Intelligent reconfigurable elevator steel structure hoistway system
By using intelligent reconfigurable structural modules and intelligent collaborative control modules of multi-agent systems, the limitations of elevator shafts in terms of adaptability, safety and intelligence have been solved, enabling flexible installation, real-time monitoring and efficient maintenance, and improving the overall performance and intelligence level of elevator shafts.
Patent Information
- Application Number
- CN202510933097.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-08
- Publication Date
- 2025-10-28
AI Technical Summary
Existing elevator shafts have limitations in terms of safe operation, installation efficiency, user experience, and intelligent collaboration. They are difficult to adapt to different building structures and spatial requirements, are difficult to monitor and maintain, and lack intelligent collaboration, resulting in installation difficulties, numerous safety hazards, energy waste, and insufficient emergency response.
It employs intelligent reconfigurable structural modules, a layout optimization module based on an improved bat algorithm, a fault prediction and health management module based on deep learning, an intelligent collaborative control module based on a multi-agent system, an augmented reality-assisted installation and maintenance module, and a blockchain-based full lifecycle data management module to achieve flexible structural adjustment, real-time monitoring, intelligent collaboration, and data security management.
It improves the adaptability and flexibility of elevator shafts, enhances safety and reliability, improves the level of intelligence, optimizes installation and maintenance processes, ensures data management and collaborative work, and reduces costs and risks.
Smart Images
Figure CN120841333A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of building engineering and elevator technology, and in particular to an intelligent reconfigurable elevator steel structure shaft system. Background Technology
[0002] In the development of elevator shafts, existing technologies have several limitations, negatively impacting elevator safety, installation efficiency, and user experience. Fixed structures lack flexibility: Traditional elevator shafts mostly use fixed steel or concrete structures. This makes it difficult for the shafts to adapt to different building structures and spatial layouts. When adding elevators to older buildings, due to their age and diverse structural forms, fixed shaft structures often cannot match the original building's structural characteristics, leading to installation difficulties and even requiring large-scale structural modifications, increasing costs and construction risks. Furthermore, for buildings with special spatial requirements or potential future functional adjustments, fixed shaft structures cannot be flexibly modified, limiting the effective use of building space. Monitoring and maintenance are difficult: Existing elevator shaft monitoring methods are limited, typically relying on periodic manual inspections to detect potential problems. However, dynamic data such as vibrations and stress changes generated during elevator operation are difficult to obtain in real time, leading to the failure to promptly detect and address some safety hazards. For example, tiny cracks in the shaft structure or loose connections are difficult to detect in the early stages; once they develop into serious problems, they may cause elevator malfunctions or even safety accidents. Furthermore, the complex internal structure of the elevator shaft limits the operating space for maintenance personnel, increasing the difficulty and cost of maintenance. Lack of intelligent collaboration: Modern buildings pursue intelligent management, but existing elevator shafts lack effective intelligent collaboration with other building systems. Equipment within the elevator shaft (such as ventilation, lighting, and safety equipment) often operates independently, unable to intelligently adjust according to the actual operating status of the elevator and changes in the building environment. For example, during elevator downtime, ventilation and lighting equipment continue to operate in fixed modes, resulting in energy waste. Simultaneously, in emergencies such as fires or earthquakes, the elevator shaft cannot quickly coordinate with the building's emergency management system, affecting the efficiency of personnel evacuation and rescue. Summary of the Invention
[0003] This invention aims to provide an intelligent, reconfigurable elevator shaft system, addressing key issues currently faced by elevator shafts in practical applications and improving their performance and applicability. It includes:
[0004] The intelligent reconfigurable structural module adopts a modular design. It adjusts the connection angle and length through intelligent connection nodes and uses shape memory alloy material to achieve structural fine-tuning. The connection angle adjustment range is ±30° and the length adjustment range is 0-200mm.
[0005] The layout optimization module based on the improved bat algorithm encodes the layout scheme into individual bat position vectors and optimizes the layout through the fitness function Fitness=w1×S+w2×U+w3×E, where S represents the structural stability index, U represents the space utilization rate, E represents the equipment operating efficiency, and w1, w2, and w3 are weight coefficients.
[0006] The deep learning-based fault prediction and health management module collects data through sensors and uses a model combining convolutional neural networks and long short-term memory networks to predict faults and assess health status.
