Equipment operation and maintenance management system and method

Abstract

The invention discloses an equipment operation and maintenance management system and method. Multi-modal real-time operation data is acquired through the multi-modal data acquisition module, so that the operation state of the target equipment in the whole life cycle can be reflected more comprehensively, and the accuracy and timeliness of fault detection are improved. The fault early warning module carries out deep analysis on the real-time operation data by using a pre-trained fault early warning model supporting multi-modal data analysis in combination with historical operation data of the target equipment and pre-configured risk early warning conditions, can accurately judge whether the real-time operation data is normal or abnormal operation data, and improves the accuracy of the real-time operation data. And a corresponding fault diagnosis result is given, so that potential faults can be found in advance, and the fault processing efficiency is improved. The fault processing module automatically matches and obtains a corresponding emergency processing scheme by using the pre-trained fault processing model and historical operation and maintenance case knowledge in the historical fault knowledge base, so that the burden of operation and maintenance personnel is reduced, and the accuracy and efficiency of fault processing are improved.

Description

Technical Field

[0001] The present application relates to the field of operation and maintenance information management, and in particular to an equipment operation and maintenance management system and method. Background Art

[0002] With the rapid development of the modern transportation industry, rail transit, as an important part of the urban public transportation system, has received increasing attention for its safety and operational efficiency. As a key element supporting the stable operation of the entire system, the operation and maintenance of rail transit equipment is directly related to the safety and travel experience of passengers. However, in the current field of rail transit equipment operation and maintenance, traditional equipment status monitoring systems face many challenges, which are mainly due to the limitations of monitoring methods and insufficient data analysis and processing capabilities.

[0003] Traditional equipment status monitoring systems mainly rely on preset fixed rules and thresholds to determine whether the equipment is in an abnormal state. Although this method can achieve basic monitoring functions to a certain extent, its limitations are obvious. First of all, fixed rules and thresholds are often difficult to fully cover the various complex working conditions and potential failure modes that the equipment may encounter, resulting in insufficient monitoring coverage. In particular, at different stages of the equipment life cycle, its performance parameters and fault characteristics will change, and traditional static monitoring methods cannot effectively adapt to such dynamic changes, and thus cannot fully capture the state changes of the equipment throughout its life cycle. This directly leads to the lag and inaccuracy of warning information, making it difficult to provide timely and accurate fault warning information to operation and maintenance personnel.

[0004] When dealing with emergencies, the shortcomings of traditional monitoring systems are more prominent. Due to the lack of an effective plan matching mechanism and rapid response capabilities, when equipment fails, manual intervention is often required for diagnosis and processing. This not only increases the workload of operation and maintenance personnel, but also prolongs the troubleshooting time and reduces operation and maintenance efficiency. At the same time, due to the inaccuracy and lag of plan matching, it may also lead to improper or over-reaction of fault handling measures, further affecting the safe and stable operation of the rail transit system.

[0005] In summary, there are many problems and shortcomings in the current traditional monitoring methods in the operation and maintenance scenarios of rail transit equipment, and there is an urgent need to introduce more advanced and intelligent monitoring technologies and methods. Summary of the invention

[0006] The present application provides an equipment operation and maintenance management system and method, which is used to solve the problems that the existing traditional detection system cannot fully capture the status changes of the equipment throughout its life cycle, has lags and inaccuracies, lacks an effective plan matching mechanism and rapid response capabilities, and has low operation and maintenance efficiency.

[0007] In a first aspect, the present application provides an equipment operation and maintenance management system, the system comprising:

[0008] A multimodal data acquisition module, used to acquire real-time operation data of a target device; wherein the real-time operation data includes data collected from multiple modes, and the multiple modes include but are not limited to at least two of the following: audio, video, and numerical value;

[0009] A fault warning module is used to obtain the real-time operation data; perform fault analysis on the real-time operation data through a pre-trained fault warning model that supports multimodal data analysis of the target device to obtain a fault diagnosis result; wherein the fault warning model combines the pre-saved historical operation data of the target device to determine the target change data in the real-time operation data, and evaluates the target change data in combination with the pre-configured risk warning conditions; if it is determined that the target change data meets the risk warning conditions, the fault diagnosis result identifies the real-time operation data as abnormal operation data; if it is determined that the target change data does not meet the risk warning conditions, the fault diagnosis result identifies the real-time operation data as normal operation data;

[0010] A fault processing module is used to perform fault processing on the abnormal operating data by using a pre-trained fault processing model that supports multimodal data analysis of the target device when the fault diagnosis result of the real-time operating data identifies the real-time operating data as abnormal operating data, so as to obtain an emergency disposal plan; wherein the fault processing model combines historical operation and maintenance case knowledge in a pre-built historical fault knowledge base to determine a target historical operation and maintenance case that matches the abnormal operating data, and obtains a saved emergency disposal plan from the target historical operation and maintenance case;

[0011] An output module is used to output the emergency response plan.

