A method, device, equipment, and system for intelligent elevator maintenance based on artificial intelligence.

Abstract

This invention discloses an intelligent elevator maintenance method, device, equipment, and system based on artificial intelligence. Information is collected by the elevator DTU terminal, uploaded to the intelligent maintenance system, and pushed to a large elevator model. The diagnostic model uses the data to inspect and locate elevators and projects requiring maintenance. The task model determines personnel, generates tasks, and instructs the automatic programming model to generate plugins. Maintenance personnel complete the maintenance according to prompts on a mobile client and upload the data. The effectiveness evaluation model evaluates and scores the performance, generates a case study, and stores it in the maintenance case study model. If the performance is satisfactory, the elevator's health status is reset. This system utilizes self-developed key technologies, such as automatic programming and large elevator model construction, to achieve intelligent maintenance, knowledge sharing, improved maintenance quality, and standardized practices. It effectively solves many drawbacks of traditional elevator maintenance methods, improves elevator maintenance efficiency, quality, and safety, ensures reliable elevator operation, and promotes the intelligent development of the elevator maintenance industry.

Description

Technical Field

[0001] This invention relates to the field of elevator maintenance technology, and in particular to a method, apparatus, equipment and system for intelligent elevator maintenance based on artificial intelligence. Background Technology

[0002] With the rapid advancement of urbanization, high-rise buildings are springing up like mushrooms after rain, and elevators have become an indispensable key vertical transportation tool in modern urban life. While providing convenient travel for people, their safety, reliability, and comfort have also become the focus of attention from all sectors of society.

[0003] Traditional elevator maintenance primarily relies on scheduled maintenance. This model has several insurmountable problems. First, scheduled maintenance often fails to consider the actual operating condition and needs of the elevator, easily leading to a significant waste of resources. For example, some elevators in good operating condition and with low usage frequency may not require comprehensive maintenance according to a fixed cycle, yet the same maintenance procedures still need to be performed, which undoubtedly consumes a large amount of human, material, and time resources.

[0004] Secondly, inefficiency is another significant drawback of the traditional maintenance model. Due to a lack of accurate fault prediction and location capabilities, maintenance personnel may spend a considerable amount of time meticulously inspecting each elevator component, even if the elevator has no substantial malfunction or only minor potential issues. This not only prolongs maintenance time but may also disrupt normal use of the building due to prolonged downtime, causing considerable inconvenience to users.

[0005] Furthermore, the lack of standardized maintenance practices is a common problem in traditional models. Different maintenance companies and personnel, due to differences in their experience, technical skills, and internal standards, exhibit varying operational procedures and quality control when performing elevator maintenance tasks. For example, regarding the inspection and maintenance of certain critical components, some maintenance personnel may only perform simple visual inspections, while others may conduct more in-depth and meticulous performance testing and adjustments. This difference may lead to varying degrees of risk for the elevator during subsequent operation.

[0006] Furthermore, the traditional maintenance process suffers from a severe lack of oversight mechanisms. During maintenance, it is difficult to monitor in real time whether maintenance personnel are operating in a standardized manner and whether they are following established standards and procedures. This makes it possible for some maintenance work to be perfunctory or involve shoddy workmanship, without the relevant management departments or elevator users being able to detect and stop it in a timely manner.

[0007] Evaluating the effectiveness of maintenance is also a major challenge in the traditional model. Due to a lack of scientific and systematic evaluation methods, it is difficult to accurately determine whether a maintenance operation has truly and effectively improved elevator performance or reduced the probability of malfunctions. Often, only a rough judgment can be made based on whether the elevator malfunctions again in the short term, rather than conducting an in-depth analysis of the impact of various data and operations during the maintenance process on the long-term operational stability of the elevator.

[0008] Finally, with the rapid increase in the number of elevators, the shortage of maintenance personnel has become increasingly prominent. Traditional maintenance models, due to their low efficiency and cumbersome processes, require a large number of maintenance personnel to meet the growing demand for elevator maintenance. However, in reality, the training rate for maintenance personnel lags far behind the growth rate of the number of elevators, further exacerbating the pressure and difficulty of elevator maintenance work, making it difficult to effectively guarantee the timeliness and quality of elevator maintenance.

[0009] In summary, traditional elevator maintenance models have numerous limitations in resource utilization, efficiency improvement, standardization, process supervision, effect evaluation, and personnel allocation, making them ill-suited to the rapidly evolving needs of the modern elevator industry. This invention addresses this critical situation by actively exploring the application of artificial intelligence technology in the elevator industry. Its aim is to effectively solve the many challenges faced by intelligent elevator maintenance, providing a more reliable and efficient guarantee for the safe and stable operation of elevators. Summary of the Invention

[0010] The purpose of this invention is to provide a method, device, equipment, and system for intelligent elevator maintenance based on artificial intelligence. It innovatively leverages the powerful capabilities of large-scale AI models to automatically program and generate maintenance plugins, a feature that endows maintenance work with a high degree of intelligence and autonomy. Even offline, the maintenance plugins can still play a crucial role, effectively overcoming the dependence of traditional maintenance on the network environment and greatly expanding the implementation scenarios and flexibility of maintenance work.