[0007] The intelligent collaborative control module based on the multi-Agent system treats the equipment in the well as agents and achieves collaborative work through a distributed communication protocol.
[0008] The augmented reality-assisted installation and maintenance module uses augmented reality technology to provide visual operation guidance for installation and maintenance personnel;
[0009] The blockchain-based full lifecycle data management module stores full lifecycle data on the blockchain and uses smart contracts to achieve data sharing and access control.
[0010] Furthermore, in the intelligent reconfigurable structure module, the intelligent connection node incorporates a micro motor, sensor, and controller, and shape memory alloy material is used to support the beams and connectors.
[0011] Furthermore, in the layout optimization module based on the improved bat algorithm, the improved bat algorithm introduces an adaptive pulse emission rate and a local search strategy. The formula for calculating the individual bat frequency is fi = fmin + (fmax - fmin) × rand(), where rand() is a random number between 0 and 1. The formula for updating the individual bat speed is Vit = Vit-1 + (Xit - Xbestt) × fi, and the formula for updating the position is Xit = Xit-1 + Vit, where Xbestt is the current global optimal position. Parameters such as the number of bats and the maximum number of iterations can be adjusted according to the actual situation.
[0012] Furthermore, in the deep learning-based fault prediction and health management module, the number of layers and nodes in the convolutional neural network and long short-term memory network can be adjusted according to data characteristics and prediction accuracy requirements. After the data is processed by the convolutional neural network to extract local features and the long short-term memory network to learn time series features, the fault prediction results and health status assessment indicators are output through a fully connected layer. Furthermore, in the intelligent collaborative control module based on a multi-agent system, each agent has perception, decision-making, and execution capabilities, and the communication protocol adopts a publish-subscribe pattern or a message queue.
[0013] Furthermore, the augmented reality-assisted installation and maintenance module displays the virtual installation location and connection method of the shaft module through AR devices during installation, and displays the health status, fault location and maintenance steps of the shaft structure in real time during maintenance.
[0014] Furthermore, the blockchain-based full lifecycle data management module utilizes the decentralized and tamper-proof characteristics of blockchain to ensure the authenticity and security of data, and enables different stakeholders to access and use data in a tiered manner through smart contracts.
[0015] Furthermore, the layout optimization module based on the improved bat algorithm compares the fitness value of the newly generated solution with that of the current global optimal solution through multiple iterations when calculating fitness. If the new solution is better, the global optimal solution is updated until the maximum number of iterations is reached and the global optimal layout scheme is output.
[0016] Furthermore, the deep learning-based fault prediction and health management module uses the cross-entropy loss function as the loss function during model training, uses the Adam optimizer to train the model, and improves model performance by adjusting parameters such as convolution kernel size, stride, and number of LSTM units.
[0017] Furthermore, the intelligent collaborative control module based on the multi-Agent system, during system operation, after each Agent senses environmental information, makes decisions and executes actions according to decision rules, and at the same time, it interacts with other Agents through communication protocols to achieve collaboration. For example, when the security protection Agent receives building fire alarm information, it notifies the lighting Agent to turn off non-emergency lighting and notifies the ventilation Agent to switch to smoke exhaust mode.
[0018] Beneficial effects:
[0019] Enhanced adaptability and flexibility: Intelligent reconfigurable structural modules enable elevator shafts to quickly adapt to changes in different building structures and spatial layouts, reducing the workload of building renovations and improving the utilization efficiency of building space.
[0020] Enhanced safety and reliability: The deep learning-based fault prediction and health management module can monitor the health status of the shaft structure in real time, detect potential faults in advance, provide maintenance personnel with timely maintenance suggestions, reduce the probability of safety accidents, and improve the reliability of elevator operation.
[0021] Enhancing the level of intelligence: The multi-Agent system intelligent collaborative control module realizes intelligent collaboration of equipment in the elevator shaft, improves energy utilization efficiency, enhances emergency response capabilities in emergency situations, and improves the level of intelligence of the elevator system.