[0012] In a second aspect, the present application also provides a device operation and maintenance management method, the method comprising:

[0013] Acquire real-time operation data of the target device; wherein the real-time operation data includes data collected from multiple modalities, and the multiple modalities include but are not limited to at least two of the following: audio, video, and numerical value;

[0014] By means of a pre-trained fault warning model supporting multi-modal data analysis of the target device, a fault analysis is performed on the real-time operation data to obtain a fault diagnosis result; wherein the fault warning model combines the pre-saved historical operation data of the target device to determine the target change data in the real-time operation data, and evaluates the target change data in combination with the pre-configured risk warning conditions; if it is determined that the target change data meets the risk warning conditions, the fault diagnosis result identifies the real-time operation data as abnormal operation data; if it is determined that the target change data does not meet the risk warning conditions, the fault diagnosis result identifies the real-time operation data as normal operation data;

[0015] In the case where the fault diagnosis result of the real-time operation data identifies that the real-time operation data is abnormal operation data, the abnormal operation data is subjected to fault processing by means of a pre-trained fault processing model that supports multimodal data analysis of the target device, so as to obtain an emergency disposal plan; wherein the fault processing model combines historical operation and maintenance case knowledge in a pre-built historical fault knowledge base to determine a target historical operation and maintenance case that matches the abnormal operation data, and obtains a saved emergency disposal plan from the target historical operation and maintenance case;

[0016] Output the emergency response plan.

[0017] The beneficial effects of this application are as follows:

[0018] 1. Through the multimodal data acquisition module, the system can simultaneously obtain the real-time operation data of the target device from multiple dimensions such as audio, video, and numerical values, so as to more comprehensively reflect the operating status of the target device throughout its life cycle and improve the accuracy and timeliness of fault detection.

[0019] 2. The fault warning module uses a pre-trained fault warning model that supports multimodal data analysis, combined with the historical operation data of the target equipment and the pre-configured risk warning conditions, to conduct in-depth analysis of the real-time operation data. It can accurately determine whether the real-time operation data is normal or abnormal operation data, and give corresponding fault diagnosis results, which helps to discover potential faults in advance, facilitates operation and maintenance personnel to quickly locate problems, improves the efficiency of fault handling, and improves the stability and reliability of the system.

[0020] 3. When the fault diagnosis result is abnormal operating data, the fault handling module will use the pre-trained fault handling model and the historical operation and maintenance case knowledge in the historical fault knowledge base to automatically match and obtain the corresponding emergency response plan, which not only reduces the burden on operation and maintenance personnel, but also improves the accuracy and efficiency of fault handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0022] Figure 1 A schematic diagram of the structure of an equipment operation and maintenance management system provided in an embodiment of the present application;

[0023] Figure 2 A schematic diagram of the workflow of an equipment operation and maintenance management system provided for this application;

[0024] Figure 3 A process diagram of a device operation and maintenance management method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0026] In order to timely and comprehensively capture the status changes of equipment throughout its life cycle, improve the efficiency and quality of equipment operation and maintenance management, and provide an effective plan matching mechanism and rapid response capabilities, the present application provides a system and method for equipment operation and maintenance management.

[0027] Embodiment 1:

[0028] This application provides a device operation and maintenance management system. Figure 1 A schematic diagram of the structure of an equipment operation and maintenance management system provided in an embodiment of the present application, the system includes:

[0029] The multimodal data acquisition module 11 is used to acquire real-time operation data of the target device; wherein the real-time operation data includes data collected from multiple modes, and the multiple modes include but are not limited to at least two of the following: audio, video, and numerical value;

[0030] A fault warning module 12 is used to obtain the real-time operation data; perform fault analysis on the real-time operation data through a pre-trained fault warning model that supports multimodal data analysis of the target device to obtain a fault diagnosis result; wherein the fault warning model combines the pre-saved historical operation data of the target device to determine the target change data in the real-time operation data, and evaluates the target change data in combination with the pre-configured risk warning conditions; if it is determined that the target change data meets the risk warning conditions, the fault diagnosis result identifies the real-time operation data as abnormal operation data; if it is determined that the target change data does not meet the risk warning conditions, the fault diagnosis result identifies the real-time operation data as normal operation data;

[0031] A fault processing module 13 is used to perform fault processing on the abnormal operating data by using a pre-trained fault processing model that supports multimodal data analysis of the target device when the fault diagnosis result of the real-time operating data identifies the real-time operating data as abnormal operating data, so as to obtain an emergency disposal plan; wherein the fault processing model combines historical operation and maintenance case knowledge in a pre-built historical fault knowledge base to determine a target historical operation and maintenance case that matches the abnormal operating data, and obtains a saved emergency disposal plan from the target historical operation and maintenance case;

[0032] The output module 14 is used to output the emergency response plan.

