Method, device, server and storage medium for identifying batch hidden danger equipment

Abstract

The present disclosure provides a method, device, server and storage medium for identifying batch hidden danger equipment, which belongs to the field of Internet technology. The method includes: generating a combination dimensional feature sequence of each key component based on the component data of each key component; identifying batch hidden danger problems of each key component in the device to be identified based on the device feature data of the device to be identified, the component feature data of each key component, and the combination dimensional feature sequence of each key component. The present disclosure obtains the component data of each key component in the device to be identified. The component data can characterize the operating status of the device to be identified. Therefore, there is no need to use manual calculation and analysis. The batch hidden danger problems in the key components can be identified based on the component data, and on this basis, the key component can be located based on the device feature data of the device to be identified and the component feature data of the key component, thereby improving the recognition accuracy and recognition efficiency.

Description

Technical Field

[0001] The present disclosure relates to the field of Internet technology, and in particular to a method, device, server, and storage medium for identifying batches of potentially hazardous equipment. Background Art

[0002] In the internet technology sector, batch defect equipment refers to equipment with batch defect issues. Batch defect issues refer to defects in equipment or its components introduced during the R&D and production stages due to lax software logic or lax production process control. With the development of internet and cloud technologies, the number of devices deployed in data centers is increasing, and at the same time, business losses caused by batch defect issues are increasing year by year. To reduce business losses caused by batch defect issues, it is necessary to promptly identify batch defect equipment.

[0003] When related technologies are used to identify batch hidden danger equipment, the following methods are mainly used: obtaining the fault data of each device that has failed; manually calculating the fault data of each device; based on the manual calculation results and experience, analyzing the fault problems of each device and identifying batch hidden danger problems of the equipment.

[0004] However, as the number of devices deployed in data centers increases, the frequency of device failures becomes increasingly higher. Limited by manual calculation methods and experience, the identification method that relies solely on manual calculation and experience not only has low recognition efficiency but also poor accuracy. Summary of the Invention

[0005] The present disclosure provides a method, device, server, and storage medium for identifying batch devices with hidden dangers, which can improve the efficiency and accuracy of identifying batch devices with hidden dangers. The technical solution is as follows:

[0006] In one aspect, a method for identifying batches of potentially hazardous equipment is provided, the method comprising:

[0007] Obtaining device characteristic data of a device to be identified, component characteristic data of each key component in the device to be identified, and component data of each key component, wherein the key component is a component that affects the normal operation of the device to be identified, the device characteristic data is used to distinguish different devices, the component characteristic data is used to distinguish different key components in the device to be identified, and the component data is used to reflect the operating status of the corresponding key component;

[0008] Generating, based on the component data of each key component, a combined dimensional feature sequence of each key component, wherein the combined dimensional feature sequence is a sequence composed of at least one component feature;

[0009] Based on the equipment feature data of the equipment to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component, batch hidden danger problems of each key component in the equipment to be identified are identified.

[0010] In another aspect, a device for identifying batch hidden danger equipment is provided, the device comprising:

[0011] an acquisition module, configured to acquire device characteristic data of a device to be identified, component characteristic data of each key component in the device to be identified, and component data of each key component, wherein the key component is a component that affects the normal operation of the device to be identified, the device characteristic data is used to distinguish different devices, the component characteristic data is used to distinguish different key components in the device to be identified, and the component data is used to reflect the operating status of the corresponding key component;

[0012] A generating module, configured to generate, based on component data of each key component, a combined dimensional feature sequence of each key component, wherein the combined dimensional feature sequence is a sequence composed of at least one component feature;

[0013] The identification module is used to identify batch hidden danger problems of each key component in the equipment to be identified based on the equipment feature data of the equipment to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component.

[0014] On the other hand, a server is provided, comprising a processor and a memory, wherein the memory stores at least one program code, and the at least one program code is loaded and executed by the processor to implement the method for identifying batch hidden danger equipment as described in the first aspect.

[0015] On the other hand, a computer-readable storage medium is provided, in which at least one program code is stored. The at least one program code is loaded and executed by a processor to implement the method for identifying batch hidden danger equipment as described in the first aspect.

[0016] On the other hand, a computer program product or computer program is provided, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the method provided in various optional implementations of the above-mentioned aspect.

[0017] The technical solutions provided by the embodiments of the present disclosure have the following beneficial effects:

[0018] The component data of each key component in the equipment to be identified is obtained. This component data can characterize the operating status of the equipment to be identified. Therefore, there is no need to use manual calculation and analysis. Batch hidden dangers in key components can be identified based on the component data. On this basis, based on the equipment feature data of the equipment to be identified and the component feature data of the key components, the key components can be located, thereby improving recognition accuracy and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1 This is an architectural diagram of a batch hidden danger equipment identification and processing system provided by an embodiment of the present disclosure;

[0021] Figure 2 This is an architectural diagram of a batch hidden danger equipment processing system module provided by an embodiment of the present disclosure;

[0022] Figure 3 This is a schematic diagram of a master-slave deployment solution for a batch hidden danger problem equipment processing system module provided by an embodiment of the present disclosure;

[0023] Figure 4 This is a functional overview diagram of a batch hidden danger problem equipment processing system module provided by an embodiment of the present disclosure;

[0024] Figure 5 This is the implementation environment involved in a method for identifying batches of potentially hazardous equipment provided by an embodiment of the present disclosure;

[0025] Figure 6 This is a flow chart of a method for identifying batch hidden danger equipment provided by an embodiment of the present disclosure;

[0026] Figure 7 This is a flow chart of another method for identifying batches of potentially hazardous equipment provided by an embodiment of the present disclosure;

[0027] Figure 8 This is a flow chart of a batch hidden danger identification model training method provided by an embodiment of the present disclosure;

[0028] Figure 9 This is a schematic diagram of the structure of a device for identifying batch hidden danger equipment provided by an embodiment of the present disclosure;

[0029] Figure 10The present invention is a server for identifying batches of devices with hidden dangers according to an exemplary embodiment. DETAILED DESCRIPTION

[0030] In order to make the objectives, technical solutions and advantages of the present disclosure more clear, the embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.

[0031] It should be understood that the terms "each," "plurality," and "any" used in the embodiments of the present disclosure include two or more, each refers to each of the corresponding plurality, and any refers to any one of the corresponding plurality. For example, if a plurality of words includes 10 words, each refers to each of the 10 words, and any refers to any one of the 10 words.