[0022] Optimized installation and maintenance processes: The augmented reality-assisted installation and maintenance module provides intuitive and accurate operation guidance for installation and maintenance personnel, reducing human error, improving the efficiency and quality of installation and maintenance, and lowering maintenance costs. Secure data management and collaboration: The blockchain-based full lifecycle data management module ensures the secure storage and traceability of elevator shaft data, promotes information sharing and collaborative work among stakeholders, and contributes to the full lifecycle management of elevator shafts. Attached Figure Description
[0023] Figure 1 AI-assisted installation flowchart;
[0024] Figure 2 Blockchain data management flowchart. Detailed Implementation
[0025] Example 1
[0026] Specific implementation methods combined Figures 1 to 2 Detailed explanation follows. Intelligent Reconfigurable Structural Modules: Employing a modular design concept, the elevator shaft is divided into multiple standard modules, which are connected via intelligent connection nodes. These intelligent connection nodes incorporate micro-motors, sensors, and controllers, automatically adjusting the connection angle and length according to building structure and spatial requirements. The connection angle adjustment range is ±30°, and the length adjustment range is 0-200mm. Key components, such as support beams and connectors, are made using shape memory alloy (SMA) material. SMA can change shape when temperature or current changes, thereby enabling fine-tuning of the shaft structure.
[0027] The layout optimization module based on the improved bat algorithm encodes the elevator shaft layout scheme (including support point locations, equipment installation locations, etc.) as position vectors for individual bats. The improved bat algorithm, building upon traditional algorithms, introduces an adaptive pulse emission rate and a local search strategy. A fitness function, Fitness = w1×S + w2×U + w3×E, is constructed based on the structural stability, space utilization, and equipment operating efficiency of the layout scheme. Here, S represents the structural stability index, U represents the space utilization rate, E represents the equipment operating efficiency, and w1, w2, and w3 are weighting coefficients. Individual bats continuously adjust their flight speed and position to search for the optimal layout scheme in the solution space.
[0028] The deep learning-based fault prediction and health management module involves installing multiple sensors in key areas of the elevator shaft (such as support beams, connecting nodes, and guide rails) to collect real-time data on vibration, stress, and displacement. A deep learning model combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks is used to process the data. The CNN extracts local features from the data, while the LSTM learns the time-series features, enabling real-time monitoring and fault prediction of the shaft structure's health status. The model input consists of multi-channel data collected by the sensors. After processing through CNN and LSTM layers, the output, via a fully connected layer, provides fault prediction results and health status assessment indicators.
[0029] The intelligent collaborative control module based on a multi-agent system treats each piece of equipment within the elevator shaft (ventilation equipment, lighting equipment, safety protection equipment, etc.) as an independent agent. Each agent possesses perception, decision-making, and execution capabilities, enabling autonomous decision-making based on its own status and environmental information. The multi-agent system achieves information exchange and collaborative work among the agents through a distributed communication protocol. For example, when there are people in the elevator car, the lighting agent and ventilation agent automatically adjust their operating modes; when an emergency occurs in the building, the safety protection agent collaborates with the building emergency management system to execute corresponding emergency operations.
[0030] Augmented Reality-Assisted Installation and Maintenance Module: Utilizing augmented reality (AR) technology, this module provides visual operational guidance for installation and maintenance personnel. During installation, the AR device allows installers to see the virtual installation location and connection method of the shaft module, improving installation accuracy and efficiency. During maintenance, the AR device can display the real-time health status of the shaft structure, fault locations, and repair steps, assisting maintenance personnel in quickly locating and resolving problems.
[0031] A blockchain-based full lifecycle data management module stores elevator shaft data throughout its entire lifecycle, from design, manufacturing, installation, operation to decommissioning, on the blockchain. Leveraging the decentralized and immutable nature of blockchain, the authenticity and security of the data are ensured. Smart contracts enable data sharing and access control, allowing different stakeholders (such as building developers, elevator manufacturers, maintenance companies, and regulatory authorities) to access and use data according to their permissions. For example, maintenance companies can view the shaft's historical maintenance records and operational data, while regulatory authorities can monitor the shaft's safety status in real time.
[0032] Innovative algorithm modeling and solving
[0033] Improved Bat Algorithm Modeling and Solution Process:
[0034] Initialization parameters: Set the bat population size N, maximum number of iterations T, frequency range fmin-fmax, pulse emission rate r0, loudness A0, and search space dimension D. Randomly initialize the position Xi and velocity Vi of individual bats, i = 1, 2, ..., N.
[0035] Fitness calculation: The fitness value of each bat individual is calculated according to the fitness function Fitness = w1×S + w2×U + w3×E.
