A fault diagnosis method based on data center
By collecting and analyzing monitoring images in the data center, identifying and diagnosing cable failures, the fault identification problem caused by cable path complexity is solved, automated fault diagnosis is achieved, and the operation efficiency and cable management effect of the data center are improved.
Patent Information
- Application Number
- CN202411178949.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-27
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-08-27
AI Technical Summary
The cable paths in the data center are complex and numerous, which makes it impossible to automatically and accurately identify and diagnose cable fault status, affecting data transmission stability and equipment operation efficiency.
By collecting monitoring images in the data center, identifying information about network cables and power cables, combining historical information for electromagnetic interference analysis and wiring harness bending recognition, wiring harness fault integration diagnosis is carried out, and automated fault identification and diagnostic results are provided.
It realizes automated identification and fault diagnosis of cables in the data center, improves data transmission stability and cable service life, promptly detects and solves cable failures, and improves the operation efficiency and maintenance convenience of the data center.
Smart Images

Figure CN118707250B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data centers, and in particular to a fault diagnosis method based on a data center. Background Art
[0002] The widespread use of multiple servers in data centers has led to the complexity and challenges of cable management. As the scale of data centers expands, the number of power cables and network cables increases dramatically, and the distribution and management of these cables in server racks becomes particularly important. Among them, when power cables and network cables are wired in parallel at a close distance, the power cables may cause electromagnetic interference to the network cables, thereby affecting the stability and reliability of data transmission. Secondly, the cables may be bent and squeezed during maintenance, causing physical damage to the cables, thereby affecting the normal operation of the equipment. Due to the complex and large number of cable paths in data centers, it is extremely difficult to manually diagnose and maintain the status of cables, and problems are often not discovered and solved in a timely manner, resulting in reduced operating efficiency of data centers. Summary of the invention
[0003] The present invention aims to solve the technical problem in the prior art that the cable paths in a data center are complex and numerous, and the cable fault status cannot be automatically and accurately identified and diagnosed, and proposes a fault diagnosis method based on a data center.
[0004] In a first aspect, the present invention provides a data center-based fault diagnosis method, comprising: collecting monitoring images of a designated area in a target data center;
[0005] Performing harness identification on the monitoring image to obtain network harness information of multiple network harnesses of multiple servers and power harness information of multiple power lines, wherein the identification is performed in combination with multiple historical harness information, the network harness information includes harness servers and network harness paths, and the multiple servers are arranged in the designated area;
[0006] According to the plurality of network cable harness information and the plurality of power cable harness information, electromagnetic interference fault analysis is performed respectively to obtain a plurality of electromagnetic interference information;
[0007] According to the plurality of network cable harness information and the plurality of power cable harness information, combined with the plurality of historical harness information, harness bending identification is performed to obtain a plurality of harness bending information;
[0008] According to the data features of the multiple servers, combined with the multiple electromagnetic interference information and the multiple wiring harness bending information, integrated wiring harness fault diagnosis analysis is performed to obtain wiring harness fault diagnosis results.
[0009] In a second aspect, the present invention provides a fault diagnosis device based on a data center, the device comprising: an image acquisition module, for acquiring monitoring images of a specified area in a target data center;
[0010] a harness identification module, used to perform harness identification on the monitoring image, and obtain network harness information of multiple network harnesses of multiple servers and power harness information of multiple power lines, wherein the identification is performed in combination with multiple historical harness information, the network harness information includes harness servers and network harness paths, and the multiple servers are arranged in the designated area;
[0011] An electromagnetic interference analysis module, used to perform electromagnetic interference fault analysis respectively according to the plurality of network cable harness information and the plurality of power cable harness information, and obtain a plurality of electromagnetic interference information;
[0012] A bending analysis module, used to identify the bending of the wiring harness according to the plurality of network cable harness information and the plurality of power cable harness information, combined with the plurality of historical wiring harness information, to obtain a plurality of wiring harness bending information;
[0013] The fault diagnosis module is used to perform integrated diagnosis and analysis of wiring harness faults according to the data features of the multiple servers, in combination with the multiple electromagnetic interference information and the multiple wiring harness bending information, to obtain wiring harness fault diagnosis results.
[0014] In a third aspect, the present invention provides an electronic device, comprising: a memory for storing a computer software program; a processor for reading and executing the computer software program, thereby implementing the data center-based fault diagnosis method of the first aspect.
[0015] In a fourth aspect, the present invention provides a non-transitory computer-readable storage medium having a computer software program stored therein, and when the computer software program is executed by a processor, the data center-based fault diagnosis method of the first aspect is implemented.
[0016] The present invention collects monitoring images of a specified area in a target data center, performs wiring harness identification on the monitoring images, obtains network cable harness information of multiple network cable harnesses of multiple servers and power cable harness information of multiple power cables, multiple servers are arranged in the specified area, and specifically identifies the type of each wiring harness, the server to which it belongs, and the path of the wiring harness, and then performs electromagnetic interference fault analysis based on the multiple network cable harness information and the multiple power cable harness information to obtain multiple electromagnetic interference information, performs wiring harness bending identification based on the multiple network cable harness information and the multiple power cable harness information in combination with multiple historical wiring harness information to obtain multiple wiring harness bending information; performs wiring harness fault integrated diagnosis analysis based on the data characteristics of multiple servers in combination with multiple electromagnetic interference information and multiple wiring harness bending information to obtain wiring harness fault diagnosis results. The present invention automatically identifies wire harnesses in a data center and identifies them in combination with historical wire harness information, with a high recognition accuracy. It then performs electromagnetic interference analysis and wire harness bending identification to perform fault diagnosis, thereby ensuring the transmission stability of the data center and the service life of the cables. It also performs integrated analysis in combination with data features of different servers, and can perform key evaluations in combination with the characteristics of server operating data, based on which it diagnoses cable fault status and provides accurate wire harness fault diagnosis results, thereby helping to promptly discover and resolve cable fault problems in the data center and improve the operating efficiency and maintenance convenience of the data center. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 A schematic diagram of a flow chart of a fault diagnosis method based on a data center provided by an embodiment of the present invention;
[0018] Figure 2 A schematic diagram of the structure of a data center-based fault diagnosis device provided by an embodiment of the present invention;
[0019] Figure 3 A functional block diagram of an electronic device provided by an embodiment of the present invention;
[0020] Figure 4 A functional block diagram of a computer-readable storage device provided by the present invention.
[0021] In the accompanying drawings, the components represented by the reference numerals are listed as follows:
[0022] Image acquisition module 11, harness identification module 12, electromagnetic interference analysis module 13, bending analysis module 14, fault diagnosis module 15, electronic device 100, memory 110, processor 120, computer program 111, computer readable storage medium 200, computer program 211. DETAILED DESCRIPTION
[0023] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.
[0024] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.
[0025] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.