[0032] First, the technologies involved in the embodiments of the present disclosure are introduced.

[0033] Cloud computing is a computing model that distributes computing tasks across a resource pool consisting of a large number of computers, enabling various application systems to access computing power, storage space, and information services as needed. The network that provides these resources is called the "cloud." To users, these resources appear infinitely scalable and can be accessed at any time, used on demand, expanded at any time, and paid for on a per-use basis.

[0034] As a provider of cloud computing infrastructure, a cloud computing resource pool (referred to as a cloud platform, generally referred to as an IaaS (Infrastructure as a Service) platform) is established. Various types of virtual resources are deployed in the resource pool for external customers to choose and use. The cloud computing resource pool mainly includes: computing devices (virtualized machines, including operating systems), storage devices, and network devices.

[0035] Based on logical functional divisions, the PaaS (Platform as a Service) layer can be deployed on top of the IaaS (Infrastructure as a Service) layer, and the SaaS (Software as a Service) layer can be deployed on top of the PaaS layer. SaaS can also be deployed directly on top of IaaS. PaaS is a platform for software execution, such as databases and web containers. SaaS is a variety of business software, such as web portals and text messaging apps. Generally speaking, SaaS and PaaS are upper layers relative to IaaS.

[0036] In the Internet field, relevant data of equipment will be collected throughout the entire life cycle of equipment, from leaving the factory, going online, to going offline, and availability monitoring and alarming will be performed based on the collected data. Usually, the relevant data collected will contain rich information, such as data center, computer room, rack, machine model, fixed assets, shelf time, component type, component model, firmware version, log data and other information. The disclosed embodiment establishes a batch hidden danger equipment identification and processing system. By parsing, pre-processing, data cleaning, feature extraction and other processing of the collected data, it can not only quickly discover batch hidden danger problems in equipment or components, but also promptly discover online batch hidden danger equipment that has not yet been processed.

[0037] Please refer to Figure 1 , which shows a batch of hidden danger equipment identification and treatment system provided by the embodiment of the present disclosure, which includes an online identification subsystem and an offline training subsystem. The online identification subsystem and the offline training subsystem will be introduced separately below.

[0038] Online identification subsystem

[0039] The online identification subsystem can collect component data of each key component in the entire network equipment, and parse the collected key component data, and then perform preprocessing, data cleaning, component feature extraction and other operations on the parsed component data, and then obtain at least one characteristic key field corresponding to the component type of different key components according to the component type of the key component. Based on at least one characteristic key field, the component feature indicated by at least one characteristic key field is extracted from the component data of the key component. For example, based on at least one characteristic key field, component features such as fixed assets, server product serial code, component type, component model, firmware version, etc. are extracted, and according to the correlation between each component feature, the extracted component features are combined into a combined dimensional feature sequence. Then, based on the combined dimension feature sequence, the offline trained batch hidden danger identification model is used to identify whether a key component in a certain device has a batch hidden danger problem. If a key component in a certain device has a batch hidden danger problem and there is currently a corresponding solution for the batch hidden danger problem, the key component can be repaired based on the solution; if there is currently no corresponding solution for the batch hidden danger problem, the repair is determined to have failed, and the relevant data of the key component is sent to the offline training subsystem to adjust the trained batch hidden danger identification model and use manual methods to re-formulate the solution.

[0040] The online identification subsystem provides an external interface for performing hidden danger analysis based on the collected component data. The external interface can receive the component data of the corresponding component input by the user and feed back the identification result of the component to the user.

[0041] The online identification subsystem mainly includes three modules: data acquisition module, batch hidden danger identification engine module and batch hidden danger problem equipment processing system module.

[0042] Among them, the data acquisition module is mainly used to collect component data and log data of each key component of the entire network equipment. The component data includes hard disk data, memory data, CPU (Central Processing Unit) data, etc., and pre-processes the collected component data, and then stores the pre-processed data in the database.

[0043] The batch hidden danger identification engine module is the core part of the online identification subsystem. It is mainly used to clean the collected component data to remove dirty data in the component data and extract relevant component features, thereby generating a combined dimension feature sequence. Then, the trained batch hidden danger identification model is used to identify batch hidden danger problems of key components in the equipment. If a key component is identified to have a batch hidden danger problem, the relevant data of the key component will be pushed to the batch hidden danger problem equipment processing system module.

[0044] The batch hidden danger problem equipment processing system module is mainly used to mark and lock equipment or key components with batch hidden danger problems. If there is a corresponding solution to the batch hidden danger problem of the equipment or key component, the equipment or key component will be repaired based on the solution. If there is no solution to the batch hidden danger problem of the equipment or key component, it is determined that the modification has failed, and then the batch hidden danger identification model is adjusted based on the relevant data of the equipment or key component, and the relevant data of the equipment or key component is sent to the technician. The technician uses manual analysis and sends the analysis results to the manufacturer of the equipment or component. The manufacturer further analyzes and confirms the existing batch hidden danger problems, and then formulates a plan for the determined batch hidden danger problem and tests it based on the formulated plan. When the test passes, the batch hidden danger problem and the corresponding solution are published for application through the online identification subsystem.

[0045] The batch hidden danger equipment processing system module can select components from existing mature and stable technology stacks and components. For example, for the front-end, Node.js, Vue, and Element UI (User Interface) can be selected; for the back-end, the Spring Cloud suite can be selected; for computing, the TDW (Tencent Distributed Warehouse) / IDEX algorithm can be selected; and for storage, a three-tier storage system consisting of memory, Redis cluster, and UDB (Universal Database) cluster can be selected. The components selected for the batch hidden danger equipment processing system module are not only lightweight, feature-rich, and enable fast and convenient development, but also consistent with the core technology stack.

[0046] See also Figure 2 The batch hidden danger problem equipment processing system module adopts a layered architecture system. This architecture system makes the entire system hierarchy clear, the performance is stable and reliable, and the different functions are decoupled. It is not only convenient for multiple people to develop in parallel, but also convenient for subsequent maintenance and upgrades.

[0047] See also Figure 2 The batch hidden danger equipment processing system module also adopts a microservices approach, splitting system management functions and business functions into microservices, thereby enabling independent deployment of these functions. The batch hidden danger equipment processing system module reuses the existing offline computing module and adds a business backend gateway. External API (Application Programming Interface) services use a unified token management center for authentication.

