Methods, devices, electronic equipment, and storage media for determining the probability of equipment failure.
By acquiring device status and link status information and using graph convolutional networks to calculate the probability of device failure, the problem of untimely and inefficient device fault detection in existing technologies is solved, and global fault self-discovery and accurate location are achieved.
Patent Information
- Application Number
- CN202310640200.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-31
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2043-05-31
AI Technical Summary
Existing equipment fault detection methods suffer from problems such as untimely detection and low efficiency. In particular, single-point fault detection methods rely on expert experience, resulting in low fault detection efficiency.
By acquiring the device status information and information transmission link status information of multiple devices to be tested, an adjacency feature weight matrix and feature status matrix of graph data are generated. Feature extraction is performed using a graph convolutional network to calculate the probability of device failure, thereby achieving global fault self-discovery and accurate location.
It improves the efficiency of equipment fault detection, enables proactive early detection of faulty equipment, supports fault tracing and precise location positioning, and avoids the problem of untimely detection in existing technologies.
Smart Images

Figure CN116662052B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of financial technology and other related technical fields. Specifically, it relates to a method, apparatus, electronic device, and storage medium for determining the probability of device failure. Background Technology
[0002] In existing technologies, the status of a device is usually detected by sending a heartbeat message to a single device. If the device can return a reply message to the heartbeat message in a timely manner, it means that the device is in a normal state. If the device fails to return a reply message to the heartbeat message in a timely manner or sends a specific event fault message, it means that the device has malfunctioned and is in an abnormal state.
[0003] However, the aforementioned fault detection methods are reactive, meaning that faults can only be detected some time after they occur. Furthermore, these methods are single-point fault detection methods, meaning they can only detect faults from one side of the equipment and require expert analysis of single-point alarm information to locate the fault and assess its impact. Therefore, existing equipment fault detection methods suffer from problems of untimely fault detection and low efficiency.
[0004] There is currently no effective solution to the above problems. Summary of the Invention
[0005] This application provides a method, apparatus, electronic device, and storage medium for determining the probability of equipment failure, so as to at least solve the technical problem of low failure detection efficiency in the prior art when performing equipment failure detection.
[0006] According to one aspect of this application, a method for determining the failure probability of a device is provided, comprising: acquiring device status information of M devices to be tested, wherein the M devices to be tested form N information transmission links, each information transmission link consists of at least two devices to be tested, and M and N are both positive integers; determining link status information corresponding to each information transmission link based on the device status information of the devices to be tested in each information transmission link; generating a first matrix based on the device status information of the M devices to be tested, and generating a second matrix based on the link status information corresponding to each information transmission link, wherein the first matrix is an adjacency feature weight matrix corresponding to graph data, the graph data includes M graph nodes and multiple edges, each graph node corresponds to a device to be tested, each edge is used to characterize the connection relationship between the graph nodes at both ends of the edge, and the second matrix is a feature state matrix corresponding to the graph data; performing a graph convolution operation on the first matrix and the second matrix to obtain a target matrix, wherein the target matrix consists of M values, each value corresponding to a device to be tested; and determining the failure probability of each device to be tested based on the values in the target matrix.
[0007] Furthermore, the method for determining the probability of device failure also includes: Step 1, obtaining the first status information of the i-th device among the M devices to be tested, wherein the first status information is used to characterize whether the i-th device to be tested successfully sends or receives a heartbeat message at a preset time, and the i-th device to be tested is any one of the M devices to be tested; Step 2, obtaining the second status information of the i-th device to be tested, wherein the second status information is used to characterize the memory usage rate of the i-th device to be tested within a preset time period; Step 3, determining the device status information corresponding to the i-th device to be tested based on the first status information and the second status information of the i-th device to be tested; Step 4, repeating the process from Step 1 to Step 3 until the device status information of the M devices to be tested is obtained.
[0008] Furthermore, the method for determining the probability of device failure also includes: determining a first score corresponding to the i-th device under test based on the first status information of the i-th device under test, wherein the score of the first score is either a first score or a second score, the first score being the score corresponding to the first score when the i-th device under test successfully sends or receives a heartbeat message at a preset time, and the second score being the score corresponding to the first score when the i-th device under test fails to send or receive a heartbeat message at a preset time; determining a target value interval from multiple preset value intervals based on the memory usage rate of the i-th device under test during a preset time period, wherein the target value interval is the value interval in which the memory usage rate is located, and each preset interval corresponds to a preset score; using the preset score corresponding to the target value interval as the second score corresponding to the i-th device under test; and determining the device status information of the i-th device under test based on the first score and the second score.
[0009] Furthermore, the method for determining the failure probability of the equipment also includes: obtaining a first preset weight corresponding to the first state information and a second preset weight corresponding to the second state information; performing a weighted summation calculation on the first score and the second score according to the first preset weight and the second preset weight to obtain the equipment state score corresponding to the i-th device to be tested, wherein the equipment state score is used to characterize the equipment state information of the i-th device to be tested.
[0010] Furthermore, the method for determining the failure probability of the equipment also includes: obtaining the equipment status score corresponding to the device under test in each information transmission link; obtaining the preset weight value corresponding to the device under test in each information transmission link, wherein the preset weight value corresponding to each device under test is used to characterize the importance of the device under test to the information transmission link in which the device under test is located; and calculating the weighted average of the equipment status scores corresponding to the devices under test in each information transmission link based on the preset weight values corresponding to the devices under test in each information transmission link to obtain the link status score corresponding to the information transmission link, wherein the link status score is used to characterize the link status information corresponding to the information transmission link.
