A method and system for inspection

Abstract

The application provides a method and system for inspection, which uses a virtual inspection robot to replace a physical inspection robot, reduces dependence on hardware devices, and is suitable for a low-cost and high-flexibility virtual inspection scene. The method performs risk assessment on inspection nodes in an inspection route through a weighted failure probability model, dynamically adjusts the inspection time of each inspection node according to the assessment result, and realizes dynamic matching of the inspection time length and the node risk. The method includes historical failure frequency, device criticality and other dimensions into the failure risk evaluation model, thereby generating a differentiated device inspection scheme, so that the inspection scheme can accurately adapt to different inspection scenes.

Description

Technical Field

[0001] This invention relates to the field of industrial automation and intelligent operation and maintenance technology, specifically to an inspection method and system. Background Technology

[0002] Currently, automated inspection of industrial equipment mainly relies on physical robots with fixed programs. These inspection solutions often employ an equal time allocation strategy, failing to differentiate between the criticality and failure probability of equipment, resulting in underutilization of inspection resources. High-risk equipment in the system typically requires manual labeling, lacking adaptive optimization capabilities. Furthermore, physical robot inspections are limited by hardware costs and mobility, making it difficult to cover complex scenarios. Summary of the Invention

[0003] This specification describes one or more embodiments of an inspection method and system to at least partially solve the above-mentioned technical problems.

[0004] Firstly, an inspection method is provided, including the following steps:

[0005] Determine the inspection route based on the inspection requirements of the target system;

[0006] Historical fault data of inspection nodes in the inspection route are collected based on a preset time window. A weighted fault probability model is used to determine the risk score of the inspection node based on the historical fault data of the inspection node. The inspection node represents a local inspection area.

[0007] Based on the risk score of the inspection node, the inspection duration of the inspection node is determined;

[0008] The failure probability of each target device in the inspection node is determined based on the failure risk assessment model.

[0009] The inspection duration of the target equipment is determined based on the failure probability of the target equipment.

[0010] The inspection route and the inspection duration of each target device in the inspection node are input into the virtual inspection robot, and the inspection plan is executed by the virtual inspection robot.

[0011] Obtain the inspection results obtained by the virtual inspection robot executing the inspection plan.

[0012] As an optional implementation of the method described in the first aspect, the risk score of the inspection node is determined using the risk assessment model based on the historical fault data of the inspection node, specifically including:

[0013] calculate:

[0014] ;

[0015] in, Indicates inspection nodes Risk score, , , These are the weighting coefficients. The non-linear adjustment factor is a constant. Indicates inspection nodes The number of historical failures that occurred within the specified time window. This indicates the inspection node determined based on the historical fault data. The severity of the fault, Indicates inspection nodes The criticality level.

[0016] Furthermore, based on the risk score of the inspection node, the inspection duration of the inspection node is determined, specifically including:

[0017] calculate:

[0018] ;

[0019] in, Indicates inspection nodes Inspection time, This is the estimated inspection time for the inspection route. N This indicates the total number of inspection nodes in the inspection route.

[0020] Furthermore, the method also includes:

[0021] Based on the inspection duration of the inspection node and the inspection node Calculate the inspection duration adaptation coefficient based on the number of historical faults occurring within the specified time window:

[0022] ;

[0023] in, Indicates inspection nodes Inspection duration adaptation coefficient, Indicates the baseline inspection duration. This represents the standard deviation of the risk score for each inspection node in the inspection route. K Indicates the adjustment coefficient;

[0024] like If the value exceeds the preset adaptation threshold, then the inspection node will be... The inspection duration has been adjusted as follows:

[0025] ;

[0026] This represents the average risk score of each inspection node in the inspection route.

[0027] As an optional implementation of the method described in the first aspect, determining the failure probability of each target device in the inspection node based on the failure risk assessment model specifically includes:

[0028] calculate:

[0029] ;

[0030] in, Indicates inspection nodes Target equipment Failure probability score , , Indicates the weighting coefficient. Indicates inspection nodes Target equipment The number of historical failures that occurred within the specified time window. This represents the total number of historical faults that occurred in all target equipment along the inspection route within the specified time window. This indicates the inspection node determined based on the historical fault data. Target equipment The severity of the fault, Indicates inspection nodes Target equipment The criticality level.