[0011] To achieve the above objectives, the present invention provides a method, apparatus, equipment, and system for intelligent elevator maintenance based on artificial intelligence:

[0012] System Architecture:

[0013] The system of this invention mainly includes a mobile phone A1, a smart maintenance client C1 installed on the mobile phone with a plugin support, a cloud-based smart maintenance system Y1 (including a maintenance server S1, a communication server Z1, and a database B1), an elevator large model X1 (covering a maintenance case model M1, a status and fault model M2, a diagnostic model M3, a task model M4, an automatic programming model M5, and an effect evaluation model M6), an elevator E1, a DTU terminal T1 connected to the elevator to upload faults, and maintenance personnel P1.

[0014] Key technologies:

[0015] Automated Programming Technology: This invention's automated programming method inputs a large number of code libraries and programming languages ​​corresponding to maintenance projects into an automated programming model for learning and training. It utilizes machine learning algorithms to understand programming rules and syntax, identify code patterns and structures, and associate maintenance projects with the code. After maintenance rules are entered, the model can automatically generate code (plugins) that meets the requirements by converting user instructions into programming language through natural language processing. This technology is the first of its kind in the industry.

[0016] Elevator large-scale model construction technology: By deploying a general large-scale model locally on a private cloud, inputting massive amounts of elevator industry data, and going through steps such as data cleaning, word segmentation and tokenization, data augmentation and standardization, sampling and weight adjustment, and combining natural language processing, multiple sub-models are constructed according to their uses, including maintenance case model M1, status and fault model M2, diagnostic model M3, task model M4, automatic programming model M5, and effect evaluation model M6.

[0017] The intelligent maintenance system is built using the following technical aspects: It handles specific maintenance tasks, employs a mainstream Web API application architecture, uses JSON format for data exchange, utilizes a relational database, and implements client-server communication via HTTPS. The development process includes surveying customer needs, determining technical solutions, programming development, testing and verification, and system deployment.

[0018] Maintenance project and code association technology: This technology utilizes an AI big data model to extract header comments from program source code and transform them into big data tags for automatic association. Programmers add comments when implementing maintenance project functions, and after coding and testing, upload the code to the AI ​​big data model. The model automatically extracts the comments and tags them, thus achieving the association. This is an innovative technology.

[0019] Maintenance rule creation technology: Based on big data tags and the latest national standards for maintenance projects, the standard text is digitized using OCR technology, and maintenance rules are automatically created using keyword matching (intelligent word segmentation) technology. Maintenance projects that do not match national standards are included in the enterprise standard. When updating, only the latest national standard needs to be uploaded, which improves the level of automation and intelligence and is a self-developed innovative technology.

[0020] Elevator intelligent maintenance methods and steps:

[0021] Elevator status and fault information upload: Elevator E1 collects status and fault information U1 in real time through DTU terminal T1 and uploads the elevator status and fault information U2 to the communication server Z1 of intelligent maintenance system X1. After receiving the information, the communication server Z1 writes the fault information into database B1 for long-term storage, and pushes the status and fault information U3 to the status and fault model M2 of elevator large model X1 for long-term storage.

[0022] Elevator maintenance needs diagnosis: The elevator large-scale model's diagnostic model M3 is inspected regularly (daily). Based on diagnostic rules (e.g., elevators with a health level <30 require maintenance, with priority given to elevators with low health levels), the elevators requiring maintenance are precisely located, and maintenance items are determined. For example, if elevator A experiences a passenger entrapment fault, a Class A fault, or a car door opening and closing fault, its health level is calculated to be 10, requiring maintenance, and the maintenance item is normal car door opening and closing. If elevator B experiences an emergency electric operation fault and its wire rope has been used less than 100,000 times, its health level is 20, requiring maintenance, and the maintenance item is replacing the wire rope. Priority is determined, such as prioritizing the maintenance of elevator A.

[0023] Maintenance Task Generation: The elevator large model's task model M4 determines maintenance personnel based on the elevators diagnosed by the diagnostic model M3 and the duty roster, and generates maintenance tasks. For example, if elevators A and B are the responsibility of maintenance personnel P1, two maintenance tasks are generated, and corresponding instructions are sent to the automatic programming M5. For instance, a plugin is generated for elevator A (containing standard maintenance procedures + checking that the car door opens and closes normally, including uploading video recordings and related interface information) and a plugin is generated for elevator B (containing standard maintenance procedures + replacing the wire rope, including uploading video recordings and related interface information).

[0024] Automatic plugin generation and push: After receiving instructions, the automatic programming model M5 of the large elevator model automatically analyzes and programs plugins that can be called by the client. After the plugins are generated, the task model M4 is notified, and the task model M4 automatically pushes the task information and plugins to the maintenance personnel P1.

[0025] On-site maintenance operation: After receiving the task, maintenance personnel P1 carries mobile phone A1, which contains the intelligent maintenance client and task information, to the elevator site. They open the client, load the plugin interface (Load), call the Run interface, and interactively complete the maintenance according to the plugin prompts, recording the process. If elevator A completes the standard maintenance procedure and checks that the car door is operating normally, elevator B will complete the standard maintenance procedure and replace the wire rope.

[0026] Maintenance Effectiveness Evaluation: After maintenance personnel P1 complete the maintenance, the plugin interface Upload is automatically called to upload the data and video recordings of the maintenance process to the elevator effectiveness evaluation model M6. The effectiveness evaluation model M6 analyzes the data and video to determine the effectiveness of the maintenance and gives a score.

[0027] Case generation and knowledge sharing: After the effectiveness evaluation model M6 completes the evaluation and scoring, it automatically generates a case and pushes it to the elevator maintenance case model M1 to provide a reference for subsequent maintenance and achieve knowledge sharing.