[0011] Furthermore, the method for determining the failure probability of the equipment also includes: generating a third matrix based on the equipment status scores corresponding to M devices to be tested, wherein each value in the third matrix is a device status score corresponding to a device to be tested; obtaining the graph adjacency matrix corresponding to the graph data, wherein the graph adjacency matrix is used to characterize the adjacency relationship between graph nodes in the graph data; calculating the product of the graph adjacency matrix and the third matrix to obtain a first matrix; and generating a second matrix based on the link status score corresponding to each information transmission link, wherein each value in the second matrix is a link status score corresponding to an information transmission link.
[0012] Furthermore, the method for determining the failure probability of the equipment also includes: obtaining the symmetric normalized matrix corresponding to the first matrix; performing a first graph convolution operation on the symmetric normalized matrix and the second matrix through the first layer of the graph convolutional network model, wherein the first graph convolution operation is used to calculate the product of the symmetric normalized matrix, the second matrix, and the first weight matrix to obtain the third matrix, and the first weight matrix is the weight matrix corresponding to the first layer of the graph convolutional network; activating the third matrix through an activation function to obtain the activation function value, wherein the activation function value is a matrix; performing a second graph convolution operation on the activation function value and the symmetric normalized matrix through the second layer of the graph convolutional network model, wherein the second graph convolution operation is used to calculate the product of the activation function value, the symmetric normalized matrix, and the second weight matrix to obtain the fourth matrix, and the second weight matrix is the weight matrix corresponding to the second layer of the graph convolutional network; generating the target matrix based on the fourth matrix.
[0013] Furthermore, the method for determining the failure probability of the equipment also includes: normalizing the fourth matrix into a probability distribution vector, wherein the probability distribution vector is represented in matrix form; and using the probability distribution vector as the target matrix.
[0014] Furthermore, the method for determining the probability of equipment failure also includes: after performing graph convolution operation based on the first matrix and the second matrix to obtain the target matrix, obtaining the device identifier of the device to be tested corresponding to the target value in the target matrix, wherein the target value is a value in the target matrix that is greater than a preset threshold; generating prompt information based on the device identifier, wherein the prompt information is used to prompt the target object to troubleshoot the device to be tested corresponding to the device identifier.
[0015] According to another aspect of this application, an apparatus for determining the probability of device failure is also provided, comprising: a device information acquisition module for acquiring device status information of M devices to be tested, wherein the M devices to be tested form N information transmission links, each information transmission link consists of at least two devices to be tested, and M and N are both positive integers; a link information acquisition module for determining the link status information corresponding to each information transmission link based on the device status information of the devices to be tested in each information transmission link; an information processing module for generating a first matrix based on the device status information of the M devices to be tested, and generating a second matrix based on the link status information corresponding to each information transmission link, wherein the first matrix is an adjacency feature weight matrix corresponding to graph data, the graph data includes M graph nodes and multiple edges, each graph node corresponds to a device to be tested, each edge is used to characterize the connection relationship between the graph nodes at both ends of the edge, and the second matrix is a feature state matrix corresponding to the graph data; a calculation module for performing graph convolution operation based on the first matrix and the second matrix to obtain a target matrix, wherein the target matrix consists of M values, each value corresponding to a device to be tested; and a result output module for determining the probability of each device to be tested failing based on the values in the target matrix.
[0016] According to another aspect of this application, a computer-readable storage medium is also provided, which stores a computer program, wherein the computer program controls the device where the computer-readable storage medium is located to perform the above-described method for determining the probability of device failure when the computer program is running.
[0017] According to another aspect of this application, an electronic device is also provided, wherein the electronic device includes one or more processors and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method for determining the probability of failure of the device described above.
[0018] In this application, the method involves acquiring the device status information of multiple devices under test and combining it with the link status information of the information transmission link in which each device under test resides. First, the device status information of M devices under test is acquired, where the M devices form N information transmission links, and each information transmission link consists of at least two devices under test. M and N are both positive integers. Then, the link status information corresponding to each information transmission link is determined based on the device status information of the devices under test in that link. Subsequently, a first matrix is generated based on the device status information of the M devices under test, and a second matrix is generated based on the link status information corresponding to each information transmission link. The first matrix is an adjacency feature weight matrix corresponding to graph data, where the graph data includes M graph nodes and multiple edges. Each graph node corresponds to one device under test, and each edge represents the connection relationship between the graph nodes at both ends of the edge. The second matrix is a feature state matrix corresponding to the graph data. Finally, a graph convolution operation is performed on the first and second matrices to obtain a target matrix. The probability of each device under test failing is determined based on the values in the target matrix, where the target matrix consists of M values, each corresponding to one device under test.
[0019] As described above, this application obtains the device status information of the device under test and the link status information of the information transmission link where the device under test is located. It then analyzes and transforms the device status information into a first matrix and the link status information into a second matrix. A graph convolution operation is performed between the first and second matrices to obtain a target matrix containing the failure probability information of the device under test, thus providing a method for determining the failure probability of a device. Furthermore, this application uses a graph convolutional network to extract features from both the device status information and the link status information, and relies on the global failure incidence rate for device failure early warning. This achieves graph-based global fault self-discovery and supports device fault backtracking and precise fault location, thereby avoiding the technical problem of untimely device fault detection in existing technologies, which leads to low device fault detection efficiency.