[0036] Bat individual update: For each bat individual, its frequency is calculated using the formula fi = fmin + (fmax - fmin) × rand(), where rand() is a random number between 0 and 1. The velocity is updated using the formula Vit = Vit-1 + (Xit - Xbestt) × fi, where Xbestt is the current global optimal position. The position is updated using the formula Xit = Xit-1 + Vit.
[0037] Local search and pulse emission: A local search is performed with probability ri, which is adaptively adjusted according to the number of iterations. If a local search is performed, a new solution is randomly generated near the current optimal solution. Simultaneously, these two parameters are adjusted according to the update rules for loudness Ai and pulse emission rate ri.
[0038] Optimal solution update: Compare the fitness value of the newly generated solution with that of the current global optimal solution. If the new solution is better, then update the global optimal solution.
[0039] Termination condition judgment: If the maximum number of iterations T is reached, the algorithm terminates and outputs the globally optimal layout scheme; otherwise, it returns to the fitness calculation step to continue iterating.
[0040] Deep learning (CNN-LSTM) model modeling and solution process:
[0041] Data Acquisition and Preprocessing: Vibration, stress, and displacement data of the elevator shaft are collected using sensors. The collected data is cleaned to remove outliers and noise. Then, normalization is performed to map the data to the [0,1] interval to meet the model input requirements.
[0042] Model Construction: A CNN-LSTM model was built. The CNN part consists of multiple convolutional layers and pooling layers. The convolutional layers use kernels of different sizes to extract local features from the data. The pooling layers are used to reduce data dimensionality and computational cost. The LSTM part receives the features output by the CNN and learns the time-series features of the data. Finally, the features output by the LSTM are integrated through a fully connected layer to output fault prediction results and health status assessment indicators.
[0043] Model Training: The preprocessed data is divided into training, validation, and test sets. Cross-entropy loss is used as the model's loss function, and the Adam optimizer is used to train the model. During training, model parameters, such as kernel size, stride, and number of LSTM units, are continuously adjusted to improve model performance.
[0044] Model evaluation: The trained model is evaluated on the test set, and its performance is measured using metrics such as accuracy, recall, and F1 score. Based on the evaluation results, the model is further optimized and adjusted until its performance reaches a satisfactory level.
[0045] Multi-Agent System Modeling and Solution Process:
[0046] Agent Design: Design a corresponding agent for each device, defining the agent's perception range, decision rules, and execution actions. For example, a lighting agent can sense the number of people in an elevator car and the light intensity. The decision rule could be to turn on the lighting equipment and adjust the brightness when there are people in the car and the light is dim; the execution action would be to control the switching on and off of the lighting equipment and adjust its brightness.
[0047] Communication Protocol Design: Design a communication protocol for the multi-agent system to ensure accurate and timely information exchange between agents. Employ publish-subscribe patterns or message queues to implement communication between agents.
[0048] System initialization: Initializes the state and environment information of each agent, enabling agents to make decisions and take actions based on initial conditions.
[0049] System Operation and Collaboration: During system operation, each agent continuously senses environmental information, makes decisions based on decision rules, and executes corresponding actions. Simultaneously, it interacts with other agents through communication protocols to achieve collaborative work. For example, when a security agent receives a building fire alarm, it uses the communication protocol to notify the lighting agent to turn off non-emergency lighting and the ventilation agent to switch to smoke extraction mode.
[0050] Example 2
[0051] Implementation of Intelligent Reconfigurable Structural Modules: In an elevator retrofit project for an old building, intelligent reconfigurable structural modules were used to adjust the connection angles and lengths of each module through intelligent connection nodes, ensuring a tight integration between the shaft and the building structure, taking advantage of the building's irregular structure. During installation, shape memory alloy components were used to fine-tune the shaft structure, ensuring its stability. Actual testing showed that the installed shaft could withstand various loads generated by elevator operation, meeting safety requirements.
[0052] Improved Bat Algorithm for Layout Optimization: In the design of an elevator shaft in a newly built office building, an improved bat algorithm was used to optimize the support point positions and equipment installation locations of the shaft. With a bat population size of 50 and a maximum iteration count of 200, after multiple iterations, the resulting layout scheme improved the structural stability of the shaft by 20%, increased space utilization by 15%, and significantly improved equipment operating efficiency.
[0053] Implementation of Deep Learning-Based Fault Prediction and Health Management: A deep learning-based fault prediction and health management system was installed in the elevator shaft of a high-rise residential building. Sensors collect data from the shaft in real time, and the deep learning model analyzes the data. In one monitoring session, the model predicted a potential loosening risk at a certain connection node a week in advance, allowing maintenance personnel to promptly reinforce it and prevent a potential safety accident.
