transformer image by a median filtering algorithm to obtain a first image, and enhance the detail information of the image by a limited contrast adaptive histogram equalization algorithm to obtain a target image.

[0110] The image judgment module is used to convert the pixel value f at the pixel point (x, y) at time t on the target image t (x, y) is matched with the mean of a preset number of Gaussian models through a matching formula. If the pixel value of the target image fails to match the background model, it is determined to be a foreground point.

[0111] The risk judgment module is used to identify the image area containing the foreground point and calculate the edge complexity of the image area. If the edge complexity exceeds a preset value, the image area is judged to be a fire point.

[0112] The fire risk module is used to determine the existence of a fire risk and issue a warning if there is a fire point in the target image and the number of occurrences of abnormal data within a preset period exceeds the warning threshold.

[0113] Edge complexity calculation formula:

[0114] Among them, A i is the edge pixel f t The curvature at (x, y), B is the mean curvature, N is the total number of edge pixels, and K represents the complexity.

[0115] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. An early warning method for power monitoring indicators, characterized in that: The method comprises: Acquire index data of the power monitoring system at a first frequency, preprocess the index data to obtain standard indexes, and calculate a comprehensive score based on the standard indexes; The comprehensive score is sampled at a second frequency to obtain a score set, the comprehensive scores in the score set grouped by a preset number are used as target groups, and the multiple target groups are sequentially tested for deviation from the characteristic curve; the characteristic curve is obtained by obtaining a preset number of normal operating indicator data in a historical database, calculating the historical indicator data by a weight formula to obtain a characteristic score, and fitting the characteristic score to obtain the characteristic curve; the first frequency is greater than the second frequency; If the deviation between the target group and the characteristic curve is higher than the preset deviation, the target group is determined to be abnormal data, the second frequency is increased to obtain a third frequency, and the third frequency is used to sample and score the comprehensive score. If the number of occurrences of abnormal data within a preset period exceeds the warning threshold, a warning is issued; If the number of occurrences of abnormal data within a preset period exceeds the warning threshold, a warning is issued, and the method further includes: Acquire a transformer image taken by an infrared thermal imager, perform noise reduction on the transformer image by using a median filtering algorithm to obtain a first image, and perform a contrast-limited adaptive histogram equalization algorithm on the first image to enhance detail information of the image to obtain a target image; The pixel value at the pixel point (x, y) at time t on the target image Matching with the mean of a preset number of Gaussian models through a matching formula, if the pixel value of the target image fails to match with the background model, it is determined to be a foreground point; Identify an image area containing a foreground point, calculate the edge complexity of the image area, and if the edge complexity exceeds a preset value, determine that the image area is a fire point; If there is a fire point in the target image, and the number of occurrences of abnormal data within a preset period exceeds the warning threshold, it is determined that there is a fire risk and a warning is issued; Edge complexity calculation formula: in, It is the edge pixel B is the mean curvature, N is the total number of edge pixels, and K represents the complexity.

2. The early warning method of power monitoring indicators according to claim 1 is characterized in that: The index data include: temperature, current, transformer internal magnetic field and external magnetic field; the standard index includes: standard magnetic field index, standard temperature index and standard current index; the standard index obtained by preprocessing the index data includes: The first indicator data is obtained by performing data cleaning on the indicator data to filter out abnormal values, and the first temperature indicator and the first current indicator in the first indicator data are normalized to obtain a standard temperature indicator and a standard current indicator; The first internal magnetic field and the first external magnetic field in the first index data are converted from the time domain to the frequency domain, the center frequency of all magnetic fields in the frequency domain is determined, the peak closest to the first external magnetic field in the first internal magnetic field is eliminated to obtain a corrected magnetic field index, the corrected magnetic field index is converted from the time domain back to the frequency domain to obtain a second magnetic field index, the second magnetic field index is normalized to obtain a standard magnetic field index.

3. The early warning method of power monitoring indicators according to claim 1 is characterized in that: The method for calculating a comprehensive score based on the standard index also includes: The standard data is calculated through a weight formula to obtain a comprehensive score; Weight formula: in, Represents the comprehensive score, Indicates the standard temperature index, Indicates the standard current index, Indicates standard magnetic field index.