[0048] See also Figure 2 The batch hidden danger equipment processing system module uses a data caching design for storage, enabling three-level storage across memory, Redis clusters, and UDB clusters. This three-level storage approach improves data query efficiency. The batch hidden danger equipment processing system module also uses a data sharding design for storage. This sharding separates configuration data from business data, thereby separating hot and cold data and preventing data contamination.

[0049] Considering that disaster recovery may occur, in order to improve the availability and consistency of services, see Figure 3 The batch hidden danger equipment processing system module adopts a remote dual-machine, active-standby deployment solution. This approach can achieve disaster recovery and high availability.

[0050] In order to facilitate users to deal with equipment or components with batch hidden danger problems, the batch hidden danger equipment processing system module also provides users with a visual query and processing interface. Based on this visual query and processing interface, users can query and handle batch hidden danger problems of key components of any equipment. Figure 4 The visual query and processing interface includes multiple functional modules, such as the hidden danger handling module, change management module, approval pending module, data statistics module, permission management module, user management module, menu management module, department management module, operation log module, user feedback module, knowledge base, and online service module. The core functional modules are the hidden danger handling module and the change management module. The hidden danger handling module is visualized on the interface as the "My Hidden Danger" option. When a click on this option is detected, the user jumps to the device hidden danger page. Based on this device hidden danger page, the user can address batch hidden danger issues in the equipment. When a click on the "I want to repair" option on this device hidden danger page is detected, the batch hidden danger issue in the equipment can be repaired. The change management function module is visualized on the interface as the "Submit Change" option. When the user wants to change the repair method for the equipment, they can click the "Submit Change" option. When a click on the "Submit Change" option is detected, the repair method for the batch hidden danger equipment is changed. The batch hidden danger equipment processing system module, by providing a visual interface, enables agile and efficient online closed-loop management of batch hidden dangers.

[0051] Offline training subsystem

[0052] The offline training subsystem mainly performs data preprocessing, data cleaning, and full-dimensional offline data statistics on the historical full data and online recognition results of each device in the entire network, thereby realizing the training of the model. During training, the offline training subsystem performs full-dimensional combination calculations on the component feature data and the device feature data in chronological order based on the historical full data and online recognition results of each device, and combines the extracted component features into a combined dimensional feature sequence based on the chronological relationship, and then the full-dimensional combination results of the component feature data and the device feature data and the combined dimensional feature sequence formed by the combination of component features constitute a combined statistical dimensional sequence data matrix, and then based on the combined statistical dimensional sequence data matrix, realize the training and adjustment process of the model. After the model training is completed, the trained batch hidden danger identification model is synchronously updated to the hidden danger identification engine module of the online identification subsystem, so that the hidden danger identification engine module can identify batch hidden danger equipment online.

[0053] Figure 5 The implementation environment of the method for identifying batch hidden danger equipment provided by the embodiment of the present disclosure is shown. Figure 5, the implementation environment includes: server 501 and server 502.

[0054] Among them, the server 501 can be Figure 1 The online identification subsystem shown in FIG. 5 is integrated with the server 501. Figure 1 The functions of the data acquisition module, batch hidden danger identification engine module and batch hidden danger problem equipment processing system module in the online identification subsystem shown are capable of online collecting the device feature data of each device in the entire network, the component feature data of each key component in each device and the component data of each key component, and processing the collected data, and then identifying the batch hidden danger equipment based on the batch hidden danger identification model trained by the server 502.

[0055] The server 502 may be an offline training subsystem, which can obtain offline the historical full data and online identification results of each device on the Internet. The historical full data includes the device feature data of each device in the historical time period, the component feature data of each key component in each device, the component data of each key component, etc., and train the batch hidden danger identification model offline based on the acquired data, and update the trained batch hidden danger identification model.

[0056] The above-mentioned server 501 and server 502 can be independent physical servers, or a server cluster or distributed system composed of multiple physical servers. They can also be cloud servers that provide basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), as well as big data and artificial intelligence platforms.

[0057] The server 501 and the server 502 may be connected directly or indirectly via wired communication or wireless communication.

[0058] The method for identifying batch devices with potential hazards provided by the embodiments of the present disclosure can be applied to the following scenarios:

[0059] In the first scenario, the business party hopes to discover batch defects in the equipment as early as possible and deal with the equipment with batch defects in a timely manner before the equipment fails, thereby avoiding losses and impacts on the business caused by batch defects.

[0060] In the second scenario, major cloud service providers and equipment manufacturers need to increase their efficiency in handling batches of equipment with potential hazards. On the one hand, this will improve their own service quality, and on the other hand, it will provide support for stable business operations to meet customer needs.

[0061] based on Figure 5In the implementation environment shown, the present disclosure provides a method for identifying batch hidden danger equipment, see Figure 6 , the method process provided by the embodiment of the present disclosure includes:

[0062] 601. Obtain device feature data of a device to be identified, component feature data of each key component in the device to be identified, and component data of each key component.

[0063] Among them, key components are components that affect the normal operation of the device to be identified, device feature data is used to distinguish different devices, component feature data is used to distinguish different key components in the device to be identified, and component data is used to reflect the operating status of the corresponding key components.

[0064] 602. Generate a combination dimension feature sequence of each key component based on the component data of each key component.

[0065] Among them, the combined dimension feature sequence is a sequence composed of at least one component feature.

[0066] 603. Based on the device feature data of the device to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component, batch hidden danger problems of each key component in the device to be identified are identified.

[0067] The method provided by the embodiment of the present disclosure obtains component data of each key component in the equipment to be identified. The component data can characterize the operating status of the equipment to be identified. Therefore, there is no need to use manual calculation and analysis. Batch hidden danger problems in key components can be identified based on the component data. On this basis, based on the equipment feature data of the equipment to be identified and the component feature data of the key component, the key component can be located, thereby improving recognition accuracy and recognition efficiency.

[0068] In another embodiment of the present disclosure, based on the component data of each key component, a combination dimension feature sequence of each key component is generated, including:

[0069] For any key component, according to the component type of the key component, obtain at least one characteristic key field corresponding to the component type;

[0070] Extracting, from component data of a key component, a component feature indicated by at least one feature key field according to at least one feature key field;

[0071] The component features indicated by at least one feature key field are combined into a combined dimension feature sequence of the key component.