[0028] Elevator Health Reset: After maintenance is completed, the elevator's health is reset to 100, and the elevator will resume normal operation.

[0029] Furthermore, an AI-based intelligent elevator maintenance method includes the following steps:

[0030] Step S1: The elevator collects and uploads status and fault information to the intelligent maintenance system in real time through the DTU terminal. The intelligent maintenance system then pushes this information to the status and fault model of the elevator's large model.

[0031] Step S2: The diagnostic model of the large elevator model is inspected according to the preset cycle. Based on the elevator operation data, historical fault records and diagnostic rules, the elevators that need maintenance and their maintenance items are located, and the maintenance priority is determined.

[0032] Step S3: The elevator large model determines the maintenance personnel based on the diagnosis results and duty roster, generates maintenance tasks, and sends instructions to the automatic programming model.

[0033] Step S4: The automatic programming model generates client-callable plugins based on instructions and its own learning and training results, and notifies the task model to push task information and plugins to maintenance personnel.

[0034] Step S5: Maintenance personnel use the mobile smart maintenance client to load the plugin and complete the maintenance according to the prompts. At the same time, they record a video of the maintenance process and upload the data and video to the effect evaluation model after completion.

[0035] Step S6: The effect evaluation model uses data analysis and image recognition technology to process the uploaded information, obtain the maintenance effect and score it, and then generate a case and push it to the maintenance case model;

[0036] Step S7: If the maintenance effect assessment is qualified, the system will automatically reset the elevator health level to 100, so that the elevator can resume normal operation.

[0037] As a further improvement to the technical solution of the present invention, the construction of the large elevator model includes the local deployment of a general large model, inputting elevator industry data, and constructing multiple sub-models through data cleaning, word segmentation and tokenization, data augmentation and standardization, sampling and weight adjustment, and natural language processing. The sub-models include at least a maintenance case model, a status and fault model, a diagnostic model, a task model, an automatic programming model, and an effect evaluation model.

[0038] As a further improvement to the technical solution of the present invention, the automatic programming model learns a large number of code libraries and programming languages ​​corresponding to maintenance projects, and automatically generates code plugins based on user instructions and preset maintenance rules. The generation process involves natural language processing to convert instructions into programming languages.

[0039] As a further improvement to the technical solution of the present invention, the maintenance projects and code are automatically associated by extracting header comments from the program source code using an AI big data model and converting them into big data tags. The maintenance rules are automatically created by using big data tags and the latest national standards for maintenance projects through OCR technology and keyword matching technology. Maintenance projects that do not match national standards constitute enterprise standards.

[0040] As a further improvement to the technical solution of the present invention, an intelligent elevator maintenance device based on artificial intelligence includes:

[0041] The information acquisition and transmission module is installed in the elevator and DTU terminal to collect elevator status and fault information and upload it to the intelligent maintenance system.

[0042] The diagnostic module, located within the large elevator model, conducts regular inspections based on elevator operation data and historical fault information to locate elevators and maintenance items that require maintenance.

[0043] The task generation module determines maintenance personnel and generates maintenance tasks based on the diagnostic results in the large elevator model.

[0044] The automatic programming module is also located within the large elevator model, generating plugins based on instructions from the task generation module.

[0045] The client module, installed on the mobile phone, receives task information and plugins, guides maintenance personnel in their work, and uploads data.

[0046] The effectiveness evaluation module, located within the large elevator model, evaluates the maintenance effectiveness and generates case studies.

[0047] The knowledge sharing module, integrated into the large elevator model, enables the sharing and exchange of maintenance cases among maintenance personnel.

[0048] As a further improvement to the technical solution of the present invention, the large elevator model also includes a maintenance case model, which is used to store cases generated by the effect evaluation module and provide a reference for subsequent maintenance.

[0049] As a further improvement to the technical solution of the present invention, the intelligent maintenance system includes a maintenance server, a communication server, and a database. The communication server receives elevator information and stores it in the database. The maintenance server processes elevator data and interacts with the large elevator model. The database stores elevator operation data, fault information, maintenance task information, plug-in information, and maintenance case information.

[0050] An AI-based intelligent elevator maintenance device includes an elevator, a DTU terminal, and a mobile phone. The elevator communicates with the intelligent maintenance system through the DTU terminal. The mobile phone has an intelligent maintenance client installed and interacts with the intelligent maintenance system. The DTU terminal continuously collects elevator status and fault information, and the mobile phone receives and displays task information and plugins to assist maintenance personnel in completing tasks and uploading data.

[0051] An AI-based intelligent elevator maintenance system includes a cloud-based intelligent maintenance system and a large-scale elevator model. The intelligent maintenance system and the large-scale elevator model work together. The intelligent maintenance system is responsible for elevator data collection, storage, and preliminary processing, while the large-scale elevator model performs in-depth analysis, diagnosis, task allocation, plugin generation, effect evaluation, and knowledge management. Together, they realize the intelligent elevator maintenance process.

[0052] As a further improvement to the technical solution of the present invention, the cloud-based intelligent maintenance system adopts the mainstream Web API application architecture, uses JSON format for data interaction, uses a relational database, and implements communication between the client and the server based on HTTPS. Its establishment process includes surveying customer needs, determining technical solutions, programming development, testing and verification, and system launch steps.