[0020] Therefore, the technical solution of this application achieves the goal of determining the failure probability of the device under test by obtaining the device status information of the device under test and the link status information of the information transmission link where the device under test is located, and performing graph analysis and graph convolution calculation on the information. This achieves the technical effect of improving the efficiency of determining the failure probability of the device, and solves the technical problem of low failure detection efficiency in the prior art when performing fault detection on the device. Attached Figure Description
[0021] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:
[0022] Figure 1 This is a flowchart of an optional method for determining the probability of device failure according to an embodiment of this application;
[0023] Figure 2 This is a flowchart of an optional method for obtaining the status information of the i-th device to be detected, according to an embodiment of this application.
[0024] Figure 3 This is a schematic diagram of an optional device for determining the probability of device failure according to an embodiment of this application;
[0025] Figure 4 This is a schematic diagram of an optional electronic device according to an embodiment of this application. Detailed Implementation
[0026] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.
[0027] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0028] It should also be noted that all relevant information (including but not limited to user device information and user personal information) and data (including but not limited to data used for display and analysis) involved in this application are information and data authorized by the user or fully authorized by all parties. For example, if there is an interface between this system and the relevant user or organization, before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain the relevant information only after receiving consent from the aforementioned user or organization.
[0029] The present application will be further described below with reference to various embodiments.
[0030] Example 1
[0031] According to an embodiment of this application, an embodiment of a method for determining the probability of device failure is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.
[0032] This application provides a device failure probability determination system for performing the device failure probability determination method in the embodiments of this application. Figure 1 This is a flowchart of an optional method for determining the failure probability of a device according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:
[0033] Step S101: Obtain the device status information of M devices to be tested.
[0034] In step S101, M devices to be tested form N information transmission links, and each information transmission link consists of at least two devices to be tested, where M and N are both positive integers.
[0035] In addition, in step S101, the device to be detected includes, but is not limited to, devices such as computers, network switches, servers, intelligent monitoring devices, and displays. Optionally, the device status information includes, but is not limited to, information on whether the device to be detected has received a heartbeat message, the memory usage rate of the device to be detected, and the CPU usage rate of the device to be detected. The device status information of a device to be detected can be represented by a device status score corresponding to that device. When calculating the device status score, the score and preset weight corresponding to each device status information can be determined based on multiple different device status information corresponding to the device to be detected, and then the device status score of the device to be detected is calculated by weighted summation.
[0036] Step S102: Determine the link status information corresponding to the information transmission link based on the device status information of the device to be detected in each information transmission link.
[0037] In step S102, the link status information includes, but is not limited to, interface IP address, subnet mask, network type, link cost, and device status information of each device to be detected included on the information transmission link. The link status information of an information transmission link can be represented by the link status score corresponding to the information transmission link. Optionally, there may be multiple devices to be detected on a certain information transmission link. In this case, when calculating the link status score of the information transmission link, a corresponding preset weight value can be set for each device to be detected in the information transmission link, and the link status score of the information transmission link can be calculated by weighted average in combination with the device status score corresponding to each device to be detected in the information transmission link. The preset weight value can be determined according to the importance of the device to be detected to the information transmission link.
[0038] Step S103: Generate a first matrix based on the device status information of M devices to be detected, and generate a second matrix based on the link status information corresponding to each information transmission link.
[0039] In step S103, the first matrix is the adjacency feature weight matrix corresponding to the graph data. The graph data includes M graph nodes and multiple edges. Each graph node corresponds to a device to be detected. Each edge is used to characterize the connection relationship between the graph nodes at both ends of the edge. The second matrix is the feature state matrix corresponding to the graph data.
[0040] Furthermore, in step S103, the graph data does not limit the number of graph nodes or the number of edges connected. The attributes or features of the graph nodes and edges can include various types of data. Therefore, the attributes of the graph nodes or edges in the graph data are comprehensive, capable of representing the actual state of the graph nodes and processing the basic indicators to a certain extent into readable fault-sensitive information. In this application, the graph data consists of M graph nodes and multiple edges. Each graph node corresponds to a device to be detected, and each edge is used to represent the connection relationship between the devices to be detected at both ends of the edge.
[0041] Optionally, the process of generating the first matrix from the device status information of the M devices to be tested is exemplified as follows: Generate matrix A based on the device status information of the M devices to be tested, where matrix A contains the status scores of the corresponding M devices; obtain the graph adjacency matrix B corresponding to the graph data (a graph adjacency matrix is a two-dimensional array that represents a graph and can be used to characterize the adjacency relationship between graph nodes in the graph data); calculate the product of matrix A and matrix B to obtain the first matrix C.
[0042] Step S104: Perform graph convolution operation on the first matrix and the second matrix to obtain the target matrix.
[0043] In step S104, the target matrix consists of M values, each value corresponding to a device to be detected.
[0044] Additionally, in step S104, the graph convolution operation process is exemplified as follows: Obtain the symmetric normalized matrix corresponding to the first matrix C, and name the symmetric normalized matrix as follows: The second matrix D is processed through a first-layer graph convolutional network to perform the first graph convolution operation, resulting in the third matrix E. The ReLU activation function is then applied to activate the third matrix E, yielding the corresponding activation matrix. Finally, the activation matrix and the symmetric normalization matrix are processed through a second-layer graph convolutional network. The second graph convolution operation is performed to obtain the fourth matrix. Finally, the fourth matrix is normalized to a probability distribution vector. The final probability distribution vector is represented in matrix form to obtain the target matrix.
[0045] Step S105: Determine the probability of each device under test malfunctioning based on the values in the target matrix.
[0046] In step S105, the target matrix contains M values, each representing a probability value. These M values correspond to the failure probabilities of the M devices under test. In this application, the failure probabilities of each device under test can be visualized by sending the probability information to a display terminal device. The display methods include, but are not limited to, charts, tables, or bar graphs. Optionally, this application can also preset a failure threshold for each device under test. When the target value of a device under test in the target matrix exceeds the preset threshold, an early warning message is generated, prompting the target object (device user or maintenance personnel) to investigate the device under test.