[0072] In another embodiment of the present disclosure, before generating the combined dimensional feature sequence of each key component based on the component data of each key component, the method further includes:

[0073] Perform data cleaning on component data of each key component;

[0074] Based on the cleaned component data of each key component, an operation of generating a combined dimensional feature sequence of each key component is performed.

[0075] In another embodiment of the present disclosure, before cleaning the component data of each key component, the method further includes:

[0076] Pre-process component data of each key component;

[0077] Based on the pre-processed component data of each key component, a data cleaning operation is performed on the component data of each key component.

[0078] In another embodiment of the present disclosure, batch hidden danger problems of each key component in the device to be identified are identified based on the device feature data of the device to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component, including:

[0079] Input the device feature data of the equipment to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component into the batch hidden danger identification model, and output the identification results of the batch hidden danger problems of each key component in the equipment to be identified;

[0080] Among them, the batch hidden danger identification model is used to identify batch hidden danger problems of each key component in the equipment based on the equipment feature data of the equipment, the component feature data of the key components and the combined dimension feature sequence of the key components.

[0081] In another embodiment of the present disclosure, the training process of the batch hidden danger identification model includes:

[0082] Obtaining sample device feature data of at least one training sample device, sample component feature data of each sample key component of at least one training sample device, and sample component data of each sample key component, wherein each key component is marked with existing batch hidden danger issues;

[0083] For any sample key component in any training sample device, determining at least one sample feature key field corresponding to the sample key component according to the component type of the sample key component;

[0084] Extracting, from the sample component data of the sample key component, a component feature indicated by at least one sample feature key field according to the at least one sample feature key field;

[0085] Combining component features indicated by at least one sample feature key field into a sample combination dimension feature sequence of sample key components;

[0086] The initial batch hidden danger identification model is trained based on the sample device feature data of at least one training sample device, the sample component feature data of at least one sample key component, and the sample combination dimension feature sequence of at least one sample key component to obtain a batch hidden danger identification model.

[0087] In another embodiment of the present disclosure, the method further comprises:

[0088] Obtaining a failure rate of each sample key component in at least one training sample device within a preset time period;

[0089] Adjusting the characteristic key fields corresponding to each sample key component in at least one training sample device according to the failure rate of each sample key component in a preset time period and a preset threshold value in at least one training sample device;

[0090] The batch hidden danger identification model is updated based on the adjusted characteristic key fields of each sample key component in at least one training sample device.

[0091] In another embodiment of the present disclosure, after identifying batch potential problems of each key component in the device to be identified based on the device feature data of the device to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component, the following further comprises:

[0092] In response to the identification result that a batch hidden danger problem exists in a specified key component in the equipment to be identified, the specified key component is repaired.

[0093] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present disclosure, and will not be described in detail here.

[0094] based on Figure 5 In the implementation environment shown in FIG, the present disclosure provides a method for identifying batch hidden danger equipment. Figure 5 In the implementation environment shown, the server 501 executes the embodiment of the present disclosure as an example, see Figure 7 , the method process provided by the embodiment of the present disclosure includes:

[0095] 701. The server obtains device feature data of the device to be identified, component feature data of each key component in the device to be identified, and component data of each key component.

[0096] Key components are those that affect the normal operation of the device being identified. These include hard drives, memory, CPUs, motherboards, network cards, power supplies, and RAID cards. Device signature data is used to distinguish different devices and includes features such as device model, device ID, and device type. Component signature data is used to distinguish different key components within the device being identified and includes features such as component name and ID. Component data reflects the operating status of the corresponding key components and includes features such as time, component type, component model, and firmware version.

[0097] During the operation of the device to be identified, the server can collect device feature data of the device to be identified, component feature data of each component in the device to be identified, component data of each key component, etc. Since the acquired device feature data, component feature data, and component data are usually in file format and generally difficult to read, it is necessary to parse the file format data to obtain structured data so that the parsed data can be read.

[0098] Based on the acquired device feature data of the device to be identified and the component feature data of each key component in the device to be identified, the server can also perform a full-dimensional combination of the device feature data of the device to be identified and the component feature data of each key component in chronological order. Table 1 shows the full-dimensional combination of the device feature data and component feature data.

[0099] Table 1

[0100]

[0101] 702. The server pre-processes component data of each key component.

[0102] Taking into account that different manufacturers have different naming methods and formats for various components in the equipment, in order to facilitate the management of the component data of each key component in the equipment, the server will also pre-process the component data of each key component of the equipment to be identified. By pre-processing the component data of each key component of the equipment to be identified, the naming method and format of the component data of each key component in the equipment to be identified can be changed. By pre-processing the equipment to be identified, the same key components in different devices have a unified naming method and a unified format.

[0103] 703. The server cleans the component data of each key component.

[0104] The data usually obtained may include some dirty data whose values are not within the given range, or are meaningless for practical purposes, or whose data formats are illegal, or whose codes are not standardized, or whose business logic is ambiguous. These dirty data are meaningless for specific business processing and will also increase the pressure of data processing. Therefore, after pre-processing the component data of each key component of the device to be identified, the server will also perform data cleaning on the component data of each key component to remove the dirty data in each key component.

[0105] 704. The server generates a combination dimension feature sequence of each key component based on the component data of each key component.

[0106] When the server generates the combined dimension feature sequence of each key component based on the component data of each key component, the following steps can be adopted:

[0107] 7041. For any key component, the server obtains at least one characteristic key field corresponding to the component type according to the component type of the key component.

[0108] Key feature fields are fields that can reflect batch potential issues in key components. These key feature fields can be obtained offline by analyzing the component characteristics of key components across a large number of devices. For any key component, the server can obtain the component type from the component data. Based on the component type, the server can then retrieve at least one key feature field corresponding to the component type from a pre-stored correspondence between component types and key feature fields.

[0109] 7042. The server extracts the component feature indicated by at least one feature key field from the component data of the key component based on at least one feature key field.

[0110] Based on the at least one characteristic key field corresponding to the acquired component type, the server may extract the component characteristics indicated by each characteristic key field from the component data of the key component based on the at least one characteristic key field. For example, if the characteristic key fields corresponding to the component type include a component model field, a component version field, a firmware version field, etc., then based on the component model field, the component characteristics indicated by the component model field are extracted from the component data of the key component; based on the component version field, the component characteristics indicated by the component version field are extracted from the component data of the key component; based on the firmware version field, the component characteristics indicated by the firmware version field are extracted from the component data of the key component, and so on.