[0054] Implementation of Multi-Agent System Intelligent Collaborative Control: A multi-agent system intelligent collaborative control module was applied in the elevator shaft of a commercial center. During off-peak hours at night, the ventilation agent and lighting agent automatically reduce their operating power according to a collaborative strategy, achieving energy savings. In a fire drill, the safety protection agent worked in conjunction with the building's emergency management system to quickly switch the elevator to fire-fighting mode, ensuring the safe evacuation of personnel.
[0055] Augmented Reality-Assisted Installation and Maintenance Implementation: During the installation and maintenance of an elevator shaft in a hotel, installation and maintenance personnel wore AR devices. During installation, the AR devices displayed virtual installation guidance information, helping installers accurately install each module, reducing installation time by 30%. During maintenance, the AR devices displayed the real-time health status and fault locations of the shaft structure, assisting maintenance personnel in quickly completing repairs, improving maintenance efficiency by 40%.
[0056] Implementation of Blockchain-Based Lifecycle Data Management: A blockchain-based lifecycle data management module has been adopted in multiple elevator shaft projects. Stakeholders such as building developers, elevator manufacturers, maintenance companies, and regulatory authorities share and manage data through the blockchain platform. Maintenance companies can view historical maintenance records and operational data of the shaft at any time, while regulatory authorities can monitor the safety status of the shaft in real time, improving the management efficiency and transparency of elevator shafts.
[0057] Example 3
[0058] Employing intelligent reconfigurable structural modules, based on the actual contours of the building's curved facade, the connection nodes utilize built-in micro-motors, sensors, and controllers to precisely adjust the connection angles of each standard module within a range of ±30°, and flexibly adjust the lengths within the 0-200mm range. Simultaneously, shape memory alloy materials are used for the support beams and connectors in key areas, allowing for subtle modifications to the shaft structure through temperature changes, ensuring the shaft perfectly conforms to the building's exterior curves.
[0059] During installation, augmented reality (AR) assistance was used, with installers wearing AR devices. These devices displayed the virtual installation locations and connection methods of the shaft modules in real time, providing visual guidance that enabled installers to quickly and accurately assemble and secure each module. Compared to traditional installation methods, this project reduced installation time by 40%, achieved higher shaft installation precision, and ensured excellent fit with the building structure. This effectively minimized damage to the historical building's appearance and structure while meeting the load requirements for safe elevator operation.
[0060] In the maintenance and management of multiple elevator shafts in a large commercial complex, a fault prediction and health management module based on deep learning and a full lifecycle data management module based on blockchain are applied to ensure the safe and stable operation of the elevator system.
[0061] Vibration, stress, and displacement sensors are installed at key locations in the elevator shaft, such as support beams, connecting nodes, and guide rails, to collect large amounts of data in real time. After cleaning and normalization preprocessing, the collected data is input into a deep learning model combining a Convolutional Neural Network (CNN) and a Long Short-Term Memory Network (LSTM). The CNN layer uses convolutional kernels of different sizes to extract local features of the data, the LSTM layer learns the time-series features of the data, and the fully connected layer outputs fault prediction results and health status assessment indicators.
[0062] For example, in one monitoring session, the model, through analysis of vibration and stress data over several consecutive days, predicted two weeks in advance that a support beam in an elevator shaft was at risk of structural strength degradation due to long-term fatigue, and generated a detailed risk report.
[0063] Meanwhile, all data related to the elevator shaft, from parameters in the design and manufacturing stages to monitoring data during installation and operation, and maintenance records—the entire lifecycle data—is stored on the blockchain. Leveraging the decentralized and immutable nature of blockchain, the data's authenticity and reliability are ensured. Maintenance companies, through permissions granted by smart contracts, can promptly access fault prediction reports and relevant historical data, quickly formulate and execute maintenance plans. Regulatory authorities can also monitor the shaft's safety status in real time through the blockchain platform. The entire process is traceable, significantly improving the efficiency and safety of elevator shaft maintenance management and ensuring the normal operation of the commercial complex.