[0031] As an optional implementation of the method described in the first aspect, the method further includes:

[0032] Obtain the inspection results of the virtual inspection robot in each round of inspection, and update the risk score of the inspection node based on the inspection results of the previous round and the current round:

[0033] ;

[0034] in, Indicates the updated inspection node i Risk score, Indicates the inspection nodes in the previous round of inspections. i Risk score, This indicates the inspection node obtained based on the current round of inspection results. i The change in risk, The forgetting factor is a positive constant less than 1.

[0035] Based on the updated risk score of the inspection node, the inspection time is reassigned to the inspection node.

[0036] As an optional implementation of the method described in the first aspect, the virtual inspection robot, during the execution of the inspection plan, collects equipment operating status data of each target device along the inspection path, and determines whether the corresponding target device has malfunctioned through an anomaly detection interpreter; the anomaly detection interpreter is represented as:

[0037] ;

[0038] in, Indicates the target device The contribution of fault characteristics Indicates the target device The set of neighboring nodes, Indicates from the target device The set of neighboring nodes The selected set of nodes

[0039] Represents a set of nodes Size, Indicates the total number of target devices. Represents a set of nodes Equipment operating status characteristics and target equipment The fusion features of the operating state characteristics, Represents a set of points The characteristics of the equipment's operating status;

[0040] when When the contribution of fault characteristics exceeds a preset threshold, the target device is determined. A malfunction has occurred.

[0041] As an optional implementation of the method described in the first aspect, the method further includes:

[0042] Based on the equipment operation status data of the target equipment collected by the virtual inspection robot during the execution of the inspection plan, a virtual mapping model of the equipment is constructed to visualize the operation status of the target equipment.

[0043] The device virtual mapping model is represented as follows:

[0044] ;

[0045] in, Indicates the target device At any moment The running status, Indicates the target device The inherent parameters, Indicates the target device At any moment The running status, Indicates the backtracking time window, Indicates the target device Environmental factors.

[0046] As an optional implementation of the method described in the first aspect, the method further includes:

[0047] Based on the inspection results obtained by the virtual inspection robot executing the inspection plan and the pre-built risk propagation model based on the force-directed graph, an inspection plan for the next round of inspection is constructed.

[0048] The risk propagation model is expressed as follows:

[0049] ;

[0050] in, This represents the total energy of all target devices in the target system. Indicates the target device To the target device The direction of risk transmission Indicates the target device With target equipment The strength of the connection between them Indicates the target device With target equipment The strength of the fault correlation between them Indicates the target device Location in the force guidance diagram Indicates the target device Location in the force guidance diagram Indicates the target device in the force guidance diagram With target equipment The distance between them.

[0051] Secondly, an inspection system is provided to implement the above-mentioned inspection method. This system includes:

[0052] The data acquisition module determines the inspection route based on the inspection requirements of the target system, and collects historical fault data of the inspection nodes in the inspection route based on a preset time window.

[0053] The node risk score calculation module is used to determine the risk score of the inspection node based on the historical fault data of the inspection node using a weighted fault probability model; the inspection node represents a local inspection area.

[0054] The node inspection time allocation module is used to determine the inspection duration of the inspection node based on the risk score of the inspection node.

[0055] The equipment failure probability calculation module is used to determine the failure probability of each target device in the inspection node based on the failure risk assessment model.

[0056] The equipment inspection time allocation module is used to determine the inspection duration of the target equipment based on the failure probability of the target equipment.

[0057] The virtual inspection robot is used to execute an inspection plan based on the inspection route and the inspection duration of each target device in the inspection node; and to push the inspection plan and inspection results to the operation and maintenance personnel's operation and maintenance terminal.

[0058] Beneficial effects: One or more embodiments of this specification provide an inspection method and system, which have the following beneficial effects:

[0059] This method uses virtual inspection robots to replace physical inspection robots, reducing reliance on hardware equipment and making it suitable for low-cost, highly flexible virtual inspection scenarios.

