CN116612121B - Transformer fault rapid detection method based on artificial intelligence - Google Patents
Transformer fault rapid detection method based on artificial intelligence Download PDFInfo
- Publication number
- CN116612121B CN116612121B CN202310891162.8A CN202310891162A CN116612121B CN 116612121 B CN116612121 B CN 116612121B CN 202310891162 A CN202310891162 A CN 202310891162A CN 116612121 B CN116612121 B CN 116612121B
- Authority
- CN
- China
- Prior art keywords
- edges
- edge
- area
- degree
- infrared image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 30
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 17
- 239000011229 interlayer Substances 0.000 claims abstract description 114
- 230000005856 abnormality Effects 0.000 claims abstract description 29
- 238000000034 method Methods 0.000 claims description 26
- 230000002159 abnormal effect Effects 0.000 claims description 20
- 238000004364 calculation method Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 9
- 239000006185 dispersion Substances 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000003708 edge detection Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 6
- 230000000739 chaotic effect Effects 0.000 claims description 4
- 230000003321 amplification Effects 0.000 claims description 3
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 3
- 238000010191 image analysis Methods 0.000 abstract description 4
- 238000004804 winding Methods 0.000 description 9
- 238000010606 normalization Methods 0.000 description 7
- 230000008859 change Effects 0.000 description 4
- 239000003086 colorant Substances 0.000 description 4
- 239000010410 layer Substances 0.000 description 4
- 238000013507 mapping Methods 0.000 description 4
- 230000005855 radiation Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000009434 installation Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000009795 derivation Methods 0.000 description 2
- 230000010365 information processing Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 239000011810 insulating material Substances 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Abstract
The invention relates to the technical field of image analysis, in particular to a transformer fault rapid detection method based on artificial intelligence, which comprises the steps of determining one type of edge, two types of edges and three types of edges to be selected at different positions through image analysis of an infrared image of a transformer to be detected; in order to facilitate the subsequent acquisition of more accurate interlayer regions, the edges to be selected are screened to obtain one type of edges; determining the degree of confusion based on the pixel gray scale characteristics inside the two kinds of edges, and updating the interlayer region based on the degree of confusion; obtaining the degree of abnormality through the area characteristics and the temperature characteristics of the new interlayer region; and judging whether the transformer to be detected has faults or not by using the abnormality degree. The invention realizes the rapid detection of the transformer faults, effectively improves the timeliness of the transformer fault detection, and is mainly applied to the field of transformer fault detection.
Description
Technical Field
The invention relates to the technical field of image analysis, in particular to a transformer fault rapid detection method based on artificial intelligence.
Background
With the rapid development of the power industry, the demand of transformers is increasing year by year. The transformer is one of the most important devices in the power system and plays a vital role in the safe and stable operation of the power system, however, some old transformers may exist in the civil and industrial fields, and the ageing and damage of insulating materials at each part of the transformer reduce the safety of the transformer and even may cause accidents such as fire and the like. Therefore, it is necessary to quickly detect the transformer failure.
The fault detection of the transformer can generally adopt a temperature difference method, the temperature difference method compares a fault area with surrounding normal areas, calculates a temperature difference value and determines the fault area. However, the transformer fault determination result corresponding to the temperature difference method has poor timeliness, for example, when a large-area high-temperature area exists in an infrared image of the transformer and the temperature difference is extremely large, the transformer fault determination result is only reported.
Disclosure of Invention
In order to solve the technical problem of poor timeliness of the existing transformer fault detection method, the invention aims to provide an artificial intelligence-based transformer fault rapid detection method, which adopts the following technical scheme:
the embodiment of the invention provides a transformer fault rapid detection method based on artificial intelligence, which comprises the following steps:
acquiring an infrared image set of a transformer to be detected, and carrying out image preprocessing on each infrared image in the infrared image set to acquire each edge to be selected, each second-class edge and each third-class edge;
screening the edges of each class to be selected according to the gray value of each pixel point on the edges of each class to be selected to obtain the edges of each class, and determining an area surrounded by a spacing area between the edges of each class and the minimum class as an interlayer area;
according to the gray value of each pixel point in each class-II edge, carrying out pixel chaotic analysis on the interior of each class-II edge, and determining the chaotic degree of the interior of each class-II edge;
updating each interlayer region according to the confusion degree to obtain new interlayer regions;
determining an abnormal degree set corresponding to the transformer to be detected according to the area and the temperature of each new interlayer region in each infrared image and the area of the region surrounded by each three types of edges;
and judging whether the transformer to be detected has faults or not according to the abnormal degree set.
Further, according to the gray value of each pixel point on each edge of the class to be selected, each edge of the class to be selected is screened to obtain each edge of the class, including:
for any one edge to be selected, sliding the constructed sliding window with the preset size on the edge to be selected according to the preset step length to obtain each sliding window area;
counting the occurrence frequency of the pixel points with different preset gray levels in each sliding window area according to the gray value of each pixel point in each sliding window area and each preset gray level;
calculating entropy values of all sliding window areas according to the occurrence frequencies of pixel points with different preset gray levels in all the sliding window areas;
if the entropy value of any sliding window area is larger than the dispersion degree threshold value, judging that the edges to be selected are one type, otherwise, judging that the edges to be selected are not one type.
