Remote maintenance methods and related equipment based on distributed experimental equipment

Abstract

This application belongs to the field of data processing, and particularly relates to a remote maintenance method and related equipment for distributed experimental equipment. The method includes: in response to a monitoring command from a distributed experimental equipment management system, acquiring equipment operation data of a target device within the system; inputting the equipment operation data into a fault intelligent diagnosis model to perform anomaly pattern recognition on the target device, thereby obtaining potential anomaly patterns matching the target device; determining equipment flow control information of the distributed experimental equipment management system based on the potential anomaly patterns; generating anomaly alarm information for the target device based on the equipment flow control information and sending it to a user on a remote client side; and in response to a remote maintenance command based on the anomaly alarm information, performing remote maintenance operations on the target device that match the remote maintenance command. This method can improve the efficiency of remote maintenance of distributed experimental equipment, reduce system operational risks, and enhance system stability and reliability.

Description

Technical Field

[0001] This application belongs to the field of data processing, and in particular relates to a remote maintenance method and related equipment based on distributed experimental equipment. Background Technology

[0002] To promote the widespread adoption of intelligent applications across various industries and sectors, distributed experimental equipment generates a large amount of data.

[0003] In related technologies, laboratories deploy various types of experimental equipment, each with different usage and maintenance requirements. This diversity and complexity makes remote maintenance difficult, especially when multiple devices need to be updated and monitored simultaneously. Furthermore, experimental equipment often requires costly maintenance and regular inspections; failure to promptly detect or address equipment problems can lead to expensive repair or replacement costs, as well as the loss of experimental data.

[0004] Therefore, there is an urgent need to design a remote maintenance solution based on distributed experimental equipment to solve at least one of the above-mentioned technical problems. Summary of the Invention

[0005] This application provides a remote maintenance method and related equipment based on distributed experimental equipment, which can improve the efficiency of remote maintenance of distributed experimental equipment, promptly detect potential risk patterns in the system, reduce system operation risks, and enhance system stability and reliability.

[0006] Firstly, this application provides a remote maintenance method for distributed experimental equipment, applied to a distributed experimental equipment management and control system. The distributed experimental equipment management and control system is used to operate and manage experimental equipment, which includes different types of equipment. The remote maintenance method for distributed experimental equipment includes:

[0007] In response to a monitoring command from the distributed experimental equipment management and control system, the system acquires the equipment operation data of the target equipment within the distributed experimental equipment management and control system.

[0008] The device operation data is input into the fault intelligent diagnosis model to perform abnormal pattern recognition of the target device, so as to obtain the potential abnormal patterns matched by the target device.

[0009] Based on the potential anomaly patterns, the equipment flow control information of the distributed experimental equipment management and control system is determined;

[0010] Based on the device flow control information, an abnormal alarm message for the target device is generated and sent to the user on the remote client side;

[0011] In response to a remote maintenance command issued in response to the abnormal alarm information, a remote maintenance operation matching the remote maintenance command is performed on the target device.

[0012] Secondly, embodiments of this application provide a distributed experimental equipment management and control system. This system is used to operate and manage experimental equipment, which includes different types of equipment. The distributed experimental equipment management and control system includes:

[0013] The monitoring unit is configured to acquire the device operation data of the target device in the distributed experimental equipment management system in response to a monitoring command of the distributed experimental equipment management system.

[0014] The identification unit is configured to input the device operation data into the fault intelligent diagnosis model to perform abnormal pattern identification of the target device in order to obtain potential abnormal patterns matching the target device.

[0015] The transfer unit is configured to determine the equipment transfer control information of the distributed experimental equipment management and control system based on the potential abnormal patterns.

[0016] The alarm unit is configured to generate abnormal alarm information for the target device based on the device flow control information and send it to the user on the remote client side;

[0017] The maintenance unit is configured to perform remote maintenance operations on the target device in response to a remote maintenance command that matches the remote maintenance command in response to the feedback of the abnormal alarm information.

[0018] Thirdly, embodiments of this application provide an intelligent computing platform, the intelligent computing platform comprising:

[0019] At least one processor, memory, and input / output unit;

[0020] The memory is used to store computer programs, the processor is used to call the computer programs stored in the memory to execute the remote maintenance method based on distributed experimental equipment in the first aspect, and the input / output unit is used to receive user input and display the output information of the computer programs stored in the memory.

[0021] Fourthly, a computer-readable storage medium is provided, comprising instructions that, when executed on a computer, cause the computer to perform the remote maintenance method of the first aspect based on a distributed experimental device.

[0022] The technical solution provided in this application embodiment can be applied to a distributed experimental equipment management and control system. This system is used to operate and manage experimental equipment, including different types of devices. In this solution, firstly, in response to a monitoring command from the distributed experimental equipment management and control system, the system acquires the equipment operation data of the target device. This provides a data foundation for subsequent intelligent detection and remote maintenance, helping to promptly identify potential abnormal patterns and provide early warnings of possible problems, thereby reducing equipment failures and maintenance costs. Next, the equipment operation data is input into a fault intelligent diagnosis model to perform abnormal pattern recognition of the target device, obtaining potential abnormal patterns matching the target device. This allows for automatic adjustment of equipment processes and control strategies based on abnormal patterns, reducing failure risks and improving equipment utilization and efficiency. Then, based on the potential abnormal patterns, the system determines the equipment flow control information of the distributed experimental equipment management and control system. This helps improve the accuracy and reliability of potential abnormal pattern recognition and reduces the impact of subjective factors on equipment management and maintenance. Finally, based on the equipment flow control information, abnormal alarm information for the target device is generated and sent to the user on the remote client side. Finally, in response to the remote maintenance command issued in response to the abnormal alarm information, a remote maintenance operation matching the remote maintenance command is performed on the target device. This allows for timely detection of abnormal situations and the implementation of corresponding measures, thereby improving the reliability and safety of equipment operation. Furthermore, performing remote maintenance operations in response to remote maintenance commands effectively reduces maintenance response time and improves maintenance efficiency.

[0023] The technical solution of this application, through intelligent fault diagnosis models, equipment flow control, and remote maintenance, helps to realize intelligent and real-time management of distributed experimental equipment management systems, improve the efficiency of remote maintenance of distributed experimental equipment, promptly identify potential risk patterns in the system, reduce system operation risks, enhance system stability and reliability, reduce maintenance costs, and provide users with a better user experience and service support. Attached Figure Description

[0024] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0025] Figure 1 This is a flowchart illustrating a remote maintenance method for distributed experimental equipment according to an embodiment of this application.

[0026] Figure 2 This is a schematic diagram of the structure of a distributed experimental equipment management and control system according to an embodiment of this application;

[0027] Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0030] To promote the widespread adoption of intelligent applications across various industries and sectors, distributed experimental equipment generates a large amount of data.