[0060] This method uses a weighted failure probability model to assess the risks of inspection nodes along the inspection route and dynamically adjusts the inspection time of each inspection node based on the assessment results, thereby achieving a dynamic match between inspection duration and node risk.

[0061] This method incorporates dimensions such as historical failure frequency and equipment criticality into the failure risk assessment model, thereby generating differentiated equipment inspection plans that can accurately adapt to different inspection scenarios.

[0062] Correspondingly, the inspection system described in this invention also has the above-mentioned technical effects. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 This is a flowchart illustrating an inspection method according to an embodiment of the present invention;

[0065] Figure 2 This is a schematic diagram of the structure of an inspection system according to an embodiment of the present invention. Detailed Implementation

[0066] First, it should be noted that the terminology used in the embodiments of this invention is for the purpose of describing specific embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” used in the embodiments of this invention and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0067] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments in this specification, and not all of the embodiments. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0068] It should be noted that the steps of the corresponding methods are not necessarily performed in the order shown and described in this specification in other embodiments. In some other embodiments, the methods may include more or fewer steps than described in this specification. Furthermore, a single step described in this specification may be broken down into multiple steps in other embodiments; and multiple steps described in this specification may be combined into a single step in other embodiments.

[0069] The inspection method and system described in this specification will be further described in detail below with reference to the accompanying drawings and specific embodiments. However, this detailed description does not constitute a limitation on the embodiments of this specification.

[0070] Please refer to Figure 1 , Figure 1 A schematic diagram illustrating the process of an inspection method is shown. For example... Figure 1 As shown, the method includes steps S100 to S112.

[0071] S100: Determine the inspection route based on the inspection requirements of the target system.

[0072] This embodiment uses a virtual inspection robot for inspection. The inspection route mentioned above refers to the system-level or software-level inspection route. The inspection route contains multiple inspection nodes, which represent local inspection areas. A local inspection area can be understood as a local topology within the target device topology of the entire system. The target devices mentioned here can include hardware devices (such as network devices, communication devices, production line equipment) and software devices (such as software components, databases, etc.). In other words, an inspection node typically contains multiple target devices.

[0073] S102: Collect historical fault data of inspection nodes in the inspection route based on a preset time window, and determine the risk score of the inspection node based on the historical fault data of the inspection node using a weighted fault probability model.

[0074] The aforementioned time windows are typically set relatively short so that the collected historical fault data can characterize the recent fault status of each inspection node. For example, the time window can be the 24 hours preceding the current moment, and historical fault data of the aforementioned inspection nodes can be collected within these 24 hours.

[0075] Specifically, historical fault records can be extracted from the enterprise's MES (Manufacturing Execution System), SCADA (Supervisory and Data Acquisition System), or equipment logs. The aforementioned historical fault data mainly includes various parameters of the target equipment included in the inspection node, specifically: the number of historical faults that occurred in the target equipment within the time window, the severity of the faults, and the criticality level of the target equipment (such as core production line equipment, auxiliary equipment), etc.

[0076] Based on parameters such as the historical number of failures, failure severity, and criticality level of the target equipment within a time window, the historical number of failures, failure severity, and criticality level of the corresponding inspection node can be determined. The historical number of failures for an inspection node refers to the total number of failures occurring in all target equipment included in that inspection node within the time window. The failure severity of an inspection node can be graded based on its historical failure count. Specifically, different failure severity levels can be set, each assigned a score to characterize the severity of the failure; a higher score indicates a more severe failure. The criticality level of an inspection node is pre-set. Specifically, a criticality score (ranging from 0 to 1) can be used to represent the criticality level of the inspection node; a higher score indicates a higher criticality level. For example, the criticality score for core equipment is 1, for secondary equipment it is 0.5, and for auxiliary equipment it is 0.1.

[0077] The expression for the weighted failure probability model described above is:

[0078] ;

[0079] in, Indicates inspection nodes Risk score, , , These are the weighting coefficients. The non-linear adjustment factor is a constant. Indicates inspection nodes The number of historical failures that occurred within the time window. This indicates the inspection nodes determined based on historical fault data. The severity of the fault, Indicates inspection nodes The criticality level.