[0033] In view of the problems that the traditional monitoring methods currently used in the field of equipment operation and maintenance cannot fully capture the status changes of the equipment throughout its life cycle, the early warning information has lags and inaccuracies, lacks an effective plan matching mechanism and rapid response capabilities, and has low operation and maintenance efficiency, this application proposes an innovative operation and maintenance system. The system aims to achieve comprehensive, accurate and efficient management of the target equipment status by integrating multi-modal data collection, intelligent analysis, real-time early warning and automated response functional modules. The following is a detailed introduction to the functions of each module in the equipment operation and maintenance management system (hereinafter referred to as the system) provided by this application:

[0034] 1. Multimodal data acquisition module 11.

[0035] The multimodal data acquisition module 11 is the front-end part of the system, which is responsible for capturing various forms of data from the target device in real time, aiming to achieve comprehensive and uninterrupted monitoring of the target device throughout its entire life cycle from installation to scrapping. These data are not limited to the traditional single form (such as numerical values), but cover a variety of "modalities", that is, different types of information representation methods. Among them, the multiple modalities include but are not limited to at least two of the following: audio, video and numerical values.

[0036] For the collection of numerical values, different types of sensors can be installed at various key locations of the target equipment, such as near heat sources, stress points, mechanical connections, etc. These sensors include but are not limited to temperature sensors, pressure sensors, vibration sensors, etc., which are used to monitor the physical parameters of the equipment in real time.

[0037] For audio collection, audio collection equipment can be installed around the target device or at key locations to capture the sound signals of the target device, thereby providing non-numerical information such as abnormal sounds of the target device.

[0038] For video acquisition, video cameras can be installed around the target device or at key locations to capture visual images of the target device, thereby providing non-numerical information such as changes in the device's appearance.

[0039] It should be noted that the multimodal data acquisition module 11 acquires the operating data of the target device in real time to ensure the timeliness and accuracy of the data.

[0040] Multimodal real-time operation data can complement each other. For example, video images can reveal damage to the appearance of equipment, while audio signals may capture abnormal sounds during mechanical operation. These can help to provide a more comprehensive understanding of the equipment status.

[0041] 2. Fault warning module 12.

[0042] The fault warning module 12 receives the real-time operation data transmitted by the multimodal data acquisition module 11, and uses the pre-trained fault warning model to perform data analysis to predict or identify whether the equipment is about to fail, thereby discovering potential faults in advance and issuing warnings in time so that operation and maintenance personnel can take measures quickly to avoid the expansion of faults. Exemplarily, after the fault warning module 12 obtains the real-time operation data, it can call a pre-trained fault warning model that supports multimodal data analysis of the target device. The real-time operation data is input into the fault warning model. The real-time operation data is subjected to fault analysis by the fault warning model to obtain a fault diagnosis result.

[0043] Among them, when the fault early warning model is used to perform fault analysis on the real-time operation data and obtain the fault diagnosis result, the fault early warning model can combine the pre-saved historical operation data of the target device to identify the data points that have changed significantly compared with the normal operation state of the target device, that is, the target change data. The fault early warning model compares and evaluates the target change data with the pre-configured risk early warning conditions. Among them, the risk early warning conditions can be the historical fault data of the same type of equipment, expert experience or industry standard formulation, which are used to define which data changes may indicate the occurrence of a fault. The fault early warning model gives the fault diagnosis result based on the result of the risk assessment. If the target change data meets the risk early warning condition, the fault diagnosis result is abnormal operating data; if the target change data meets the risk early warning condition, the fault diagnosis is normal operating data.

[0044] In a possible implementation, the pre-configured risk warning conditions may include one or more of the following types:

[0045] Type 1: determining whether the value of the target change data is within a pre-configured value range corresponding to the target change data.

[0046] This condition involves verifying the specific value of the target change data. The system can set one or more value ranges for different types of data, which are usually determined based on the normal operating parameters of the equipment, historical fault data, industry standards or expert experience. When the value of the determined target change data exceeds these preset ranges, the system will consider that these data may represent an abnormal equipment state, and thus trigger further fault analysis or early warning mechanisms.

[0047] For example, if the target change data is the temperature value of the target device, and the temperature should be between 20°C and 60°C during normal operation, then any temperature value outside this range will be considered to meet the risk warning condition.

[0048] Type 2: Determine whether the target change data meets any pre-configured abnormal pattern.

[0049] In addition to direct value range verification, the system may also predefine a series of abnormal patterns based on historical data and expert knowledge. These abnormal patterns may include specific data fluctuation patterns, interruptions in periodic changes, sudden abnormal peaks or valleys, etc. When the determined target change data meets any of these abnormal patterns, the system will consider that the target device may have some form of failure or abnormal condition, and issue an early warning accordingly.

[0050] For example, if the target change data is a vibration audio signal of the target device, and a certain type of mechanical failure is known to cause a specific frequency component to appear in the vibration audio signal, then when this frequency component appears in the vibration audio signal, it is considered to meet the abnormal pattern condition.

[0051] Type 3: Determine whether the change rate of the target change data exceeds the pre-configured rate threshold:

[0052] The change rate of certain parameters of the target device (such as temperature, pressure, speed, etc.) usually remains within a relatively stable range during normal operation. When the change rate of these parameters suddenly increases or decreases and exceeds the system preset rate threshold, it may mean that the target device is experiencing some form of drastic change or abnormal process. In this case, the system will also regard the target change data as a potential risk signal and trigger the subsequent fault warning mechanism.