[0053] The present invention has the following beneficial effects:

[0054] 1. Intelligent maintenance level

[0055] By innovatively leveraging the powerful capabilities of large-scale artificial intelligence models to automatically program and generate maintenance plugins, this feature endows maintenance work with a high degree of intelligence and autonomy. Even offline, the maintenance plugins can still play a crucial role, effectively overcoming the traditional dependence on network environments for maintenance and greatly expanding the implementation scenarios and flexibility of maintenance work.

[0056] With precise guidance from intelligent plug-ins, maintenance personnel can quickly pinpoint key maintenance priorities and critical operational steps, significantly reducing the substantial time wasted due to blind troubleshooting and experience differences in traditional maintenance models. Maintenance tasks that previously might have taken hours or even longer can be drastically reduced to tens of minutes or even less with the assistance of intelligent plug-ins, effectively improving maintenance efficiency. This increased efficiency not only reduces elevator downtime and minimizes disruption to people using elevators within the building, but also directly reduces the labor and material costs associated with prolonged maintenance, such as reducing overtime for maintenance personnel and unnecessary wear and tear on maintenance tools and spare parts.

[0057] From a customer experience perspective, the elevator can resume normal operation more quickly, which greatly improves customers' satisfaction and perception of the convenience of using the elevator, and enhances their trust and recognition of the elevator management and maintenance service providers.

[0058] 2. Knowledge Sharing Level

[0059] This invention constructs an efficient knowledge-sharing ecosystem. Case data generated during each elevator maintenance process, including elevator fault type, maintenance operation details, plug-in information used, and the final maintenance effect evaluation, is automatically uploaded to the maintenance case model within the large elevator model. This initiative enables maintenance personnel nationwide to conveniently access these valuable practical experiences and knowledge outcomes.

[0060] For inexperienced novice maintenance personnel, studying these rich case studies allows them to quickly understand the handling methods and techniques for different types of elevators in various fault situations, accelerating their professional skill growth and improvement, shortening the training cycle, and enabling them to independently undertake maintenance tasks more quickly while ensuring work quality. For experienced senior maintenance personnel, these case data also provide a platform for mutual exchange and learning. They can discover new maintenance ideas and methods, further optimize their own maintenance strategies, and also gain a more comprehensive and in-depth understanding of the overall development trends and common problems in the elevator industry.

[0061] From the perspective of the macro development of the elevator maintenance industry, this knowledge-sharing mechanism promotes the dissemination and inheritance of technology and experience within the industry, breaks down information barriers between regions and enterprises, is conducive to the overall improvement and balanced development of the industry's technical level, and injects strong momentum into the sustainable development of the elevator maintenance industry.

[0062] 3. Improve the quality of maintenance.

[0063] By conducting in-depth analysis of detailed data generated during maintenance and recording of maintenance operations, this invention achieves accurate assessment and effective monitoring of maintenance quality. At the data level, it can precisely monitor changes in the performance parameters of various elevator components before and after maintenance, such as the stability of the elevator's operating speed, the response time and braking force of the braking system, and voltage and current fluctuations in the electrical system, thereby determining whether the maintenance operations have truly achieved the expected optimization and repair effects.

[0064] In terms of video analysis, it allows for direct observation of the operational standardization and completeness of maintenance personnel. For example, it enables observation of whether key components are disassembled, cleaned, installed, and debugged according to standard procedures, and whether any important inspection steps are missed. This multi-dimensional analysis and evaluation method is more accurate and scientific than the traditional method of subjectively judging maintenance quality based solely on the subsequent operation of the elevator.

[0065] Once any problems or deficiencies in maintenance quality are discovered, timely feedback can be provided to maintenance personnel for rectification and optimization, thus forming a closed-loop quality control system. This ensures that every maintenance operation can effectively improve the elevator's operational performance and safety, reduce the probability of elevator malfunctions, and provide passengers with a safer and more reliable elevator riding environment.

[0066] 4. Standardization of maintenance

[0067] This invention, based on rigorous standard maintenance procedures and accurate diagnostic data, generates a unified maintenance plugin through automatic programming, fundamentally solving the industry-wide problem of inconsistent maintenance standards. Regardless of the region or the maintenance personnel performing the elevator maintenance task, using the plugin generated by this invention ensures comprehensive and meticulous maintenance and upkeep of the elevator according to the same standard procedures. Attached Figure Description

[0068] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0069] Figure 1 This is a system architecture diagram of an embodiment of the present invention. Detailed Implementation

[0070] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention, but are not intended to limit the present invention.

[0071] It should be noted that all directional indicators (such as up, down, left, right, front, back, upper end, lower end, top, bottom, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicator will also change accordingly.

[0072] In this invention, unless otherwise explicitly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral part; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0073] Furthermore, in this invention, descriptions involving "first," "second," etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature. Additionally, the technical solutions of various embodiments can be combined with each other, but only on the basis of being achievable by those skilled in the art. If the combination of technical solutions is contradictory or impossible to implement, such a combination should be considered non-existent and not within the scope of protection claimed by this invention.

[0074] The present invention will be further described in detail below with reference to the accompanying drawings.

[0075] Reference Figure 1 In actual operation, this AI-based intelligent elevator maintenance system involves close collaboration among its components, following the detailed implementation steps below:

[0076] First, ensure a stable and reliable connection between elevator E1 and DTU terminal T1. The communication server Z1 has efficient and stable network reception and transmission capabilities. Database B1 is running normally and has sufficient storage space to record various types of elevator data. The large elevator model X1 has completed accurate training based on massive amounts of elevator industry data and can accurately perform complex operations such as diagnosis, task allocation, and automatic programming. The intelligent maintenance client C1 is fully installed and functioning normally on mobile phone A1. Maintenance personnel P1 have received system operation training and are familiar with the entire maintenance process.