[0047] As described above, this application obtains the device status information of the device under test and the link status information of the information transmission link where the device under test is located. It then analyzes and transforms the device status information into a first matrix and the link status information into a second matrix. A graph convolution operation is performed between the first and second matrices to obtain a target matrix containing the target value of the failure probability of the device under test, thus providing a method for determining the failure probability of a device. Furthermore, this application extracts features from both the device status information and the link status information using a graph convolutional network, and relies on the global failure incidence rate for device failure early warning. This achieves the technical effect of graph-based global fault self-discovery and supports fault backtracking and precise fault location of the device under test, thereby avoiding the technical problem of low device fault detection efficiency in existing technologies.
[0048] Therefore, the technical solution of this application achieves the goal of determining the failure probability of the device under test by obtaining the device status information of the device under test and the link status information of the information transmission link where the device under test is located, and performing graph analysis and graph convolution calculation on the information. This achieves the technical effect of improving the efficiency of determining the failure probability of the device, and solves the technical problem of low failure detection efficiency in the prior art when performing fault detection on the device.
[0049] Figure 2 This is a flowchart of an optional method for obtaining the status information of the i-th device to be detected, according to an embodiment of this application, such as... Figure 2 As shown, the method includes the following steps:
[0050] Step S201: Obtain the first status information of the i-th device among the M devices to be tested.
[0051] In step S201, the first status information is used to characterize whether the i-th device to be detected successfully sends or receives a heartbeat message at a preset time. The i-th device to be detected is any one of the M devices to be detected.
[0052] It should be noted that the preset time can be customized according to different types of devices under test. For example, if device 1 under test is a server, an average reception time a1 can be calculated based on the server's historical heartbeat message reception time records. If the heartbeat message sending time of device 1 under test is b1, then the preset reception time of device 1 under test is b1+a1. If device 2 under test is a monitor, an average reception time a2 can be calculated based on the monitor's historical heartbeat message reception time records. If the heartbeat message sending time of device 2 under test is b2, then the preset reception time of device 2 under test is b2+a2. The values of a1 and a2 can be different.
[0053] Step S202: Obtain the second status information of the i-th device to be tested.
[0054] In step S202, the second state information is used to characterize the memory usage rate of the i-th device to be detected within a preset time period.
[0055] Optionally, the second state information can also be used to characterize the CPU utilization rate of the i-th device under test within a preset time period. The preset time period can be defined and set according to different types of devices under test, and the preset time periods can be different for different types of devices under test.
[0056] Furthermore, the second state information can also be used to characterize the i-th device under test within a preset time period, including but not limited to memory usage, CPU usage, and I / O resource utilization. When characterizing multiple device state information, the weights of the memory usage, CPU usage, and I / O resource utilization of the i-th device under test can be preset. The device state information can be represented by a device state score, and a weighted average of the multiple device state scores of the i-th device under test can be calculated to obtain the second state information of the i-th device under test.
[0057] Step S203: Determine the device status information corresponding to the i-th device under test based on the first status information and the second status information of the i-th device under test.
[0058] Specifically, in step S203, the first preset weight corresponding to the first state information of the i-th device to be tested and the second preset weight corresponding to the second state information can be set in advance, and the first score corresponding to the first state information of the i-th device to be tested and the second score corresponding to the second state information can be obtained, and the first score and the second score can be calculated by weighted summation.
[0059] Finally, repeat the process from S201 to S203 until the device status information of M devices to be tested is obtained.
[0060] In one optional embodiment, the device failure probability determination system can determine a first score corresponding to the i-th device under test based on the first state information of the i-th device under test. The first score is either a first value or a second value. The first value is the score corresponding to the first score when the i-th device under test successfully sends or receives a heartbeat message at a preset time, and the second value is the score corresponding to the first score when the i-th device under test fails to send or receive a heartbeat message at the preset time. Furthermore, the device failure probability determination system can also determine a target value interval from multiple preset value intervals based on the memory usage rate of the i-th device under test over a preset time period. The target value interval is the value interval in which the memory usage rate falls, and each preset interval corresponds to a preset score. Then, the device failure probability determination system uses the preset score corresponding to the target value interval as the second score corresponding to the i-th device under test. Finally, the device failure probability determination system determines the device state information of the i-th device under test based on the first score and the second score.
[0061] For example, if the i-th device under test successfully sends (or receives) a heartbeat message at a preset time, the score corresponding to the first score is 100. If it fails to send (or receive) a heartbeat message at the preset time, the score corresponding to the first score is 50. When the memory utilization rate of the i-th device under test is <50%, the corresponding second score is 100. When 50% ≤ memory utilization rate <70%, the corresponding second score is 80. When 70% ≤ memory utilization rate <90%, the corresponding second score is 60. When 90% ≤ memory utilization rate ≤100%, the corresponding second score is 40.
[0062] In one optional embodiment, the device failure probability determination system can first obtain a first preset weight corresponding to the first state information and a second preset weight corresponding to the second state information; then, the device failure probability determination system performs a weighted summation calculation on the first score and the second score according to the first preset weight and the second preset weight to obtain the device state score corresponding to the i-th device to be detected, wherein the device state score is used to characterize the device state information of the i-th device to be detected.