[0053] For example, if the target change data is the pressure value of the target device, and under normal circumstances the pressure change should be relatively gentle, but suddenly there is a sharp rise or fall, and the rate of change exceeds a preset threshold, then it can be considered that it meets the rate change condition.

[0054] It should be noted that the pre-configured risk warning condition may include only one of the above types, or may include multiple types, for example, type 1 and type 2, type 2 and type 3, type 1 and type 3. When multiple types are included, it is necessary to determine that the risk warning condition of each type is not met before determining that the target change data does not meet the risk warning condition. If it is determined that any type of risk warning condition is met, it is determined that the target change data meets the risk warning condition.

[0055] In a possible implementation, the fault warning module 12 can pre-process the real-time operation data before performing fault analysis on the real-time operation data through the fault warning model to obtain the fault diagnosis result, so as to improve the data quality and make the data more suitable for subsequent analysis and model processing, thereby improving the accuracy and efficiency of fault warning. Among them, the pre-processing includes but is not limited to one or more of the following: normalization, denoising, and structured processing. During the pre-processing process, one or more methods can be selected according to the characteristics of the data and the requirements of the model. For example, for numerical values ​​containing a lot of noise, denoising can be performed; for video and audio data, feature extraction and structured processing can be performed; for numerical data, normalization can be performed to ensure data consistency.

[0056] In one example, a basic fault warning model is trained based on sample operation data in a fault warning sample set related to the target device and its corresponding reference fault diagnosis results to obtain a trained fault warning model that supports multimodal data analysis of the target device; wherein the sample operation data is based on data collected from the historical fault knowledge base, and the sample operation data includes data collected from multiple modalities on the target device.

[0057] In order to obtain a fault warning model that supports multimodal data analysis of the target device, in this application, it is necessary to clarify the device type of the target device to be monitored and warned, such as a key machine on an industrial production line, a server cluster in a data center, or a transformer in a smart grid, transportation equipment, etc. Then, for the device type, a historical fault knowledge base of the device type is constructed. The historical fault knowledge base includes a large number of maintenance records, fault reports, historical operation data, historical fault cases, emergency disposal plans, etc. From the pre-constructed historical fault knowledge base, the historical operation data is determined as sample operation data, that is, a fault warning sample set related to the target device is collected, so as to train the basic fault warning model through the sample operation data in the fault warning sample set. Among them, the sample operation data covers multiple modes, that is, data obtained from different angles, different sensors or different data sources, such as temperature, vibration, current, voltage, sound, etc. These multimodal data can more comprehensively reflect the operating status of the equipment.

[0058] For each collected sample operation data, its corresponding reference fault diagnosis result needs to be clarified. For example, the reference fault diagnosis result of the sample operation data is manually labeled to indicate whether the sample operation data is normal operation data or abnormal operation data.

[0059] In one possible implementation, since data of different modalities may differ in dimension, sampling rate, noise level, etc., the sample operation data may be preprocessed, including cleaning (removing outliers, missing values), normalization (making data of different modalities at the same level), feature extraction, and other steps.

[0060] For any sample operation data in the fault warning sample set related to the target device, the sample operation data is input into the basic fault warning model. The sample operation data is subjected to fault analysis by the basic fault warning model to determine the expected fault diagnosis result. Based on the expected fault diagnosis result and the reference fault diagnosis result, a loss value is determined. Based on the loss value, the parameters in the basic fault warning model are adjusted to obtain a trained fault warning model that supports multimodal data analysis of the target device.

[0061] Since the fault warning sample set contains a large amount of sample operation data, the above operation can be performed for each sample operation data. When the preset convergence condition is met, the training of the fault warning model is completed.

[0062] In a possible implementation, when there are enough sample operation data in the fault warning sample set, the basic warning model can be an untrained model. When the number of sample operation data in the fault warning sample set cannot meet the training requirements, the basic warning model can be a model that has been trained with a large number of natural language samples and has natural language understanding capabilities, and the basic warning model is fine-tuned through the sample operation data in the fault warning sample set.

[0063] Among them, satisfying the preset convergence condition may be that the number of iterations for training the basic fault warning model reaches the set maximum number of iterations, or the change in the determined loss value remains within a very small range in multiple consecutive iterations, etc. In specific implementation, the setting can be flexibly performed and is not specifically limited here.

[0064] As a possible implementation method, when performing model training, the sample operation data in the fault warning sample set can be divided into training samples and test samples. The basic fault warning model is first trained based on the training samples, and then the reliability of the trained fault warning model is verified based on the test samples.

[0065] 3. Fault handling module 13.