[0111] 7043. The server combines the component features indicated by at least one feature key field into a combined dimensional feature sequence of key components.

[0112] Based on the component features indicated by at least one acquired feature key field, the server combines the correlations between the component features and the component features indicated by at least one feature key field to obtain a combined dimensional feature sequence for the key component. This combined dimensional feature sequence includes at least one component feature. The correlations between the component features refer to relationships belonging to the same key component, meaning that the component features of the same key component are combined into a combined dimensional feature sequence. Table 2 shows the combined dimensional feature sequence obtained by combining at least one component feature.

[0113] Table 2

[0114]

[0115] 705. The server identifies batch hidden dangers of each key component in the device to be identified based on the device feature data of the device to be identified, the component feature data of each key component, and the combined dimension feature sequence of each key component.

[0116] The disclosed embodiment provides a batch hidden danger identification model, which is used to identify batch hidden danger problems of each key component in the equipment based on the equipment feature data of the equipment, the component feature data of the key components and the combined dimensional feature sequence of the key components. Based on the batch hidden danger identification model, the server inputs the equipment feature data of the equipment to be identified, the component feature data of each key component and the combined dimensional feature sequence of each key component into the batch hidden danger identification model, and outputs the identification results of the batch hidden danger problems of each key component in the equipment to be identified. The identification result is whether there is a batch hidden danger problem for each key component in the equipment to be identified. If any key component does not have a batch hidden danger problem, the relevant data of the key component will not be processed. If the key component has a batch hidden danger problem, the key component will be subsequently repaired according to the specific batch hidden danger problem.

[0117] In another embodiment of the present disclosure, the server identifies batch hidden danger problems for each key component in the device to be identified based on the device feature data of the device to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component. If the identification result is that the specified key component in the device to be identified has a batch hidden danger problem, in response to the identification result that the specified key component in the device to be identified has a batch hidden danger problem, the server repairs the specified key component. If there is a corresponding solution for the batch hidden danger problem, the specified key component is successfully repaired based on the existing solution. If there is no corresponding solution for the batch hidden danger problem, it is determined that the repair of the specified key component has failed. In response to the failure to repair the specified key component, the server may send the component data of the specified key component to the offline training subsystem, and the offline training subsystem will adjust and update the batch hidden danger identification model. Among them, the specified key component can be any key component in the device to be identified that has a batch hidden danger problem.

[0118] The method provided by the embodiment of the present disclosure obtains component data of each key component in the equipment to be identified. The component data can characterize the operating status of the equipment to be identified. Therefore, there is no need to use manual calculation and analysis. Batch hidden danger problems in key components can be identified based on the component data. On this basis, based on the equipment feature data of the equipment to be identified and the component feature data of the key component, the key component can be located, thereby improving recognition accuracy and recognition efficiency.

[0119] based on Figure 5 In the implementation environment shown in FIG, the present disclosure provides a training method for a batch hidden danger identification model. Figure 5 In the implementation environment shown, the server 502 executes the embodiment of the present disclosure as an example, see Figure 8 , the method process provided by the embodiment of the present disclosure includes:

[0120] 801. The server obtains sample device feature data of at least one training sample device, sample component feature data of each sample key component of at least one training sample device, and sample component data of each sample key component.

[0121] The server may obtain device feature data of at least one device from the internet and use the obtained device feature data of the at least one device as sample device feature data for at least one training sample device. The server may also obtain component feature data of each key component of the at least one device from the internet and use the obtained component feature data of each key component of the at least one device as sample component feature data for each sample key component of the at least one training sample device, and obtain sample component data for each sample key component. Each sample key component is annotated with a batch potential risk issue, and each annotated batch potential risk issue corresponds to a label value.

[0122] Of course, in addition to obtaining the above data from the Internet, the server can also receive relevant data of key components that failed to be repaired sent by the online recognition subsystem, and perform model training based on the received relevant data.

[0123] Furthermore, the server will also perform data preprocessing, data cleaning and other operations on the sample device feature data of at least one training sample device, the sample component feature data of each sample key component of at least one training sample device, and the sample component data of each sample key component. After performing the data preprocessing, data cleaning and other operations, the server executes step 802.

[0124] 802. For any sample key component in any training sample device, the server determines at least one sample feature key field corresponding to the sample key component according to the component type of the sample key component.

[0125] For any sample key component in any training sample device, the server can obtain the component type of the sample key component from the component data of the sample key component, and then determine at least one sample feature key field corresponding to the sample key component based on the component type of the sample key component.

[0126] 803. The server extracts, based on at least one sample feature key field, a component feature indicated by at least one sample feature key field from the sample component data of the sample key component.

[0127] Based on the acquired at least one sample feature key field, the server extracts the component feature indicated by each sample feature key field from the sample component data of the sample key component.

[0128] 804. The server combines the component features indicated by at least one sample feature key field into a sample combination dimensional feature sequence of sample key components.

[0129] Based on the extracted component features indicated by each sample feature key field, the server may combine the component features indicated by at least one sample feature key field into a sample combination dimensional feature sequence of sample key components according to the correlation between the component features.

[0130] 805. The server trains the initial batch hidden danger identification model based on the sample device feature data of at least one training sample device, the sample component feature data of at least one sample key component, and the sample combination dimension feature sequence of at least one sample key component to obtain a batch hidden danger identification model.

[0131] The server can construct a target loss function for the initial batch hidden danger identification model in advance, and set initial values for the model parameters of the initial batch hidden danger identification model. Based on the set initial values of each parameter, the predicted value of each sample key component in each training sample device can be determined. By inputting the predicted value and labeled value of each sample key component in each training sample device into the target loss function, the function value of the target loss function can be calculated.

[0132] If the objective loss function value does not meet the threshold condition, the model parameters of the initial batch hidden danger identification model are adjusted, and the objective loss function value is continuously calculated until the obtained function value meets the threshold condition. The threshold condition can be set according to the processing accuracy.