Further, according to the gray value of each pixel point in each class two edge, performing pixel confusion analysis on each class two edge to determine the degree of confusion in each class two edge, including:
for any two kinds of edges, determining the maximum gray value, the minimum gray value and the gray mean square error in the two kinds of edges according to the gray value of each pixel point in the two kinds of edges;
determining a first difference value between a maximum gray value and a minimum gray value in the two kinds of edges, and determining the ratio of the first difference value to the gray mean square error as the instability degree of the two kinds of edges;
determining the number of different preset gray levels in the two kinds of edges as the discontinuity degree in the two kinds of edges, and carrying out numerical amplification treatment on the discontinuity degree to obtain a new discontinuity degree;
the product of the degree of instability and the new degree of discontinuity is determined as the degree of confusion inside the corresponding class two edge.
Further, updating each interlayer region according to the confusion degree to obtain new each interlayer region, including:
if the degree of confusion in any two kinds of edges is larger than a confusion threshold, judging that all pixel points in the two kinds of edges belong to an interlayer region where the positions of the pixel points are located, otherwise, judging that all pixel points in the two kinds of edges do not belong to the interlayer region where the positions of the pixel points are located, and dividing all pixel points in the two kinds of edges into target interlayer regions; the target interlayer region is an interlayer region adjacent to the interlayer region where the two types of edge positions are located, and the temperature value of the target interlayer region is higher than that of the interlayer region where the two types of edge positions are located.
Further, the calculation formula of the abnormality degree is:
wherein AD is the degree of abnormality corresponding to the transformer to be detected,for the temperature value of the 1 st interlayer region in the next infrared image +.>For the temperature value of the 1 st interlayer region in the previous infrared image, +.>For the area of the 1 st interlayer region in the next infrared image +.>For the area of the region surrounded by the kth three kinds of edges in the next infrared image,/>For the area of the 1 st interlayer region in the previous infrared image,/for>For the area of the region surrounded by the kth three types of edges in the previous infrared image, K is the total number of the three types of edges in each infrared image, and +.>For the temperature value of the s-th interlayer region in the next infrared image, +.>For the temperature value of the s-th interlayer region in the previous infrared image, +.>For the latterArea of the s-th interlayer region in the infrared image,/->The area of the S-th interlayer region in the previous infrared image is given, and S is the total number of interlayer regions in each infrared image.
Further, performing image preprocessing on each infrared image in the infrared image set to obtain each to-be-selected type of edge, each type of edge and each type of three types of edges, including:
graying any infrared image in the infrared image set to obtain a gray image;
performing image enhancement processing on the gray level image to obtain a new gray level image;
performing edge detection on the new gray level image to obtain each temperature edge;
fitting the temperature edges to obtain closed temperature edges;
and determining each edge to be selected, each second-class edge and each third-class edge according to the position of each pixel point on each closed temperature edge.
Further, determining each of the one type of edge to be selected, each of the two types of edges, and each of the three types of edges according to the position of each pixel point on each of the closed temperature edges, including:
selecting an edge with the largest area of the surrounding area and the highest temperature from each closed temperature edge as a first target edge, and determining the product of the area of the surrounding area of the first target edge and a preset adjustment factor as an area threshold value of the edges to be selected; determining the edges of the closed temperature, which are surrounded by the edges in each closed temperature edge and have the area larger than an area threshold, as the edges to be selected;
and selecting the edge with the largest perimeter from all the closed temperature edges as a second target edge, determining the closed temperature edges outside the to-be-selected type of edges positioned inside the area surrounded by the second target edge as a type of edges, and determining the closed temperature edges outside the area surrounded by the target edges as three types of edges.
The invention has the following beneficial effects:
the invention provides a transformer fault rapid detection method based on artificial intelligence, which comprises the steps of firstly collecting infrared images of a transformer to be detected, and carrying out image preprocessing operation on each infrared image to obtain each edge to be selected, each edge of two types and each edge of three types in order to obtain more accurate image confidence and improve the accuracy of fault detection results; the false edges exist in the edges to be selected under the interference of image factors, and in order to obtain more accurate interlayer areas, screening treatment is carried out on the edges to be selected to obtain the edges of each type; in order to clarify the attribution of the area surrounded by the two kinds of edges, the disorder degree of the pixels is analyzed and determined by analyzing the gray value of each pixel point in the two kinds of edges, and then each interlayer area is updated according to the disorder degree, so that the new interlayer areas are obtained, and compared with the interlayer areas, the accuracy of the new interlayer areas is higher, which is helpful for improving the accuracy of the subsequent abnormality degree determined based on the characteristics of the interlayer areas; compared with a temperature difference method, the transformer fault detection is carried out according to the abnormal degree determined by the area of each interlayer region, the temperature and the area of the region surrounded by the three types of edges, so that the timeliness of the transformer fault detection is effectively improved, the transformer fault quick detection is facilitated, the occurrence of serious fire and other accidents of the transformer due to poor timeliness is avoided, and the method is mainly applied to the field of transformer fault detection.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for rapidly detecting faults of a transformer based on artificial intelligence.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description is given below of the specific implementation, structure, features and effects of the technical solution according to the present invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
Application scenario for which this embodiment is aimed: during the operation of the transformer, the temperature of a specific part of the winding of the transformer is often higher Wen Xianxiang, the temperature of the specific part is higher than that of other parts of the transformer, and the temperature change condition of the winding part can be monitored and analyzed through infrared images shot by the infrared monitoring device. Specifically, the embodiment provides a method for rapidly detecting a fault of a transformer based on artificial intelligence, as shown in fig. 1, including the following steps:
s1, acquiring an infrared image set of a transformer to be detected, and carrying out image preprocessing on each infrared image in the infrared image set to obtain each edge to be selected, each edge of two types and each edge of three types.