[0026] like Figure 1 As shown, an embodiment of the present invention provides a fault diagnosis method based on a data center, including:
[0027] S10: Collect monitoring images of a designated area in the target data center;
[0028] In the embodiment of the present application, the target data center is the data center to be diagnosed for faults, in which multiple data servers are deployed, and multiple data servers are connected through wire harnesses to provide data services, and the designated area is the area for fault monitoring in the target data center, in which multiple servers and other equipment are deployed. There are many cables for multiple servers, and when maintaining the servers, the cables may be moved away from their original wiring positions, causing electromagnetic interference between them, and some cables may be bent.
[0029] In the embodiment of the present application, in order to realize the automatic identification and fault diagnosis of cable faults, fault diagnosis analysis is performed by image technology. In the designated area of the data center, high-resolution surveillance cameras are deployed to collect images in real time. These images cover the server racks and cable wiring areas, providing detailed visual data. Through these images, the cable layout and status in the area can be fully understood, providing basic data support for subsequent analysis and processing.
[0030] S20: performing harness identification on the monitoring image to obtain network harness information of multiple network harnesses of multiple servers and power harness information of multiple power lines, wherein the identification is performed in combination with multiple historical harness information, the network harness information includes harness servers and network harness paths, and the multiple servers are arranged in the designated area;
[0031] In an embodiment of the present application, wire harness identification is performed on the collected monitoring images, wherein the collected monitoring images are first analyzed and segmented using image processing technology to identify the cable images in the images, and then the network cable harnesses and power cable harnesses of multiple servers are identified, and the path information of each harness is identified.
[0032] In order to improve the accuracy and reliability of identification, this process combines multiple historical wiring harness information for identification. The historical wiring harness information includes the cable wiring data collected and identified when the cable fault diagnosis of the specified area was last performed.
[0033] Exemplarily, the historical wiring harness information may also be the historical wiring harness information of multiple wiring harnesses recorded after the wiring harnesses of multiple servers have been sorted out, wherein each historical wiring harness information includes the wiring harness type of each wiring harness, the server to which the wiring harness belongs, and the wiring harness path, the wiring harness types include power cables and network cables, and the wiring harness path includes the continuous wiring harness position information of each wiring harness on the floor and racks of the data center, such as the position information of a wiring harness from beginning to end in a specified area.
[0034] Through the combination of image recognition and deep learning technology, the automatic identification and classification of cables in the data center is realized. The historical wiring harness information is used for auxiliary analysis to improve the accuracy and stability of identification and ensure the effective management of the complex cable wiring environment in the data center.
[0035] Step S20 in the method provided in the embodiment of the present application includes:
[0036] Performing wire beam image segmentation on the monitoring image to obtain a plurality of wire beam images;
[0037] According to the recognition history of the server harness, a sample harness image set is collected, and the harness in each sample harness image is labeled with type, attribution and path to obtain a sample harness information set, wherein each sample harness information includes a harness type, a harness server and a harness path, and the harness type is a network harness or a power harness;
[0038] Using the sample wire harness image set and the sample wire harness information set to train an integrated wire harness identification channel, wherein the integrated wire harness identification channel includes a plurality of wire harness identification branches;
[0039] Inputting the plurality of wire harness images into the plurality of wire harness identification branches respectively, identifying and obtaining a plurality of wire harness type sets, a plurality of wire harness server sets and a plurality of wire harness path sets;
[0040] Selecting the wiring harness types with the highest occurrence frequency in the plurality of wiring harness type sets respectively to obtain a plurality of wiring harness types;
[0041] fusing the plurality of wire beam path sets to obtain a plurality of fused wire beam paths;
[0042] According to the plurality of wiring harness server sets, a plurality of wiring harness server probability sets are calculated and obtained, a plurality of historical wiring harness paths and a plurality of historical wiring harness servers in a plurality of historical wiring harness information are combined, a distance from the plurality of fused wiring harness paths is calculated, a plurality of historical wiring harness server probability sets are obtained by allocation, a plurality of historical wiring harness server probability sets are combined, a server corresponding to the maximum plurality of server probabilities is screened, and a plurality of wiring harness servers are obtained;
[0043] In combination with the plurality of harness types, the plurality of fused harness paths and the plurality of harness servers, network harness information of the plurality of network harnesses of the plurality of servers and power harness information of the plurality of power lines are obtained.
[0044] In the embodiment of the present application, firstly, the monitoring image of the designated area is segmented into wire harness images, and the image of the wire harness part in the monitoring image is segmented by image processing technology to obtain multiple wire harness images. Among them, by identifying the cable features in the image, the cable is separated from the background to generate an independent wire harness image.
[0045] Exemplarily, semantic segmentation in deep learning is used to train a wire harness image segmentation model for segmenting a wire harness diagram, sample monitoring images are collected, and then the background image pixels in the sample monitoring images are labeled and classified, and the pixels of the wire harness image are labeled and classified, and then the wire harness image segmentation model is supervised and trained, and the model parameters are adjusted according to the deviation of the model output until the use requirements are met, for example, the accuracy of the wire harness image reaches 90%, and the trained wire harness image segmentation model is used to segment the wire harness images in the monitoring diagram to obtain multiple wire harness images. Optionally, the wire harness image segmentation can also be performed based on other methods such as edge detection in the prior art.
[0046] For a line beam image with a repeated portion, pixels of the repeated portion may be added to the line beam images of two line beams to form a plurality of line beam images.
[0047] Furthermore, based on the multiple segmented wire harness images, wire harness information is recognized.
[0048] Among them, according to the recognition history of the server harness, such as the data of the harness fault diagnosis in the data center, a set of sample harness images is collected. These sample images contain the cable data that have been marked before. In each sample harness image, the cable is marked with type (network cable or power cable), belonging (connected server) and path (the path where the cable passes) to obtain a detailed set of sample harness information. The sample harness information set includes the type of each sample cable, the connected server and the specific wiring path. This information provides basic data for training the recognition model.
[0049] Among them, type labeling can be based on the color and thickness of the wire harness, attribution labeling can be based on the record number on the wire harness, and path labeling can construct a coordinate system in a specified area, dividing the wire harness in each sample wire harness image into multiple parts, and labeling the path position according to the coordinate values of the multiple parts in the coordinate system, so as to obtain a sample wire harness information set.
[0050] An integrated harness recognition channel is trained using a set of sample harness images and a set of sample harness information. The integrated harness recognition channel is a deep learning model that contains multiple harness recognition branches. The training data of the multiple harness recognition branches are different. Different multiple information recognitions can be performed on the input harness images to obtain the probability that each harness image belongs to different types and multiple servers, so as to identify multiple harness paths. By integrating multiple recognition results, the recognition accuracy of harness information can be improved.
[0051] The step of "using the sample wire harness image set and the sample wire harness information set to train the integrated wire harness recognition channel" in the method provided in the embodiment of the present application includes:
[0052] Randomly selecting a preset proportion of sample data from the sample wire harness image set and the sample wire harness information set to train a first wire harness recognition branch;
[0053] Continue to randomly select a preset proportion of sample data from the sample wire harness image set and the sample wire harness information set to train other multiple wire harness recognition branches;
[0054] The plurality of wire harness recognition branches that have been trained are combined to obtain an integrated wire harness recognition channel.