[0133] The parameter values of each parameter when the threshold condition is met are obtained, and the initial batch hidden danger identification model corresponding to the parameter values of each parameter when the threshold condition is met is used as the trained batch hidden danger identification model. The batch hidden danger identification model is used to identify batch hidden danger issues of each key component in the equipment based on the equipment feature data, the component feature data of the key components, and the combined dimension feature sequence of the key components.

[0134] In another embodiment of the present disclosure, the server will also update the trained batch hidden danger identification model. The following steps may be used for the update:

[0135] In the first step, the server obtains the failure rate of each sample key component in at least one training sample device within a preset time period.

[0136] The server obtains the number of each key component in each training sample device within a preset time period, and obtains the number of failures of each key component within the preset time period. Based on the number of failures of each key component within the preset time period and the number of each key component in each training sample device within the preset time period, the server determines the failure rate of each sample key component in the training sample device within the preset time period. The preset time period can be 1 week, 2 weeks, 3 weeks, etc. Table 3 shows the failure rates of the sample key components within different time periods.

[0137] Table 3

[0138]

[0139]

[0140] In the second step, the server adjusts the characteristic key fields corresponding to each sample key component in at least one training sample device according to the failure rate of each sample key component in the at least one training sample device within a preset time period and a preset threshold.

[0141] The preset thresholds are failure rate benchmarks, used to measure the failure rate of each key component within a preset time period. The preset thresholds include a first threshold, a second threshold, and a third threshold. The first threshold can be represented by L1, the second threshold can be represented by L2, and the third threshold can be represented by L3. The first threshold is greater than the second threshold, and the second threshold is greater than the third threshold. The first, second, and third thresholds can be obtained based on statistical data. Table 4 shows the failure rate benchmarks for different key components.

[0142] Table 4

[0143]

[0144] The server compares the failure rate of each sample key component in at least one training sample device within a preset time period with a preset threshold value. When the failure rate of each sample key component in at least one training sample device within a preset time period is greater than the preset threshold value, the characteristic key fields corresponding to each sample key component in at least one training sample device are readjusted, for example, new characteristic key fields are added, and then the batch hidden danger identification model is trained based on the characteristic key fields corresponding to the readjusted sample key fields. When performing a benchmark comparison based on a preset threshold value, the comparison method includes a current benchmark comparison and a recent benchmark comparison, wherein the current benchmark comparison is to compare the current failure rate of each key component with the failure rate benchmark. The recent benchmark comparison is to calculate the average failure rate of the previous 4 periods, and compare the average failure rate of the previous 4 periods with the failure rate benchmark. In the embodiment of the present disclosure, 1 period can be 1 week.

[0145] In another embodiment of the present disclosure, the server can also compare the failure rate of each sample key component within a preset time period with its own failure rate. For example, when the statistical period is weekly, the average weekly failure rate of the key component in the past 12 weeks can be calculated, and the failure rate of the key component in this week can be compared with the average weekly failure rate; when the statistical period is monthly, the average monthly failure rate of the key component in the past 1 year can be calculated, and the failure rate of the key component in this month can be compared with the average monthly failure rate.

[0146] In another embodiment of the present disclosure, the server may also perform a similar comparison of the failure rates of the same sample key components in different training sample devices within the same time period. Similar comparison methods include current period similar comparison, recent period similar comparison, and recent period similar comparison. The current period similar comparison can compare the failure rates of the same sample key components in different training sample devices within the current period, the recent period similar comparison can compare the failure rates of the same sample key components in different training sample devices within four historical periods, and the historical period similar comparison can compare the failure rates of the same sample key components in different training sample devices within 12 historical periods.

[0147] By performing benchmark comparisons, self-comparisons, and similar comparisons, the server can determine the status of key components, including normal, warning, and alarm. For example, when performing benchmark comparisons based on preset thresholds, if the failure rate of a key component is greater than L3, the key component status is determined to be warning; if the failure rate of the key component is greater than L1, the key component status is determined to be alarm.

[0148] Table 5 shows a schematic diagram of the comparison results of the above-mentioned benchmark comparison, self-comparison and similar comparison.

[0149] Table 5

[0150]

[0151] In the third step, the server updates the batch hidden danger identification model based on the adjusted feature key fields of each sample key component in at least one training sample device.

[0152] See also Figure 9 The present disclosure provides a device for identifying batch hidden danger equipment, the device comprising:

[0153] Acquisition module 901 is used to acquire device characteristic data of the device to be identified, component characteristic data of each key component in the device to be identified, and component data of each key component. Key components are components that affect the normal operation of the device to be identified. The device characteristic data is used to distinguish different devices. The component characteristic data is used to distinguish different key components in the device to be identified. The component data is used to reflect the operating status of the corresponding key components.

[0154] A generating module 902 is configured to generate a combined dimension feature sequence of each key component based on the component data of each key component, wherein the combined dimension feature sequence is a sequence composed of at least one component feature;

[0155] The identification module 903 is used to identify batch hidden dangers of each key component in the equipment to be identified based on the equipment feature data of the equipment to be identified, the component feature data of each key component, and the combined dimension feature sequence of each key component.

[0156] In another embodiment of the present disclosure, a generation module 902 is used to obtain, for any key component, at least one characteristic key field corresponding to the component type according to the component type of the key component; extract, from the component data of the key component, component features indicated by at least one characteristic key field according to the at least one characteristic key field; and combine the component features indicated by at least one characteristic key field into a combined dimensional feature sequence of the key component.

[0157] In another embodiment of the present disclosure, the apparatus further comprises:

[0158] Data cleaning module, used to clean component data of each key component;

[0159] The generation module is used to generate a combination dimension feature sequence of each key component based on the component data after cleaning of each key component.

[0160] In another embodiment of the present disclosure, the device further comprises:

[0161] A preprocessing module is used to preprocess component data of each key component;

[0162] The cleaning module is used to clean the component data of each key component based on the pre-processed component data of each key component.

[0163] In another embodiment of the present disclosure, the identification module 903 is configured to input device feature data of the device to be identified, component feature data of each key component, and a combined dimensional feature sequence of each key component into a batch hidden danger identification model, and output an identification result of the batch hidden danger problem of each key component in the device to be identified;

[0164] Among them, the batch hidden danger identification model is used to identify batch hidden danger problems of each key component in the equipment based on the equipment feature data of the equipment, the component feature data of the key components and the combined dimension feature sequence of the key components.