[0080] In this weighted failure probability model , , These are trainable weight coefficients. For training the weighted failure probability model, historical failure parameters of different devices can be collected, including the number of historical failures, the severity of the failures, and the criticality level of the device. Then, based on the historical failure parameters of the devices, historical failure parameters of the inspection nodes are constructed, including the number of historical failures, the severity of the failures, and the criticality level of the inspection node. These historical failure parameters of the inspection nodes are used as training samples. Each training sample is scored based on expert experience; a higher score indicates a greater failure probability, and this score serves as the label for the training sample. These training samples are used to train the aforementioned weighted failure probability model, iteratively updating the weight coefficients. , , This continues until a weighted failure probability model that meets the requirements is obtained.

[0081] S104: Determine the inspection duration of the inspection node based on the risk score of the inspection node.

[0082] After determining the inspection route, an initial budget value for the overall inspection time of the route can be assigned. Then, based on the budget value and the proportion of the risk score of each inspection node to the total risk score of all inspection nodes along the route, the inspection time for each node can be determined. Specifically, the expression for calculating the inspection time of an inspection node is as follows:

[0083] ;

[0084] in, Indicates inspection nodes Inspection time, This is the budgeted inspection time for the inspection route. N This indicates the total number of inspection nodes in the inspection route.

[0085] In some implementations, the suitability of the inspection time allocation scheme can be determined by constructing a reverse analysis model.

[0086] Specifically, it can be based on the inspection duration of the inspection nodes and the inspection nodes themselves. Calculate the inspection duration adaptation coefficient based on the number of historical faults that occurred within the above time window:

[0087] ;

[0088] in, Indicates inspection nodes Inspection duration adaptation coefficient, Indicates the baseline inspection duration. This represents the standard deviation of the risk score for each inspection node along the inspection route. K This represents the adjustment coefficient.

[0089] like If the value exceeds the preset adaptation threshold, a time allocation strategy based on risk entropy will be adopted to allocate inspection nodes. The inspection duration has been adjusted as follows:

[0090] ;

[0091] This represents the average risk score of each inspection node along the inspection route.

[0092] By employing the aforementioned inspection duration adjustment strategy, inspection times can be dynamically adjusted based on the historical number of equipment failures, ensuring that high-risk equipment receives more attention. This approach allows for the rational allocation of inspection time within limited resources, improving inspection efficiency. Furthermore, by analyzing the rate of change of the inspection duration adaptation coefficient at each inspection node, the trend of equipment failures at that node can be predicted, enabling preventative maintenance.

[0093] S106: Determine the failure probability of each target device in the inspection node based on the failure risk assessment model.

[0094] The expression for the fault risk assessment model is:

[0095] ;

[0096] in, Indicates inspection nodes Target equipment Failure probability score , , Indicates the weighting coefficient. Indicates inspection nodes Target equipment The number of historical failures that occurred within the time window. This represents the total number of historical faults that occurred in all target equipment along the inspection route within the time window. This indicates the inspection nodes determined based on historical fault data. Target equipment The severity of the fault, Indicates inspection nodes Target equipment The criticality level.

[0097] S108: Determine the inspection duration of the target equipment based on the failure probability of the target equipment.

[0098] Specifically, the inspection time of the target equipment can be calculated using the following formula:

[0099] ;

[0100] in, Indicates inspection nodes Target equipment Inspection duration, J Indicates inspection nodes The total number of target devices.

[0101] S110: Input the inspection route and the inspection duration of each target device in the inspection node into the virtual inspection robot, and execute the inspection plan through the virtual inspection robot.

[0102] The virtual inspection robot described in this embodiment is a virtual inspection robot that runs in the form of software algorithm modules on an industrial control server or cloud platform. The virtual inspection robot can interface with the enterprise system's PLC (Programmable Logic Controller), MES and other systems through API or OPCUA protocol to obtain the equipment operation status data of each target device in real time.