[0066] The main function of the fault processing module 13 is to automatically start and apply a series of predefined strategies and algorithms to analyze and process these abnormal data when the system detects that the real-time operation data is abnormal, that is, when the real-time operation data indicates that the target device is in an abnormal operating state, so as to generate an effective emergency disposal plan to guide the operation and maintenance personnel to solve the problem quickly and accurately. Exemplarily, when the fault warning module 12 determines that the fault diagnosis result of the real-time operation data identifies the real-time operation data as abnormal operating data, it will send the real-time operation data to the fault processing module 13. After the fault processing module 13 obtains the abnormal operating data, it can call a pre-trained fault processing model that supports multimodal data analysis of the target device. The abnormal operating data is input into the fault processing model. The input abnormal operating data can be fault processed by the fault processing model to obtain an emergency disposal plan.

[0067] Specifically, one of the core capabilities of the fault handling model is its integration with the historical fault knowledge base. The historical fault knowledge base is a database containing information such as various fault cases, fault causes, impact scope, and emergency response plans that have occurred in the past. When the fault handling model receives abnormal operating data, it will combine these historical case knowledge and use the configured matching algorithm (such as similarity calculation, cluster analysis, etc.) to find the historical operation and maintenance case that is most similar to the abnormal operating data. Once the target historical operation and maintenance case that matches the target abnormal operating data is found, the fault handling model will extract the saved emergency response plan from the target historical operation and maintenance case. Among them, the emergency response plan can include a series of specific operating steps, required tools or resources, expected effects and other information, which directly guide the operation and maintenance personnel on how to quickly respond to and solve problems. The operation and maintenance personnel can take quick action based on the guidance of these cases to restore the normal operating state of the equipment and reduce the impact of the fault on the overall performance of the system.

[0068] In a possible implementation, a basic fault handling model is trained based on sample abnormal operating data and its corresponding reference emergency handling solutions in a fault handling sample set related to the target device to obtain a trained fault handling model that supports multimodal data analysis of the target device; wherein the sample abnormal operating data and its corresponding reference emergency handling solutions are collected based on the historical fault knowledge base, and the sample abnormal operating data include data collected from multiple modalities on the target device.

[0069] It should be noted that the process of training the basic fault handling model is similar to the process of training the basic fault prediction model mentioned above. The specific process can refer to the training process of the basic fault prediction model mentioned above, and will not be described in detail here.

[0070] 4. Output module 14.

[0071] The output module 14 is responsible for presenting the analyzed, matched and determined emergency response plan to the operation and maintenance personnel or related systems in a clear and easy-to-understand manner. The emergency response plan can be presented in a variety of forms, including text description, chart, flow chart, etc., depending on the complexity of the emergency response plan and the needs of the recipient. For example, for a simple fault, only a short text description may be required; for a complex fault, a detailed description may be required in combination with a flow chart, a list of steps and key considerations.

[0072] In a possible implementation, in order to ensure the readability and operability of the emergency response plan, the output module 14 also needs to format and layout the emergency response plan. For example, select appropriate fonts, font sizes, colors, spacing, etc., and reasonably arrange the layout and structure of the content. Through good formatting and layout, the emergency response plan can be made more intuitive and clear, which is convenient for operation and maintenance personnel to quickly understand and execute.

[0073] In a possible implementation, the output module 14 supports multiple output channels to meet the needs of different scenarios. For example, it can directly display the emergency response plan on the monitoring screen for on-site operation and maintenance personnel to view; it can also print the plan into a paper document for off-site personnel to refer to; it can also send the plan to relevant personnel via email, text messages, etc. This multi-channel output capability improves the flexibility and accessibility of the emergency response plan.

[0074] In some cases, the output module 14 can also support customized output to personalize the emergency response plan according to the needs of different users or different devices. For example, plans with different levels of detail can be provided for operation and maintenance personnel with different skill levels. Through customized output, it can ensure that the emergency response plan is closer to actual needs and improve its practicality and effectiveness.

[0075] In a possible implementation, the output module 14 also has feedback and recording functions. After the operation and maintenance personnel have executed the emergency response plan, they can provide a feedback channel through the output module 14 to feed back the execution results, problems encountered and other information to the system. At the same time, the output module 14 should also record the output of each emergency response plan, including key information such as output time, receiving personnel, and execution results, so as to facilitate subsequent analysis and improvement.

[0076] In one example, the output module 14 is also used to send a notification message indicating that the target device has a fault, so as to promptly convey the fault information to relevant personnel so that they can take corresponding measures, for example, through various means such as email, text messages, instant messaging tools, and monitoring system interfaces, so as to select a suitable communication channel according to the recipient's preferences and actual conditions.

[0077] In some cases, the output module 14 can also set different priorities for notification messages according to the urgency and impact scope of the target historical operation and maintenance cases, so as to help the recipient decide the order of response according to the priority and ensure that critical faults are handled in a timely manner.

[0078] It should be noted that the output module 14 can be implemented in a variety of ways, depending on the system architecture and technology selection, for example, a graphical user interface (GUI), such as using a web page, desktop application, etc.; a command line interface (CLI); file output, such as PDF, Word document, TXT text, etc.; message push, such as email, SMS, instant messaging tools, etc. Early warning mechanism: Once the evaluation results show that the equipment has a potential failure risk, the system will automatically trigger the early warning mechanism. Early warning methods may include SMS notifications, email alerts, system pop-ups, etc., to ensure that the operation and maintenance personnel can receive early warning information in a timely manner.