Claims
1. An intelligent reconfigurable elevator steel structure shaft system, characterized in that, include: The intelligent reconfigurable structural module adopts a modular design. It adjusts the connection angle and length through intelligent connection nodes and uses shape memory alloy material to achieve structural fine-tuning. The connection angle adjustment range is ±30° and the length adjustment range is 0-200mm. The layout optimization module based on the improved bat algorithm encodes the layout scheme into individual bat position vectors and optimizes the layout through the fitness function Fitness=w1×S+w2×U+w3×E, where S represents the structural stability index, U represents the space utilization rate, E represents the equipment operating efficiency, and w1, w2, and w3 are weight coefficients. The deep learning-based fault prediction and health management module collects data through sensors and uses a model combining convolutional neural networks and long short-term memory networks to predict faults and assess health status. The intelligent collaborative control module based on the multi-Agent system treats the equipment in the well as agents and achieves collaborative work through a distributed communication protocol. The augmented reality-assisted installation and maintenance module uses augmented reality to provide visual operation guidance for installation and maintenance personnel; The blockchain-based full lifecycle data management module stores full lifecycle data on the blockchain and uses smart contracts to achieve data sharing and access control.
2. The intelligent reconfigurable elevator steel structure shaft system according to claim 1, characterized in that, In the intelligent reconfigurable structural module, the intelligent connection node has a built-in micro motor, sensor and controller, and shape memory alloy material is used to support the beams and connectors.
3. The intelligent reconfigurable elevator steel structure shaft system according to claim 1, characterized in that, In the layout optimization module based on the improved bat algorithm, the improved bat algorithm introduces an adaptive pulse emission rate and a local search strategy. The formula for calculating the frequency of an individual bat is fi = fmin + (fmax - fmin) × rand(), where rand() is a random number between 0 and 1. The formula for updating the speed of an individual bat is Vit = Vit-1 + (Xit - Xbestt) × fi, and the formula for updating the position is Xit = Xit-1 + Vit, where Xbestt is the current global optimal position. Parameters such as the number of bats and the maximum number of iterations can be adjusted according to the actual situation.
4. The intelligent reconfigurable elevator steel structure shaft system according to claim 1, characterized in that, In the deep learning-based fault prediction and health management module, the number of layers and nodes of the convolutional neural network and long short-term memory network can be adjusted according to the data characteristics and prediction accuracy requirements. After the data is processed by the convolutional neural network to extract local features and the long short-term memory network to learn time series features, the fault prediction results and health status assessment indicators are output through the fully connected layer.
5. The intelligent reconfigurable elevator steel structure shaft system according to claim 1, characterized in that, In the intelligent collaborative control module based on the multi-Agent system, each agent has perception, decision-making, and execution capabilities, and the communication protocol adopts a publish-subscribe pattern or a message queue.
6. The intelligent reconfigurable elevator steel structure shaft system according to claim 1, characterized in that, The augmented reality-assisted installation and maintenance module displays the virtual installation location and connection method of the shaft module through AR devices during installation, and displays the health status, fault location and maintenance steps of the shaft structure in real time during maintenance.
7. The intelligent reconfigurable elevator steel structure shaft system according to claim 1, characterized in that, The blockchain-based full lifecycle data management module utilizes the decentralized and tamper-proof characteristics of blockchain to ensure the authenticity and security of data, and enables different stakeholders to access and use data in a tiered manner through smart contracts.
8. The intelligent reconfigurable elevator steel structure shaft system according to claim 1, characterized in that, The layout optimization module based on the improved bat algorithm compares the fitness value of the newly generated solution with that of the current global optimal solution through multiple iterations when calculating fitness. If the new solution is better, the global optimal solution is updated until the maximum number of iterations is reached and the global optimal layout scheme is output.
9. The intelligent reconfigurable elevator steel structure shaft system according to claim 1, characterized in that, The deep learning-based fault prediction and health management module uses the cross-entropy loss function as the loss function during model training, uses the Adam optimizer to train the model, and improves model performance by adjusting parameters such as convolution kernel size, stride, and number of LSTM units.
10. The intelligent reconfigurable elevator steel structure shaft system according to claim 1, characterized in that, The intelligent collaborative control module based on the multi-Agent system, during system operation, after each agent senses environmental information, makes decisions and executes actions according to decision rules, and at the same time, it interacts with other agents through communication protocols to achieve collaboration. For example, when the security protection agent receives building fire alarm information, it notifies the lighting agent to turn off non-emergency lighting and notifies the ventilation agent to switch to smoke exhaust mode.