[0103] The virtual inspection robot is primarily responsible for automatically executing inspection tasks based on inspection requirements, recording inspection results, and generating reports. The virtual robot performs the following specific tasks during the inspection process:

[0104] 1. Virtual robots can obtain equipment operation status data of target equipment from PLCs, sensors and other devices in the enterprise system, and can also obtain key parameters in the production process through the MES system.

[0105] 2. Data Analysis:

[0106] The virtual robot uses a preset analysis algorithm to analyze the operating status data of the target device and identify whether the target device has any abnormalities or malfunctions.

[0107] For example, a virtual robot can use an anomaly detection interpreter to determine whether a target device has malfunctioned. The anomaly detection interpreter is represented as:

[0108] ;

[0109] in, Indicates the target device The contribution of fault characteristics Indicates the target device The set of neighboring nodes, Indicates from the target device The set of neighboring nodes The selected set of nodes Represents a set of nodes Size, Indicates the total number of target devices. Represents a set of nodes Equipment operating status characteristics and target equipment The fusion features of the operating state characteristics, Represents a set of points The equipment operating status characteristics are obtained by extracting features from the equipment operating status data of the target equipment. The specific feature extraction method can be adaptively selected according to requirements, and this embodiment does not impose any restrictions on it.

[0110] when When the contribution of fault characteristics exceeds a preset threshold, the target device is determined. A malfunction has occurred.

[0111] 3. Equipment status visualization and fault prediction

[0112] Based on the equipment operation status data of the target equipment collected by the virtual inspection robot during the execution of the inspection plan, a virtual mapping model of the equipment is constructed to visualize the operation status of the target equipment.

[0113] Specifically, the expression for the aforementioned device virtual mapping model is:

[0114] ;

[0115] in, Indicates the target device At any moment The running status, Indicates the target device The inherent parameters, Indicates the target device At any moment The running status, Indicates the backtracking time window, Indicates the target device Environmental factors.

[0116] This virtual equipment mapping model can predict the status of target equipment at a future point in time. This function is crucial for developing inspection plans because it helps identify which target equipment may require more attention or maintenance in the future. The virtual equipment mapping model analyzes the operating status and environmental factors of target equipment, predicting potential failures. This provides an important basis for optimizing inspection plans, as it helps inspectors identify potential fault points in advance, enabling preventative maintenance. Inspectors can optimize their inspection plans based on the predictions from the virtual equipment mapping model, ensuring high-risk equipment is prioritized for inspection, thereby improving inspection efficiency and equipment reliability. As the operating status and environmental factors change over time, the virtual equipment mapping model can dynamically update this data, allowing inspectors to dynamically adjust their inspection plans to adapt to changes in equipment status.

[0117] S112: Obtain the inspection results obtained by the virtual inspection robot executing the inspection plan.

[0118] The aforementioned inspection results include equipment operation status data of inspection nodes and / or target equipment collected by the virtual inspection robot during the inspection process, as well as the inspection node fault status and / or target equipment fault status determined based on the equipment operation status data of inspection nodes and / or target equipment.

[0119] In scenarios involving multiple rounds of inspections along the same inspection route, the inspection results of the virtual inspection robot in each round can be obtained. Based on the inspection results of the previous round and the current round, a sliding time window is used to update the risk score of the inspection node.

[0120] ;

[0121] in, Indicates the updated inspection node i Risk score, Indicates the inspection nodes in the previous round of inspections. i Risk score, This indicates the inspection node obtained based on the current round of inspection results. i The change in risk, The forgetting factor is a positive constant less than 1 used to adjust the impact of historical data on the current risk score.

[0122] Based on the updated risk scores of the inspection nodes, inspection time is reassigned to the inspection nodes.

[0123] A sliding time window is used to update the risk score of the inspection node. By retaining a certain amount of recent historical data to calculate the current risk score, it ensures that the risk score reflects the latest inspection results. By introducing a forgetting factor, the influence of historical data on the current risk score can be reduced, making the latest inspection results have a greater impact on the risk score.

[0124] In some implementations, in order to analyze and visualize the fault propagation characteristics between target devices for inspection personnel to observe, a risk propagation model based on force-directed graphs can be constructed to describe the fault propagation relationship between target devices.