4. The early warning method of power monitoring indicators according to claim 1 is characterized in that: The early warning threshold is greater than the warning threshold, and the second frequency is increased to obtain the third frequency, which includes: If the number of occurrences of abnormal data within the preset period does not exceed the warning threshold, the third frequency returns to the second frequency sampling, and the corrector is started to correct the sensor parameters.

5. An early warning system for power monitoring indicators, characterized in that: The system includes a preprocessing module, a deviation detection module, and an early warning module: The preprocessing module is used to obtain index data of the power monitoring system at a first frequency, preprocess the index data to obtain standard indexes, and calculate a comprehensive score based on the standard indexes; The deviation detection module is used to sample the comprehensive score at a second frequency to obtain a score set, take the comprehensive scores in the score set in groups of a preset number as target groups, and sequentially perform deviation detection on the multiple target groups with a characteristic curve; the characteristic curve is obtained by obtaining a preset number of normal operation indicator data in a historical database, calculating the historical indicator data by a weight formula to obtain a characteristic score, and fitting the characteristic score to obtain a characteristic curve; the first frequency is greater than the second frequency; The early warning module is used to determine that the target group is abnormal data if the deviation between the target group and the characteristic curve is higher than a preset deviation, increase the second frequency to obtain a third frequency, sample the comprehensive score with the third frequency, and issue an early warning if the number of occurrences of abnormal data within a preset period exceeds an early warning threshold; The early warning module also includes an image processing module, an image judgment module, a risk judgment module and a fire risk module: The image processing module is used to obtain a transformer image taken by an infrared thermal imager, perform noise reduction on the transformer image by a median filtering algorithm to obtain a first image, and perform a contrast-limited adaptive histogram equalization algorithm on the first image to enhance detail information of the image to obtain a target image; The image judgment module is used to convert the pixel value at the pixel point (x, y) at time t on the target image Matching with the mean of a preset number of Gaussian models through a matching formula, if the pixel value of the target image fails to match with the background model, it is determined to be a foreground point; The risk judgment module is used to identify the image area containing the foreground point, calculate the edge complexity of the image area, and if the edge complexity exceeds a preset value, judge the image area as a fire point; The fire risk module is used to determine that there is a fire risk and issue an early warning if there is a fire point in the target image and the number of occurrences of abnormal data within a preset period exceeds the early warning threshold; Edge complexity calculation formula: in, It is the edge pixel B is the mean curvature, N is the total number of edge pixels, and K represents the complexity.

6. The early warning system for power monitoring indicators according to claim 5 is characterized in that: The preprocessing module includes an index processing module and a magnetic field module: The indicator processing module is used to obtain first indicator data by performing data cleaning on the indicator data to filter out abnormal values, and to normalize the first temperature indicator and the first current indicator in the first indicator data to obtain a standard temperature indicator and a standard current indicator; The magnetic field module is used to convert the first internal magnetic field and the first external magnetic field in the first indicator data from the time domain to the frequency domain, determine the center frequency of all magnetic fields in the frequency domain, eliminate the peak closest to the first external magnetic field in the first internal magnetic field to obtain a corrected magnetic field index, convert the corrected magnetic field index from the time domain back to the frequency domain to obtain a second magnetic field index, and normalize the second magnetic field index to obtain a standard magnetic field index.

7. The early warning system for power monitoring indicators according to claim 5, characterized in that: The preprocessing module includes a weight calculation module: The weight calculation module is used to calculate the standard data through the weight formula to obtain a comprehensive score; Weight formula: in, Represents the comprehensive score, Indicates the standard temperature index, Indicates the standard current index, Indicates standard magnetic field index.

8. The early warning system for power monitoring indicators according to claim 5, characterized in that: The early warning module includes a warning module: The warning module is used to return the third frequency to the second frequency for sampling if the number of occurrences of abnormal data within a preset period does not exceed the warning threshold, and start the corrector to correct the sensor parameters.

Images