[0077] When the elevator is in operation, the DTU terminal T1 continuously collects elevator status and fault information, covering various aspects such as elevator speed, car position, door system status, and electrical system parameters. For example, the DTU terminal T1 acquires elevator speed data every few seconds and checks whether the elevator's safety circuit is functioning correctly. Once an anomaly is detected, such as abnormal fluctuations in elevator speed or a broken safety circuit, this status and fault information U1 is immediately processed and uploaded to the communication server Z1 of the intelligent maintenance system X1 via a preset communication protocol.

[0078] After receiving information from the elevator, the communication server Z1 quickly writes the fault information into database B1 for long-term storage, so as to facilitate subsequent querying and analysis. At the same time, the status and fault information U3 is pushed to the status and fault model M2 of the elevator's large model X1 in real time. The status and fault model M2 will perform in-depth analysis and recording of this data to build a dynamic profile of the elevator's operating status.

[0079] The elevator's large-scale diagnostic model M3 initiates inspection tasks according to a preset daily inspection cycle. It makes judgments based on a comprehensive and scientific set of diagnostic rules, considering factors such as historical elevator fault data, recent operating parameter trends, and wear prediction models for various components. Taking an elevator in a commercial building as an example, if an elevator has frequently experienced door system fault alarms in the past month, and its door operator's operating current fluctuations recently exceed the normal range by 30%, combined with the door operator's service life and number of operations, the diagnostic rules calculate the elevator's health level to be only 20 (health level <30 indicates maintenance is required). The maintenance item is precisely determined to be a comprehensive overhaul and adjustment of the door system, including performance testing of the door operator motor, sensitivity checks of the door locks, and cleaning and lubrication of the door tracks. After identifying elevators requiring maintenance, priority is determined based on factors such as health level values, fault severity, and the building's usage requirements. In a large hotel with multiple elevators, if two elevators require maintenance at the same time, and one elevator frequently poses a risk of entrapment and has a health level of only 10, while the other elevator only has a minor car lighting malfunction and a health level of 25, then the diagnostic model M3 will prioritize the elevator with a health level of 10 as the high-priority maintenance task to minimize the impact on the safety and user experience of hotel guests.

[0080] Next, the elevator's task model M4, based on the results from the diagnostic model M3 and the pre-set maintenance personnel duty roster, accurately determines the personnel responsible for maintenance. Assuming a maintenance team is responsible for elevator maintenance in multiple areas, task model M4 will comprehensively consider factors such as the maintenance personnel's geographical location, current workload, and experience with specific elevator types when assigning tasks. For example, for an elevator of a specific brand and model with a complex electrical fault, task model M4 will prioritize assigning the task to maintenance personnel P1 who are skilled in repairing the electrical systems of that brand of elevators, are located nearby, and have a relatively low current workload. After determining the maintenance personnel, task model M4 generates detailed maintenance task information and simultaneously issues corresponding instructions to the automatic programming M5. For example, regarding the elevator door system fault in the aforementioned commercial building, the instruction to automatic programming M5 is: "Generate a plugin for this elevator, containing standard maintenance procedures + comprehensive door system inspection (including door motor performance testing, door lock sensitivity checking, door track cleaning and lubrication), requiring video upload, with the loading interface being Load, the running interface being Run, and the upload interface being Upload."

[0081] Upon receiving instructions, the elevator's large-scale automated programming model M5 quickly initiates its automatic analysis program. Based on the previously learned codebase and algorithm model, and combined with the specific requirements of the maintenance project, it rapidly generates client-callable plugins. For example, the automated programming model M5 selects relevant code snippets from the codebase for door operator motor performance testing, combines them with standard code procedures for door lock sensitivity checks and operational instructions for door track cleaning and lubrication, integrates and optimizes them according to a preset plugin architecture, generating a complete and functionally defined maintenance plugin. Once the plugin is generated, it automatically notifies the task model M4, which immediately pushes the task information and plugin to the intelligent maintenance client C1 on the maintenance personnel P1's mobile phone A1 via the intelligent maintenance system.

[0082] After receiving the task notification, maintenance personnel P1, carrying mobile phone A1, head to the elevator site. Upon arrival, they open the intelligent maintenance client C1 and click "Load" on the plugin interface. Client C1 automatically downloads and loads the corresponding plugin from its local cache or the server. Once loaded, the maintenance personnel clicks "Run" on the interface, and client C1 displays the maintenance steps and requirements on the phone screen with a clear and intuitive interactive interface. Taking elevator door system maintenance as an example, client C1 first prompts the maintenance personnel to turn off the elevator power and set warning signs. It then progressively demonstrates the operation method for testing the door motor performance, such as connecting testing instruments, starting the motor, and reading parameters like current and speed, requiring the maintenance personnel to input the test data into the client. In the door lock sensitivity check, the client explains in detail how to use specialized tools to simulate door lock opening and closing actions, check the door lock's response time and locking force, and provides standard data ranges for comparison and judgment. Throughout the maintenance process, the maintenance personnel complete each operation sequentially according to the plugin prompts and record the maintenance process on their mobile phone.