[0063] Optionally, the first preset weight and the second preset weight can be preset by the user according to the status of the device under test. For example, if the i-th device under test fails to send a heartbeat message at a preset time, the score corresponding to the first score is 50, and the memory utilization rate of the i-th device under test is 30%, then the score corresponding to the second score is 100. The first preset weight is 60%, and the second preset weight is 40%. In this case, the device status score corresponding to the i-th device under test is 50*60%+100*40%=70.
[0064] In one optional embodiment, the device failure probability determination system can determine the link status information corresponding to each information transmission link based on the device status information of the device under test in each information transmission link. The steps are as follows: First, the device failure probability determination system obtains the device status score corresponding to the device under test in each information transmission link; next, the device failure probability determination system obtains the preset weight value corresponding to the device under test in each information transmission link, wherein the preset weight value corresponding to each device under test is used to characterize the importance of the device under test to the information transmission link in which the device under test is located; then, the device failure probability determination system calculates a weighted average of the device status scores corresponding to the devices under test in each information transmission link based on the preset weight values corresponding to the devices under test in each information transmission link, to obtain the link status score corresponding to the information transmission link, wherein the link status score is used to characterize the link status information corresponding to the information transmission link.
[0065] Optionally, each information transmission link can correspond to multiple devices of different categories to be tested. The preset weight value corresponding to each device to be tested can be evaluated according to the importance of the device to be tested to the information transmission link. For example, there are three devices to be tested in the current link: monitor A1, switch A2, and computer A3. At this time, the switch has the greatest impact on the current information transmission link, while the monitor, as the display terminal, has the least impact. Therefore, the preset weight values of A1, A2, and A3 can be set to 20%, 50%, and 30%, respectively. Assuming that the obtained device status scores of A1, A2, and A3 are 70, 100, and 60, respectively, the score of the current information transmission link is 70*20% + 100*50% + 60*30% = 82.
[0066] In an optional embodiment, the device failure probability determination system can generate a first matrix based on the device status information of M devices to be tested, and generate a second matrix based on the link status information corresponding to each information transmission link. The steps include: First, the device failure probability determination system can generate a third matrix based on the device status scores corresponding to the M devices to be tested, wherein each value in the third matrix is a device status score corresponding to a device to be tested; Second, the device failure probability determination system can obtain a graph adjacency matrix corresponding to the graph data, wherein the graph adjacency matrix is used to characterize the adjacency relationship between graph nodes in the graph data; Calculate the product of the graph adjacency matrix and the third matrix to obtain the first matrix; Finally, the device failure probability determination system can generate a second matrix based on the link status score corresponding to each information transmission link, wherein each value in the second matrix is a link status score corresponding to an information transmission link, that is:
[0067] C = AB
[0068] Among them, the third matrix A is generated based on the device status information scores of M devices to be tested; the second matrix D is generated based on the link status scores corresponding to each information transmission link; B is the graph adjacency matrix corresponding to the graph data; and the first matrix C is equal to the product of the graph adjacency matrix and the third matrix.
[0069] In one optional embodiment, the device failure probability determination system performs a graph convolution operation on a first matrix and a second matrix to obtain a target matrix, including: obtaining the symmetric normalized matrix corresponding to the first matrix, i.e.:
[0070]
[0071] in, Z is the symmetric normalized matrix corresponding to the first matrix C, and Z is the degree matrix.
[0072] The first layer of the graph convolutional network performs a first graph convolution operation on the symmetric normalized matrix and the second matrix. This first graph convolution operation calculates the product of the symmetric normalized matrix, the second matrix, and the first weight matrix to obtain the third matrix. The first weight matrix is the weight matrix corresponding to the first layer of the graph convolutional network. The first layer graph convolution operation is as follows:
[0073]
[0074] Where D is the second matrix generated based on the link state score corresponding to each information transmission link, W0 is the first weight matrix of the first layer graph convolutional network, E is the third matrix, and GCN 1st Characterize the first layer of the graph convolutional network.
[0075] The third matrix is activated using an activation function to obtain the activation function value, which is a matrix. A second graph convolution operation is then performed on the activation function value and the symmetric normalization matrix using the second layer of the graph convolutional network model. This second graph convolution operation calculates the product of the activation function value, the symmetric normalization matrix, and the second weight matrix, resulting in the fourth matrix. The second weight matrix is the weight matrix corresponding to the second layer of the graph convolutional network. Finally, the target matrix is generated based on the fourth matrix.
[0076]
[0077] Here, ReLU is the activation function, which is used in the second layer of the graph convolutional network GCN. 2nd The ReLU activation function operates on neurons in the graph convolutional network, mapping the inputs to the outputs of the neurons. The introduction of the ReLU activation function also increases the nonlinearity of the graph convolutional network model. In the above formula, F is the fourth matrix; W1 is the second weight matrix of the second layer of the graph convolutional network. This is the activation function value.
[0078] In an optional embodiment, the device failure probability determination system can generate a target matrix using a fourth matrix, the steps of which include: normalizing the fourth matrix into a probability distribution vector, wherein the probability distribution vector is represented in matrix form; and using the probability distribution vector as the target matrix, i.e.:
[0079]
[0080] Specifically, the probability distribution vectorization of the fourth matrix F can be achieved using the softmax function, also known as the normalization exponential function. This function can represent the results of multi-classification in the form of probabilities, where G is the target matrix.
[0081] In an optional embodiment, after the device failure probability determination system performs graph convolution operation based on the first matrix and the second matrix to obtain the target matrix, the method further includes: first, the device failure probability determination system obtains the device identifier of the device to be detected corresponding to the target value in the target matrix, wherein the target value is a value in the target matrix that is greater than a preset threshold; second, the device failure probability determination system generates prompt information based on the device identifier, wherein the prompt information is used to prompt the target object to troubleshoot the device to be detected corresponding to the device identifier.