[0055] In the embodiment of the present application, a preset proportion of sample data is randomly selected from the sample wire harness image set and the sample wire harness information set to train the first wire harness recognition branch. Specifically, a portion of images is randomly selected from the sample wire harness image set, and corresponding annotation information (including wire harness type, wire harness server, and wire harness path) is extracted from the corresponding sample wire harness information set. Exemplarily, a preset proportion of 30%-70% of sample data is randomly selected, and these sample data are used to supervise the training of the first wire harness recognition branch so that the first wire harness recognition branch can recognize the characteristic information of the wire harness.
[0056] Exemplarily, based on the convolutional neural network in deep learning in the prior art, a first wiring harness recognition branch is constructed, and a preset proportion of sample wiring harness images and sample wiring harness information is used to train the first wiring harness recognition branch. The sample wiring harness information is used as supervisory data, and the model parameters of the first wiring harness recognition branch are adjusted according to the deviation between the output of the first wiring harness recognition branch and the sample wiring harness information. For example, according to whether the output wiring harness type is consistent with the wiring harness type in the sample wiring harness information, and the deviation between the output wiring harness path and the sample wiring harness path, the parameters of the first wiring harness recognition branch are adjusted until the output is consistent with the supervisory data, and multiple rounds of iterative training are performed for each group of data until the accuracy meets the requirements, for example, 90%, and the training is completed.
[0057] Furthermore, a preset proportion of sample data is randomly selected from the sample harness image set and the sample harness information set to train other multiple harness recognition branches. Each recognition branch randomly selects a different subset from the sample set for training, ensuring that each branch can learn the characteristics of cables from different data distributions. Through training based on different subsets, the generalization ability and robustness of the large model formed by multiple harness recognition branches, i.e., the integrated harness recognition channel, is improved.
[0058] The trained integrated wire harness recognition channel is constructed by combining the trained multiple wire harness recognition branches, for example, connecting the input layer of the first wire harness recognition branch of the output layer of the trained multiple wire harness recognition branches.
[0059] The embodiment of the present application can not only effectively improve the accuracy of cable identification through multi-branch training and integrated recognition methods, but also improve the adaptability of the model to complex cable wiring environments, reduce the deviations and errors that may occur in a single model, and improve the convergence efficiency of each branch training. By randomly selecting a preset proportion of sample data for training, it is ensured that each recognition branch can learn cable features from different data perspectives, thereby improving the performance and reliability of the integrated harness recognition channel as a whole.
[0060] In an embodiment of the present application, multiple harness images are input into multiple harness identification branches respectively, and multiple harness identification branches identify the harness type, the server to which they belong, and the harness path for each harness image. The output results of multiple harness identification branches are collected to obtain multiple harness type sets, multiple harness server sets, and multiple harness path sets for multiple harness images. Each harness type set includes the type classification results (such as network cable harness or power cable harness) of the cables identified in a harness image by multiple identification branches, the harness server set includes the connected servers (such as server A, server B, etc.) obtained by multiple identification branches for one harness image, and the harness path set includes the detection results of the cable path obtained by multiple identification branches for one harness image, specifically including the path position information of the harness.
[0061] Furthermore, in the plurality of harness type sets, the harness types with the highest frequency of occurrence are selected. Specifically, for each cable, by counting the classification results of all identification branches, the harness type with the highest frequency of occurrence is selected to ensure the accuracy and stability of the final classification result.
[0062] Multiple wire beam path sets are fused, and multiple fused wire beam paths are obtained by integrating the path recognition results of each recognition branch. For example, based on multiple path position information of a position of a wire beam in the wire beam path recognized by each recognition branch, the coordinate values in the multiple path position information can be averaged to obtain the fused path position information of the position of the wire beam. In this way, the path position information of multiple positions of each wire beam is fused and calculated to obtain multiple fused wire beam paths of multiple wire beams.
[0063] Exemplarily, three wire harness identification branches identify the path position information of the middle position of a wire harness, and the coordinate values obtained in the ground coordinate system of the specified area are (101, 76), (105, 65) and (97, 72) respectively. Then, the wire harness path position information of the middle position of the wire harness is rounded to (101, 71). In this way, the fused wire harness path of each wire harness can be calculated to obtain the path position distribution of the wire harness in the specified area.
[0064] Furthermore, based on multiple harness server sets, multiple harness server probability sets are calculated. For example, the proportion of different servers in each harness server set is calculated as the harness server probability. If a harness server set includes 10 harness-owned servers and server A appears 7 times, the harness server probability of server A is 0.7.
[0065] By counting the server identification results of each identification branch, the probability distribution of each wire harness belonging to different multiple servers is calculated to obtain multiple wire harness server probability sets. Among them, the server to which the wire harness belongs is analyzed based on the wire harness image features. The server identification information on the wire harness may be partially blocked, resulting in a low accuracy rate. Therefore, combined with multiple historical wire harness information, statistical analysis of the wire harness belonging server is performed to improve the analysis accuracy of the wire harness belonging server.
[0066] The step of the method provided in the embodiment of the present application, "calculating and obtaining multiple beam server probability sets according to the multiple beam server sets, combining multiple historical beam paths and multiple historical beam servers in multiple historical beam information, calculating the distances to the multiple fused beam paths, allocating and obtaining multiple historical beam server probability sets, combining the multiple beam server probability sets, screening the servers corresponding to the largest multiple server probabilities, and obtaining multiple beam servers", includes:
[0067] Calculate and obtain multiple wire harness server probability sets according to the appearance ratios of multiple servers in the multiple wire harness server sets;
[0068] Acquire multiple historical harness paths and multiple historical harness servers in multiple historical harness information of the last harness fault diagnosis;
[0069] Calculating the distances between the multiple fused wire harness paths and the multiple historical wire harness paths respectively to obtain multiple path distance sets, and combining the multiple historical wire harness servers to obtain multiple historical wire harness server probability sets, wherein the size of the historical wire harness server probability is negatively correlated with the size of the distance;
[0070] The plurality of wire bundle server probability sets and the plurality of historical wire bundle server probability sets are combined to calculate and obtain a plurality of weighted server probability sets, and servers corresponding to the largest plurality of weighted server probabilities are respectively selected as a plurality of wire bundle servers.
[0071] In an embodiment of the present application, based on multiple wiring harness server sets, multiple wiring harness server probability sets are calculated, and the proportion of occurrence of different servers in each wiring harness server set is specifically calculated as the wiring harness server probability to obtain a wiring harness server probability set, wherein the greater the wiring harness server probability of a certain wiring harness server, the greater the probability that the wiring harness belongs to the server.
[0072] Get multiple historical harness paths and multiple historical harness servers in multiple historical harness information of the last harness fault diagnosis. The historical harness information is the previously recorded cable path location data, including the historical path of the cable and the connected server information. The multiple historical harness servers are consistent with the current multiple servers, including the server numbers.