[0079] 5. Storage module.

[0080] The storage module is responsible for comprehensively and systematically saving the full-link data of the target device, which covers all information related to the device throughout its entire life cycle, starting from the installation of the target device until the device is scrapped. The types of this full-link data include, but are not limited to: real-time operation data of the target device, fault diagnosis results of the target device, emergency response plans corresponding to abnormal operation data of the target device, maintenance records of the target device, fault information of the target device, and scrap information of the target device.

[0081] This storage module can be used to establish a data traceability mechanism, allowing operation and maintenance personnel to view the historical data of the equipment at any time, providing data support for fault analysis and optimization.

[0082] The beneficial effects of this application are as follows:

[0083] 1. Through the multimodal data acquisition module 11, the system can simultaneously obtain the real-time operation data of the target device from multiple dimensions such as audio, video, and numerical values, so as to more comprehensively reflect the operation status of the target device throughout its life cycle and improve the accuracy and timeliness of fault detection.

[0084] 2. The fault warning module 12 uses a pre-trained fault warning model that supports multimodal data analysis, combined with the historical operation data of the target equipment and the pre-configured risk warning conditions, to conduct an in-depth analysis of the real-time operation data. It can accurately determine whether the real-time operation data is normal or abnormal operation data, and give corresponding fault diagnosis results, which helps to discover potential faults in advance, facilitates operation and maintenance personnel to quickly locate problems, improves the efficiency of fault handling, and improves the stability and reliability of the system.

[0085] 3. When the fault diagnosis result is abnormal operating data, the fault handling module 13 will use the pre-trained fault handling model and the historical operation and maintenance case knowledge in the historical fault knowledge base to automatically match and obtain the corresponding emergency response plan, which not only reduces the burden on operation and maintenance personnel, but also improves the accuracy and efficiency of fault handling.

[0086] Embodiment 2:

[0087] Figure 2 A schematic diagram of the workflow of an equipment operation and maintenance management system provided for this application, the process includes:

[0088] S201: The equipment operation and maintenance management system obtains a pre-built historical fault knowledge base related to the target equipment.

[0089] The historical fault knowledge base stores historical fault cases of devices of the same device type as the target device, and emergency response plans are stored in the historical fault cases.

[0090] S202: The equipment operation and maintenance management system obtains a pre-trained fault prediction model and a fault processing model that support multi-modal data analysis of the target equipment.

[0091] In one possible implementation, a basic fault warning model is trained based on sample operation data in a fault warning sample set related to a target device and its corresponding reference fault diagnosis results to obtain a trained fault warning model that supports multimodal data analysis of the target device; wherein the sample operation data is collected based on a historical fault knowledge base, and the sample operation data includes data collected from multiple modalities on the target device.

[0092] In one possible implementation, a basic fault handling model is trained based on sample abnormal operating data and its corresponding reference emergency handling solutions in a fault handling sample set related to a target device to obtain a trained fault handling model that supports multimodal data analysis of the target device; wherein the sample abnormal operating data and its corresponding reference emergency handling solutions are collected based on a historical fault knowledge base, and the sample abnormal operating data include data collected from multiple modalities on the target device.

[0093] S203: The multimodal data acquisition module of the equipment operation and maintenance management system obtains real-time operation data of the target equipment.

[0094] The real-time operation data includes data collected from multiple modalities, and the multiple modalities include but are not limited to at least two of the following: audio, video, and numerical values.

[0095] S204: The multimodal data acquisition module transmits the real-time operation data to the fault warning module of the equipment operation and maintenance management system.

[0096] S205: The fault warning module pre-processes the received real-time operation data.

[0097] The preprocessing includes but is not limited to normalization, denoising, and structural processing.

[0098] S206: The fault warning module performs fault analysis on the real-time operation data through a pre-trained fault warning model that supports multi-modal data analysis of the target device to obtain a fault diagnosis result.

[0099] Among them, the fault warning model combines the pre-saved historical operation data of the target equipment, determines the target change data in the real-time operation data, and evaluates the target change data in combination with the pre-configured risk warning conditions; if it is determined that the target change data meets the risk warning conditions, the fault diagnosis result identifies the real-time operation data as abnormal operation data; if it is determined that the target change data does not meet the risk warning conditions, the fault diagnosis result identifies the real-time operation data as normal operation data.

[0100] S207: When the fault diagnosis result of the real-time operation data indicates that the real-time operation data is abnormal operation data, the fault prediction module notifies the output module of the equipment operation and maintenance management system of the fault diagnosis result, and executes S212.

[0101] S208: When the fault diagnosis result of the real-time operation data identifies the real-time operation data as abnormal operation data, the fault prediction module transmits the abnormal operation data to the fault processing module of the equipment operation and maintenance management system.