[0031] In related technologies, laboratories deploy various types of experimental equipment, each with different usage and maintenance requirements. This diversity and complexity makes remote maintenance difficult, especially when multiple devices need to be updated and monitored simultaneously. Furthermore, experimental equipment often requires costly maintenance and regular inspections; failure to promptly detect or address equipment problems can lead to expensive repair or replacement costs, as well as the loss of experimental data.

[0032] Therefore, there is an urgent need to design a remote maintenance solution based on distributed experimental equipment to solve at least one of the above-mentioned technical problems.

[0033] This application provides a remote maintenance method and related equipment based on distributed experimental equipment.

[0034] Specifically, the remote maintenance scheme based on distributed experimental equipment can be applied to a distributed experimental equipment management and control system. This system is used to operate and manage experimental equipment, which includes various types of devices. In this scheme, firstly, in response to monitoring commands from the distributed experimental equipment management and control system, the system acquires the equipment operation data of the target device. This provides a data foundation for subsequent intelligent detection and remote maintenance, helping to promptly identify potential abnormal patterns and provide early warnings of possible problems, thereby reducing equipment failures and maintenance costs. Next, the equipment operation data is input into a fault intelligent diagnosis model to perform abnormal pattern identification of the target device, obtaining potential abnormal patterns matching the target device. This allows for automatic adjustment of equipment processes and control strategies based on abnormal patterns, reducing failure risks and improving equipment utilization and efficiency. Then, based on the potential abnormal patterns, the system determines the equipment flow control information of the distributed experimental equipment management and control system. This helps improve the accuracy and reliability of potential abnormal pattern identification and reduces the impact of subjective factors on equipment management and maintenance. Finally, based on the equipment flow control information, abnormal alarm information for the target device is generated and sent to the user on the remote client side. Finally, in response to the remote maintenance command issued in response to the abnormal alarm information, a remote maintenance operation matching the remote maintenance command is performed on the target device. This allows for timely detection of abnormal situations and the implementation of corresponding measures, thereby improving the reliability and safety of equipment operation. Furthermore, performing remote maintenance operations in response to remote maintenance commands effectively reduces maintenance response time and improves maintenance efficiency.

[0035] In remote maintenance solutions based on distributed experimental equipment, intelligent fault diagnosis models, equipment flow control, and remote maintenance operations help to achieve intelligent and real-time management of the distributed experimental equipment management system, improve the efficiency of remote maintenance of distributed experimental equipment, promptly identify potential risk patterns in the system, reduce system operation risks, enhance system stability and reliability, reduce maintenance costs, and provide users with a better user experience and service support.

[0036] The remote maintenance solution based on distributed experimental equipment provided in this application can also be executed by an electronic device, such as a server, server cluster, or cloud server. This electronic device can also be a terminal device such as a mobile phone, computer, tablet computer, wearable device, or dedicated device (such as a dedicated terminal device with a remote maintenance system based on distributed experimental equipment). These electronic devices can also carry the chips described in the above embodiments. Alternatively, these electronic devices can also install a service program for executing the remote maintenance solution based on distributed experimental equipment.

[0037] Figure 1A schematic diagram illustrating a remote maintenance method for distributed experimental equipment provided in this application embodiment is shown below. Figure 1 The method includes the following steps:

[0038] 101. In response to the monitoring command of the distributed experimental equipment management and control system, acquire the equipment operation data of the target equipment in the distributed experimental equipment management and control system.

[0039] In this embodiment, the distributed experimental equipment management system is used to operate and manage experimental equipment. This experimental equipment includes different types of devices.

[0040] In this embodiment, the remote maintenance method based on the distributed test equipment safety management system involves a relatively complex system architecture, requiring full consideration of communication, data transmission, and security between the server and client. The following is an introduction to the system's technical background and related equipment:

[0041] Optionally, the distributed testing equipment safety management system is designed based on a client / server (C / S) architecture. C / S architecture separates the user interface from the application logic. Users communicate with the server through the client, and the server processes requests and returns results. This architecture is typically used for applications requiring complex logic processing and large amounts of data storage. Microservice architecture is an architectural pattern that breaks down an application into a series of small, independently deployed services. Each microservice can be developed, deployed, and scaled independently, communicating through lightweight communication protocols, thus achieving a highly cohesive and loosely coupled system architecture. QT is a cross-platform C++ application development framework that can be used to develop graphical user interface applications. It provides rich GUI components and tools, simplifying the development process of cross-platform applications while offering good performance and scalability. Distributed testing equipment refers to testing equipment deployed at different locations or nodes. These devices can collaborate to complete testing tasks and communicate and exchange data through a network. Distributed testing equipment typically has high flexibility and scalability, and can be dynamically configured and deployed as needed. The server is the core of the entire system, undertaking the implementation of various functional modules and data processing tasks. In this system, the server adopts a microservice architecture, comprising multiple independent functional modules such as operational status monitoring, audit log collection and analysis, ledger management, and resource management. Each module can be deployed and expanded independently, interacting through a lightweight communication mechanism. The client serves as the user interface, responsible for receiving user input, displaying information, and sending requests to the server. Developed using QT, the client includes modules for user authentication, equipment management, information management, test parameter management, file management, software management, and remote control. Through communication with the server, the client enables user interaction and remote maintenance. In summary, the remote maintenance method for the distributed test equipment security management system combines C / S architecture, microservice architecture, and QT development technology, achieving real-time management and intelligent control of distributed test equipment, providing strong support for the security, stability, and reliability of the test equipment.

[0042] As an optional embodiment, in 101, in response to a monitoring command to the distributed experimental equipment management and control system, acquiring the equipment operation data of the target equipment in the distributed experimental equipment management and control system can be implemented as follows:

[0043] 201. Obtain the experimental tasks matched with the target devices in the distributed experimental equipment management system; wherein, the experimental tasks matched with different types of devices have different task requirements;

[0044] 202. Based on the task requirements of the target device, a remote control task for the target device is constructed through a dynamic control model; wherein, the remote control task is used to assist the target device in achieving the experimental task;

[0045] 203. Synchronize the remote control task of the device to the target device in the distributed experimental equipment management system, and receive the monitoring instructions generated by the target device in response to the remote control task of the device;

[0046] 204. Based on the monitoring instructions, collect the equipment operation data of the target equipment in real time during the execution of the experimental task.

[0047] Specifically, in section 201, by acquiring experimental tasks that match the target equipment, it can be ensured that different types of equipment are managed and controlled according to their task requirements. The experimental task requirements of different equipment may vary, and accurately matching experimental tasks helps improve the efficiency and reliability of equipment operation. In section 202, based on the task requirements of the target equipment, a dynamic control model is constructed to achieve targeted remote control tasks for the equipment. This task model design can better assist the target equipment in completing experimental tasks, improving experimental efficiency and accuracy. In section 203, synchronizing remote control tasks to the target equipment enables the system to monitor and regulate equipment operation in real time, improving the accuracy and real-time nature of control. Simultaneously, receiving monitoring commands from the equipment helps obtain feedback information and adjust control strategies in a timely manner. In section 204, based on monitoring commands, real-time collection of equipment operation data during the execution of experimental tasks helps to comprehensively understand the equipment's operating status, performance, and potential anomalies. This allows for early detection of problems and the implementation of corresponding measures, ensuring the smooth progress of experimental tasks.