[0125] The risk propagation model is expressed as follows:

[0126] ;

[0127] in, This represents the total energy of all target devices in the target system. Indicates the target device To the target device The direction of risk transmission Indicates the target device With target equipment The strength of the connection between them Indicates the target device With target equipment The strength of the fault correlation between them Indicates the target device Location in the force guidance diagram Indicates the target device Location in the force guidance diagram Indicates the target device in the force guidance diagram With target equipment The distance between them.

[0128] In the force-directed graph, each node represents a target device. The force-directed algorithm is used to calculate the position of each target device in the force-directed graph, so that nodes with high association strength are closer together in the graph.

[0129] By constructing a risk propagation model, the strength of fault correlation between target devices can be visualized, helping inspection personnel identify critical equipment and potential fault propagation paths. Inspection personnel can determine inspection priorities based on the strength of fault correlation between target devices to plan inspection routes, allowing virtual inspection robots to prioritize devices with high correlation strength. Furthermore, by analyzing the correlation strength between target devices, fault propagation paths can be predicted, enabling preventative maintenance.

[0130] Corresponding to the inspection method described above, this embodiment also provides an inspection system for implementing the inspection method described above. Please refer to... Figure 2 , Figure 2 A schematic diagram of an inspection system is shown. Figure 2 As shown, the system includes:

[0131] The data acquisition module determines the inspection route based on the inspection requirements of the target system and collects historical fault data of the inspection nodes in the inspection route based on a preset time window.

[0132] The node risk score calculation module is used to determine the risk score of the inspection node based on the historical fault data of the inspection node using a weighted fault probability model; the inspection node represents a local inspection area.

[0133] The node inspection time allocation module is used to determine the inspection duration of the inspection node based on the risk score of the inspection node.

[0134] The equipment failure probability calculation module is used to determine the failure probability of each target device in the inspection node based on the pre-trained failure risk assessment model.

[0135] The equipment inspection time allocation module is used to determine the inspection duration of the target equipment based on the failure probability of the target equipment.

[0136] The virtual inspection robot is used to execute inspection plans based on the inspection route and the inspection duration of each target device in the inspection node; and to push the inspection plan and inspection results to the operation and maintenance personnel's operation and maintenance terminals.

[0137] The specific principles of each module in the above inspection system can be found in [reference needed]. Figure 1 The corresponding steps in the inspection method shown will not be repeated here.

[0138] It is understood that the structures illustrated in the embodiments of this specification do not constitute a specific limitation on the system of the embodiments of this specification. In other embodiments of the specification, the above system may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0139] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the apparatus embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0140] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0141] It should be noted that the above examples are merely specific embodiments of the present invention, and the present invention is obviously not limited to the above embodiments, with many similar variations. All modifications that can be directly derived or conceived by those skilled in the art from the content disclosed in this invention should fall within the protection scope of this invention.

Claims

1. A method of inspection, characterized by, The method comprises the steps of: determining an inspection route based on an inspection requirement of a target system; collecting historical fault data of an inspection node in the inspection route based on a preset time window, the inspection node representing a local inspection area; determining a risk score of the inspection node according to the historical fault data of the inspection node by using a weighted fault probability model; ； wherein, represents a risk score of the inspection node , is a weight coefficient, is a nonlinear adjustment factor, which is a constant, represents a risk score of the inspection node , represents a failure severity of the inspection node determined according to the historical failure data, represents a criticality level of the inspection node ; determining an inspection duration of the inspection node according to the risk score of the inspection node; ； wherein, represents the inspection time of the inspection node , is a budget value of the inspection time length of the inspection route, N represents the total number of inspection nodes in the inspection route; based on a patrol duration of the patrol node and the patrol node a number of historical faults occurring in the time window, calculate a patrol duration adaptation coefficient: ； wherein, represents a patrol node a patrol duration adaptation coefficient of the patrol node, represents a reference patrol duration, represents a standard deviation of risk scores of the patrol nodes in the patrol route, K represents an adjustment coefficient; If greater than a preset adaptation threshold, the patrol time length of the patrol node is adjusted to be: ； a mean value representing risk scores of the inspection nodes in the inspection route; determining a fault probability of each target device in the inspection node according to a fault risk evaluation model; determining an inspection duration of the target device based on the fault probability of the target device; inputting the inspection route and the inspection duration of each target device in the inspection node into a virtual inspection robot, and executing an inspection scheme by the virtual inspection robot; obtaining an inspection result obtained by the virtual inspection robot executing the inspection scheme.