[0102] S209: The fault processing module performs fault processing on the abnormal operation data through a pre-trained fault processing model that supports multi-modal data analysis of the target device to obtain an emergency handling plan.

[0103] Among them, the fault handling model combines the historical operation and maintenance case knowledge in the pre-built historical fault knowledge base, determines the target historical operation and maintenance case that matches the abnormal operation data, and obtains the saved emergency response plan from the target historical operation and maintenance case.

[0104] S210: The fault processing module notifies the output module of the emergency handling plan.

[0105] S211: The output module outputs the emergency response plan and executes S203.

[0106] S212: The output module outputs a notification message indicating that a fault exists in the target device.

[0107] In a possible implementation, the equipment operation and maintenance management system also includes a storage module to comprehensively and systematically save the full-link data of the target equipment through the storage module. The full-link data starts from the installation of the target equipment and continues until the equipment is scrapped, covering all information related to the equipment throughout its life cycle. Among them, the types of the full-link data include but are not limited to: real-time operation data of the target equipment, fault diagnosis results of the target equipment, emergency response plans corresponding to abnormal operation data of the target equipment, maintenance records of the target equipment, fault information of the target equipment, and scrapping information of the target equipment.

[0108] Embodiment 3:

[0109] Based on the same inventive concept, the present application also provides a device operation and maintenance management method. Figure 3 A process diagram of a device operation and maintenance management method provided in an embodiment of the present application, the process includes:

[0110] S301: Acquire real-time operation data of a target device; wherein the real-time operation data includes data collected from multiple modalities, and the multiple modalities include but are not limited to at least two of the following: audio, video, and numerical value.

[0111] S302: Perform fault analysis on the real-time operation data through a pre-trained fault warning model that supports multimodal data analysis of the target device to obtain a fault diagnosis result; wherein the fault warning model combines the pre-saved historical operation data of the target device to determine the target change data in the real-time operation data, and evaluates the target change data in combination with the pre-configured risk warning conditions; if it is determined that the target change data meets the risk warning conditions, the fault diagnosis result identifies the real-time operation data as abnormal operation data; if it is determined that the target change data does not meet the risk warning conditions, the fault diagnosis result identifies the real-time operation data as normal operation data.

[0112] S303: When the fault diagnosis result of the real-time operation data identifies the real-time operation data as abnormal operation data, the abnormal operation data is fault processed by a pre-trained fault processing model that supports multimodal data analysis of the target device to obtain an emergency response plan; wherein the fault processing model combines historical operation and maintenance case knowledge in a pre-built historical fault knowledge base to determine a target historical operation and maintenance case that matches the abnormal operation data, and obtains a saved emergency response plan from the target historical operation and maintenance case.

[0113] S304: Output the emergency response plan.

[0114] The device operation and maintenance management method provided in the present application is applied to computer equipment, which can be an intelligent terminal, such as a computer, a robot, etc., or a server, such as an application server, a business server, etc.

[0115] In some possible implementations, the method further comprises:

[0116] The full-link data of the target device is stored; wherein the full-link data includes all data associated with the target device during the period from the installation of the target device to the scrapping of the target device.

[0117] In some possible implementations, before performing fault analysis on the real-time operation data using a pre-trained fault warning model that supports multimodal data analysis of the target device to obtain a fault diagnosis result, the method further includes:

[0118] The real-time operation data is preprocessed; wherein the preprocessing includes but is not limited to one or more of the following: normalization, denoising, and structured processing.

[0119] In some possible implementations, determining that the target change data meets the risk warning condition includes one or more of:

[0120] Determining that the value of the target change data is not within a pre-configured value range corresponding to the target change data;

[0121] Determining that the target change data conforms to any pre-configured abnormal pattern;

[0122] It is determined that a change rate of the target change data exceeds a preconfigured rate threshold.

[0123] In some possible implementations, when the fault diagnosis result of the real-time operation data identifies that the real-time operation data is abnormal operation data, the method further includes:

[0124] Output a notification message for prompting that a fault exists in the target device.

[0125] In some possible implementations, a basic fault warning model is trained based on sample operation data in a fault warning sample set related to the target device and its corresponding reference fault diagnosis results to obtain a trained fault warning model that supports multimodal data analysis of the target device; wherein the sample operation data is based on data collected from the historical fault knowledge base, and the sample operation data includes data collected from multiple modalities on the target device.

[0126] In certain possible implementations, a basic fault handling model is trained based on sample abnormal operating data and its corresponding reference emergency handling solutions in a fault handling sample set related to the target device to obtain a trained fault handling model that supports multimodal data analysis of the target device; wherein the sample abnormal operating data and its corresponding reference emergency handling solutions are collected based on the historical fault knowledge base, and the sample abnormal operating data include data collected from multiple modalities on the target device.

[0127] Since the principle of solving the problem by the above method is similar to that of the equipment operation and maintenance management system, the implementation of the above method can refer to the embodiment of the system, and the repeated parts will not be repeated.