[0048] Through steps 201 to 204, the system achieves steps such as experimental task matching, dynamic control model construction, remote equipment control task synchronization, and real-time data acquisition, providing a more efficient and intelligent distributed experimental equipment management system. Precisely matching experimental tasks, building targeted control models, real-time monitoring and adjustment of equipment operation, comprehensive understanding of equipment status, and timely response to anomalies help improve equipment operating efficiency and reliability, achieving refined and automated equipment management, and providing users with a better experimental experience and service support.

[0049] In this embodiment, the dynamic control model is a control model that dynamically adjusts to changes in real-time data and system state. In a distributed experimental equipment control system, the dynamic control model can dynamically adjust the remote control tasks of the equipment according to the task requirements and real-time conditions of the target equipment, so as to ensure that the equipment can effectively complete the predetermined tasks and adapt to the constantly changing working environment.

[0050] Optionally, the dynamic control model is updated and adjusted based on real-time feedback data and monitoring commands to ensure that the model remains consistent with the actual state of the target equipment. By continuously collecting equipment operating data, analyzing equipment performance indicators, and monitoring changes in equipment status, the dynamic control model can promptly identify problems and challenges in equipment operation and adjust remote control tasks accordingly to improve the system's adaptability, stability, and efficiency.

[0051] The key to dynamic control models lies in their ability to respond promptly to changes in equipment status and adjustments to task requirements, enabling real-time control decisions and operational guidance. By implementing dynamic control models, distributed experimental equipment control systems can better adapt to complex working environments and task requirements, improving equipment operating efficiency, reliability, and flexibility, thereby providing users with a superior service experience.

[0052] As an optional example, assume that the dynamic control model includes at least the following structure: a feature recognition layer, a feature fusion layer, and a dynamic construction layer. Based on this, the process of constructing the remote control task corresponding to the target device through the dynamic control model based on the task requirements of the target device can be implemented as follows:

[0053] Through the feature recognition layer, the operation and maintenance requirements features of the target device are extracted from the task requirements of the target device;

[0054] Through the feature fusion layer, multi-dimensional feature fusion and conflict detection are performed on the operation and maintenance requirements of the target device to obtain the comprehensive operation and maintenance characteristics of the target device.

[0055] Through the feature fusion layer, based on the pre-established management and control task components and the spatial mapping relationship between the task components and the requirement dimension, the comprehensive operation and maintenance characteristics of the target device are dynamically constructed to obtain the remote management and control tasks of the target device.

[0056] In a distributed experimental equipment management scenario, a dynamic control model can be used to build a system with intelligent control capabilities, enabling remote monitoring and control of each experimental device. In this scenario, each experimental device may have different operational needs, such as device status, usage frequency, and energy consumption. The feature recognition layer can extract these features from the monitoring data of each experimental device, such as online status, temperature, humidity, and energy consumption. The feature fusion layer fuses the multi-dimensional features extracted from each experimental device, such as fusing real-time status information with historical data and usage patterns, while detecting potential conflicts between features to ensure the accuracy of the comprehensive features. Based on pre-established control task components and the spatial mapping relationship between task components and requirement dimensions, the comprehensive operational features are processed in the dynamic construction layer to dynamically construct remote control tasks. For example, based on the comprehensive characteristics of each experimental device and the experimental progress, the operating parameters or modes of the experimental devices can be dynamically adjusted to improve equipment utilization and experimental efficiency. Through this dynamic control model, intelligent remote management and control of distributed experimental equipment can be achieved, dynamically adjusting control strategies according to real-time data and requirements to improve equipment utilization and efficiency while reducing management costs. This intelligent control system can be applied to various scenarios such as laboratory management and industrial production, providing users with more intelligent and efficient management solutions.

[0057] For example, a dynamic control model based on fuzzy logic control is used to implement a virtual control system for the temperature of a distributed experimental device. The input is assumed to be the difference between the current temperature and the desired temperature, and the output is the magnitude of the heating or cooling control force. The dynamic control model comprises three main parts: fuzzification (definition of the fuzzy set), a fuzzy rule base, and defuzzification (defuzzifying the fuzzy output to obtain the actual control force).

[0058] The first step is fuzzification: The input variable is, for example, the deviation (e), which is the difference between the current temperature and the desired temperature. The deviation can be divided into three fuzzy sets: negative deviation (NB), zero deviation (ZE), and positive deviation (PB). The output variable is, for example, the control force (Δu), which can be divided into three fuzzy sets: cooling (CO), unchanged (NO), and heating (HE). The second step is the fuzzy rule base, which defines the relationship between the deviation and the control force based on experience and rules. For example, Rule 1: IF e is PB THEN Δu is CO; Rule 2: IF e is ZE THEN Δu is NO; Rule 3: IF e is NB THEN Δu is HE. The third step is defuzzification, which involves defuzzifying the control force obtained from the fuzzy rule base. Methods such as weighted averaging can be used to obtain the final control force. This is a dynamic control model based on fuzzy logic control. In practical applications, it is necessary to design fuzzy sets, rule bases, and defuzzification methods according to specific problems and system characteristics in order to achieve effective control of the system.

[0059] 102. Input the equipment operation data into the fault intelligent diagnosis model to perform abnormal pattern recognition of the target equipment, so as to obtain the potential abnormal patterns matched by the target equipment.

[0060] As an optional embodiment, it is assumed that the intelligent fault diagnosis model includes at least the following structure: a data preprocessing layer, a fault feature extraction layer, an adaptive control layer, and an abnormal pattern recognition layer.

[0061] Based on the above model structure, in step 102, inputting the equipment operation data into the intelligent fault diagnosis model to perform abnormal pattern recognition of the target equipment, so as to obtain the potential abnormal patterns matching the target equipment, can be achieved through the following steps:

[0062] 301, The device operation data is preprocessed through a data preprocessing layer; the preprocessing includes at least: data cleaning, data normalization, and data smoothing.

[0063] 302. Through the fault feature extraction layer, statistical features, frequency domain features, and time domain features are extracted from the equipment operation data and fused into fault analysis features.

[0064] 303. Through the adaptive control layer, the model parameters and operating strategies of the abnormal pattern recognition layer are dynamically adjusted according to the real-time monitored equipment status and environmental changes, so as to adapt to different equipment working conditions and corresponding fault modes.

[0065] 304. Through the abnormal pattern recognition layer, the fault analysis features are pattern recognized based on an adaptive recognition strategy to obtain potential abnormal pattern information of the target device.