[0082] It should be noted that the preset threshold values for different devices under test can be different. For example, the preset threshold for device 1 can be set to 4 / 5, and the preset threshold for device 2 can be set to 2 / 5. If the target values for device 1 and device 2 in the target matrix are 9 / 10 and 1 / 5 respectively, then a prompt message will be generated for device 1, and device 2 will be judged to be operating normally and will not trigger an alarm. Furthermore, the methods for generating device prompt messages include, but are not limited to, email linkage, voice broadcast, and SMS notifications.
[0083] In this application, the method of acquiring the device status information of multiple devices to be tested and combining it with the link status information of the information transmission link where each device to be tested is located is adopted. The probability of each device to be tested failing is obtained through graph data analysis and graph convolution calculation, thereby achieving the purpose of proactively discovering faulty devices in advance and improving the timeliness of faulty device discovery. This achieves the technical effect of improving the detection efficiency of device faults and solves the technical problem of low fault detection rate when performing fault detection on devices in the prior art.
[0084] Furthermore, since this application also combines the information transmission link where each device under test is located to analyze the failure probability of each device under test, it achieves the technical effect of comprehensively analyzing the failure probability of the device under test from both the information transmission link side and the single device side, which is conducive to improving the accuracy of the analysis of device failure probability.
[0085] Example 2
[0086] This embodiment provides a device for determining the probability of equipment failure, wherein each implementation unit / module in the device corresponds to each implementation step in Embodiment 1.
[0087] Figure 3 This is a schematic diagram of an optional device for determining the probability of device failure, provided according to an embodiment of this application. Figure 3 As shown, it includes: a device information acquisition module 301, a link information acquisition module 302, an information processing module 303, a calculation module 304, and a result output module 305.
[0088] Optionally, the device information acquisition module is used to acquire device status information of M devices to be tested, wherein the M devices to be tested form N information transmission links, and each information transmission link consists of at least two devices to be tested, where M and N are both positive integers. The link information acquisition module is used to determine the link status information corresponding to each information transmission link based on the device status information of the devices to be tested in each information transmission link. The information processing module is used to generate a first matrix based on the device status information of the M devices to be tested, and to generate a second matrix based on the link status information corresponding to each information transmission link, wherein the first matrix is an adjacency feature weight matrix corresponding to the graph data, the graph data includes M graph nodes and multiple edges, each graph node corresponds to a device to be tested, and each edge is used to characterize the connection relationship between the graph nodes at both ends of the edge, and the second matrix is a feature state matrix corresponding to the graph data. The calculation module is used to perform graph convolution operation based on the first matrix and the second matrix to obtain a target matrix, wherein the target matrix consists of M values, each value corresponding to a device to be tested. The result output module is used to determine the probability of each device to be tested malfunctioning based on the values in the target matrix.
[0089] Optionally, the device information acquisition module includes: a first execution submodule, a second execution submodule, a third execution submodule, and a fourth execution submodule. The first execution submodule is used to execute step 1, acquiring the first status information of the i-th device among the M devices to be tested, wherein the first status information indicates whether the i-th device successfully sends or receives a heartbeat message at a preset time, and the i-th device can be any one of the M devices to be tested; the second execution submodule is used to execute step 2, acquiring the second status information of the i-th device to be tested, wherein the second status information indicates the memory usage rate of the i-th device to be tested within a preset time period; the third execution submodule is used to execute step 3, determining the device status information corresponding to the i-th device based on the first and second status information of the i-th device to be tested; the fourth execution submodule is used to execute step 4, repeating steps 1 to 3 until the device status information of the M devices to be tested is obtained.
[0090] Optionally, the third execution submodule includes: a first determining unit, a second determining unit, a scoring processing unit, and a third determining unit. The first determining unit is used to determine a first score corresponding to the i-th device under test based on the first status information of the i-th device under test. The first score is either a first value or a second value. The first value is the score corresponding to the first score when the i-th device under test successfully sends or receives a heartbeat message at a preset time, and the second value is the score corresponding to the first score when the i-th device under test fails to send or receive a heartbeat message at the preset time. The second determining unit is used to determine a target value interval from multiple preset value intervals based on the memory usage rate of the i-th device under test over a preset time period. The target value interval is the value interval in which the memory usage rate falls, and each preset interval corresponds to a preset score. The scoring processing unit is used to use the preset score corresponding to the target value interval as the second score corresponding to the i-th device under test. The third determining unit is used to determine the device status information of the i-th device under test based on the first score and the second score.
[0091] Optionally, the third determining unit includes: a first acquisition subunit and a first calculation subunit. The first acquisition subunit is used to acquire a first preset weight corresponding to the first state information and a second preset weight corresponding to the second state information; the first calculation subunit is used to perform a weighted summation calculation on the first score and the second score according to the first preset weight and the second preset weight to obtain the device state score corresponding to the i-th device to be tested, wherein the device state score is used to characterize the device state information of the i-th device to be tested.
[0092] The link information acquisition module includes a first acquisition submodule, a second acquisition submodule, and a calculation submodule. The first acquisition submodule acquires the device status score corresponding to the device under test in each information transmission link. The second acquisition submodule acquires a preset weight value corresponding to the device under test in each information transmission link, wherein the preset weight value of each device under test represents the importance of the device under test to the information transmission link in which it resides. The calculation submodule calculates a weighted average of the device status scores corresponding to the devices under test in each information transmission link based on the preset weight values, thereby obtaining the link status score corresponding to the information transmission link. The link status score represents the link status information of the information transmission link.