The method comprises the first step of obtaining an infrared image set of a transformer to be detected.
In this embodiment, through the infrared detection device installed on the upper portion of the side wall of the transformer oil tank to be detected, the infrared radiation information of the primary transformer winding part is acquired at fixed intervals, the acquired infrared radiation information is transmitted to the information processing module outside the oil tank, the information processing module converts the received infrared radiation information into an integral infrared image, and meanwhile, the temperature of each part is marked in the integral infrared image, so that the subsequent primary temperature judgment on the integral infrared image is facilitated. Wherein the temperature information in the overall infrared image may be presented in the form of numbers or colors. When the temperature of the winding area in the whole infrared image is subjected to severe temperature change, namely the temperature exceeds a preset temperature threshold, the temperature threshold can be set to 90, temperature analysis is carried out on a plurality of infrared images continuously shot by the winding area, the infrared images continuously shot form an infrared image set, and each infrared image in the infrared image set is an infrared image arranged in sequence according to shooting time. The infrared image set at least comprises 2 infrared images, the acquisition interval time between adjacent infrared images can be 10 minutes, and the temperature threshold can be set by an implementer according to specific practical conditions.
The installation position of the infrared detection device should be selected near the part where the transformer winding can be observed, for example, a transformer winding inlet and outlet, a grounding wire port and the like; meanwhile, adverse effects on the tolerance capability of the transformer caused by the installation of the infrared detection device need to be avoided, for example, the installation of the infrared detection device in the area with the highest pressure level of the oil tank is not feasible; in order to improve the accuracy of the collected infrared radiation information, the installed infrared detection device needs to be reinforced and sealed.
And secondly, carrying out image preprocessing on each infrared image in the infrared image set to obtain each edge to be selected, each edge of the second class and each edge of the third class.
A first substep, for any one infrared image in the infrared image set, graying the infrared image to obtain a gray image; performing image enhancement processing on the gray level image to obtain a new gray level image; performing edge detection on the new gray level image to obtain each temperature edge; fitting the temperature edges to obtain the closed temperature edges.
In this embodiment, in order to facilitate analysis of image information of an infrared image, a weighted average method is used to perform graying processing on the infrared image, so as to obtain a gray image. Because the gray level image can gray the colors corresponding to different temperatures in the infrared image, the colors on the infrared image are mapped onto the gray level image according to the preset sequence of red, yellow, green, blue and purple by adopting a color mapping method, and it is worth to say that the normalization processing can be carried out by considering the variables before and after mapping during the mapping. And meanwhile, in order to ensure that the obtained gray image is closer to the image seen by human eyes, gamma correction is carried out on the gray image, namely, the contrast of the gray image is improved through gamma gray stretching adjustment, and the gray image subjected to image enhancement processing is determined to be a new gray image. And detecting by using a Canny edge detection algorithm to obtain each temperature edge in the new gray level image, and fitting the scattered edge points in the new gray level image by using a least square method to obtain each closed temperature edge in order to make each temperature edge more continuous. The implementation processes of the weighted average method, the color mapping method, the gamma correction, the Canny edge detection algorithm and the least square method are all the prior art, and are not in the scope of the present invention, and are not described in detail herein.
And a second sub-step, determining each edge of the class to be selected, each edge of the class and each edge of the class three according to the position of each pixel point on each closed temperature edge.
It should be noted that the edges to be selected are edges containing one layer, and the edges to be selected are mainly formed by the influence of layering of temperature areas, which can reflect layering conditions of different temperature areas, and the temperature of the temperature areas shows that the temperature of the central area is highest, and the diffusion temperature is lower towards the surrounding; the second-class edges are small closed polygonal edges positioned in an interlayer region formed by adjacent edges to be selected, the distribution of the second-class edges in the interlayer region is irregular, and the second-class edges are formed by disordered gray distribution caused by uneven temperature distribution, so that the partial region is also identified as an edge during edge detection; there are also small temperature edges around the largest candidate class of edges, which are identified as three classes of edges. Based on the image characteristics of the to-be-selected type of edges, the type of edges and the type of edges, classifying and processing each closed temperature edge, the specific implementation steps can include:
in this embodiment, selecting an edge with the largest area of the surrounding area and the highest temperature from each closed temperature edge as a first target edge, and determining the product of the area of the surrounding area of the first target edge and a preset adjustment factor as an area threshold value of the edges to be selected; determining the edges of the closed temperature, which are surrounded by the edges in each closed temperature edge and have the area larger than an area threshold, as the edges to be selected; and selecting the edge with the largest perimeter from all the closed temperature edges as a second target edge, determining the closed temperature edges outside the to-be-selected type of edges positioned inside the area surrounded by the second target edge as a type of edges, and determining the closed temperature edges outside the area surrounded by the target edges as three types of edges.
It should be noted that, the edge encloses the area, that is, the number of pixels in the area, the edge is formed by a single pixel, the perimeter of the edge, that is, the number of pixels on the edge, the preset adjustment factor may be set to 1/9, and the value of the preset adjustment factor may be set by an implementer according to a specific practical situation, which is not limited herein specifically.
Thus, the embodiment obtains each edge of the to-be-selected type, each edge of the second type and each edge of the third type in each infrared image.
S2, screening the edges of each class to be selected according to the gray value of each pixel point on the edges of each class to be selected to obtain the edges of each class, and determining an area surrounded by a spacing area between the edges of each class and the smallest class as an interlayer area.