[0066] In this embodiment, the potential abnormal mode information is used to indicate candidate abnormal behaviors of the target device. For example, the potential abnormal mode information can help identify possible faults or abnormalities in the device, thereby enabling preventative maintenance or fault diagnosis. Below are some examples of possible potential abnormal mode information: 1. Abnormal waveform shape: The waveform data collected by the device's sensors exhibits a special shape within certain time periods, differing from the waveform shape during normal operation, which may indicate a fault in some components of the device. 2. Frequency variation: Abnormal changes in certain frequency components during device operation may indicate a change in the vibration frequency of certain components, possibly due to wear or loosening. 3. Abnormal step response: When the device is subjected to external stimuli or changes in input signals, the step response differs from the normal response, which may indicate an abnormality in the device's control system. 4. Specific pattern recognition: Specific fault modes are identified, such as unstable operating modes or frequent occurrences of similar fault behaviors under certain operating conditions. 5. Abnormal trend: The monitored trend of changes in key device parameters is inconsistent with historical data, which may indicate that the device is in a potential fault state, requiring timely intervention and repair. 6. Frequent occurrence of anomalies: The frequent occurrence of the same anomaly or fault in a recent period may indicate a potential malfunction in a component or system of the equipment.

[0067] This information on potential abnormal patterns can be obtained by analyzing and extracting features from equipment operating data. It can then be applied to abnormal pattern recognition and fault diagnosis, helping maintenance or repair personnel to promptly identify and address potential equipment anomalies, thereby improving equipment reliability and operating efficiency.

[0068] Specifically, in step 301, the data preprocessing layer is mainly used to preprocess the equipment operation data. Data cleaning removes noise, outliers, or missing values ​​to ensure data integrity and accuracy; data normalization scales the data to a uniform size, eliminating the influence of different dimensions; and data smoothing smooths the data to make changes more stable, which is beneficial for subsequent feature extraction and analysis. In step 302, the fault feature extraction layer extracts various features from the preprocessed equipment operation data and integrates them into fault analysis features. Statistical features, such as mean, variance, skewness, and kurtosis, describe the statistical properties of the data; frequency domain features extract frequency domain information using methods such as Fourier transform; and time domain features analyze time-series data to extract features that change over time. In step 303, the adaptive control layer dynamically adjusts the model parameters and operating strategies of the anomaly pattern recognition layer based on real-time monitored equipment status and environmental changes to adapt to different equipment operating conditions and fault modes, ensuring the robustness and efficiency of the diagnosis. In step 304, the anomaly pattern recognition layer performs pattern recognition on the previously extracted fault analysis features based on an adaptive recognition strategy to obtain potential anomaly pattern information of the target device. At this layer, machine learning, pattern matching, or other algorithms are used to analyze the features, identify and define the device's anomaly patterns, and provide accurate fault diagnosis results.

[0069] Through the above-mentioned structure and steps, the intelligent fault diagnosis model can comprehensively analyze equipment operation data, extract fault characteristics, and adaptively adjust the diagnosis strategy according to real-time conditions. Finally, through the abnormal pattern recognition layer, it can realize the abnormal pattern recognition of the target equipment, providing accurate and reliable support for equipment maintenance and fault prediction.

[0070] In an optional embodiment, in step 303 above, the adaptive control layer dynamically adjusts the model parameters and operating strategies of the abnormal pattern recognition layer based on real-time monitored equipment status and environmental changes to adapt to different equipment operating conditions and corresponding fault modes. This can be achieved as follows:

[0071] 401. A probe detection unit is used to identify risk probes on equipment status data and environmental change data to obtain identification results;

[0072] 402. If a first risk probe exists in the equipment status data and / or environmental change data, the model learning rate is adjusted to ensure the speed at which the abnormal pattern recognition layer learns potential abnormal patterns. The first risk probe is used to indicate the phenomenon that the convergence speed does not meet the set threshold when processing certain fault modes. The detection dimension and detection value corresponding to the first risk probe are determined by the target equipment type.

[0073] 403. If a second risk probe exists in the device status data and / or environmental change data, the number of hidden layer nodes is increased or decreased to accommodate the newly added potential abnormal patterns learned by the abnormal pattern recognition layer; the detection dimension and detection value corresponding to the second risk probe are determined by the newly added potential abnormal patterns.

[0074] 404. If a third risk probe is present in the device status data and / or environmental change data, the number of iterations will be dynamically adjusted to improve the robustness of the anomaly pattern recognition layer. The detection dimension and detection value corresponding to the third risk probe are determined by the robustness detection results of the anomaly pattern recognition layer.

[0075] In this embodiment, a probe detection unit and a risk probe are introduced to dynamically adjust the model parameters and operating strategy of the abnormal pattern recognition layer to adapt to different device operating conditions and fault modes: The following is a detailed description of the probe detection unit, the risk probe, and the corresponding steps:

[0076] The probe detection unit is responsible for monitoring equipment status data and environmental change data to identify potential risk probes. It can perform detection based on real-time monitoring data and preset rules or thresholds, indicating whether adjustments to the parameters and strategies of the abnormal pattern recognition layer are needed.

[0077] Risk probes are potential risk identifiers identified by the probe detection unit, used to indicate content that needs dynamic adjustment to ensure the effectiveness of abnormal pattern recognition. In this embodiment, it includes a first risk probe, a second risk probe, and a third risk probe, corresponding to different adjustment dimensions and strategies. In step 402, when the first risk probe is detected, the model learning rate of the abnormal pattern recognition layer is adjusted to improve the learning speed and ensure that potential abnormal patterns can be learned more quickly. In step 403, when the second risk probe is detected, the number of hidden layer nodes is increased or decreased to adapt to newly learned potential abnormal patterns and improve the model's adaptability. In step 404, when the third risk probe is detected, the number of iterations of the abnormal pattern recognition layer is dynamically adjusted to improve the model's robustness and ensure accurate identification of potential abnormal patterns of the device.

[0078] By introducing probe detection units and risk probes, and dynamically adjusting the parameters and strategies of the abnormal pattern recognition layer based on the detection results of different risk probes, it is possible to adapt to changes in equipment operating conditions and fault modes more flexibly and efficiently, thereby improving the accuracy and reliability of abnormal pattern recognition.

[0079] Specifically, in section 401, a probe detection unit is used to identify risk probes from equipment status data and environmental change data. Before obtaining the identification results, the equipment types in the distributed experimental equipment management system can be monitored in real time, and the types of risk probes deployed in the probe detection unit can be dynamically adjusted based on the monitoring results. Then, the dynamically adjusted probe detection unit is distributed to different devices or corresponding management nodes to achieve real-time monitoring of each device in the distributed experimental equipment management system.