[0083] After maintenance personnel P1 completes all maintenance operations, they click the "Upload" plugin interface. At this time, mobile phone A1 uploads the data recorded during the maintenance process (such as various test parameters, information on replaced parts, etc.) and video recordings to the elevator performance evaluation model M6 via the network. Upon receiving the data and video recordings, the performance evaluation model M6 processes them using advanced data analysis algorithms and image recognition technology. For example, by analyzing the door operator motor performance test data, it determines whether the motor's output power and efficiency have returned to normal ranges, compares the door lock sensitivity check data with standard values, and uses image recognition technology to perform a compliance check on the maintenance operations in the video recordings, such as checking whether the maintenance personnel used tools correctly and performed operations according to the prescribed sequence. Based on these analysis results, the performance evaluation model M6 accurately determines the effectiveness of this maintenance and scores it according to preset scoring criteria.

[0084] After the effectiveness evaluation model M6 completes the evaluation and scoring, it automatically organizes the detailed information of this maintenance into a case study and pushes it to the elevator maintenance case study model M1. For example, it integrates the elevator's basic information (brand, model, installation location, etc.), fault details, maintenance operation details, plug-in usage, effectiveness evaluation data, and final score into a complete case study and stores it in the maintenance case study model M1 for maintenance personnel nationwide to refer to in subsequent work.

[0085] Finally, once the maintenance is completed and the effectiveness evaluation is satisfactory, the system automatically resets the elevator's health status to 100, and the elevator returns to normal operation, awaiting the next status monitoring and maintenance cycle.

[0086] Implementation Cases

[0087] Case 1: Maintenance of an elevator malfunction in a high-rise residential building

[0088] In a high-rise residential complex, one of the elevators suddenly experienced abnormal shaking and flickering car lights during operation. The DTU terminal T1 quickly collected this status information and uploaded it to the intelligent maintenance system. The communication server Z1 received and stored the data, then pushed it to the status and fault model M2. The diagnostic model M3, after analysis and combining it with the elevator's historical operating data, determined that the problem likely stemmed from a fault in the elevator's electrical system, involving abnormal inverter parameters and loose wiring connections. The calculated health level of the elevator was 15, indicating an urgent need for maintenance.

[0089] Based on the duty roster, task model M4 determines that experienced maintenance personnel P1 will be responsible for the elevator maintenance task, and issues an instruction to automatic programming model M5: "Generate a plug-in for this elevator, the content of which is standard maintenance procedure + electrical system inspection (including inverter parameter inspection and adjustment, line connection tightness inspection, and electrical component performance testing). Video recording is required. The loading interface is Load, the running interface is Run, and the uploading interface is Upload."

[0090] After the automatic programming model M5 generates the plugin, P1 receives the task and proceeds to the site. The plugin is loaded via the intelligent maintenance client C1, and the operation is performed according to the prompts. During the inspection of the inverter parameters, several key parameters were found to be outside the normal range; adjustments were made according to the plugin's instructions. During the inspection of wiring connections, a loose power cable was found; after tightening, electrical component performance tests were conducted, and the test data was entered into the client. The entire process was recorded and uploaded to the effect evaluation model M6.

[0091] The performance evaluation model M6 analyzed the data and video recordings, confirming that the inverter parameters had returned to normal, the electrical components were performing stably, and the maintenance operation was in accordance with regulations. The maintenance was rated 90 points, and the case was pushed to the maintenance case model M1 for other maintenance personnel to refer to and learn from. After that, the elevator's health was reset to 100, and it resumed normal operation.

[0092] Case 2: Elevator Group Maintenance Dispatch in a Shopping Mall

[0093] A large shopping mall has multiple elevators. During a routine inspection, a large-scale elevator model revealed that three elevators had varying degrees of problems. One elevator's doors experienced a jamming motion when closing, resulting in a health rating of 25; another elevator made slight noise during operation, with a health rating of 35; and the third elevator had a leveling accuracy deviation, resulting in a health rating of 20.

[0094] Diagnostic model M3 prioritizes elevators with leveling accuracy deviations based on fault conditions and health status, as this issue may affect passenger safety and convenience when entering and exiting the elevator. Task model M4 assigns maintenance personnel P1, who is skilled in elevator mechanical system debugging, to be responsible for the maintenance task of this elevator based on the location and skill expertise of the maintenance personnel, and generates corresponding plug-in instructions.

[0095] Upon arrival at the site, P1 inspected, adjusted, and lubricated the elevator's traction system, guide rails, and other components related to leveling accuracy, following the plugin's instructions. After completing the maintenance, data and video recordings were uploaded. Once the effectiveness evaluation model M6 passed the evaluation, the case was stored in the maintenance case model M1. Simultaneously, task model M4 continued to schedule maintenance tasks for the other two elevators, sequentially completing the intelligent and efficient maintenance work for the entire shopping mall's elevator group. This ensured the safe and stable operation of the shopping mall's elevators and improved the customer's elevator experience.

[0096] Through the above implementation methods and case studies, the entire intelligent elevator maintenance system has achieved intelligent operation throughout the entire process, from elevator status monitoring, fault diagnosis, task allocation, plug-in generation, on-site maintenance to effect evaluation and knowledge sharing, effectively improving the efficiency, quality and standardization of elevator maintenance.