[0128] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. An equipment operation and maintenance management system, characterized in that: The system comprises: A multimodal data acquisition module, used to acquire real-time operation data of a target device; wherein the real-time operation data includes data collected from multiple modes, and the multiple modes include but are not limited to at least two of the following: audio, video, and numerical value; A fault warning module is used to obtain the real-time operation data; perform fault analysis on the real-time operation data through a pre-trained fault warning model that supports multimodal data analysis of the target device to obtain a fault diagnosis result; wherein the fault warning model combines the pre-saved historical operation data of the target device to determine the target change data in the real-time operation data, and evaluates the target change data in combination with the pre-configured risk warning conditions; if it is determined that the target change data meets the risk warning conditions, the fault diagnosis result identifies the real-time operation data as abnormal operation data; if it is determined that the target change data does not meet the risk warning conditions, the fault diagnosis result identifies the real-time operation data as normal operation data; A fault processing module is used to perform fault processing on the abnormal operating data by using a pre-trained fault processing model that supports multimodal data analysis of the target device when the fault diagnosis result of the real-time operating data identifies the real-time operating data as abnormal operating data, so as to obtain an emergency disposal plan; wherein the fault processing model combines historical operation and maintenance case knowledge in a pre-built historical fault knowledge base to determine a target historical operation and maintenance case that matches the abnormal operating data, and obtains a saved emergency disposal plan from the target historical operation and maintenance case; An output module is used to output the emergency response plan.

2. The system according to claim 1, characterized in that The system further comprises: a storage module; The storage module is used to store the full-link data of the target device; wherein the full-link data includes all data associated with the target device during the period from the installation of the target device to the scrapping of the target device.

3. The system according to claim 1, characterized in that The fault warning module is also used to perform fault analysis on the real-time operation data through a pre-trained fault warning model that supports multimodal data analysis of the target device, so as to pre-process the real-time operation data before obtaining a fault diagnosis result; wherein the preprocessing includes but is not limited to one or more of the following: normalization, denoising, and structured processing.

4. The system according to claim 1, characterized in that Determining that the target change data meets the risk warning condition includes one or more of: Determining that the value of the target change data is not within a pre-configured value range corresponding to the target change data; Determining that the target change data conforms to any pre-configured abnormal pattern; It is determined that a change rate of the target change data exceeds a preconfigured rate threshold.

5. The system according to claim 1, wherein: The output module is further configured to output a notification message for prompting that a fault exists in the target device when the fault diagnosis result of the real-time operation data identifies the real-time operation data as abnormal operation data.

6. The system according to claim 1, wherein: Based on the sample operation data in the fault warning sample set related to the target device and its corresponding reference fault diagnosis results, the basic fault warning model is trained to obtain a trained fault warning model that supports multimodal data analysis of the target device; wherein the sample operation data is based on data collected from the historical fault knowledge base, and the sample operation data includes data collected from multiple modalities on the target device.

7. The system according to claim 1, characterized in that Based on the sample abnormal operating data and the corresponding reference emergency handling solutions in the fault handling sample set related to the target device, the basic fault handling model is trained to obtain a trained fault handling model that supports multimodal data analysis of the target device; wherein the sample abnormal operating data and the corresponding reference emergency handling solutions are collected based on the historical fault knowledge base, and the sample abnormal operating data include data collected from multiple modalities on the target device.

8. A method for equipment operation and maintenance management, characterized in that: The method comprises: Acquire real-time operation data of the target device; wherein the real-time operation data includes data collected from multiple modalities, and the multiple modalities include but are not limited to at least two of the following: audio, video, and numerical value; By means of a pre-trained fault warning model supporting multi-modal data analysis of the target device, a fault analysis is performed on the real-time operation data to obtain a fault diagnosis result; wherein the fault warning model combines the pre-saved historical operation data of the target device to determine the target change data in the real-time operation data, and evaluates the target change data in combination with the pre-configured risk warning conditions; if it is determined that the target change data meets the risk warning conditions, the fault diagnosis result identifies the real-time operation data as abnormal operation data; if it is determined that the target change data does not meet the risk warning conditions, the fault diagnosis result identifies the real-time operation data as normal operation data; In the case where the fault diagnosis result of the real-time operation data identifies that the real-time operation data is abnormal operation data, the abnormal operation data is subjected to fault processing by means of a pre-trained fault processing model that supports multimodal data analysis of the target device, so as to obtain an emergency disposal plan; wherein the fault processing model combines historical operation and maintenance case knowledge in a pre-built historical fault knowledge base to determine a target historical operation and maintenance case that matches the abnormal operation data, and obtains a saved emergency disposal plan from the target historical operation and maintenance case; Output the emergency response plan.

9. The method according to claim 8, characterized in that The method further comprises: The full-link data of the target device is stored; wherein the full-link data includes all data associated with the target device during the period from the installation of the target device to the scrapping of the target device.

10. The method according to claim 8, characterized in that In the case where the fault diagnosis result of the real-time operation data identifies that the real-time operation data is abnormal operation data, the method further includes: Output a notification message for prompting that a fault exists in the target device.

Images