[0080] As an optional embodiment, in 304, the fault analysis features are subjected to pattern recognition based on an adaptive recognition strategy through an anomaly pattern recognition layer to obtain potential anomaly pattern information of the target device, which can be implemented as follows:

[0081] Calculate the predicted probability distribution value of each feature point in the fault analysis feature; based on the predicted probability distribution value, generate a predicted probability distribution map of each feature point and its relationship with surrounding feature points; stack the predicted probability distribution maps of each feature point and its relationship with surrounding feature points in three-dimensional space to form a spatial probability distribution prediction network for the fault analysis feature; adopt an adaptive identification strategy that matches the type of the target device to perform adaptive detection on the spatial probability distribution prediction network of the fault analysis feature to obtain candidate probability prediction values ​​of potential abnormal patterns in the fault analysis feature; wherein, the correspondence between the feature point distribution pattern of the spatial probability distribution prediction network and the type of potential abnormal pattern is determined by a pre-input mapping network table in the adaptive identification strategy; make a comprehensive decision on the candidate probability prediction values ​​of potential abnormal patterns to obtain potential abnormal patterns that match the target device.

[0082] In distributed experimental equipment management scenarios, anomaly pattern recognition technology can help monitor and diagnose the operating status of various experimental devices, promptly detect potential abnormal behaviors, and improve equipment reliability and operating efficiency. The following is a detailed description of an embodiment for this scenario. For the operating data of each experimental device, the probability distribution prediction value of fault analysis features is first calculated, for example, by predicting the distribution of feature points using methods such as probability density estimation. Based on the predicted probability distribution values, a probability distribution map of each feature point and the relationships between them is generated. This map reflects the spatial distribution of feature points, which helps in analyzing the correlation between features. Then, the feature points and their probability distribution maps are stacked in three-dimensional space to construct a spatial probability distribution network of fault analysis features, demonstrating the spatial distribution and correlation of features. Next, using an adaptive recognition strategy matching the type of each experimental device, the spatial probability distribution network of fault analysis features is adaptively detected to identify possible anomaly patterns. Through a pre-input mapping network, the correspondence between the distribution pattern of feature points in the spatial probability distribution network and the type of potential anomaly pattern is found. Then, a comprehensive decision is made based on the candidate probability prediction values ​​of potential anomaly patterns to obtain the potential anomaly pattern matching the target device. By comprehensively considering the probability distribution, spatial correlation, and abnormal pattern detection results of each feature point, the final abnormal pattern prediction result is determined. Through the implementation of the above steps, feature analysis and abnormal pattern recognition can be performed on each experimental device in a distributed experimental equipment management scenario. This allows for timely identification of potential abnormal behaviors of the devices, providing important references for equipment maintenance management and fault prediction, and ensuring the stable operation and safety of the experimental equipment.

[0083] In this embodiment of the application, the correspondence between the feature point distribution pattern of the spatial probability distribution prediction network and the potential anomaly pattern type includes one of the following:

[0084] Concentrated anomaly patterns result in the following distributions of feature points in a spatial probability distribution prediction network: specific feature points are abnormally concentrated around one or a few feature points; diffuse anomaly patterns result in the following distributions of feature points in a spatial probability distribution prediction network: specific feature points are abnormally scattered throughout the feature point space; continuous anomaly patterns result in the following distributions of feature points in a spatial probability distribution prediction network: specific feature points exhibit continuous abnormal changes; discontinuous anomaly patterns result in the following distributions of feature points in a spatial probability distribution prediction network: specific feature points or sets of specific feature points exhibit discontinuous anomalies; clustering anomaly patterns result in the following distributions of feature points in a spatial probability distribution prediction network: specific feature points exhibit clustering anomaly patterns in space.

[0085] Specifically, the correspondence between the feature point distribution pattern of the spatial probability distribution prediction network and the potential anomaly pattern type can be determined based on specific fault analysis characteristics and equipment type. Generally, different types of equipment faults will exhibit certain patterns and correspondences in the distribution pattern of feature points. Here are some possible examples of these correspondences:

[0086] Concentrated anomaly patterns: Certain fault modes may cause specific anomalies to concentrate around one or a few characteristic points, forming concentrated anomaly patterns. For example, in vibration signals, a certain frequency component may be excessively high or experience a sudden abnormal increase.

[0087] Scattered anomaly patterns: Other failure modes may cause anomalies to be scattered throughout the feature point space, forming scattered anomaly patterns. For example, in a temperature change curve, the overall temperature distribution shows irregular changes rather than being concentrated at a single point.

[0088] Continuous abnormal mode: Some fault modes may manifest as continuous abnormal changes between characteristic points, forming a continuous abnormal mode. For example, in a current waveform, the current value continuously increases or decreases over a certain period of time.

[0089] Intermittent anomaly patterns: Some fault modes may cause intermittent anomalies in specific feature points or regions, forming intermittent anomaly patterns. For example, in spectrum analysis, abnormal values ​​may appear intermittently in a certain frequency band.

[0090] Clustering Anomalies: Certain failure modes may cause feature points to exhibit anomalous clustering patterns in space. For example, multiple anomalous clusters may form in the feature point space, each cluster representing a different failure type.

[0091] By analyzing the spatial probability distribution of fault analysis features and predicting the correspondence between the feature point distribution patterns and anomaly pattern types in the network, we can provide guidance for adaptive identification strategies, helping to identify potential anomaly patterns and ultimately obtain the matching anomaly patterns of the target device. In practical applications, corresponding anomaly pattern identification strategies can be designed according to specific devices and fault types to improve the accuracy and adaptability of the model.

[0092] 103. Based on the potential anomaly patterns, determine the equipment flow control information of the distributed experimental equipment management and control system.

[0093] In a distributed experimental equipment management system, equipment flow control information can be related to administrators or the permissions of management terminals. This equipment flow control information involves the administrators' access and operation permissions to the experimental equipment and may be affected by potential abnormal behavior.

[0094] In a distributed experimental equipment management system, based on identified potential anomaly patterns, the management system can determine equipment flow control information according to factors such as the severity and scope of the anomaly, to ensure the safety and normal operation of the experimental equipment. When a potential anomaly is detected in the experimental equipment, the system can automatically adjust the permissions of management personnel, such as restricting access or operation permissions for specific personnel or teams to potentially affected equipment, to reduce the possibility of further risks and losses. For management terminals in the distributed experimental equipment management system, the system can dynamically adjust their permission settings according to identified potential anomaly patterns, restricting or expanding the scope of operation of the management terminals on the equipment, ensuring that the equipment is properly managed and controlled in abnormal situations. The management system continuously monitors the operating status and anomalies of the experimental equipment, adjusts the equipment flow control information in real time based on the latest identification results to adapt to changes in equipment status, and takes timely measures to ensure the safe and stable operation of the equipment.

[0095] Through the above steps, the experimental equipment management system can identify potential abnormal patterns and promptly determine appropriate equipment flow control information, ensuring that managers and management terminals can effectively manage and control the equipment in abnormal situations, thereby improving the efficiency and safety of experimental equipment management.

[0096] 104. Based on the device flow control information, generate abnormal alarm information for the target device and send it to the user on the remote client side.

[0097] 105. In response to the remote maintenance command for the feedback of the abnormal alarm information, perform a remote maintenance operation on the target device that matches the remote maintenance command.