[0165] In another embodiment of the present disclosure, an apparatus for training a batch hidden danger identification model includes:

[0166] an acquisition module, configured to acquire sample device feature data of at least one training sample device, sample component feature data of each sample key component of at least one training sample device, and sample component data of each sample key component, wherein each sample key component is annotated with existing batch hidden danger issues;

[0167] a determination module configured to determine, for any sample key component in any training sample device, at least one sample feature key field corresponding to the sample key component according to the component type of the sample key component;

[0168] an extraction module, configured to extract, from the sample component data of the sample key component, a component feature indicated by at least one sample feature key field based on the at least one sample feature key field;

[0169] a combining module, configured to combine component features indicated by at least one sample feature key field into a sample combination dimension feature sequence of the sample key component;

[0170] The training module is used to train the initial batch hidden danger identification model based on the sample device feature data of at least one training sample device, the sample component feature data of at least one sample key component and the sample combination dimension feature sequence of at least one sample key component to obtain a batch hidden danger identification model.

[0171] In another embodiment of the present disclosure, the apparatus for training a batch hidden danger identification model further includes:

[0172] The acquisition module is further configured to acquire the failure rate of each sample key component in at least one training sample device within a preset time period;

[0173] An adjustment module, configured to adjust a characteristic key field corresponding to each sample key component in at least one training sample device according to a failure rate of each sample key component in the at least one training sample device within a preset time period and a preset threshold;

[0174] An updating module is used to update the batch hidden danger identification model based on the adjusted feature key fields of each sample key component in at least one training sample device.

[0175] In another embodiment of the present disclosure, the apparatus further comprises:

[0176] In response to the identification result that a batch hidden danger problem exists in a specified key component in the equipment to be identified, the specified key component is repaired.

[0177] In summary, the device provided by the embodiment of the present disclosure obtains component data of each key component in the equipment to be identified. The component data can characterize the operating status of the equipment to be identified. Therefore, there is no need to use manual calculation and analysis. Batch hidden dangers in key components can be identified based on the component data. On this basis, based on the equipment feature data of the equipment to be identified and the component feature data of the key components, the key components can be located, thereby improving the recognition accuracy and recognition efficiency.

[0178] Figure 10 A server for identifying batches of devices with hidden dangers is shown according to an exemplary embodiment. Figure 10 Server 1000 includes a processing component 1022, which further includes one or more processors, and memory resources represented by memory 1032 for storing instructions executable by processing component 1022, such as applications. The applications stored in memory 1032 may include one or more modules, each corresponding to a set of instructions. Furthermore, processing component 1022 is configured to execute instructions to perform the functions performed by the server in the aforementioned method for identifying batches of potentially hazardous devices.

[0179] The server 1000 may also include a power supply component 1026 configured to perform power management of the server 1000, a wired or wireless network interface 1050 configured to connect the server 1000 to the network, and an input / output (I / O) interface 1058. The server 1000 may operate based on an operating system stored in the memory 1032, such as Windows Server 2000. TM , Mac OS X TM , Unix TM ,Linux TM , FreeBSD TM or similar.

[0180] The server provided by the embodiment of the present disclosure obtains component data of each key component in the device to be identified. The component data can characterize the operating status of the device to be identified. Therefore, there is no need to use manual calculation and analysis. Batch hidden danger problems in key components can be identified based on the component data. On this basis, based on the device feature data of the device to be identified and the component feature data of the key component, the key component can be located, thereby improving recognition accuracy and recognition efficiency.

[0181] The present disclosure provides a computer-readable storage medium, wherein the storage medium stores at least one program code, and the at least one program code is loaded and executed by a processor to implement Figure 6 or Figure 7The computer-readable storage medium may be non-transitory. For example, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, or an optical data storage device.

[0182] The computer-readable storage medium provided by the embodiments of the present disclosure obtains component data of each key component in the device to be identified. The component data can characterize the operating status of the device to be identified. Therefore, batch hidden dangers in the key components can be identified based on the component data without the need for manual calculation and analysis. On this basis, the key component can be located based on the device feature data of the device to be identified and the component feature data of the key component, thereby improving recognition accuracy and recognition efficiency.

[0183] Those skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or by a program to instruct the relevant hardware, and the program may be stored in a computer-readable storage medium, which may be a read-only memory, a disk, or an optical disk, etc.

[0184] The above description is merely an optional embodiment of the present disclosure and is not intended to limit the present disclosure. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present disclosure shall be included in the scope of protection of the present disclosure.

Claims

1. A method for identifying batch hidden danger equipment, characterized in that: The method comprises: Obtaining device characteristic data of the device to be identified, component characteristic data of each key component in the device to be identified, and component data of each key component, wherein the key component is a component that affects the normal operation of the device to be identified, the device characteristic data is used to distinguish different devices, the component characteristic data is used to distinguish different key components in the device to be identified, and the component data is used to reflect the operating status of the corresponding key component, and the component data includes the shelf time, component type, component model, and firmware version of the key component; For any key component, based on the component type of the key component, obtain at least one characteristic key field corresponding to the component type. The characteristic key field is a field that can reflect batch potential problems in each key component and is obtained by statistically analyzing the component characteristics of each key component of multiple devices offline. The failure rate of the key component in the multiple devices used for statistics meets a preset threshold. extracting, from component data of the key component according to the at least one feature key field, component features indicated by each of the at least one feature key field; Combining the component features indicated by the at least one feature key field into a combined dimensional feature sequence of the key component, wherein the combined dimensional feature sequence is a sequence consisting of at least one component feature; Based on the equipment feature data of the equipment to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component, batch hidden danger problems are identified for each key component in the equipment to be identified. The batch hidden danger problems refer to defects introduced by the key components during the research and development and production processes.

2. The method according to claim 1, characterized in that For any key component, before obtaining at least one characteristic key field corresponding to the component type according to the component type of the key component, the method further includes: Performing data cleaning on component data of each key component; An operation of generating a combined dimensional feature sequence of each key component is performed based on the cleaned component data of each key component.

3. The method according to claim 2, characterized in that Before performing data cleaning on the component data of each key component, the method further includes: Preprocessing component data of each key component; According to the pre-processed component data of each key component, an operation of data cleaning is performed on the component data of each key component.