2. The method of claim 1, wherein, The method further comprises the steps of: calculating: ； in, Indicates inspection nodes Target equipment Failure probability score Indicates the weighting coefficient. Indicates inspection nodes Target equipment The number of historical failures that occurred within the specified time window. This represents the total number of historical faults that occurred in all target equipment along the inspection route within the specified time window. This indicates the inspection node determined based on the historical fault data. Target equipment The severity of the fault, Indicates inspection nodes Target equipment The criticality level.

3. The method of claim 1, wherein, The method further comprises the steps of: obtaining an inspection result of the virtual inspection robot in each round of inspection, updating the risk score of the inspection node based on the inspection result of the last round and the inspection result of the current round, and ； wherein, represents the risk score of the inspection node after the update, i represents the risk score of the inspection node in the last round of inspection, i represents the risk change amount of the inspection node according to the result of the current round of inspection, i represents a forgetting factor, which is a positive constant less than 1.​​​ redistributing the inspection time for the inspection node according to the updated risk score of the inspection node.

4. The method of claim 1, wherein, In the process of executing the inspection scheme, the virtual inspection robot collects device running state data of each target device in the inspection path, and determines whether the corresponding target device has a fault by using an abnormality detection interpreter. ； wherein, represents a fault feature contribution degree of a target device , represents a neighbor node set of a target device , represents a part node set selected from a neighbor node set of a target device , represents a part node set selected from a neighbor node set of a target device Represents a set of nodes Size, Indicates the total number of target devices. Represents a set of nodes Equipment operating status characteristics and target equipment The fusion features of the operating state characteristics, Represents a set of points The characteristics of the equipment's operating status; When greater than a preset fault feature contribution degree threshold, determining that the target device is faulty.

5. The method of claim 1, wherein, The method further comprises the steps of: constructing a device virtual mapping model based on the device running state data of the target device collected by the virtual inspection robot in the process of executing the inspection scheme, so as to realize visualization of the running state of the target device. The device virtual mapping model is represented as: ； wherein, represents a target device at a time point of an operating state, represents an intrinsic parameter of a target device at a time point of an operating state, represents an intrinsic parameter of a target device at a time point of an operating state, represents a backtracking time window, represents an environmental factor of a target device 6. The method of claim 1, wherein, The method further comprises the steps of: constructing an inspection scheme for the next round of inspection according to the inspection result obtained by the virtual inspection robot executing the inspection scheme and a risk propagation model based on a force-directed graph which is constructed in advance; The risk propagation model is represented as: ； wherein, represents the total energy of all target devices in the target system, represents the target device to which the risk propagates, represents the connection strength between the target device and the target device , represents the failure association strength between the target device and the target device , represents the position of the target device in the force directed graph, represents the position of the target device in the force directed graph, represents the distance between the target device and the target device in the force directed graph.​ 7. A patrol system for implementing the patrol method according to any one of claims 1 to 6, characterized by, comprising: a data collection module, configured to determine an inspection route based on an inspection requirement of a target system, and collect historical fault data of an inspection node in the inspection route based on a preset time window; a node risk score calculation module, configured to determine a risk score of the inspection node according to the historical fault data of the inspection node by using a weighted fault probability model; the inspection node represents a local inspection area; a node inspection time allocation module, configured to determine an inspection duration of the inspection node according to the risk score of the inspection node; a device fault probability calculation module, configured to determine a fault probability of each target device in the inspection node according to a fault risk evaluation model; a device inspection time allocation module, configured to determine an inspection duration of the target device based on the fault probability of the target device; a virtual inspection robot, configured to execute an inspection scheme according to the inspection route and the inspection duration of each target device in the inspection node; and push the inspection scheme and the inspection result to an operation and maintenance terminal of an operation and maintenance personnel.

Images