[0098] In the distributed experimental equipment management scenario, steps 104 and 105 involve the process of generating abnormal alarm information for the target equipment and implementing remote maintenance operations. These steps aim to respond promptly to equipment anomalies and ensure the effective execution of remote maintenance instructions. In step 104, based on equipment flow control information, the system analyzes potential anomaly patterns and generates corresponding abnormal alarm information. This information includes the nature, severity, and scope of impact of the anomaly, so as to notify relevant personnel or the system. The generated abnormal alarm information will be sent to users on the remote client side, informing them of potential anomalies in the target equipment. Users can monitor and track the equipment status in real time through the client to take necessary measures to respond to anomalies in a timely manner. In step 105, after receiving the abnormal alarm information, the user on the remote client side will formulate remote maintenance instructions based on the specific situation and send them to the management system. The management system will execute remote maintenance operations matching the received remote maintenance instructions on the target equipment. This may include operations such as restarting the equipment, remote diagnostics, and remote calibration to resolve equipment anomalies or malfunctions. After the remote maintenance operations are completed, the system will provide feedback on the operation execution results to the user on the remote client side, allowing the user to understand the effect of the operation and the change in equipment status.

[0099] Through steps 104 and 105, the distributed experimental equipment management system can generate corresponding alarm information and perform effective remote maintenance operations when equipment anomalies are detected. This improves the efficiency of equipment fault handling, reduces maintenance costs, ensures continuous and stable equipment operation, and enhances user-equipment interaction and real-time monitoring capabilities, thereby improving overall operational efficiency and safety.

[0100] In this embodiment, in response to a monitoring command from the distributed experimental equipment management system, the system acquires the equipment operation data of the target equipment within the system. This provides a data foundation for subsequent intelligent detection and remote maintenance, helping to promptly identify potential abnormal patterns and provide early warnings of possible problems, thereby reducing equipment failures and maintenance costs. Next, the equipment operation data is input into a fault intelligent diagnosis model to perform abnormal pattern identification of the target equipment, obtaining potential abnormal patterns matching the target equipment. This allows for automatic adjustment of equipment processes and control strategies based on abnormal patterns, reducing failure risks and improving equipment utilization and efficiency. Then, based on the potential abnormal patterns, the system determines the equipment flow control information of the distributed experimental equipment management system. This helps improve the accuracy and reliability of potential abnormal pattern identification and reduces the impact of subjective factors on equipment management and maintenance. Furthermore, based on the equipment flow control information, abnormal alarm information for the target equipment is generated and sent to the user on the remote client side. Finally, in response to a remote maintenance command based on the abnormal alarm information, a remote maintenance operation matching the remote maintenance command is performed on the target equipment. This allows for timely detection of abnormal situations and the implementation of corresponding measures, improving the reliability and safety of equipment operation. Meanwhile, responding to remote maintenance commands and performing remote maintenance operations can effectively reduce maintenance response time and improve maintenance efficiency. In this embodiment, through intelligent fault diagnosis models, equipment flow control, and remote maintenance, it helps to achieve intelligent and real-time management of the distributed experimental equipment management system, improve the remote maintenance efficiency of distributed experimental equipment, promptly identify potential risk patterns in the system, reduce system operational risks, enhance system stability and reliability, reduce maintenance costs, and provide users with a better user experience and service support.

[0101] In another embodiment of this application, a distributed experimental equipment management and control system is also provided. This system is used to operate and manage experimental equipment, which includes different types of equipment. (See also...) Figure 2 The distributed experimental equipment management and control system includes the following units:

[0102] The monitoring unit is configured to acquire the device operation data of the target device in the distributed experimental equipment management system in response to a monitoring command of the distributed experimental equipment management system.

[0103] The identification unit is configured to input the device operation data into the fault intelligent diagnosis model to perform abnormal pattern identification of the target device in order to obtain potential abnormal patterns matching the target device.

[0104] The transfer unit is configured to determine the equipment transfer control information of the distributed experimental equipment management and control system based on the potential abnormal patterns.

[0105] The alarm unit is configured to generate abnormal alarm information for the target device based on the device flow control information and send it to the user on the remote client side;

[0106] The maintenance unit is configured to perform remote maintenance operations on the target device in response to a remote maintenance command that matches the remote maintenance command in response to the feedback of the abnormal alarm information.

[0107] Further optionally, the monitoring unit, in response to a monitoring command to the distributed experimental equipment management system, acquires the equipment operation data of the target equipment in the distributed experimental equipment management system, and is configured to:

[0108] Obtain the experimental tasks matched with the target devices in the distributed experimental equipment management system; wherein, the experimental tasks matched with different types of devices have different task requirements;

[0109] Based on the task requirements of the target device, a remote control task for the target device is constructed through a dynamic control model; wherein, the remote control task is used to assist the target device in achieving the experimental task.

[0110] The system synchronizes the remote control task of the equipment with the target equipment in the distributed experimental equipment management system, and receives the monitoring instructions generated by the target equipment in response to the remote control task.

[0111] Based on the monitoring instructions, real-time data on the operation of the target device during the execution of the experimental task is collected.

[0112] Further optionally, the dynamic control model includes at least the following structure: a feature recognition layer, a feature fusion layer, and a dynamic construction layer;

[0113] The monitoring unit, based on the task requirements of the target device, constructs the corresponding remote management and control task for the target device through a dynamic management and control model, and is configured as follows:

[0114] Through the feature recognition layer, the operation and maintenance requirements features of the target device are extracted from the task requirements of the target device;

[0115] Through the feature fusion layer, multi-dimensional feature fusion and conflict detection are performed on the operation and maintenance requirements of the target device to obtain the comprehensive operation and maintenance characteristics of the target device.

[0116] Through the feature fusion layer, based on the pre-established management and control task components and the spatial mapping relationship between the task components and the requirement dimension, the comprehensive operation and maintenance characteristics of the target device are dynamically constructed to obtain the remote management and control tasks of the target device.

[0117] Further optionally, the fault intelligent diagnosis model includes at least the following structure: a data preprocessing layer, a fault feature extraction layer, an adaptive control layer, and an abnormal pattern recognition layer;

[0118] The identification unit inputs the device operation data into the fault intelligent diagnosis model to perform abnormal pattern identification of the target device, so as to obtain the potential abnormal patterns matching the target device, and is configured as follows:

[0119] The device operation data is preprocessed through a data preprocessing layer; the preprocessing includes at least: data cleaning, data normalization, and data smoothing.

[0120] Through the fault feature extraction layer, statistical features, frequency domain features, and time domain features are extracted from the equipment operation data and fused into fault analysis features.

[0121] Through the adaptive control layer, the model parameters and operating strategies of the abnormal pattern recognition layer are dynamically adjusted according to the real-time monitored equipment status and environmental changes, so as to adapt to different equipment working conditions and corresponding fault modes.

[0122] The fault analysis features are pattern-recognized using an anomaly pattern recognition layer based on an adaptive recognition strategy to obtain potential anomaly pattern information of the target device; the potential anomaly pattern information is used to indicate candidate abnormal behaviors of the target device.