It should be noted that, the edges to be selected may be false edges caused by fitting, in order to obtain more accurate edges, image features of each edge to be selected are analyzed, false edges in each edge to be selected are removed, edges to be selected other than the false edges are determined as edges, and the determining steps of the edges to be selected may include:
the first step, for any one of the edges to be selected, sliding the constructed sliding window with the preset size on the edge to be selected according to the preset step length to obtain each sliding window area.
In the present embodiment, the preset size is constructed asAnd let->Sliding on the edges of the type to be selected according to a preset step length 4 in a sliding mode that a pixel point is randomly selected on the edges of the type to be selected as a central pixel point, and the central pixel point corresponds to the sliding window +.>And starting sliding at the sliding window position, and returning to the starting point after finishing one circle of sliding, so that each sliding window area generated in the sliding process can be obtained. The preset size and the preset step length may be set by the practitioner according to specific practical situations, and are not particularly limited herein.
And secondly, counting the occurrence frequency of the pixel points with different preset gray levels in each sliding window area according to the gray value of each pixel point in each sliding window area and each preset gray level.
In this embodiment, the gray values 0 to 255 are divided into 26 gray intervals, and the gray intervals are preset gray levels, for example, the first preset gray level is 0 to 9, and then the 26 th preset gray level is 250 to 255. In order to analyze the dispersion degree of the pixel points in the window area, the occurrence frequency of the pixel points of each gray level in the window area is counted for any one window area. For the feature case, if the last sliding window area on the edge of the candidate class is less than 7, the sliding window area is discarded. The magnitude of the preset gray level may be set by the practitioner according to a specific practical situation, and is not particularly limited herein.
Thirdly, calculating entropy values of the sliding window areas according to the occurrence frequencies of pixel points with different preset gray levels in the sliding window areas.
In this embodiment, the calculation formula of the entropy value of each sliding window area may be:
wherein, the liquid crystal display device comprises a liquid crystal display device,for the to-be-selected oneEntropy of the p-th sliding window area on the class edge, c is the serial number of the preset gray level, 26 is the total number of the preset gray level,/or->For the occurrence frequency of the pixel point with the c-th preset gray level in the p-th sliding window area on the edge of the to-be-selected type, < >>Is 2-base->Logarithmic (log).
It is worth to be noted that, the calculation formula of the entropy value is the prior art, and the entropy value can represent the dispersion degree of gray distribution in the sliding window area, so that false edges to be selected can be effectively judged by calculating the entropy value.
And fourthly, if the entropy value of any sliding window area is larger than the dispersion degree threshold value, judging that the edges to be selected are one type, otherwise, judging that the edges to be selected are not one type.
It should be noted that, the width of the interior of the temperature region in the infrared image is wide and narrow, for most of the temperature regions, there is a portion with a narrower width, and the gray scale distribution in the interior of the sliding window region corresponding to the portion with a narrower width is more dispersed, for example, compared with the sliding window region spanning two temperature regions, the gray scale distribution in the interior of the sliding window region spanning three temperature regions with a narrower width of two edges is more dispersed, that is, when the gray scale dispersion degree of any one sliding window region on the edges to be selected is greater, the edges to be selected are determined to be the real edges. Based on the characteristics, judging whether the edges to be selected are real edges or not through the gray level distribution characteristics of the sliding window areas on the edges to be selected, wherein the specific steps are as follows:
when the entropy value of any sliding window area on the edges of the to-be-selected type is larger than the scattering degree threshold value, the reliability degree that the edges of the to-be-selected type are true edges is high, the edges of the to-be-selected type can be judged to be the edges of the type, otherwise, the edges of the to-be-selected type are judged not to be the edges of the type. And judging one type of edge for each type of edge to be selected, and screening out all types of edges. The dispersion degree threshold may be set to be 1.5 times of an average value of entropy values of all sliding window areas on the edges of the candidate class, and the dispersion degree threshold may be set by an implementer according to a specific practical situation, which is not specifically limited herein.
After one type of edge in each infrared image is obtained, the area surrounded by the interval area between adjacent types of edges and the smallest type of edge is determined as an interlayer area for facilitating subsequent image analysis. The interlayer region is a temperature region, and can represent a temperature increment condition and an area expansion condition.
To this end, the present embodiment obtains the respective interlayer regions in each infrared image.
S3, according to the gray value of each pixel point in each class-II edge, performing pixel confusion analysis on each class-II edge, and determining the confusion degree in each class-II edge.
It should be noted that, if the internal temperature of the area surrounded by the two kinds of edges is changed, if the internal confusion degree of the two kinds of edges is high, namely the internal instability of the two kinds of edges, it is indicated that the area surrounded by the two kinds of edges is not completely converted into a temperature area with a higher level, namely the color of the area surrounded by the two kinds of edges is not completely converted into the color of the interlayer area of the inner layer; if the degree of disorder inside the two kinds of edges is low, namely the inside of the two kinds of edges is stable, the surrounding area of the two kinds of edges is similar to the temperature area of the higher level, namely the color of the surrounding area of the two kinds of edges is close to the color of the interlayer area of the inner layer. Therefore, in order to analyze the attribution situation of the pixels in the edges of each class, the degree of confusion in the edges of each class needs to be determined, and the degree of confusion in the edges of each class is determined by the gray value of each pixel point in the edges of each class, which comprises the following specific implementation steps:
first, for any two kinds of edges, determining the maximum gray value, the minimum gray value and the gray mean square error of the two kinds of edges according to the gray value of each pixel point in the two kinds of edges.