[0093] Optionally, the information processing module includes: a third matrix generation submodule, a third acquisition submodule, a first calculation submodule, and a second matrix generation submodule. The third matrix generation submodule is used to generate a third matrix based on the device status scores corresponding to M devices to be detected, wherein each value in the third matrix is a device status score corresponding to a device to be detected; the third acquisition submodule is used to acquire the graph adjacency matrix corresponding to the graph data, wherein the graph adjacency matrix is used to represent the adjacency relationship between graph nodes in the graph data; the first calculation submodule is used to calculate the product of the graph adjacency matrix and the third matrix to obtain a first matrix; the second matrix generation submodule is used to generate a second matrix based on the link status score corresponding to each information transmission link, wherein each value in the second matrix is a link status score corresponding to an information transmission link.
[0094] Optionally, the calculation module includes: a fourth acquisition submodule, a first graph convolution operation submodule, an activation submodule, a second graph convolution operation submodule, and a target matrix generation submodule. The system comprises the following components: a fourth acquisition submodule, used to acquire the symmetric normalized matrix corresponding to the first matrix; a first graph convolution operation submodule, used to perform a first graph convolution operation on the symmetric normalized matrix and the second matrix through the first layer of the graph convolutional network model, wherein the first graph convolution operation calculates the product of the symmetric normalized matrix, the second matrix, and the first weight matrix to obtain the third matrix, and the first weight matrix is the weight matrix corresponding to the first layer of the graph convolutional network; an activation submodule, used to activate the third matrix through an activation function to obtain the activation function value, wherein the activation function value is a matrix; a second graph convolution operation submodule, used to perform a second graph convolution operation on the activation function value and the symmetric normalized matrix through the second layer of the graph convolutional network model, wherein the second graph convolution operation calculates the product of the activation function value, the symmetric normalized matrix, and the second weight matrix to obtain the fourth matrix, and the second weight matrix is the weight matrix corresponding to the second layer of the graph convolutional network; and a target matrix generation submodule, used to generate a target matrix based on the fourth matrix.
[0095] Optionally, the target matrix generation submodule includes a matrix processing unit and a vector processing unit. The matrix processing unit normalizes the fourth matrix into a probability distribution vector, which is represented in matrix form. The vector processing unit uses the probability distribution vector as the target matrix.
[0096] The device for determining the probability of equipment failure further includes an acquisition module and an information generation module. The acquisition module acquires the device identifier of the device to be tested corresponding to the target value in the target matrix, where the target value is a value in the target matrix that is greater than a preset threshold. The information generation module generates prompt information based on the device identifier, which prompts the target object to troubleshoot the device to be tested corresponding to the device identifier.
[0097] Example 3
[0098] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform the device failure probability determination method of any one of the above by executing the executable instructions.
[0099] Figure 4 This is a schematic diagram of an electronic device provided according to an embodiment of this application, such as... Figure 4 As shown, this application provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the device failure probability determination method in Embodiment 1 above.
[0100] Example 4
[0101] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored computer program, wherein, when the computer program is running, it controls the device where the computer-readable storage medium is located to execute the device failure probability determination method in Embodiment 1 above.
[0102] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0103] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0104] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.
[0105] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0106] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0107] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.
[0108] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.
Claims
1. A method of determining a failure probability of a device, characterized by, include: Obtain device status information of M devices to be tested, wherein the M devices to be tested form N information transmission links, each information transmission link consists of at least two devices to be tested, and M and N are both positive integers; The link status information corresponding to each information transmission link is determined based on the device status information of the device to be detected in each information transmission link. A first matrix is generated based on the device status information of the M devices to be detected, and a second matrix is generated based on the link status information corresponding to each information transmission link. The first matrix is an adjacency feature weight matrix corresponding to the graph data. The graph data includes M graph nodes and multiple edges. Each graph node corresponds to a device to be detected, and each edge is used to characterize the connection relationship between the graph nodes at both ends of the edge. The second matrix is a feature status matrix corresponding to the graph data. A graph convolution operation is performed on the first matrix and the second matrix to obtain the target matrix, wherein the target matrix consists of M values, each value corresponding to a device to be detected; The probability of each device under test failing is determined based on the values in the target matrix. The process of generating a first matrix based on the device status information of the M devices to be tested, and generating a second matrix based on the link status information corresponding to each information transmission link, includes: generating a third matrix based on the device status scores corresponding to the M devices to be tested, wherein each value in the third matrix is a device status score corresponding to a device to be tested; obtaining a graph adjacency matrix corresponding to the graph data, wherein the graph adjacency matrix is used to characterize the adjacency relationship between graph nodes in the graph data; calculating the product of the graph adjacency matrix and the third matrix to obtain the first matrix; and generating the second matrix based on the link status score corresponding to each information transmission link, wherein each value in the second matrix is a link status score corresponding to an information transmission link.
2. The method of claim 1, wherein, Obtain the device status information of M devices to be tested, including: Step 1: Obtain the first status information of the i-th device among the M devices to be tested, wherein the first status information is used to characterize whether the i-th device to be tested successfully sends or receives a heartbeat message at a preset time, and the i-th device to be tested is any one of the M devices to be tested; Step 2: Obtain the second status information of the i-th device to be tested, wherein the second status information is used to characterize the memory usage rate of the i-th device to be tested within a preset time period; Step 3: Determine the device status information corresponding to the i-th device under test based on the first status information and the second status information of the i-th device under test; Step 4: Repeat steps 1 to 3 until the device status information of the M devices to be tested is obtained.