[0123] Further optionally, the identification unit, through an adaptive control layer, dynamically adjusts the model parameters and operating strategy of the abnormal pattern identification layer according to real-time monitored equipment status and environmental changes, to adapt to different equipment operating conditions and corresponding fault modes, and is configured as follows:

[0124] A probe detection unit is used to identify risk probes from equipment status data and environmental change data to obtain identification results;

[0125] If a first risk probe is present in the device status data and / or environmental change data, the model learning rate is adjusted to ensure the speed at which the abnormal pattern recognition layer learns potential abnormal patterns. The first risk probe is used to indicate the phenomenon that the convergence speed does not meet the set threshold when processing certain fault modes. The detection dimension and detection value corresponding to the first risk probe are determined by the target device type.

[0126] If a second risk probe exists in the device status data and / or environmental change data, the number of hidden layer nodes is increased or decreased to accommodate the newly added potential abnormal patterns learned by the abnormal pattern recognition layer; the detection dimension and detection value corresponding to the second risk probe are determined by the newly added potential abnormal patterns.

[0127] If a third risk probe is present in the device status data and / or environmental change data, the number of iterations is dynamically adjusted to improve the robustness of the anomaly pattern recognition layer; the detection dimension and detection value corresponding to the third risk probe are determined by the robustness detection results of the anomaly pattern recognition layer.

[0128] Further optionally, the identification unit, employing a probe detection unit, performs risk probe identification on device status data and environmental change data to obtain the identification result, and is further configured as follows:

[0129] The device types in the distributed experimental equipment management system are monitored in real time, and the types of risk probes deployed in the probe detection unit are dynamically adjusted based on the monitoring results.

[0130] The dynamically adjusted probe detection units are distributed to different devices or corresponding control nodes to achieve real-time monitoring of each device in the distributed experimental equipment control system.

[0131] Further optionally, the identification unit, through an anomaly pattern recognition layer, performs pattern recognition on the fault analysis features based on an adaptive identification strategy to obtain potential anomaly pattern information of the target device, and is configured as follows:

[0132] Calculate the predicted probability distribution of each feature point in the fault analysis features;

[0133] Based on the probability distribution prediction values, generate a probability distribution prediction map of each feature point and its relationship with surrounding feature points.

[0134] The probability distribution prediction maps of each feature point and its relationship with surrounding feature points are stacked in three-dimensional space to form the spatial probability distribution prediction network of the fault analysis features.

[0135] An adaptive identification strategy matching the type of the target device is adopted to adaptively detect the spatial probability distribution prediction network of the fault analysis features to obtain the candidate probability prediction values ​​of potential abnormal patterns in the fault analysis features; wherein, the correspondence between the feature point distribution pattern of the spatial probability distribution prediction network and the type of potential abnormal pattern is determined by the mapping network table pre-input in the adaptive identification strategy.

[0136] A comprehensive decision is made based on the predicted candidate probabilities of potential abnormal patterns to obtain the potential abnormal patterns that match the target device.

[0137] Further, optionally, the correspondence between the feature point distribution pattern of the spatial probability distribution prediction network and the potential anomaly pattern type is as follows:

[0138] The concentrated anomaly pattern leads to the distribution of feature points in the spatial probability distribution prediction network as follows: specific feature points are abnormally concentrated around one or a few feature points.

[0139] The diffuse anomaly pattern leads to the distribution of feature points in the spatial probability distribution prediction network as follows: specific feature points are abnormally scattered throughout the feature point space.

[0140] Continuous anomalous patterns result in the distribution of feature points in the spatial probability distribution prediction network being such that there are continuous anomalous changes between specific feature points;

[0141] Discontinuous anomaly patterns cause the distribution of feature points in the spatial probability distribution prediction network to be discontinuous anomalies in specific feature points or sets of specific feature points;

[0142] Clustering anomalies cause the distribution of feature points in a spatial probability distribution prediction network to exhibit anomalies in clustering patterns in space.

[0143] In this embodiment, through intelligent fault diagnosis models, equipment flow control, and remote maintenance, it is helpful to realize intelligent and real-time management of the distributed experimental equipment management system, improve the remote maintenance efficiency of distributed experimental equipment, promptly detect potential risk patterns in the system, reduce system operation risks, enhance system stability and reliability, reduce maintenance costs, and provide users with a better user experience and service support.

[0144] In another embodiment of this application, an intelligent computing platform is also provided, including: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0145] Memory, used to store computer programs;

[0146] When the processor executes a program stored in memory, it implements the remote maintenance method based on distributed experimental equipment as described in the method embodiment.

[0147] The communication bus 1140 mentioned in the above electronic device may be a peripheral component interconnection standard.

[0148] (Peripheral Component Interconnect, PCI) bus or extended industry standard architecture

[0149] (Extended Industry Standard Architecture, EISA) bus, etc. This communication bus 1140 can be divided into address bus, data bus, control bus, etc.

[0150] This application provides a remote maintenance method for a distributed experimental device used to construct a low-power computing unit.

[0151] For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0152] The communication interface 1120 is used for communication between the above-mentioned electronic device and other devices.

[0153] The memory 1130 may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0154] The aforementioned processor 1110 can be a general-purpose processor, including artificial intelligence processors, graphics processing units (GPUs), machine learning units (MLUs), central processing units (CPUs), network processors (NPs), etc.; it can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0155] Accordingly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed, can implement the steps that can be performed by an electronic device in the above method embodiments.