And secondly, determining a first difference value between a maximum gray value and a minimum gray value in the two kinds of edges, and determining the ratio of the first difference value and the gray mean square error as the instability degree of the two kinds of edges.
Thirdly, determining the number of different preset gray levels in the two kinds of edges as the discontinuity degree in the two kinds of edges, and carrying out numerical amplification treatment on the discontinuity degree to obtain a new discontinuity degree.
And fourthly, determining the product of the instability degree and the new discontinuity degree as the confusion degree of the inner part of the corresponding class-II edge.
As an example, the calculation formula of the degree of confusion inside the class-two edges may be:
wherein C is the degree of confusion inside the edges of the class II,is the maximum gray value inside the edges of the class ii,n is the number of pixel points in the two kinds of edges, N is the number of pixel points in the two kinds of edges, +.>Gray value of nth pixel point inside class II edge +.>For the gray average value inside the edges of class II, +.>For gray mean square error inside class II edges, +.>For the first difference inside the edges of the class +.>For the degree of instability inside the edges of class II, e is a natural constant, +.>For adjusting parameters +.>For the degree of discontinuity inside the edges of the class +.>Is a new degree of discontinuity inside the class-two edges.
In the calculation formula of the degree of confusion, the degree of discontinuity inside the edges of the class II is relatedThe more different preset gray levels exist in the two kinds of edges, the worse the continuity of the two kinds of edges is, namely the higher the degree of discontinuity is, the more the number of the preset gray levels with the frequency of non-0 in the two kinds of edges is, and the more the preset gray levels exist in the two kinds of edges; regarding the instability degree of the interior of the two kinds of edges, the greater the instability degree, the more unstable the gray level distribution of the interior of the two kinds of edges is, the larger the gray level value fluctuation exists, and when the gray level distribution of the interior of the two kinds of edges is relatively unstable, the first difference value is relatively larger than the gray level mean square error; the new degree of discontinuity, the greater the new degree of discontinuity, the adjustment parameter +.>Taking experience value as 0.1, adjusting parameter +.>Is to adjust monotonically increasing function->Is a convergence speed of (a). Wherein the adjustment parameter->Can be implemented by the implementer according to the specific embodimentSetting the situation.
Thus, the embodiment obtains the degree of confusion inside the edges of each class.
And S4, updating each interlayer region according to the degree of confusion, and obtaining new each interlayer region.
In this embodiment, based on the degree of confusion inside the edges of each class two, each interlayer region is updated, which is helpful to obtain an updated interlayer region. The method comprises the following steps: and comparing the degree of confusion in the two kinds of edges with a confusion threshold, if the degree of confusion in any two kinds of edges is larger than the confusion threshold, judging that all the pixels in the two kinds of edges belong to an interlayer region where the positions of the pixels are located, otherwise, judging that all the pixels in the two kinds of edges do not belong to the interlayer region where the positions of the pixels are located, and dividing all the pixels in the two kinds of edges into target interlayer regions. The target interlayer region is an interlayer region adjacent to the interlayer region where the two types of edge positions are located, and the temperature value of the target interlayer region is higher than that of the interlayer region where the two types of edge positions are located.
For example, arranging the interlayer regions in sequence from outside to inside, if the gray level distribution inside the second-class edge is disordered and unstable for any second-class edge of the s-th interlayer region, that is, the degree of disorder is greater than a disorder threshold value, explaining that the temperature inside the region surrounded by the second-class edge is occurring, each pixel point in the region surrounded by the second-class edge should belong to the s-th interlayer region, and then, when calculating the area of the s-th interlayer region, each pixel point in the region surrounded by the second-class edge needs to be considered; if the gray level distribution inside the two kinds of edges is stable and continuous, the condition that the surrounding area of the two kinds of edges is obviously different from other areas of the s-th interlayer area is shown, the degree of confusion is not more than a confusion threshold value, each pixel point in the surrounding area of the two kinds of edges belongs to the s+1th interlayer area, and then when the area of the s-th interlayer area is calculated, each pixel point in the surrounding area of the two kinds of edges is not considered, and each pixel point in the surrounding area of the two kinds of edges is added into the s+1th interlayer area.
To this end, the present embodiment obtains new respective interlayer regions in each infrared image.
S5, determining an abnormal degree set corresponding to the transformer to be detected according to the area and the temperature of each new interlayer region in each infrared image and the area of the region surrounded by each three types of edges.
It is noted that, by observing the infrared image of the winding region, different temperature regions exhibit different colors, and as time advances, the temperature and color of each temperature region change, and the area of each temperature region gradually expands. The relative relationship of the individual temperature zones remains unchanged during the time advance, yet a fixed number of individual temperature zones is set. However, whether the transformer fails or not cannot be judged only based on the temperature data of the transformer, and the abnormal degree of the transformer is determined by analyzing the expansion condition and the temperature increment condition of different temperature grade regions in adjacent high temperature regions. It should be noted that the new respective interlayer regions in each infrared image may also be arranged in an outside-in order.