[0097] In summary, the present invention has the following beneficial effects:

[0098] I. Intelligent Maintenance Level

[0099] This innovative approach leverages the powerful capabilities of large-scale artificial intelligence models to automatically program and generate maintenance plugins, granting maintenance work a high degree of intelligence and autonomy. Even offline, these plugins remain crucial, effectively overcoming the traditional dependence on network conditions and significantly expanding the scenarios and flexibility of maintenance operations. For example, in elevators in older buildings with weak network coverage or in remote areas, maintenance personnel can still smoothly carry out their work using pre-downloaded plugins, avoiding maintenance delays caused by network issues.

[0100] With precise guidance from intelligent plug-ins, maintenance personnel can quickly pinpoint key maintenance priorities and critical operational steps, significantly reducing the substantial time wasted due to blind troubleshooting and experience differences in traditional maintenance models. Maintenance tasks that previously might have taken hours or even longer can be drastically reduced to tens of minutes or even less with the assistance of intelligent plug-ins, effectively improving maintenance efficiency. This increased efficiency not only reduces elevator downtime and minimizes disruption to people using elevators within the building, but also directly reduces the labor and material costs associated with prolonged maintenance, such as reducing overtime for maintenance personnel and unnecessary wear and tear on maintenance tools and spare parts.

[0101] From a customer experience perspective, the elevator can resume normal operation more quickly, which greatly improves customers' satisfaction and perception of the convenience of using the elevator, and enhances their trust and recognition of the elevator management and maintenance service providers.

[0102] II. Knowledge Sharing Level

[0103] This invention constructs an efficient knowledge-sharing ecosystem. Case data generated during each elevator maintenance process, including elevator fault type, maintenance operation details, plug-in information used, and the final maintenance effect evaluation, is automatically uploaded to the maintenance case model within the large elevator model. This initiative enables maintenance personnel nationwide to conveniently access these valuable practical experiences and knowledge outcomes.

[0104] For inexperienced novice maintenance personnel, studying these rich case studies allows them to quickly understand the handling methods and techniques for different types of elevators in various fault situations, accelerating their professional skill growth and improvement, shortening the training cycle, and enabling them to independently undertake maintenance tasks more quickly while ensuring work quality. For experienced senior maintenance personnel, these case data also provide a platform for mutual exchange and learning. They can discover new maintenance ideas and methods, further optimize their own maintenance strategies, and also gain a more comprehensive and in-depth understanding of the overall development trends and common problems in the elevator industry.

[0105] From the perspective of the macro development of the elevator maintenance industry, this knowledge-sharing mechanism promotes the dissemination and inheritance of technology and experience within the industry, breaks down information barriers between regions and enterprises, is conducive to the overall improvement and balanced development of the industry's technical level, and injects strong momentum into the sustainable development of the elevator maintenance industry.

[0106] III. Improving the quality of maintenance

[0107] By conducting in-depth analysis of detailed data generated during maintenance and recording of maintenance operations, this invention achieves accurate assessment and effective monitoring of maintenance quality. At the data level, it can precisely monitor changes in the performance parameters of various elevator components before and after maintenance, such as the stability of the elevator's operating speed, the response time and braking force of the braking system, and voltage and current fluctuations in the electrical system, thereby determining whether the maintenance operations have truly achieved the expected optimization and repair effects.

[0108] In terms of video analysis, it allows for direct observation of the operational standardization and completeness of maintenance personnel. For example, it enables observation of whether key components are disassembled, cleaned, installed, and debugged according to standard procedures, and whether any important inspection steps are missed. This multi-dimensional analysis and evaluation method is more accurate and scientific than the traditional method of subjectively judging maintenance quality based solely on the subsequent operation of the elevator.

[0109] Once any problems or deficiencies in maintenance quality are discovered, timely feedback can be provided to maintenance personnel for rectification and optimization, thus forming a closed-loop quality control system. This ensures that every maintenance operation can effectively improve the elevator's operational performance and safety, reduce the probability of elevator malfunctions, and provide passengers with a safer and more reliable elevator riding environment.

[0110] IV. Standardization of Maintenance and Repair

[0111] This invention, based on rigorous standard maintenance procedures and accurate diagnostic data, generates a unified maintenance plugin through automatic programming, fundamentally solving the industry-wide problem of inconsistent maintenance standards. Regardless of the region or the maintenance personnel performing the elevator maintenance task, using the plugin generated by this invention ensures comprehensive and meticulous maintenance and upkeep of the elevator according to the same standard procedures.

[0112] For example, regarding elevator door system maintenance, the plugin clearly specifies a series of standard operating procedures and technical requirements, from inspecting and lubricating the mechanical components of the door operator to testing the electrical performance of the door locks and verifying the safety circuit. Maintenance personnel only need to follow the plugin's instructions to complete the operations sequentially. This standardized maintenance model effectively avoids the problem of inconsistent maintenance quality caused by differences in individual experience, regional customs, or inconsistent internal standards within companies. This ensures that elevators can receive uniform and standardized maintenance and upkeep nationwide, greatly improving the overall stability and safety of elevator operation and laying a solid foundation for the standardized management and healthy development of the elevator industry.