Claims

1. A remote maintenance method based on distributed experimental equipment, characterized in that, A method is applied to a distributed experimental equipment management and control system, which is used to operate and manage experimental equipment, including different types of equipment; the method includes: In response to a monitoring command from the distributed experimental equipment management and control system, the system acquires the equipment operation data of the target equipment within the distributed experimental equipment management and control system. The device operation data is input into the fault intelligent diagnosis model to perform abnormal pattern recognition of the target device in order to obtain the potential abnormal patterns matched by the target device. The fault intelligent diagnosis model includes at least the following structure: data preprocessing layer, fault feature extraction layer, adaptive control layer, and abnormal pattern recognition layer. Based on the potential anomaly patterns, the equipment flow control information of the distributed experimental equipment management and control system is determined; Based on the device flow control information, an abnormal alarm message for the target device is generated and sent to the user on the remote client side; In response to a remote maintenance command for the feedback of the abnormal alarm information, perform a remote maintenance operation on the target device that matches the remote maintenance command; The step of responding to a monitoring command from the distributed experimental equipment management system and acquiring the equipment operation data of the target equipment in the distributed experimental equipment management system includes: Obtain the experimental tasks matched with the target devices in the distributed experimental equipment management system; wherein, the experimental tasks matched with different types of devices have different task requirements; Based on the task requirements of the target device, a remote control task for the target device is constructed through a dynamic control model; wherein, the remote control task is used to assist the target device in performing experimental tasks; the dynamic control model includes at least the following structure: feature recognition layer, feature fusion layer, and dynamic construction layer; The system synchronizes the remote control task of the equipment with the target equipment in the distributed experimental equipment management system, and receives the monitoring instructions generated by the target equipment in response to the remote control task. Based on the monitoring instructions, real-time data on the target device's operation during the execution of the experimental task is collected. Through the adaptive control layer, the model parameters and operating strategies of the anomaly pattern recognition layer are dynamically adjusted based on real-time monitoring of equipment status and environmental changes to adapt to different equipment operating conditions and corresponding fault modes, including: A probe detection unit is used to identify risk probes from equipment status data and environmental change data to obtain identification results; If a first risk probe is present in the device status data and / or environmental change data, the model learning rate is adjusted to ensure the speed at which the abnormal pattern recognition layer learns potential abnormal patterns. The first risk probe is used to indicate the phenomenon that the convergence speed does not meet the set threshold when processing certain fault modes. The detection dimension and detection value corresponding to the first risk probe are determined by the target device type. If a second risk probe exists in the device status data and / or environmental change data, the number of hidden layer nodes is increased or decreased to accommodate the newly added potential abnormal patterns learned by the abnormal pattern recognition layer; the detection dimension and detection value corresponding to the second risk probe are determined by the newly added potential abnormal patterns. If a third risk probe is present in the device status data and / or environmental change data, the number of iterations is dynamically adjusted to improve the robustness of the anomaly pattern recognition layer; the detection dimension and detection value corresponding to the third risk probe are determined by the robustness detection results of the anomaly pattern recognition layer. Through the anomaly pattern recognition layer, pattern recognition is performed on fault analysis features based on an adaptive recognition strategy to obtain potential anomaly pattern information of the target device, including: Calculate the predicted probability distribution of each feature point in the fault analysis features; Based on the probability distribution prediction values, generate a probability distribution prediction map of each feature point and its relationship with surrounding feature points. The probability distribution prediction maps of each feature point and its relationship with surrounding feature points are stacked in three-dimensional space to form the spatial probability distribution prediction network of the fault analysis features. An adaptive identification strategy matching the type of the target device is adopted to adaptively detect the spatial probability distribution prediction network of the fault analysis features to obtain the candidate probability prediction values ​​of potential abnormal patterns in the fault analysis features; wherein, the correspondence between the feature point distribution pattern of the spatial probability distribution prediction network and the type of potential abnormal pattern is determined by the mapping network table pre-input in the adaptive identification strategy. A comprehensive decision is made based on the predicted candidate probabilities of potential abnormal patterns to obtain the potential abnormal patterns that match the target device.

2. The remote maintenance method based on distributed experimental equipment according to claim 1, characterized in that, The process of constructing remote device management tasks based on the target device's task requirements using a dynamic management model includes: Through the feature recognition layer, the operation and maintenance requirements features of the target device are extracted from the task requirements of the target device; Through the feature fusion layer, multi-dimensional feature fusion and conflict detection are performed on the operation and maintenance requirements of the target device to obtain the comprehensive operation and maintenance characteristics of the target device. Through the feature fusion layer, based on the pre-established management and control task components and the spatial mapping relationship between the task components and the requirement dimension, the comprehensive operation and maintenance characteristics of the target device are dynamically constructed to obtain the remote management and control tasks of the target device.

3. The remote maintenance method based on distributed experimental equipment according to claim 1, characterized in that, The step of inputting the equipment operation data into the fault intelligent diagnosis model to perform abnormal pattern recognition of the target equipment, so as to obtain the potential abnormal patterns matching the target equipment, includes: The device operation data is preprocessed through a data preprocessing layer; the preprocessing includes at least: data cleaning, data normalization, and data smoothing. Through the fault feature extraction layer, statistical features, frequency domain features, and time domain features are extracted from the equipment operation data and fused into fault analysis features. Through the adaptive control layer, the model parameters and operating strategies of the abnormal pattern recognition layer are dynamically adjusted according to the real-time monitored equipment status and environmental changes, so as to adapt to different equipment working conditions and corresponding fault modes. The fault analysis features are pattern-recognized using an anomaly pattern recognition layer based on an adaptive recognition strategy to obtain potential anomaly pattern information of the target device; the potential anomaly pattern information is used to indicate candidate abnormal behaviors of the target device.

4. The remote maintenance method based on distributed experimental equipment according to claim 1, characterized in that, Before the step of using a probe detection unit to perform risk probe identification on equipment status data and environmental change data to obtain the identification results, the method further includes: The device types in the distributed experimental equipment management system are monitored in real time, and the types of risk probes deployed in the probe detection unit are dynamically adjusted based on the monitoring results. The dynamically adjusted probe detection units are distributed to different devices or corresponding control nodes to achieve real-time monitoring of each device in the distributed experimental equipment control system.

5. The remote maintenance method based on distributed experimental equipment according to claim 1, characterized in that, The correspondence between the feature point distribution pattern of the spatial probability distribution prediction network and the potential anomaly pattern type includes: The concentrated anomaly pattern leads to the distribution of feature points in the spatial probability distribution prediction network as follows: specific feature points are abnormally concentrated around one or a few feature points. The diffuse anomaly pattern leads to the distribution of feature points in the spatial probability distribution prediction network as follows: specific feature points are abnormally scattered throughout the feature point space. Continuous anomalous patterns result in the distribution of feature points in the spatial probability distribution prediction network being such that there are continuous anomalous changes between specific feature points; Discontinuous anomaly patterns cause the distribution of feature points in the spatial probability distribution prediction network to be discontinuously anomalous in specific feature points or sets of specific feature points; Clustering anomalies cause the distribution of feature points in a spatial probability distribution prediction network to exhibit anomalies in clustering patterns in space.

6. A distributed experimental equipment management and control system, used to execute the method of claim 1, characterized in that, The distributed experimental equipment management and control system is used to operate and manage experimental equipment, which includes different types of equipment. The distributed experimental equipment management and control system includes: The monitoring unit is configured to acquire the device operation data of the target device in the distributed experimental equipment management system in response to a monitoring command of the distributed experimental equipment management system. The identification unit is configured to input the device operation data into the fault intelligent diagnosis model to perform abnormal pattern identification of the target device in order to obtain potential abnormal patterns matching the target device. The transfer unit is configured to determine the equipment transfer control information of the distributed experimental equipment management and control system based on the potential abnormal patterns. The alarm unit is configured to generate abnormal alarm information for the target device based on the device flow control information and send it to the user on the remote client side; The maintenance unit is configured to perform remote maintenance operations on the target device in response to a remote maintenance command that matches the remote maintenance command in response to the feedback of the abnormal alarm information.

7. An intelligent computing platform, characterized in that, The intelligent computing platform includes: At least one processor, memory, and input / output unit; The memory is used to store computer programs, the processor is used to call the computer programs stored in the memory to execute the remote maintenance method based on distributed experimental equipment as described in any one of claims 1 to 5, and the input / output unit is used to receive user input and display the output information of the computer programs stored in the memory.

Images