As an example, the calculation formula of the degree of abnormality may be:
wherein D is the degree of abnormality corresponding to the transformer to be detected,for the new temperature value of the 1 st interlayer region in the next infrared image +.>For the new temperature value of the 1 st interlayer region in the previous infrared image +.>For the area of the new 1 st interlayer region in the next infrared image +.>The k three kinds of edges in the next infrared image are surroundedIs defined by the area of the region of (c),for the area of the new 1 st interlayer region in the previous infrared image +.>For the area of the region surrounded by the kth three types of edges in the previous infrared image, K is the total number of the three types of edges in each infrared image, and +.>For the temperature value of the new s-th interlayer region in the next infrared image +.>For the temperature value of the new s-th interlayer region in the previous infrared image, +.>For the area of the new s-th interlayer region in the next infrared image +.>The area of the new S-th interlayer region in the previous infrared image is given, and S is the total number of interlayer regions in each infrared image.
In the calculation formula of the degree of abnormality,can be used to characterize the degree of temperature difference of the new interlayer region 1 in two adjacent infrared images,/or->Can be used for representing the sum of the expansion area difference of a new interlayer region 1 in two adjacent infrared images and the expansion area difference of the region surrounded by all three types of edges; the new interlayer region 1 is the interlayer region corresponding to the largest type edge, and the region surrounded by the three types of edges is regarded as a partial region connected with the new interlayer region 1, so that when the degree of abnormality of the new interlayer region 1 is calculated, all three types of two adjacent infrared images are processedThe expansion area difference of the edge enclosing area is added to the expansion area difference of the new interlayer area 1, and the abnormality degree of the new interlayer area 1 can be obtained by calculating the product of the temperature difference degree of the new interlayer area 1 and the expansion area difference degree; />The method can represent the cumulative sum of the abnormal degrees from the new 2 nd interlayer region to the new S th interlayer region in two adjacent infrared images; the greater the temperature difference and the expansion area difference between two adjacent infrared images are, the greater the degree of abnormality of the winding part of the transformer to be detected is, and the two infrared images are in positive correlation; and the integral abnormal degree of the transformer to be detected at present is represented by calculating the accumulated sum of the abnormal degrees of all interlayer areas in two adjacent infrared images.
It should be noted that, the high temperature information of the transformer acquired in this embodiment belongs to the initial stage of the high temperature phenomenon, and the temperature of the transformer in this period is in the temperature rising stage, so the temperature of the temperature region of the next infrared image will be higher than the temperature of the temperature region corresponding to the previous infrared image. The phenomenon that the temperature of the temperature area suddenly drops does not belong to the transformer fault phenomenon, and similarly, the phenomenon that the corresponding area of each temperature area contracts does not belong to the transformer fault phenomenon.
Thus, the embodiment obtains each abnormal degree of the transformer to be detected, and forms an abnormal degree set from each abnormal degree.
And S6, judging whether the transformer to be detected has faults or not according to the abnormal degree set.
In this embodiment, based on each degree of abnormality in the set of degrees of abnormality, the operation state of the transformer to be detected is analyzed, and the specific implementation steps may include: for example, normalizing each degree of abnormality in the degree of abnormality set, determining the normalized degree of abnormality as a fault determination index, if any one fault determination index is greater than a fault determination threshold, determining that the transformer to be detected is faulty, otherwise, determining that the transformer to be detected is not faulty. The implementation method of numerical normalization includes, but is not limited to: logarithm normalization, exponential normalization, trigonometric or inverse trigonometric functions normalization, etc.; the fault determination threshold value can be obtained through historical data calculation, the empirical value of the fault determination threshold value at this time can be 0.7, and the operation data information corresponding to different types of transformers is different, so the fault determination threshold value is not specifically limited here.
For another example, a scattered point coordinate system is constructed based on each degree of abnormality in the degree of abnormality set and the corresponding time point of each degree of abnormality, and fitting is performed on all scattered points on the scattered point coordinate system by using quadratic polynomial fitting, so that a fitting curve can be obtained. And carrying out first-order derivation on the fitted curve to obtain a first derivative of each time point, determining the first derivative of each time point as a fault judgment index, comparing each fault judgment index with a fault judgment threshold value, and judging that the transformer to be detected, which is larger than the time point corresponding to the fault judgment threshold value, is faulty. The failure determination threshold at this time may be a value corresponding to 1.3 times the average value of all the failure determination indexes, or may be set according to history experience. It is worth to say that the change conditions of different abnormal degrees can be clarified through first-order derivation, and the operation state of the transformer to be detected can be further analyzed conveniently.
For another example, through a calculation process of the abnormal degree corresponding to the transformer to be detected, the abnormal degree of each interlayer region in the two adjacent infrared images can be obtained, and the fault detection of the transformer to be detected can be performed by using the abnormal degree of each new interlayer region in the two adjacent infrared images, specifically: and carrying out normalization processing on the abnormal degree of each new interlayer region, determining the abnormal degree of the normalization processing as a fault judgment index, comparing the fault judgment index of each new interlayer region with a fault judgment threshold value, judging that the corresponding new interlayer region has temperature abnormality when the fault judgment index of any new interlayer region is larger than the fault judgment threshold value, and judging that the transformer to be detected at the moment has faults. It is worth to say that, through analyzing the degree of abnormality of each new intermediate layer region, can effectively locate the specific temperature region that breaks down, be convenient for follow-up staff to carry out equipment maintenance processing to specific fault temperature region. The empirical value of the fault determination threshold at this time may still be 0.7.
It is worth to say that at least one degree of abnormality exists in the degree of abnormality collection, when only one degree of abnormality exists, it is indicated that two infrared images are collected to conduct transformer abnormality analysis, the duration of single transformer fault detection is short, and when 4 degrees of abnormality exist, it is indicated that five infrared images are collected to conduct transformer abnormality analysis, the duration of single transformer fault detection is long, and an implementer can determine the duration of single transformer fault detection according to specific practical conditions.