[0073] Among them, the change of the harness path may be caused by touching during the server maintenance process, so according to the historical harness path in the historical harness information, the distance between the current multiple fused harness paths and the historical harness paths is analyzed, and the probability that the current multiple harnesses belong to different servers can be analyzed. For example, the smaller the distance between a fused harness path and a historical harness path, the greater the probability that the harness belongs to the server to which the historical harness path belongs.
[0074] Further, the distances between the multiple fused wire bundle paths and the multiple historical wire bundle paths are calculated respectively to obtain multiple path distance sets. For example, the distances between the path position information of the multiple wire bundle parts in each fused wire bundle path and the path position information of the multiple wire bundle parts in the multiple historical wire bundle paths can be calculated, and the maximum distance or the mean distance is retained as the path distance to obtain multiple path distance sets. Each path distance set includes the distance between each fused wire bundle path and the multiple historical wire bundle paths.
[0075] Further, according to the multiple historical wire harness servers corresponding to the multiple historical wire harness paths, according to the size of the path distance in the multiple path distance sets, multiple historical wire harness server probability sets are allocated and calculated, and the size of the historical wire harness server probability is negatively correlated with the size of the distance. Exemplarily, the ratio of the path distance between the wire harness and each historical wire harness path to the sum of the path distance sets is calculated, and then 1 is subtracted from the ratio to obtain the historical wire harness server probability of the wire harness and the historical server corresponding to the historical wire harness path, so that multiple historical wire harness server probability sets are allocated and calculated.
[0076] Optionally, the path distance set may also be obtained by calculating the difference between the currently identified fused bundle path and the historical bundle paths (eg, using Euclidean distance or hash distance).
[0077] Furthermore, by combining multiple bundle server probability sets and multiple historical bundle server probability sets, weighted calculation is performed on the bundle server probability and the historical bundle server probability that each bundle belongs to the same server to obtain a weighted server probability, and then multiple weighted server probability sets are calculated. For example, the weight of the bundle server probability is 0.6, and the weight of the historical bundle server probability is 0.4.
[0078] In multiple weighted server probability sets, servers corresponding to the largest weighted server probabilities are selected respectively as servers to which the wiring bundle belongs, and then multiple servers corresponding to the largest multiple weighted server probabilities are obtained as multiple wiring bundle servers.
[0079] By combining the multiple harness types, multiple fused harness paths, and multiple harness servers obtained through analysis, network cable harness information of multiple network cable harnesses belonging to multiple servers and power cable harness information of multiple power cables can be obtained.
[0080] In this way, through image processing technology, combined with historical harness recognition record data, the final connection server of each cable is determined, ensuring the reliability and consistency of the recognition results. Through the above steps, the combination of current recognition results and historical data can achieve efficient recognition and accurate classification of cable connection servers. The use of probability calculation and path distance analysis methods improves the accuracy and stability of recognition, providing reliable data support for cable management and fault diagnosis in data centers. This method can not only improve the intelligence level of data centers, but also reduce maintenance costs and ensure the efficient operation of data centers.
[0081] S30: performing electromagnetic interference fault analysis respectively according to the plurality of network cable harness information and the plurality of power cable harness information to obtain a plurality of electromagnetic interference information;
[0082] In the embodiment of the present application, when the network cable harness is close to the power cable harness, it will be subject to electromagnetic interference from the power cable. Based on this, according to the harness paths of the harnesses in the multiple network cable harness information and the multiple power cable harness information, electromagnetic interference fault analysis of the multiple network cable harnesses is performed respectively to obtain multiple electromagnetic interference information.
[0083] Among them, electromagnetic interference (EMI) refers to the interference caused by the electromagnetic field generated by the power line when transmitting current to the signal transmission of the nearby network cable, which may cause instability of data transmission and increase of bit error rate. The closer the distance between the network cable and the power line, the greater the electromagnetic interference.
[0084] Step S30 in the method provided in the embodiment of the present application includes:
[0085] According to the plurality of network cable harness information and the harness paths in the plurality of power cable harness information, obtaining the interference distance between the network cable harness of each server and the closest power cable harness;
[0086] Electromagnetic interference fault analysis is performed according to the multiple interference distances to obtain multiple electromagnetic interference levels as multiple electromagnetic interference information.
[0087] In an embodiment of the present application, based on the bundle paths within multiple network cable bundle information and multiple power cable bundle information, the interference distance between the network cable bundle of each server and the closest power cable bundle is obtained, wherein the network cable bundle of each server and the closest power cable bundle may be the power cable bundle of other servers.
[0088] For example, the spatial position model of the network cable harness and the power cable harness of each server can be constructed based on the harness paths in the multiple network cable harness information and the multiple power cable harness information. The harness path data of the network cable and the power cable are further used to calculate the distance between each network cable and the nearest power cable to obtain the interference distance. The distance between the path position of multiple parts in each network cable harness information and the path position of multiple parts of other power cable harnesses can also be calculated, and the minimum distance is selected as the interference distance.
[0089] After obtaining the interference distance between each server's network cable harness and the closest power cable harness, the next step is to perform electromagnetic interference fault analysis based on multiple interference distances.
[0090] Exemplarily, the interference intensity corresponding to each interference distance can be calculated based on the inversely proportional relationship between the electromagnetic interference intensity and the square of the distance. It is also possible to combine the electromagnetic interference detection method in the prior art to detect the electromagnetic interference intensity at different interference distances, build a classification table of interference distances and electromagnetic interference intensity, classify the current interference distance, and obtain the interference intensity. According to the calculated interference intensity, electromagnetic interference fault analysis is performed to obtain multiple electromagnetic interference levels. These interference levels are divided by setting a threshold value for the interference intensity. For example, the interference intensity can be divided into 10 levels of 1-10, and the larger the number, the greater the interference intensity, so as to facilitate the identification and handling of electromagnetic interference problems of different severity.
[0091] Finally, multiple electromagnetic interference levels are used as multiple electromagnetic interference information, and the multiple electromagnetic interference information includes the electromagnetic interference level of each network cable. The multiple electromagnetic interference information provides maintenance personnel with a detailed interference risk assessment and provides a basis for subsequent wiring harness fault analysis.
[0092] S40: performing harness bending identification according to the plurality of network cable harness information and the plurality of power cable harness information in combination with the plurality of historical harness information to obtain a plurality of harness bending information;
[0093] In an embodiment of the present application, based on the harness paths within multiple network cable harness information and multiple power cable harness information, combined with multiple historical harness information, harness bending identification is performed to identify possible bending failures in the harness. Bending can cause harness aging and breakage, etc., affecting the operating stability of the data center.
[0094] Among them, according to the harness paths in multiple network cable harness information and multiple power cable harness information, the path curvature and angle change of each harness can be analyzed. Curvature refers to the degree of bending of the cable on the path, and the angle change refers to the angle between the segments in the cable path. Large curvature and angle change usually indicate that the cable is bent. Then analyze the static bending fault in the current harness path. Static bending can cause the harness to age and become damaged.
[0095] And combining multiple historical harness paths in multiple historical harness information, analyze the dynamic bending of multiple harnesses from the last harness fault diagnosis to the present time. Dynamic bending will produce impact force on the harness, which may cause breakage and damage.
[0096] Combined with the static bending and dynamic bending conditions, the specific bending failure conditions of multiple wire harnesses are analyzed to obtain the bending information of multiple wire harnesses.
[0097] Step S40 in the method provided in the embodiment of the present application includes:
[0098] Acquire multiple network cable harness paths and multiple power cable harness paths in multiple network cable harness information and multiple power cable harness information;
[0099] According to the multiple network cable harness paths, a plurality of static network cable bending degrees are obtained, according to the multiple power cable harness paths, a plurality of static power cable bending degrees are obtained, and a plurality of static harness bending degrees are obtained by weighted calculation;
[0100] Acquire multiple historical network cable harness paths and multiple historical power cable harness paths in the multiple historical harness information;
[0101] Analyze and obtain multiple changed network cable bending degrees according to the multiple historical network cable harness paths and the multiple network cable harness paths, analyze and obtain multiple changed power cable bending degrees according to the multiple historical power cable harness paths and the multiple power cable harness paths, and obtain multiple changed harness bending degrees by weighted calculation;
[0102] According to the plurality of static harness bending degrees and the plurality of variable harness bending degrees, a plurality of harness bending levels are respectively obtained by classification as a plurality of harness bending information.
[0103] In the embodiment of the present application, firstly, multiple network cable harness paths and multiple power cable harness paths in multiple network cable harness information and multiple power cable harness information are obtained. Multiple network cable harness paths and multiple power cable harness paths include the starting point, end point and intermediate position information of multiple harnesses in the specified area, providing basic data for subsequent bending analysis.
[0104] According to multiple network cable harness paths, multiple static network cable curvatures are obtained. The static network cable curvature refers to the degree of curvature of each network cable at the current moment. By calculating the curvature and angle change of each network cable path, its static curvature is quantified. Exemplarily, the maximum change angle of the harness path within each network cable harness path is calculated, for example, the bending angle of the network cable harness path within a length of 5 cm, that is, the angle at the starting point and the angle at the end point of the network cable harness path within 5 cm. The maximum bending angle is used as the static network cable curvature, and then multiple static network cable curvatures are obtained.
[0105] Similarly, based on multiple power line harness paths, multiple static power line bending degrees are obtained, and the same method or other harness bending degree analysis methods in the prior art are used to calculate the bending degree of each power line as multiple static power line bending degrees.
[0106] Next, the static power cord bending degree and static network cord bending degree of each server's power cord and network cord are weightedly calculated to obtain multiple static wire bundle bending degrees of multiple server wire bundles, reflecting the comprehensive degree of static bending of server network cables and power cords. The weighted calculation can adjust the bending degree according to the importance of the cable or its location to ensure the accuracy and practicality of the result. For example, the weights of the static power cord bending degree and the static network cord bending degree are set to 0.6 and 0.4 respectively.
[0107] Furthermore, multiple historical network cable harness paths and multiple historical power cable harness paths are obtained from multiple historical harness information, which are the historical power cable harness paths and network cable harness paths recorded for the multiple power cable harnesses and network cable harnesses during the last harness fault diagnosis. The harness paths of each network cable and power cable in the historical time are recorded, providing a reference for the change analysis of the current path.
[0108] According to multiple historical network cable harness paths and multiple current network cable harness paths, multiple changed network cable curvatures are analyzed and obtained. The changed network cable curvature refers to the bending change that occurs in the current harness path compared with the historical harness path. By comparing the curvature and angle changes of the current and historical paths, the changed curvature of the network cable is quantified. Exemplarily, based on the historical network cable harness path, the method in the above steps is used to obtain its historical static network cable curvature, and then the difference between the static network cable curvature of the same current network cable harness is calculated as the changed network cable curvature. Among them, if the network cable harness is not touched during data center maintenance and the path changes, the changed network cable curvature is 0 or less.
[0109] Similarly, based on multiple historical power line harness paths and multiple power line harness paths, multiple variable power line bending degrees are analyzed and calculated. Then, the variable network cable bending degree and the variable power line bending degree of each server are weighted to obtain multiple variable harness bending degrees. The weight can be the same as the weight of the static power line bending degree and the static network cable bending degree.
[0110] In this way, the changing bundle bending degrees and static bundle bending degrees of multiple power cable bundles and multiple network cable bundles of multiple servers are obtained, which respectively reflect the current static bending conditions of the bundles of each server and the changing bending conditions from historical time to the present.
[0111] Furthermore, the wire harness bending levels are classified according to multiple static wire harness bending degrees and multiple variable wire harness bending degrees, respectively, and multiple wire harness bending levels are obtained by classification. Exemplarily, different intervals can be set according to the size of the static wire harness bending degree, and different static wire harness bending levels can be set, for example, the static wire harness bending degree is 0 degrees to 30 degrees as level 1, 30 degrees to 60 degrees as level 2, etc., and then the corresponding static wire harness bending level is obtained according to the interval in which the static wire harness bending degree falls. Optionally, based on the same method or other classification methods in the prior art, the variable wire harness bending level is obtained according to the variable wire harness bending degree interval in which the variable wire harness bending degree falls, and the static wire harness bending level and the variable wire harness bending level are weighted and calculated to obtain the final wire harness bending level as the wire harness bending information. In this way, the bending failure of the wire harnesses of multiple servers can be obtained, that is, multiple wire harness bending information, which can be used as the data basis for subsequent diagnosis of the wire harness failure.
[0112] Through the above steps, the static and variable bending analysis of cables is systematically performed to generate detailed harness bending information. This information includes the location, angle and severity level of the bend, helping data center maintenance personnel to promptly detect and diagnose harness bending faults, repair possible physical damage, and ensure stable operation and efficient management of the data center.
[0113] S50: performing integrated diagnosis and analysis of wiring harness faults according to the data features of the multiple servers in combination with the multiple electromagnetic interference information and the multiple wiring harness bending information to obtain wiring harness fault diagnosis results.
[0114] In the embodiment of the present application, after analyzing and obtaining electromagnetic interference information and harness bending information reflecting electromagnetic interference fault conditions and bending fault conditions of multiple server harnesses, the importance of multiple servers is analyzed according to the data characteristics of multiple servers. Further combining multiple electromagnetic interference information and multiple harness bending information, an integrated harness fault diagnosis analysis is performed to analyze and diagnose the overall harness fault conditions of multiple servers in a specified area, wherein the importance of multiple servers is taken into consideration, so that a harness fault diagnosis result that better reflects the degree of harness fault can be obtained.
[0115] Step S50 in the method provided in the embodiment of the present application includes:
[0116] Acquire data characteristics of the multiple servers, and analyze to obtain multiple server grades;
[0117] According to the sizes of the multiple server levels, performing integrated weighted calculation on the multiple electromagnetic interference information and the multiple wire harness bending information to obtain weighted electromagnetic interference information and weighted wire harness bending information;
[0118] The wire harness fault level of the designated area is calculated and obtained according to the weighted electromagnetic interference information and the weighted wire harness bending level as a wire harness fault diagnosis result.
[0119] In an embodiment of the present application, data features of multiple servers are obtained and analyzed to obtain multiple server levels. Exemplarily, servers are graded according to the importance identification of the data stored or interacted in the data center to obtain multiple server levels. The server level reflects the importance and operating status of the server, for example, 5 levels from 1 to 5, the higher the level, the higher the importance. The importance identification of the data stored or interacted in the data center can be identified according to the importance or scale of the project to which the data running in the server belongs in advance.
[0120] Optionally, key performance indicators and operating status information can also be collected from each server in the data center, including CPU utilization, memory usage, network traffic, historical fault records, etc. Based on these data features, the servers are graded to obtain multiple server grades. For example, the greater the server operating load, the higher the server grade.
[0121] According to the size of multiple server levels, multiple electromagnetic interference information and multiple harness bending information are integrated and weighted. Exemplarily, the ratio of each server level to the sum of multiple server levels is calculated as the weight, and the weighted electromagnetic interference information and weighted harness bending level perform weighted calculation on the electromagnetic interference level and harness bending level in multiple electromagnetic interference information and multiple harness bending information to obtain the weighted electromagnetic interference level and weighted harness bending level as weighted electromagnetic interference information and weighted harness bending information. The electromagnetic interference information and weighted harness bending information respectively reflect the overall electromagnetic interference situation and harness bending fault situation of multiple servers in the specified area of the data center.
[0122] According to the weighted electromagnetic interference information and the weighted harness bending information, the harness fault level of the specified area is calculated as the harness fault diagnosis result. Exemplarily, the weighted electromagnetic interference level and the weighted harness bending level are weighted to obtain the harness fault level, wherein the weights of the weighted electromagnetic interference level and the weighted harness bending level reflect the importance of the harness fault caused by electromagnetic interference and the harness fault caused by bending in the overall harness fault, and those skilled in the art can set them by themselves, for example, the weights of the weighted electromagnetic interference level and the weighted harness bending level are 0.4 and 0.6 respectively.
[0123] The calculated harness fault level is used as the harness fault diagnosis result, which reflects the overall harness fault situation in the specified area of the data center and conducts targeted analysis based on the importance of different servers.
[0124] Through the above steps, the integrated diagnosis and analysis of wiring harness faults is systematically performed to generate detailed wiring harness fault diagnosis results. The wiring harness fault diagnosis results can help data center maintenance personnel to promptly discover and handle potential wiring harness faults. For example, when the wiring harness fault level is greater than a certain fault threshold, an alarm can be issued to call maintenance personnel to perform wiring harness maintenance, such as wiring harness arrangement and replacement. It also provides a scientific basis for the cable management and maintenance of the data center, ensuring the efficient and stable operation of the data center.
[0125] The embodiments of the present application achieve the following technical effects through the above specific implementation methods:
[0126] The present invention collects monitoring images of a specified area in a target data center, performs wiring harness identification on the monitoring images, obtains network cable harness information of multiple network cable harnesses of multiple servers and power cable harness information of multiple power cables, multiple servers are arranged in the specified area, and specifically identifies the type of each wiring harness, the server to which it belongs, and the path of the wiring harness, and then performs electromagnetic interference fault analysis based on the multiple network cable harness information and the multiple power cable harness information to obtain multiple electromagnetic interference information, performs wiring harness bending identification based on the multiple network cable harness information and the multiple power cable harness information in combination with multiple historical wiring harness information to obtain multiple wiring harness bending information; performs wiring harness fault integrated diagnosis analysis based on the data characteristics of multiple servers in combination with multiple electromagnetic interference information and multiple wiring harness bending information to obtain wiring harness fault diagnosis results. The present invention automatically identifies wire harnesses in a data center and identifies them in combination with historical wire harness information, with a high recognition accuracy. It then performs electromagnetic interference analysis and wire harness bending identification to perform fault diagnosis, thereby ensuring the transmission stability of the data center and the service life of the cables. It also performs integrated analysis in combination with data features of different servers, and can perform key evaluations in combination with the characteristics of server operating data, based on which it diagnoses cable fault status and provides accurate wire harness fault diagnosis results, thereby helping to promptly discover and resolve cable fault problems in the data center and improve the operating efficiency and maintenance convenience of the data center.
[0127] like Figure 2 As shown, based on the same inventive concept as the fault diagnosis method based on a data center provided in Example 1, an embodiment of the present invention provides a fault diagnosis device based on a data center, the device comprising:
[0128] An image acquisition module 11 is used to acquire monitoring images of a specified area in a target data center;
[0129] A harness identification module 12 is used to perform harness identification on the monitoring image, obtain network harness information of multiple network harnesses of multiple servers and power harness information of multiple power lines, wherein the identification is performed in combination with multiple historical harness information, the network harness information includes harness servers and network harness paths, and the multiple servers are arranged in the designated area;
[0130] The electromagnetic interference analysis module 13 is used to perform electromagnetic interference fault analysis according to the plurality of network cable harness information and the plurality of power cable harness information to obtain a plurality of electromagnetic interference information;
[0131] A bending analysis module 14, configured to perform wiring harness bending identification based on the plurality of network cable harness information and the plurality of power cable harness information in combination with the plurality of historical wiring harness information, and obtain a plurality of wiring harness bending information;
[0132] The fault diagnosis module 15 is used to perform integrated diagnosis and analysis of wiring harness faults according to the data features of the multiple servers, in combination with the multiple electromagnetic interference information and the multiple wiring harness bending information, to obtain wiring harness fault diagnosis results.
[0133] Furthermore, the wire harness recognition module 12 is further used to: perform wire harness image segmentation on the monitoring image to obtain a plurality of wire harness images;
[0134] According to the recognition history of the server harness, a sample harness image set is collected, and the harness in each sample harness image is labeled with type, attribution and path to obtain a sample harness information set, wherein each sample harness information includes a harness type, a harness server and a harness path, and the harness type is a network harness or a power harness;
[0135] Using the sample wire harness image set and the sample wire harness information set to train an integrated wire harness identification channel, wherein the integrated wire harness identification channel includes a plurality of wire harness identification branches;
[0136] Inputting the plurality of wire harness images into the plurality of wire harness identification branches respectively, identifying and obtaining a plurality of wire harness type sets, a plurality of wire harness server sets and a plurality of wire harness path sets;
[0137] Selecting the wiring harness types with the highest occurrence frequency in the plurality of wiring harness type sets respectively to obtain a plurality of wiring harness types;
[0138] fusing the plurality of wire beam path sets to obtain a plurality of fused wire beam paths;
[0139] According to the plurality of wiring harness server sets, a plurality of wiring harness server probability sets are calculated and obtained, a plurality of historical wiring harness paths and a plurality of historical wiring harness servers in a plurality of historical wiring harness information are combined, a distance from the plurality of fused wiring harness paths is calculated, a plurality of historical wiring harness server probability sets are obtained by allocation, a plurality of historical wiring harness server probability sets are combined, a server corresponding to the maximum plurality of server probabilities is screened, and a plurality of wiring harness servers are obtained;
[0140] In combination with the plurality of harness types, the plurality of fused harness paths and the plurality of harness servers, network harness information of the plurality of network harnesses of the plurality of servers and power harness information of the plurality of power lines are obtained.
[0141] The sample wire harness image set and the sample wire harness information set are used to train an integrated wire harness recognition channel, including:
[0142] Randomly selecting a preset proportion of sample data from the sample wire harness image set and the sample wire harness information set to train a first wire harness recognition branch;
[0143] Continue to randomly select a preset proportion of sample data from the sample wire harness image set and the sample wire harness information set to train other multiple wire harness recognition branches;
[0144] The plurality of wire harness recognition branches that have been trained are combined to obtain an integrated wire harness recognition channel.
[0145] Among them, according to the multiple wiring harness server sets, multiple wiring harness server probability sets are calculated, multiple historical wiring harness paths and multiple historical wiring harness servers in multiple historical wiring harness information are combined, the distances to the multiple fusion wiring harness paths are calculated, multiple historical wiring harness server probability sets are allocated, and the servers corresponding to the largest multiple server probabilities are screened in combination with the multiple wiring harness server probability sets, including:
[0146] Calculate and obtain multiple wire harness server probability sets according to the appearance ratios of multiple servers in the multiple wire harness server sets;
[0147] Acquire multiple historical harness paths and multiple historical harness servers in multiple historical harness information of the last harness fault diagnosis;
[0148] Calculating the distances between the multiple fused wire harness paths and the multiple historical wire harness paths respectively to obtain multiple path distance sets, and combining the multiple historical wire harness servers to obtain multiple historical wire harness server probability sets, wherein the size of the historical wire harness server probability is negatively correlated with the size of the distance;
[0149] The plurality of wire bundle server probability sets and the plurality of historical wire bundle server probability sets are combined to calculate and obtain a plurality of weighted server probability sets, and servers corresponding to the largest plurality of weighted server probabilities are respectively selected as a plurality of wire bundle servers.
[0150] Furthermore, the electromagnetic interference analysis module 13 is also used for:
[0151] According to the plurality of network cable harness information and the harness paths in the plurality of power cable harness information, obtaining the interference distance between the network cable harness of each server and the closest power cable harness;
[0152] Electromagnetic interference fault analysis is performed according to the multiple interference distances to obtain multiple electromagnetic interference levels as multiple electromagnetic interference information.
[0153] Furthermore, the bending analysis module 14 is also used for:
[0154] Acquire multiple network cable harness paths and multiple power cable harness paths in multiple network cable harness information and multiple power cable harness information;
[0155] According to the multiple network cable harness paths, a plurality of static network cable bending degrees are obtained, according to the multiple power cable harness paths, a plurality of static power cable bending degrees are obtained, and a plurality of static harness bending degrees are obtained by weighted calculation;
[0156] Acquire multiple historical network cable harness paths and multiple historical power cable harness paths in the multiple historical harness information;
[0157] Analyze and obtain multiple changed network cable bending degrees according to the multiple historical network cable harness paths and the multiple network cable harness paths, analyze and obtain multiple changed power cable bending degrees according to the multiple historical power cable harness paths and the multiple power cable harness paths, and obtain multiple changed harness bending degrees by weighted calculation;
[0158] According to the plurality of static harness bending degrees and the plurality of variable harness bending degrees, a plurality of harness bending levels are respectively obtained by classification as a plurality of harness bending information.
[0159] Furthermore, the fault diagnosis module 15 is also used for:
[0160] Acquire data characteristics of the multiple servers, and analyze to obtain multiple server grades;
[0161] According to the sizes of the multiple server levels, performing integrated weighted calculation on the multiple electromagnetic interference information and the multiple wire harness bending information to obtain weighted electromagnetic interference information and weighted wire harness bending information;
[0162] The wire harness fault level of the designated area is calculated and obtained according to the weighted electromagnetic interference information and the weighted wire harness bending level as a wire harness fault diagnosis result.
[0163] See also Figure 3 , Figure 3 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, an embodiment of the present invention provides an electronic device 100, including a memory 110, a processor 120, and a computer program 111 stored in the memory 110 and executable on the processor 120. When the processor 120 executes the computer program 111, the steps of the method in Example 1 are implemented.
[0164] See also Figure 4 , Figure 4 A schematic diagram of an embodiment of a computer-readable storage medium provided in an embodiment of the present invention. Figure 4 As shown, this embodiment provides a computer-readable storage medium 200 on which a computer program 211 is stored. When the computer program 211 is executed by a processor, the steps of the method in Embodiment 1 are implemented.
[0165] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0166] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0167] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0168] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0169] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0170] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0171] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A fault diagnosis method based on a data center, characterized in that: The method comprises: Collect surveillance images of designated areas within the target data center; Performing harness identification on the monitoring image to obtain network harness information of multiple network harnesses of multiple servers and power harness information of multiple power lines, wherein the identification is performed in combination with multiple historical harness information, the network harness information includes harness servers and network harness paths, and the multiple servers are arranged in the designated area; According to the plurality of network cable harness information and the plurality of power cable harness information, electromagnetic interference fault analysis is performed respectively to obtain a plurality of electromagnetic interference information; According to the plurality of network cable harness information and the plurality of power cable harness information, combined with the plurality of historical harness information, harness bending identification is performed to obtain a plurality of harness bending information; According to the data features of the multiple servers, combined with the multiple electromagnetic interference information and the multiple wiring harness bending information, a wiring harness fault integrated diagnosis analysis is performed to obtain a wiring harness fault diagnosis result; According to the data characteristics of the multiple servers, combined with the multiple electromagnetic interference information and the multiple harness bending information, the harness fault integrated diagnosis analysis is performed, including: Acquire data characteristics of the multiple servers, and analyze to obtain multiple server grades; According to the sizes of the multiple server levels, performing integrated weighted calculation on the multiple electromagnetic interference information and the multiple wire harness bending information to obtain weighted electromagnetic interference information and weighted wire harness bending information; The wire harness fault level of the designated area is calculated and obtained according to the weighted electromagnetic interference information and the weighted wire harness bending information as a wire harness fault diagnosis result.
2. The fault diagnosis method based on data center according to claim 1, characterized in that: Performing harness recognition on the monitoring image to obtain network harness information of multiple network harnesses of multiple servers and power harness information of multiple power lines includes: Performing wire beam image segmentation on the monitoring image to obtain a plurality of wire beam images; According to the recognition history of the server harness, a sample harness image set is collected, and the harness in each sample harness image is labeled with type, attribution and path to obtain a sample harness information set, wherein each sample harness information includes a harness type, a harness server and a harness path, and the harness type is a network harness or a power harness; Using the sample wire harness image set and the sample wire harness information set to train an integrated wire harness identification channel, wherein the integrated wire harness identification channel includes a plurality of wire harness identification branches; Inputting the plurality of wire harness images into the plurality of wire harness identification branches respectively, identifying and obtaining a plurality of wire harness type sets, a plurality of wire harness server sets and a plurality of wire harness path sets; Selecting the wiring harness types with the highest occurrence frequency in the plurality of wiring harness type sets respectively to obtain a plurality of wiring harness types; fusing the plurality of wire beam path sets to obtain a plurality of fused wire beam paths; According to the plurality of wiring harness server sets, a plurality of wiring harness server probability sets are calculated and obtained, a plurality of historical wiring harness paths and a plurality of historical wiring harness servers in a plurality of historical wiring harness information are combined, a distance from the plurality of fused wiring harness paths is calculated, a plurality of historical wiring harness server probability sets are obtained by allocation, a plurality of historical wiring harness server probability sets are combined, a server corresponding to the maximum plurality of server probabilities is screened, and a plurality of wiring harness servers are obtained; Combining the multiple harness types, the multiple fused harness paths, and the multiple harness servers, obtaining network harness information of multiple network harnesses of multiple servers and power harness information of multiple power lines; Among them, according to the multiple wiring harness server sets, multiple wiring harness server probability sets are calculated, multiple historical wiring harness paths and multiple historical wiring harness servers in multiple historical wiring harness information are combined, the distances to the multiple fusion wiring harness paths are calculated, and multiple historical wiring harness server probability sets are allocated to obtain multiple historical wiring harness server probability sets, and the servers corresponding to the largest multiple server probabilities are screened in combination with the multiple wiring harness server probability sets to obtain multiple wiring harness servers, including: Calculate and obtain multiple wire harness server probability sets according to the appearance ratios of multiple servers in the multiple wire harness server sets; Acquire multiple historical harness paths and multiple historical harness servers in multiple historical harness information of the last harness fault diagnosis; Calculating the distances between the multiple fused wire harness paths and the multiple historical wire harness paths respectively to obtain multiple path distance sets, and combining the multiple historical wire harness servers to obtain multiple historical wire harness server probability sets, wherein the size of the historical wire harness server probability is negatively correlated with the size of the distance; The plurality of wire bundle server probability sets and the plurality of historical wire bundle server probability sets are combined to calculate and obtain a plurality of weighted server probability sets, and servers corresponding to the largest plurality of weighted server probabilities are respectively selected as a plurality of wire bundle servers.
3. The fault diagnosis method based on data center according to claim 2, characterized in that: The sample wire harness image set and the sample wire harness information set are used to train an integrated wire harness recognition channel, including: Randomly selecting a preset proportion of sample data from the sample wire harness image set and the sample wire harness information set to train a first wire harness recognition branch; Continue to randomly select a preset proportion of sample data from the sample wire harness image set and the sample wire harness information set to train other multiple wire harness recognition branches; The plurality of wire harness recognition branches that have been trained are combined to obtain an integrated wire harness recognition channel.
4. The fault diagnosis method based on data center according to claim 1, characterized in that: According to the plurality of network cable harness information and the plurality of power cable harness information, electromagnetic interference fault analysis is performed respectively to obtain a plurality of electromagnetic interference information, including: According to the plurality of network cable harness information and the harness paths in the plurality of power cable harness information, obtaining the interference distance between the network cable harness of each server and the closest power cable harness; Electromagnetic interference fault analysis is performed according to the multiple interference distances to obtain multiple electromagnetic interference levels as multiple electromagnetic interference information.
5. The fault diagnosis method based on data center according to claim 1, characterized in that: According to the plurality of network cable harness information and the plurality of power cable harness information, combined with the plurality of historical harness information, harness bending identification is performed to obtain a plurality of harness bending information, including: Acquire multiple network cable harness paths and multiple power cable harness paths in multiple network cable harness information and multiple power cable harness information; According to the multiple network cable harness paths, a plurality of static network cable bending degrees are obtained, according to the multiple power cable harness paths, a plurality of static power cable bending degrees are obtained, and a plurality of static harness bending degrees are obtained by weighted calculation; Acquire multiple historical network cable harness paths and multiple historical power cable harness paths in the multiple historical harness information; Analyze and obtain multiple changed network cable bending degrees according to the multiple historical network cable harness paths and the multiple network cable harness paths, analyze and obtain multiple changed power cable bending degrees according to the multiple historical power cable harness paths and the multiple power cable harness paths, and obtain multiple changed harness bending degrees by weighted calculation; According to the plurality of static harness bending degrees and the plurality of variable harness bending degrees, a plurality of harness bending levels are respectively obtained by classification as a plurality of harness bending information.
6. A fault diagnosis device based on a data center, characterized in that: The device comprises: An image acquisition module, used to acquire monitoring images of a specified area in a target data center; a harness identification module, used to perform harness identification on the monitoring image, and obtain network harness information of multiple network harnesses of multiple servers and power harness information of multiple power lines, wherein the identification is performed in combination with multiple historical harness information, the network harness information includes harness servers and network harness paths, and the multiple servers are arranged in the designated area; An electromagnetic interference analysis module, used to perform electromagnetic interference fault analysis respectively according to the plurality of network cable harness information and the plurality of power cable harness information, and obtain a plurality of electromagnetic interference information; A bending analysis module, used to identify the bending of the wiring harness according to the plurality of network cable harness information and the plurality of power cable harness information, combined with the plurality of historical wiring harness information, to obtain a plurality of wiring harness bending information; A fault diagnosis module, configured to perform integrated diagnosis and analysis of wiring harness faults according to the data features of the multiple servers, in combination with the multiple electromagnetic interference information and the multiple wiring harness bending information, to obtain wiring harness fault diagnosis results; According to the data characteristics of the multiple servers, combined with the multiple electromagnetic interference information and the multiple harness bending information, the harness fault integrated diagnosis analysis is performed, including: Acquire data characteristics of the multiple servers, and analyze to obtain multiple server grades; According to the sizes of the multiple server levels, performing integrated weighted calculation on the multiple electromagnetic interference information and the multiple wire harness bending information to obtain weighted electromagnetic interference information and weighted wire harness bending information; The wire harness fault level of the designated area is calculated and obtained according to the weighted electromagnetic interference information and the weighted wire harness bending information as a wire harness fault diagnosis result.
7. An electronic device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the data center-based fault diagnosis method described in any one of claims 1-5.
8. A non-transitory computer-readable storage medium, characterized in that: The storage medium stores a computer software program, and when the computer software program is executed by the processor, the fault diagnosis method based on the data center as described in any one of claims 1 to 5 is implemented.