[0113] The technical solutions provided by the embodiments of the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the embodiments of the present invention. The descriptions of the embodiments above are only for helping to understand the principles of the embodiments of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the embodiments of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. An intelligent elevator maintenance method based on artificial intelligence, characterized in that, Includes the following steps: Step S1: The elevator collects and uploads status and fault information to the intelligent maintenance system in real time through the DTU terminal. The intelligent maintenance system then pushes this information to the status and fault model of the elevator's large model. Step S2: The diagnostic model of the large elevator model is inspected according to the preset cycle. Based on the elevator operation data, historical fault records and diagnostic rules, the elevators that need maintenance and their maintenance items are located, and the maintenance priority is determined. Step S3: The elevator large model determines the maintenance personnel based on the diagnosis results and duty roster, generates maintenance tasks, and sends instructions to the automatic programming model. Step S4: The automatic programming model generates client-callable plugins based on instructions and the code library and algorithm model built from previous learning and training, combined with the specific requirements of the maintenance project. The task model is then notified to push task information and the plugins to the maintenance personnel. The plugins include a Load interface, a Run interface, and an Upload interface. Step S5: Maintenance personnel use the mobile smart maintenance client to load the plugin by clicking the Load interface and run the Run interface. The maintenance steps and requirements are displayed on the client's interactive interface. The maintenance personnel complete the maintenance according to the prompts and record the maintenance process video. After completion, the data and video are uploaded to the effect evaluation model by clicking the Upload interface. Step S6: The effect evaluation model uses data analysis and image recognition technology to process the uploaded information, obtain the maintenance effect and score it, and then generate a case and push it to the maintenance case model; Step S7: If the maintenance effect assessment is qualified, the system will automatically reset the elevator health level to 100, so that the elevator can resume normal operation.

2. The elevator intelligent maintenance method based on artificial intelligence according to claim 1, characterized in that: The construction of the large elevator model includes the local deployment of a general large model. After inputting elevator industry data, multiple sub-models are constructed through data cleaning, word segmentation and tokenization, data augmentation and standardization, sampling and weight adjustment, and natural language processing. The sub-models include at least a maintenance case model, a status and fault model, a diagnostic model, a task model, an automatic programming model, and an effect evaluation model.

3. The elevator intelligent maintenance method based on artificial intelligence according to claim 1, characterized in that: The automatic programming model learns from a large number of code libraries and programming languages ​​corresponding to maintenance projects, and automatically generates code plugins based on user instructions and preset maintenance rules. The generation process involves natural language processing to convert instructions into programming languages.

4. The elevator intelligent maintenance method based on artificial intelligence according to claim 3, characterized in that: The maintenance projects and code are automatically associated by extracting header comments from the program source code using an AI big data model and converting them into big data tags. Maintenance rules are automatically created by matching the latest national standards for maintenance projects with big data tags using OCR technology and keyword matching technology. Maintenance projects that do not match national standards constitute enterprise standards.

5. An AI-based intelligent elevator maintenance device that applies the AI-based intelligent elevator maintenance method as described in any one of claims 1-4, characterized in that, include: The information acquisition and transmission module is installed in the elevator and DTU terminal to collect elevator status and fault information and upload it to the intelligent maintenance system. The diagnostic module, located within the large elevator model, conducts regular inspections based on elevator operation data and historical fault information to locate elevators and maintenance items that require maintenance. The task generation module is set in the large elevator model. It determines maintenance personnel and generates maintenance tasks based on the diagnostic results. An automatic programming module is installed within the large elevator model and generates plugins based on instructions from the task generation module. The client module, installed on the mobile phone, receives task information and plugins, guides maintenance personnel in their work, and uploads data. The effectiveness evaluation module, located within the large elevator model, evaluates the maintenance effectiveness and generates case studies. The knowledge sharing module, integrated into the large elevator model, enables the sharing and exchange of maintenance cases among maintenance personnel.

6. The elevator intelligent maintenance device based on artificial intelligence according to claim 5, characterized in that: The elevator model also includes a maintenance case model, which stores cases generated by the effect evaluation module and provides a reference for subsequent maintenance.

7. The elevator intelligent maintenance device based on artificial intelligence according to claim 5, characterized in that: The intelligent maintenance system includes a maintenance server, a communication server, and a database. The communication server receives elevator information and stores it in the database. The maintenance server processes elevator data and interacts with the large elevator model. The database stores elevator operation data, fault information, maintenance task information, plug-in information, and maintenance case information.

8. An AI-based intelligent elevator maintenance device that applies the AI-based intelligent elevator maintenance method as described in any one of claims 1-4, characterized in that, This includes elevators, DTU terminals, and mobile phones. The elevators communicate with the intelligent maintenance system through the DTU terminals, and the mobile phones have intelligent maintenance clients installed and interact with the intelligent maintenance system. The DTU terminals continuously collect elevator status and fault information, and the mobile phones receive and display task information and plugins to assist maintenance personnel in completing tasks and uploading data.

9. An AI-based intelligent elevator maintenance system applying the AI-based intelligent elevator maintenance method as described in any one of claims 1-4, characterized in that, This includes a cloud-based intelligent maintenance system and a large-scale elevator model. The intelligent maintenance system and the large-scale elevator model work together. The intelligent maintenance system is responsible for collecting, storing, and initially processing elevator data, while the large-scale elevator model performs in-depth analysis, diagnosis, task allocation, plugin generation, effect evaluation, and knowledge management. Together, they realize the intelligent elevator maintenance process.

10. The intelligent elevator maintenance system based on artificial intelligence as described in claim 9, wherein the cloud-based intelligent maintenance system adopts a mainstream Web API application architecture, uses JSON format for data interaction, uses a relational database, and implements communication between the client and server based on HTTPS. Its establishment process includes surveying customer needs, determining technical solutions, programming development, testing and verification, and system launch steps.

Images