Thus, the embodiment realizes the fault detection of the transformer to be detected.
The invention provides a transformer fault rapid detection method based on artificial intelligence. The invention obviously improves the timeliness of the fault detection of the transformer while ensuring the accuracy of the fault judgment result, and is beneficial to avoiding serious accidents of the transformer caused by poor timeliness.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention and are intended to be included within the scope of the invention.
Claims (7)
1. The transformer fault rapid detection method based on artificial intelligence is characterized by comprising the following steps of:
acquiring an infrared image set of a transformer to be detected, and carrying out image preprocessing on each infrared image in the infrared image set to acquire each edge to be selected, each second-class edge and each third-class edge;
screening the edges of each class to be selected according to the gray value of each pixel point on the edges of each class to be selected to obtain the edges of each class, and determining an area surrounded by a spacing area between the edges of each class and the minimum class as an interlayer area;
according to the gray value of each pixel point in each class-II edge, carrying out pixel chaotic analysis on the interior of each class-II edge, and determining the chaotic degree of the interior of each class-II edge;
updating each interlayer region according to the confusion degree to obtain new interlayer regions;
determining an abnormal degree set corresponding to the transformer to be detected according to the area and the temperature of each new interlayer region in each infrared image and the area of the region surrounded by each three types of edges;
and judging whether the transformer to be detected has faults or not according to the abnormal degree set.
2. The method for rapidly detecting faults of a transformer based on artificial intelligence according to claim 1, wherein the step of screening each class of edges to obtain each class of edges according to the gray value of each pixel point on each class of edges to be selected comprises the steps of:
for any one edge to be selected, sliding the constructed sliding window with the preset size on the edge to be selected according to the preset step length to obtain each sliding window area;
counting the occurrence frequency of the pixel points with different preset gray levels in each sliding window area according to the gray value of each pixel point in each sliding window area and each preset gray level;
calculating entropy values of all sliding window areas according to the occurrence frequencies of pixel points with different preset gray levels in all the sliding window areas;
if the entropy value of any sliding window area is larger than the dispersion degree threshold value, judging that the edges to be selected are one type, otherwise, judging that the edges to be selected are not one type.
3. The method for rapidly detecting the faults of the transformer based on the artificial intelligence according to claim 1, wherein the step of carrying out the pixel confusion analysis on the insides of the two kinds of edges according to the gray value of each pixel point in the insides of the two kinds of edges to determine the degree of confusion in the insides of the two kinds of edges comprises the following steps:
for any two kinds of edges, determining the maximum gray value, the minimum gray value and the gray mean square error in the two kinds of edges according to the gray value of each pixel point in the two kinds of edges;
determining a first difference value between a maximum gray value and a minimum gray value in the two kinds of edges, and determining the ratio of the first difference value to the gray mean square error as the instability degree of the two kinds of edges;
determining the number of different preset gray levels in the two kinds of edges as the discontinuity degree in the two kinds of edges, and carrying out numerical amplification treatment on the discontinuity degree to obtain a new discontinuity degree;
the product of the degree of instability and the new degree of discontinuity is determined as the degree of confusion inside the corresponding class two edge.
4. The method for quickly detecting faults of a transformer based on artificial intelligence according to claim 1, wherein updating each interlayer region according to the degree of confusion to obtain new each interlayer region comprises:
if the degree of confusion in any two kinds of edges is larger than a confusion threshold, judging that all pixel points in the two kinds of edges belong to an interlayer region where the positions of the pixel points are located, otherwise, judging that all pixel points in the two kinds of edges do not belong to the interlayer region where the positions of the pixel points are located, and dividing all pixel points in the two kinds of edges into target interlayer regions; the target interlayer region is an interlayer region adjacent to the interlayer region where the two types of edge positions are located, and the temperature value of the target interlayer region is higher than that of the interlayer region where the two types of edge positions are located.
5. The method for rapidly detecting faults of transformers based on artificial intelligence according to claim 1, wherein the calculation formula of the degree of abnormality is as follows:
wherein D is the degree of abnormality corresponding to the transformer to be detected,for the new temperature value of the 1 st interlayer region in the next infrared image +.>For the new temperature value of the 1 st interlayer region in the previous infrared image +.>For the area of the new 1 st interlayer region in the next infrared image +.>For the area of the region surrounded by the kth three kinds of edges in the next infrared image,/>For the area of the new 1 st interlayer region in the previous infrared image +.>For the area of the region surrounded by the kth three types of edges in the previous infrared image, K is the total number of the three types of edges in each infrared image, and +.>For the temperature value of the new s-th interlayer region in the next infrared image +.>For the temperature value of the new s-th interlayer region in the previous infrared image, +.>For the area of the new s-th interlayer region in the next infrared image +.>The area of the new S-th interlayer region in the previous infrared image is given, and S is the total number of interlayer regions in each infrared image.
6. The method for quickly detecting faults of a transformer based on artificial intelligence according to claim 1, wherein the step of performing image preprocessing on each infrared image in the infrared image set to obtain each one type of edge to be selected, each two types of edges and each three types of edges comprises the steps of:
graying any infrared image in the infrared image set to obtain a gray image;
performing image enhancement processing on the gray level image to obtain a new gray level image;
performing edge detection on the new gray level image to obtain each temperature edge;
fitting the temperature edges to obtain closed temperature edges;
and determining each edge to be selected, each second-class edge and each third-class edge according to the position of each pixel point on each closed temperature edge.