4. The method according to claim 1, wherein The step of identifying batch potential problems of each key component in the equipment to be identified based on the equipment feature data of the equipment to be identified, the component feature data of each key component, and the combined dimension feature sequence of each key component includes: Input the device feature data of the device to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component into the batch hidden danger identification model, and output the identification results of the batch hidden danger problems of each key component in the device to be identified; The batch hidden danger identification model is used to identify batch hidden danger problems of each key component in the equipment based on the equipment feature data of the equipment, the component feature data of the key components and the combined dimension feature sequence of the key components.

5. The method according to claim 4, characterized in that The training process of the batch hidden danger identification model includes: Obtaining sample device feature data of at least one training sample device, sample component feature data of each sample key component of the at least one training sample device, and sample component data of each sample key component, wherein each sample key component is marked with an existing batch hidden danger problem; For any sample key component in any training sample device, determining at least one sample feature key field corresponding to the sample key component according to the component type of the sample key component; extracting, from the sample component data of the sample key component, a component feature indicated by the at least one sample feature key field according to the at least one sample feature key field; Combining the component features indicated by the at least one sample feature key field into a sample combination dimension feature sequence of the sample key component; The initial batch hidden danger identification model is trained based on the sample device feature data of the at least one training sample device, the sample component feature data of the at least one sample key component, and the sample combination dimensional feature sequence of the at least one sample key component to obtain the batch hidden danger identification model.

6. The method according to claim 5, characterized in that The method further comprises: Obtaining a failure rate of each sample key component in the at least one training sample device within a preset time period; Adjusting the characteristic key fields corresponding to each sample key component in the at least one training sample device according to the failure rate of each sample key component in the at least one training sample device within a preset time period and a preset threshold; The batch hidden danger identification model is updated based on the adjusted characteristic key fields of each sample key component in the at least one training sample device.

7. The method according to any one of claims 1 to 6, characterized in that After identifying batch potential problems of each key component in the equipment to be identified based on the equipment feature data of the equipment to be identified, the component feature data of each key component, and the combined dimension feature sequence of each key component, the method further includes: In response to the identification result that a batch hidden danger problem exists in a specified key component in the equipment to be identified, the specified key component is repaired.

8. A device for identifying batch hidden danger equipment, characterized in that: The device comprises: an acquisition module, configured to acquire device characteristic data of a device to be identified, component characteristic data of each key component in the device to be identified, and component data of each key component, wherein the key component is a component that affects the normal operation of the device to be identified, the device characteristic data is used to distinguish different devices, the component characteristic data is used to distinguish different key components in the device to be identified, and the component data is used to reflect the operating status of the corresponding key component, and the component data includes the shelf time, component type, component model, and firmware version of the key component; A generation module is configured to obtain, for any key component, at least one characteristic key field corresponding to the component type according to the component type of the key component, wherein the characteristic key field is a field that is obtained by performing offline statistics on the component characteristics of each key component of multiple devices and can reflect batch hidden danger problems in each key component, and the failure rate of the key component in the multiple devices used for statistics meets a preset threshold; based on the at least one characteristic key field, extract the component characteristics indicated by each of the at least one characteristic key field from the component data of the key component; and combine the component characteristics indicated by the at least one characteristic key field into a combined dimensional feature sequence of the key component, wherein the combined dimensional feature sequence is a sequence composed of at least one component characteristic; An identification module is used to identify batch hidden danger problems of each key component in the equipment to be identified based on the equipment feature data of the equipment to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component. The batch hidden danger problem refers to defects introduced by the key components during the research and development and production process.

9. The device according to claim 8, characterized in that The device further comprises: A data cleaning module, used for cleaning the component data of each key component; A generating module is used to generate a combination dimension feature sequence of each key component according to the component data of each key component after cleaning.

10. The device according to claim 9, characterized in that The device further comprises: A preprocessing module, used for preprocessing component data of each key component; The cleaning module is used to perform data cleaning operations on the component data of each key component according to the pre-processed component data of each key component.

11. The device according to claim 8, characterized in that The identification module is configured to input the device feature data of the device to be identified, the component feature data of each key component, and the combined dimensional feature sequence of each key component into the batch hidden danger identification model, and output an identification result of the batch hidden danger problem of each key component in the device to be identified; The batch hidden danger identification model is used to identify batch hidden danger problems of each key component in the equipment based on the equipment feature data of the equipment, the component feature data of the key components and the combined dimension feature sequence of the key components.

12. The device according to claim 11, characterized in that The device for training the batch hidden danger identification model includes: an acquisition module, configured to acquire sample device feature data of at least one training sample device, sample component feature data of each sample key component of the at least one training sample device, and sample component data of each sample key component, wherein each sample key component is annotated with an existing batch hidden danger problem; a determination module configured to determine, for any sample key component in any training sample device, at least one sample feature key field corresponding to the sample key component according to the component type of the sample key component; an extraction module, configured to extract, from the sample component data of the sample key component, component features indicated by the at least one sample feature key field based on the at least one sample feature key field; a combining module, configured to combine the component features indicated by the at least one sample feature key field into a sample combination dimension feature sequence of the sample key component; A training module is used to train the initial batch hidden danger identification model based on the sample device feature data of the at least one training sample device, the sample component feature data of the at least one sample key component, and the sample combination dimensional feature sequence of the at least one sample key component to obtain the batch hidden danger identification model.

13. The device according to claim 12, characterized in that The device further comprises: The acquisition module is further configured to acquire the failure rate of each sample key component in the at least one training sample device within a preset time period; an adjustment module, configured to adjust a characteristic key field corresponding to each sample key component in the at least one training sample device according to a failure rate of each sample key component in the at least one training sample device within a preset time period and a preset threshold; An updating module is configured to update the batch hidden danger identification model based on the adjusted characteristic key fields of each sample key component in the at least one training sample device.

14. The device according to any one of claims 8 to 13, characterized in that The identification module is further configured to repair the designated key component in response to an identification result indicating that the designated key component in the device to be identified has a batch hidden danger problem.

15. A server, characterized in that: The server includes a processor and a memory, wherein at least one program code is stored in the memory, and the at least one program code is loaded and executed by the processor to implement the method for identifying batch hidden danger equipment according to any one of claims 1 to 7.

16. A computer-readable storage medium, characterized in that At least one program code is stored in the storage medium, and the at least one program code is loaded and executed by the processor to implement the method for identifying batch hidden danger equipment according to any one of claims 1 to 7.

Images