3. The method of claim 2, wherein, Based on the first state information and the second state information of the i-th device to be tested, the device state information corresponding to the i-th device to be tested is determined, including: A first score is determined based on the first status information of the i-th device under test, wherein the score of the first score is a first score or a second score. The first score is the score corresponding to the first score when the i-th device under test successfully sends or receives the heartbeat message at the preset time, and the second score is the score corresponding to the first score when the i-th device under test fails to send or receive the heartbeat message at the preset time. Based on the memory usage rate of the i-th device to be tested during the preset time period, a target value interval is determined from multiple preset value intervals, wherein the target value interval is the value interval in which the memory usage rate is located, and each preset interval corresponds to a preset score; The preset score corresponding to the target value range is used as the second score corresponding to the i-th device to be tested; The device status information of the i-th device to be tested is determined based on the first score and the second score.
4. The method of claim 3, wherein, The device status information of the i-th device to be tested is determined based on the first score and the second score, including: Obtain a first preset weight corresponding to the first state information and a second preset weight corresponding to the second state information; The first score and the second score are weighted and summed according to the first preset weight and the second preset weight to obtain the device status score corresponding to the i-th device to be tested, wherein the device status score is used to characterize the device status information of the i-th device to be tested.
5. The method of claim 4, wherein, The link status information corresponding to each information transmission link is determined based on the device status information of the device to be detected in each information transmission link, including: Obtain the device status score corresponding to the device to be detected in each information transmission link; Obtain the preset weight value corresponding to the device to be detected in each information transmission link, wherein the preset weight value corresponding to each device to be detected is used to characterize the importance of the device to be detected to the information transmission link in which the device to be detected is located; Based on the preset weight value corresponding to the device to be detected in each information transmission link, the device status score corresponding to the device to be detected in the information transmission link is calculated by weighted average to obtain the link status score corresponding to the information transmission link. The link status score is used to characterize the link status information corresponding to the information transmission link.
6. The method of claim 1, wherein, A graph convolution operation is performed on the first matrix and the second matrix to obtain the target matrix, including: Obtain the symmetric normalized matrix corresponding to the first matrix; The first graph convolution operation is performed on the symmetric normalized matrix and the second matrix through the first layer of the graph convolution network model. The first graph convolution operation is used to calculate the product of the symmetric normalized matrix, the second matrix and the first weight matrix to obtain the third matrix. The first weight matrix is the weight matrix corresponding to the first layer of the graph convolution network. The third matrix is activated by an activation function to obtain an activation function value, wherein the activation function value is a matrix. The second graph convolution operation is performed on the activation function value and the symmetric normalization matrix through the second layer of the graph convolution network model. The second graph convolution operation is used to calculate the product of the activation function value, the symmetric normalization matrix and the second weight matrix to obtain the fourth matrix. The second weight matrix is the weight matrix corresponding to the second layer graph convolution network. The target matrix is generated based on the fourth matrix.
7. The method of claim 6, wherein, Generating the target matrix based on the fourth matrix includes: The fourth matrix is normalized into a probability distribution vector, wherein the probability distribution vector is represented in matrix form; The probability distribution vector is used as the target matrix.
8. The method of claim 1, wherein, After performing graph convolution operations on the first matrix and the second matrix to obtain the target matrix, the method further includes: Obtain the device identifier of the device to be detected corresponding to the target value in the target matrix, wherein the target value is a value in the target matrix that is greater than a preset threshold; A prompt message is generated based on the device identifier, wherein the prompt message is used to prompt the target object to troubleshoot the device to be tested corresponding to the device identifier.
9. A device for determining the probability of equipment failure, characterized in that, include: The device information acquisition module is used to acquire device status information of M devices to be tested, wherein the M devices to be tested form N information transmission links, and each information transmission link consists of at least two devices to be tested, and M and N are both positive integers; The link information acquisition module is used to determine the link status information corresponding to each information transmission link based on the device status information of the device to be detected in each information transmission link. The information processing module is used to generate a first matrix based on the device status information of the M devices to be detected, and to generate a second matrix based on the link status information corresponding to each information transmission link. The first matrix is an adjacency feature weight matrix corresponding to the graph data. The graph data includes M graph nodes and multiple edges. Each graph node corresponds to a device to be detected, and each edge is used to characterize the connection relationship between the graph nodes at both ends of the edge. The second matrix is a feature status matrix corresponding to the graph data. The calculation module is used to perform graph convolution operation on the first matrix and the second matrix to obtain the target matrix, wherein the target matrix consists of M values, each value corresponding to a device to be detected; The results output module is used to determine the probability of each device under test failing based on the values in the target matrix. The information processing module includes: a third matrix generation submodule, used to generate a third matrix based on the device status scores corresponding to the M devices to be detected, wherein each value in the third matrix is a device status score corresponding to a device to be detected; a third acquisition submodule, used to acquire the graph adjacency matrix corresponding to the graph data, wherein the graph adjacency matrix is used to characterize the adjacency relationship between graph nodes in the graph data; a first calculation submodule, used to calculate the product of the graph adjacency matrix and the third matrix to obtain the first matrix; and a second matrix generation submodule, used to generate a second matrix based on the link status score corresponding to each information transmission link, wherein each value in the second matrix is a link status score corresponding to an information transmission link.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the method for determining the probability of failure of the device as described in any one of claims 1 to 8.
11. An electronic device, comprising: It includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method for determining the probability of failure of the device according to any one of claims 1 to 8.
Citation Information
Patent Citations
Multi-source timing data fault diagnosis method and medium based on graph neural network
JP7004364B1
Fault root cause determination method and device, and storage medium and electronic device
WO2023029654A1