7. The method for rapidly detecting faults of transformers based on artificial intelligence of claim 6, wherein determining each one class of edges to be selected, each two classes of edges and each three classes of edges according to the position of each pixel point on each closed temperature edge comprises:
selecting an edge with the largest area of the surrounding area and the highest temperature from each closed temperature edge as a first target edge, and determining the product of the area of the surrounding area of the first target edge and a preset adjustment factor as an area threshold value of the edges to be selected; determining the edges of the closed temperature, which are surrounded by the edges in each closed temperature edge and have the area larger than an area threshold, as the edges to be selected;
and selecting the edge with the largest perimeter from all the closed temperature edges as a second target edge, determining the closed temperature edges outside the to-be-selected type of edges positioned inside the area surrounded by the second target edge as a type of edges, and determining the closed temperature edges outside the area surrounded by the target edges as three types of edges.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310891162.8A CN116612121B (en) | 2023-07-20 | 2023-07-20 | Transformer fault rapid detection method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310891162.8A CN116612121B (en) | 2023-07-20 | 2023-07-20 | Transformer fault rapid detection method based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116612121A CN116612121A (en) | 2023-08-18 |
| CN116612121B true CN116612121B (en) | 2023-10-27 |
Family
ID=87680439
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310891162.8A Active CN116612121B (en) | 2023-07-20 | 2023-07-20 | Transformer fault rapid detection method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116612121B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104408459A (en) * | 2014-10-27 | 2015-03-11 | 国家电网公司 | Image identification method applied to power equipment monitoring |
| CN110073301A (en) * | 2017-08-02 | 2019-07-30 | 强力物联网投资组合2016有限公司 | The detection method and system under data collection environment in industrial Internet of Things with large data sets |
| CN112233073A (en) * | 2020-09-30 | 2021-01-15 | 国网山西省电力公司大同供电公司 | Real-time detection method for infrared thermal imaging abnormity of power transformation equipment |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111751763B (en) * | 2020-06-08 | 2021-08-10 | 武汉大学 | Power transformer winding fault diagnosis method based on GSMallat-NIN-CNN network |
-
2023
- 2023-07-20 CN CN202310891162.8A patent/CN116612121B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104408459A (en) * | 2014-10-27 | 2015-03-11 | 国家电网公司 | Image identification method applied to power equipment monitoring |
| CN110073301A (en) * | 2017-08-02 | 2019-07-30 | 强力物联网投资组合2016有限公司 | The detection method and system under data collection environment in industrial Internet of Things with large data sets |
| CN112233073A (en) * | 2020-09-30 | 2021-01-15 | 国网山西省电力公司大同供电公司 | Real-time detection method for infrared thermal imaging abnormity of power transformation equipment |
Non-Patent Citations (2)
| Title |
|---|
| 基于改进SSD的电力设备红外图像异常自动检测方法;王旭红;李浩;樊绍胜;蒋志鹏;;电工技术学报(S1);全文 * |
| 结合多元信息的变压器散热器红外故障识别;陈宏凯;孙国繁;赵晓光;谭民;;华中科技大学学报(自然科学版)(S1);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116612121A (en) | 2023-08-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109186813B (en) | Temperature sensor self-checking device and method | |
| US11140197B2 (en) | Method and apparatus for DDoS attack detection | |
| US11736062B2 (en) | Method and apparatus for determining operating state of photovoltaic array, device and storage medium | |
| CN117093879B (en) | Intelligent operation management method and system for data center | |
| JP5091604B2 (en) | Distribution evaluation method, product manufacturing method, distribution evaluation program, and distribution evaluation system | |
| CN113111596B (en) | Carbon peak-reaching real-time visual measuring and calculating method based on electric brain center | |
| CN116839682B (en) | Cable processing and manufacturing real-time monitoring system based on Internet of things | |
| CN117407679B (en) | Data acquisition method and system of intelligent end screen sensor | |
| CN116304958A (en) | Intelligent monitoring system and method for underground water treatment abnormality | |
| CN117312997B (en) | Intelligent diagnosis method and system for power management system | |
| CN117060409B (en) | Automatic detection and analysis method and system for power line running state | |
| CN110704987A (en) | Bearing abnormal state evaluation method based on similar working conditions of failure data mining | |
| CN116612121B (en) | Transformer fault rapid detection method based on artificial intelligence | |
| CN116227788A (en) | Fault diagnosis and evaluation method and system for substation power equipment | |
| CN112685216A (en) | Equipment abnormity monitoring system and method based on trend analysis | |
| CN111813644A (en) | System performance evaluation method and device, electronic equipment and computer readable medium | |
| CN108508860B (en) | Process industrial production system data monitoring method based on coupling relation | |
| CN117390573B (en) | Wind turbine generator operation abnormality early warning method based on time sequence prediction | |
| CN111145895A (en) | Abnormal data detection method and terminal equipment | |
| CN114662058B (en) | Wireless station monitoring method and device | |
| KR20200106791A (en) | Early warning device and method of power plant using switching technology of multiple-predictive models | |
| CN116559713A (en) | Intelligent monitoring method and device for power supply of communication base station | |
| CN116309344A (en) | Insulator abnormality detection method, device, equipment and storage medium | |
| CN117725541B (en) | Intelligent monitoring and fault diagnosis system for running state of annealing furnace | |
| CN114784974B (en) | Real-time early warning system of smart power grids |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |