Intelligent control system and method for weak current equipment based on internet of things

Abstract

The application discloses a weak current equipment intelligent control system and method based on an Internet of Things, and relates to the technical field of weak current equipment control.The application constructs a path health score function through topology modeling, path scoring, health feature analysis and supervised learning training, realizes intelligent selection and switching of a main path, and continuously monitors node states during task execution, identifies abnormal nodes and dynamically replaces paths to guarantee control continuity.If path switching fails, the application automatically traces fault root causes, isolates high-risk nodes, and links an operation and maintenance platform to intervene in processing, thereby significantly improving the adaptability, self-healing capability and engineering operation efficiency of the weak current control system.

Description

Technical Field

[0001] This invention relates to the field of low-voltage equipment control technology, specifically an intelligent control system and method for low-voltage equipment based on the Internet of Things. Background Technology

[0002] In existing IoT control systems, the operation of low-voltage equipment relies on a fixed communication topology and control link structure. Once a node in the system fails or communication is interrupted, the entire control path will be interrupted, which will seriously affect the normal operation of critical equipment. Especially in high-reliability scenarios such as hospitals, the continuous operation of low-voltage equipment such as lighting, security, and emergency broadcasting is crucial. However, traditional systems lack real-time perception of node status and a rapid recovery mechanism after a failure, making the control link vulnerable.

[0003] Current low-voltage control architectures mostly adopt unidirectional control paths or master-slave switching mechanisms, which have problems such as response lag, unstable switching, and heavy manual intervention in the reconfiguration process. The overall robustness and automatic recovery capability of the system are insufficient. Although some systems support node monitoring, they lack the self-healing capability for dynamic reconfiguration of control links. The system cannot achieve continuous control when facing complex dynamic environments, and there are serious stability risks.

[0004] Therefore, it is necessary to build an intelligent control mechanism that supports node health monitoring, path reconstruction, and adaptive adjustment of control logic, so that when any control node or communication unit fails, the system can automatically detect, replan the control chain, and quickly restore the control function, thereby significantly improving the reliability, stability, and self-recovery capability of the weak current control network. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent control system and method for low-voltage electrical equipment based on the Internet of Things, so as to solve the problems raised in the prior art.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a smart control method for low-voltage electrical equipment based on the Internet of Things, the method comprising:

[0007] Step S100: Set the low-voltage equipment as nodes, summarize the physical connection relationship and control logic channel of each node, generate the topology diagram of the low-voltage control network, and generate the control path set corresponding to different task types in different functional areas through historical records.

[0008] Step S200: By obtaining the node status table, calculate the path score of the historical records. Based on the average path score of each control path, obtain the main control path. At the same time, obtain the historical records of each control path, calculate the path stability index, communication latency index, and node health index of the historical records, and construct a path health feature set and fit it with the task execution results to generate a path health scoring function.

[0009] Step S300: Set a monitoring cycle, calculate the health score of each stage when executing the current task, and determine whether the path switching determination process is triggered. If the path switching determination process is not triggered, continue to collect data. If the path switching determination process is triggered, filter the candidate control paths and synchronize the candidate control paths that meet the conditions to the control center.

[0010] Step S400: In the control center, determine the new primary control path, activate the policy copy of the new primary control path, set the initial duration of the switchover, and determine whether the new primary control path meets the preset consistency conditions. If it does, continue to execute the new primary control path; otherwise, trigger the fault handling mechanism.

[0011] Step S500: During path switching and task recovery, nodes marked as abnormal to be confirmed are identified. By monitoring the running status data of the nodes marked as abnormal to be confirmed and drawing a fluctuation feature map, it is determined whether the nodes marked as abnormal to be confirmed are high-risk nodes. If they are high-risk nodes, maintenance personnel are reminded to handle them.

[0012] Furthermore, step S100 includes:

[0013] Step S101: Set all low-voltage electrical equipment deployed in the hospital building as nodes, and obtain the physical connection relationship and control logic channel of each low-voltage electrical equipment by parsing the wiring diagram, control system configuration table and communication link log;

[0014] Step S102: Group devices with the same superior control node into the same functional group, construct the logical connection path between each node, and generate the topology diagram of the weak current control network.

[0015] Step S103: In the topology diagram, collect the basic attributes of each node, including node type, unique identification information, communication parameters, and control load type. Write the basic attributes into the node's status table to generate the status table of all nodes.

[0016] Step S104: In the control center, obtain the historical records of task scheduling, extract the task types and functional areas corresponding to the weak current equipment executing the scheduling from the historical records, map the corresponding relationship of task paths, and generate a set of feasible control paths in combination with the current topology.

[0017] By analyzing the wiring diagrams, control system configuration tables, and communication link logs inside the hospital building, the physical connection relationships and control channel mapping relationships between various low-voltage devices are automatically extracted. This avoids the traditional method of relying on manual configuration and realizes the structured and digital management of the node relationships of the low-voltage system, effectively improving modeling efficiency and accuracy.

[0018] By classifying devices with the same superior control node by function, a functional grouping system based on control attribution logic is constructed, which makes the subsequent deployment of control strategies and path planning more in line with the actual control scenarios of the devices, and improves the aggregation and effectiveness of control strategy execution.

[0019] By collecting the basic attributes of each low-voltage node, including node type, identification information, communication parameters, load type, etc., and uniformly writing them into the status table, the structured modeling and topology visualization of the entire low-voltage control network are realized, providing high-quality structural support for subsequent system status perception, health assessment and path scheduling.

[0020] By acquiring historical task scheduling records from the control center, extracting task types, scheduling devices and their respective functional areas, and combining them with the current topology, a mapping relationship between historical tasks and feasible control paths is generated, thereby constructing a set of task paths oriented towards functional areas. This set can provide input for modules such as path scoring, health analysis, and path switching, enhancing the system's dynamic adaptability.

[0021] By implementing the method of this invention, the system can automatically complete network topology modeling, state table construction and path set generation without human intervention, providing data and structural support for subsequent control path scheduling, risk assessment and fault tolerance processing, thereby significantly improving the overall operational controllability, management transparency and intelligence of the low-voltage control system.

[0022] Furthermore, step S200 includes:

[0023] Step S201: Obtain the historical record set of a certain task type executed in a certain functional area, extract the control path used to execute the task type in the historical record, classify the historical records of the same control path, collect the status table of each node involved in the historical record, extract the communication parameters in the status table, perform normalization processing, and calculate the path score according to the following formula:

[0024] ;

[0025] Where A represents the path score, B da C is represented as the normalized value of the a-th communication parameter in the d-th node. a Let D be the weight of the a-th communication parameter.d Let d represent the weight of the d-th node, b represent the total number of communication parameters, and e represent the total number of nodes.

[0026] Step S202: Obtain the path score corresponding to all historical records in a certain control path, calculate the average path score of the control path, sort the control paths used to execute a certain task type in a certain functional area from high to low according to the average path score, obtain a set of candidate control paths, and set the control path with the highest average path score as the main control path.

[0027] Step S203: Obtain the historical records of each control path, collect the task execution result, communication latency, and node communication parameters for each historical record, count the number of successfully executed tasks as H1 and the total number of historical records as H2, calculate the path stability index as P1 = H1 / H2, obtain the average communication latency of all communication links, where each communication link is the connection between every two nodes, and calculate the communication latency index as... , where t h Let be the average communication delay of the h-th communication link, and j be the total number of communication links. The node's communication parameters are normalized, and the node's health score is calculated using the following node health score formula:

[0028] ;

[0029] Where E represents the node's health score, B f C is represented as the normalized value of the f-th communication parameter. f The weight of the f-th communication parameter is represented by the sum of the health scores of all nodes. The average health score is calculated and set as the node health index.

[0030] Step S204: Normalize and summarize the path stability index, communication latency index, and node health index of the historical records to construct a path health feature set. Use the path health feature set of each historical record as input and the corresponding task execution result as label to construct a supervised learning sample set. Fit and train the following path health scoring function using the least squares method to train the weights of the path health scoring function. The path health scoring function is: K=Q1×P1+Q2×P2+Q3×P3, where K represents the path health score, and Q1, Q2, and Q3 represent the weights of the trained path stability index, communication latency index, and node health index, respectively.

[0031] By classifying, analyzing, and modeling the performance data of each control path in the historical task execution records, a multi-dimensional evaluation system including path scoring, stability indicators, communication latency indicators, and node health indicators was established. This system makes the quality of control paths measurable and quantifiable, providing a scientific basis for path selection, scheduling, and switching.

[0032] In the path health assessment process, not only are the overall stability of the path, such as success rate and average delay, statistically analyzed, but also node-level communication parameters are introduced as input to construct a node health scoring mechanism. This can accurately reflect potential local hazards in the path, improve the system's ability to identify fine-grained anomalies, and enhance the comprehensiveness and sensitivity of the control path assessment.

[0033] The method summarizes key indicators such as path stability, communication latency, and node health into a unified path health feature set, and combines them with historical task execution results as labels to construct a supervised learning dataset. This makes the path scoring process no longer dependent on static configuration or empirical weights, but based on actual running results for model training, resulting in higher scene adaptability and real-time accuracy.

[0034] By supervising the learning model of the health feature set, this invention uses the least squares method to fit and train the path health scoring function to obtain the dynamic weights of different indicators under different task scenarios, so that the scoring results can be accurately matched to the task type, improving the adaptability and control reliability of the main path selection.

[0035] By utilizing the path health score output by the scoring function, the system can dynamically select the path with the highest score as the main control path and construct a candidate path set for scheduling and invocation, thereby forming a path fault tolerance mechanism based on health judgment, which effectively improves the self-healing capability and task continuity assurance capability of the weak current control network in complex environments.

[0036] Furthermore, step S300 includes:

[0037] Step S301: Set a preset monitoring period, obtain the current task type and functional area, determine the current main control path, collect the communication parameters of all nodes on the main control path according to the preset monitoring period, perform normalization processing, and calculate the node health score according to the node health score formula.

[0038] Step S302: Preset a health score threshold, compare the health score of each node with the health score threshold, if the health score of a node is lower than the health score threshold, then set the node as an unhealthy node. If any node in the main control path is an unhealthy node, the system triggers a path switching determination process. If there are no unhealthy nodes in the main control path, the system continues to collect data according to the preset monitoring cycle.

[0039] Step S303: After the system triggers the path switching determination process, it traverses the candidate control path set, obtains the communication parameters of each node in the candidate control path, calculates the health score of each node, removes candidate control paths with unhealthy nodes, calculates the path stability index, communication latency index, and node health index of the remaining candidate control paths, and inputs them into the path health scoring function to calculate the path health score of the remaining candidate control paths. The remaining candidate control paths are then sorted from high to low according to the path health score, and the sorted set of remaining candidate control paths is synchronized to the control center.

[0040] Through a preset monitoring cycle, the system can collect the communication parameters of all nodes on the main control path in real time, and analyze them based on standardized processing and health scoring functions to build a quantitative model of the main path's operating status. This enables the system to continuously perceive the path's health status during task execution, effectively improving the system's ability to perceive its operating status.

[0041] By comparing node health scores with preset thresholds, the operating status of individual nodes in the main path can be independently evaluated. When the health score of a node is lower than the threshold, it is considered a potential fault risk, thereby triggering the path switching mechanism. Compared with the traditional "whole path anomaly judgment", it has higher sensitivity and local fault identification capability, reducing the risk of task interruption.

[0042] When the main path is at risk, the system can promptly initiate the path switching judgment process. By performing a secondary screening of the node health status of candidate paths, paths with unhealthy nodes are excluded, thereby improving the overall availability of candidate paths and ensuring that the switching path has higher execution reliability.

[0043] By calculating the path stability, communication latency, and node health metrics of the remaining candidate paths and inputting them into a trained path health scoring function, the system can comprehensively score and prioritize the candidate paths, select the optimal path, and synchronize it to the control center. This mechanism ensures that path switching is scientific, data-driven, and adaptable to various scenarios.

[0044] By implementing a complete closed-loop mechanism from problem discovery to fault avoidance through real-time monitoring, intelligent identification of abnormal nodes, initiation of healthy path screening, optimal path sorting, and synchronization with the control center, a highly automated and self-healing mechanism is achieved, significantly improving the task continuity assurance capability of the low-voltage control network in the event of sudden node anomalies or communication fluctuations.

[0045] Furthermore, step S400 includes:

[0046] Step S401: In the control center, select the control path with the highest path health score from the remaining candidate control path set and set it as the new main control path. Call the policy template of the current task type through the control center, generate a control policy copy based on the target control node in the new main control path, and distribute the policy copy to the target control node through the rapid deployment mechanism. Mark the policy copy as pending activation.

[0047] Step S402: Extract task context information from the node executing the current task in the original main control path. The task context information includes task status variables, sensor feedback status, and the output of the last action. Write the extracted task context information into the task context middleware for caching, and transmit the task context information synchronously to the target control node in the new main control path that deploys the strategy copy through the control channel.

[0048] Step S403: Set the policy copy in the new master control path to the active state, resume task execution, preset the initial switching duration, and perform output consistency verification by comparing the control commands output by the new master control path with the output of the original master control path in the corresponding state. If the output consistency verification meets the preset consistency conditions, the task execution of the new master control path will continue to be maintained. If the output consistency verification does not meet the preset consistency conditions, the fault handling mechanism will be triggered.

[0049] Step S404: When the fault handling mechanism is triggered, determine whether the original primary control path is available. If the original primary control path is available, switch back to the original primary control path and reactivate the original control policy copy to resume task execution. If the original primary control path is unavailable, select the path with the second highest path health score as the new primary control path based on the remaining candidate control path set, and repeat the steps of deploying control policy copy, synchronizing task context, and activating policy copy until task recovery is completed.

[0050] By automatically selecting the path with the highest health score from the remaining candidate paths and calling the corresponding policy template based on the task type, a new control policy copy is generated and distributed to the target node. This enables rapid replacement of the main path and policy deployment without interrupting task execution, significantly improving the system's response speed and policy adaptability.

[0051] During path switching, the task context of the original path, including state variables, sensor feedback, and the last action, is extracted and synchronously transmitted to the new path. This effectively avoids the context breakpoint problem caused by path switching, achieves seamless migration of task state, and provides necessary support for the continuous execution of complex tasks.

[0052] In the initial stage of activation of the new path policy copy, the system compares the control output of the new and old main paths under the same context state to make a consistency judgment. If the verification passes, the new path is confirmed to be valid. If it fails, the fault handling mechanism is triggered immediately to effectively prevent the risk of miscontrol caused by policy mismatch or path distortion and improve the robustness and security of the system during the switching process.

[0053] When the verification of the new primary path fails, the system will first determine the availability of the original primary path. If it is available, it will roll back and restore the original control strategy. If it is unavailable, it will try the second-best path in order according to the candidate path score, and repeat the deployment, synchronization and activation steps until the task is successfully restored, thus forming a complete closed-loop path switching and task recovery mechanism.

[0054] It can implement control and protection strategies that ensure uninterrupted tasks, unbroken strategies, and recoverable paths even in abnormal situations such as node anomalies and communication fluctuations. This gives the low-voltage control network a high degree of adaptability, fault tolerance, and continuous control capability, making it particularly suitable for scenarios with extremely high requirements for control stability and continuity, such as hospitals, rail transit, and data centers.

[0055] Furthermore, step S500 includes:

[0056] Step S501: During path switching and task recovery, if several output consistency verification failures, repeated rollbacks of control tasks, and path selection failures are detected, the system obtains the association relationship between the abnormal triggering node and its upstream and downstream nodes based on path switching records, abnormal logs, and control path topology, constructs a set of suspected faulty nodes, and marks the abnormal triggering node as an abnormal node to be confirmed.

[0057] Step S502: Monitor the abnormal node to be confirmed, continuously collect the running status data of the abnormal node to be confirmed, the running status data includes communication delay, instruction response time, control output offset, upstream and downstream communication success rate, and draw the fluctuation feature map of the running status data within the preset judgment period.

[0058] Step S503: Based on the fluctuation characteristic map of the running status data, calculate the fluctuation slope of the running status data, preset the fluctuation slope threshold, and when the fluctuation slope exceeds the fluctuation slope threshold, update the abnormal node to be confirmed as a high-risk node and perform logical isolation operation. At the same time, push the high-risk node to the operation and maintenance management platform to prompt maintenance personnel to intervene for diagnosis and handling.

[0059] When the control path repeatedly fails during switching and task recovery, the system no longer stays on the surface path switching logic, but actively backtracks the path switching records and exception logs, and combines the control path topology to extract the set of key nodes related to the exception trigger, so as to initially lock and mark the root cause node of the exception, laying the foundation for subsequent refined diagnosis.

[0060] For nodes that are yet to be confirmed as abnormal, the system can continuously collect their operational status data, including multi-dimensional performance indicators such as communication latency, command response time, control output offset, and upstream and downstream communication success rate, and construct an operational status fluctuation graph. The system can comprehensively reflect the node performance fluctuation trend and realize the advance monitoring and quantitative assessment of potential risks.

[0061] By calculating the slope of the fluctuation of the node's running status data and comparing it with a preset threshold, this method can identify whether the node has a continuous abnormal fluctuation trend in the time series. Compared with the traditional anomaly judgment method based on a single threshold, this method has stronger trend analysis capability and dynamic anomaly recognition capability, and can discover potential risk nodes earlier.

[0062] Once a high-risk node is identified, the system can immediately perform logical isolation operations to block its participation in the current control path and prevent it from continuously affecting task execution. At the same time, the status of the node will be pushed to the operation and maintenance management platform to prompt maintenance personnel to conduct on-site intervention and physical investigation, thereby realizing a closed-loop operation and maintenance support system from automatic identification → automatic isolation → manual intervention.

[0063] This enables low-voltage control systems to possess intelligent operation and maintenance capabilities such as self-diagnosis, self-identification, and self-isolation, effectively reducing systemic problems such as frequent path switching and task recovery failures caused by node failures, improving the system's resilience in fault environments, and providing structured technical support for the engineering operation and maintenance of large-scale complex control networks.

[0064] To better implement the above methods, an intelligent control system for low-voltage equipment based on the Internet of Things is also proposed. The system includes a topology modeling module, a control path analysis module, a task monitoring module, a switching judgment module, and an abnormal node identification module.

[0065] Topology modeling module: Sets low-voltage equipment as nodes, summarizes the physical connection relationship and control logic channel of each node, generates a topology diagram of the low-voltage control network, and generates a set of control paths corresponding to different task types in different functional areas through historical records.

[0066] Control path analysis module: By acquiring the status table of the nodes, the path score of the historical records is calculated. Based on the average path score of each control path, the main control path is obtained. At the same time, the historical records of each control path are acquired, and the path stability index, communication latency index, and node health index of the historical records are calculated. The path health feature set is constructed and fitted with the task execution results for training to generate the path health scoring function.

[0067] Task monitoring module: It has a preset monitoring cycle, calculates the health score of each stage when the current task is executed, and determines whether the path switching determination process is triggered. If the path switching determination process is not triggered, it continues to collect data. If the path switching determination process is triggered, it filters the candidate control paths and synchronizes the candidate control paths that meet the conditions to the control center.

[0068] Switching judgment module: In the control center, determine the new primary control path, activate the policy copy of the new primary control path, set the initial duration of the switch, and determine whether the new primary control path meets the preset consistency conditions. If it does, continue to execute the new primary control path; if it does not, trigger the fault handling mechanism.

[0069] Abnormal Node Identification Module: During path switching and task recovery, nodes marked as abnormal nodes to be confirmed are identified. By monitoring the running status data of the nodes marked as abnormal nodes to be confirmed and drawing fluctuation feature maps, it is determined whether the nodes to be confirmed are high-risk nodes. If they are high-risk nodes, maintenance personnel are alerted to take action.

[0070] Furthermore, the topology modeling module includes topology structure units and control path set units:

[0071] Topology Unit: All low-voltage electrical equipment deployed in the hospital building is set as nodes. By parsing the wiring diagram, control system configuration table and communication link log, the physical connection relationship and control logic channel of each low-voltage electrical equipment are obtained. Equipment with the same superior control node is classified into the same functional group, the logical connection path between each node is constructed, and the topology diagram of the low-voltage control network is generated.

[0072] Control path set unit: In the topology diagram, basic attributes of each node are collected, including node type, unique identification information, communication parameters, and control load type. These basic attributes are written into the node's status table to generate a status table for all nodes. In the control center, the historical records of task scheduling are obtained, and the task types and functional areas corresponding to the low-voltage equipment executing the scheduling are extracted from the historical records. The task path correspondence is mapped, and combined with the current topology, a feasible control path set is generated.

[0073] Furthermore, the task monitoring module includes a path switching determination unit and a unit for determining remaining candidate control paths:

[0074] Path switching determination unit: It has a preset monitoring period, acquires the currently executing task type and functional area, and determines the current main control path. According to the preset monitoring period, it collects the communication parameters of all nodes on the main control path and performs normalization processing. It calculates the node's health score according to the node health score formula, presets a health score threshold, and compares the health score of each node with the health score threshold. If the health score of a node is lower than the health score threshold, the node is set as an unhealthy node. If any node in the main control path is an unhealthy node, the system triggers the path switching determination process. If there are no unhealthy nodes in the main control path, the system continues to collect data according to the preset monitoring period.

[0075] Determine the remaining candidate control path units: After the system triggers the path switching determination process, it traverses the candidate control path set, obtains the communication parameters of each node in the candidate control path, calculates the health score of each node, eliminates candidate control paths with unhealthy nodes, calculates the path stability index, communication latency index, and node health index of the remaining candidate control paths, and inputs them into the path health scoring function to calculate the path health score of the remaining candidate control paths. The remaining candidate control paths are then sorted from high to low according to the path health score, and the sorted set of remaining candidate control paths is synchronized to the control center.

[0076] Furthermore, the abnormal node identification module includes units for marking abnormal nodes to be confirmed and high-risk node units:

[0077] Marking unconfirmed abnormal node units: During path switching and task recovery, if several output consistency verification failures, repeated rollbacks of control tasks, and path selection failures are detected, the system obtains the association relationship between the abnormal trigger node and its upstream and downstream nodes based on path switching records, abnormal logs, and control path topology, constructs a set of suspected fault nodes, and marks the abnormal trigger node as an unconfirmed abnormal node.

[0078] High-risk node unit: Monitors nodes awaiting confirmation of abnormality, continuously collects operational status data of these nodes, including communication latency, command response time, control output offset, and upstream / downstream communication success rate. Within a preset judgment period, it plots a fluctuation characteristic map of the operational status data, calculates the fluctuation slope based on the map, and sets a preset fluctuation slope threshold. When the fluctuation slope exceeds the threshold, the node awaiting confirmation of abnormality is updated to a high-risk node, and logical isolation is performed. Simultaneously, the high-risk node is pushed to the operation and maintenance management platform, prompting maintenance personnel to intervene for diagnosis and handling.

[0079] Compared with the prior art, the beneficial effects of the present invention are:

[0080] 1. Unlike traditional low-voltage control systems that rely solely on static path configuration, this invention constructs a complete topology diagram and path health scoring mechanism, and combines historical records to evaluate the stability, latency, and node health of each control path, thereby achieving dynamic path optimization and automatic switching during task execution, effectively avoiding the risk of task interruption due to a single path failure.

[0081] 2. This invention is the first to integrate multi-dimensional parameters such as path stability, communication latency, and node health into a unified path scoring model, which has trainable and adaptive characteristics, realizing the transformation of path selection from "rule-driven" to "data-driven", and is more flexible and generalizable than strategies based on fixed thresholds or manual configuration.

[0082] 3. Traditional systems often rely on manual troubleshooting or passive alarms. This invention is based on a suspected fault node identification mechanism based on the slope of the operating status fluctuation. It can reverse the set of problem nodes from abnormal events such as path switching failures and make judgments by combining time series data graphs. This allows for early exposure and proactive isolation of equipment problems, greatly improving system stability and fault diagnosis efficiency.

[0083] 4. Compared with existing systems that lack closed-loop linkage mechanisms in path switching and fault handling, this invention constructs a complete task scheduling closed loop through the combination of steps throughout the entire process. It can not only adaptively select the optimal control path, but also perform multi-level responses such as rapid recovery, alarm, and isolation when anomalies occur, thereby realizing the autonomous evolution capability of the control system. Attached Figure Description

[0084] Figure 1 This is a flowchart illustrating an intelligent control method for low-voltage electrical equipment based on the Internet of Things according to the present invention.

[0085] Figure 2 This is a schematic diagram of the structure of an intelligent control system for low-voltage equipment based on the Internet of Things according to the present invention. Detailed Implementation

[0086] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0087] Please see Figure 1 and Figure 2This invention provides a technical solution: an intelligent control method for low-voltage electrical equipment based on the Internet of Things, the method comprising:

[0088] Step S100: Set the low-voltage equipment as nodes, summarize the physical connection relationship and control logic channel of each node, generate the topology diagram of the low-voltage control network, and generate the control path set corresponding to different task types in different functional areas through historical records.

[0089] Step S100 includes:

[0090] Step S101: Set all low-voltage electrical equipment deployed in the hospital building as nodes, and obtain the physical connection relationship and control logic channel of each low-voltage electrical equipment by parsing the wiring diagram, control system configuration table and communication link log;

[0091] Step S102: Group devices with the same superior control node into the same functional group, construct the logical connection path between each node, and generate the topology diagram of the weak current control network.

[0092] Step S103: In the topology diagram, collect the basic attributes of each node, including node type, unique identification information, communication parameters, and control load type. Write the basic attributes into the node's status table to generate the status table of all nodes.

[0093] Step S104: In the control center, obtain the historical records of task scheduling, extract the task types and functional areas corresponding to the weak current equipment executing the scheduling from the historical records, map the corresponding relationship of task paths, and generate a set of feasible control paths in combination with the current topology.

[0094] For example, taking the inpatient building of a top-tier general hospital as a specific application scenario, the building is equipped with multiple low-voltage electronic systems, including intelligent lighting systems, security monitoring systems, elevator control systems, call intercom systems, environmental monitoring systems, etc. The terminal devices under each system are connected to the controller through physical communication lines, forming a complex low-voltage control network.

[0095] First, all low-voltage electrical devices in the hospital that have controllability or can be controlled are set as network nodes, including lighting control modules, cameras, elevator control boards, temperature and humidity sensors, call hosts, etc. By analyzing the hospital's intelligent construction wiring diagrams, control system configuration lists and communication link logs, the physical connection relationships and logical channel relationships between these nodes are obtained. For example, it is found that the bedside lamps in the 5F ward area are connected to the 5F lighting controller through the KNX bus and then converge to the central control host.

[0096] Based on the hierarchical structure of the control system, devices with the same upper-level control node are divided into the same functional group. For example, all bedside lamps and corridor lights connected to the 5F lighting controller are classified as "5F lighting subsystem". On this basis, the system generates a control logic path from the control center to each terminal node through logical deduction and connection tracing, and finally obtains a structured weak current control network topology diagram.

[0097] Collect basic attributes for each node in the topology diagram, including: node type such as switch, sensor, controller; unique identification information such as MAC address, device number; communication parameters such as communication protocol, baud rate; and control load type such as lighting, audio / video, and paging.

[0098] After data collection is completed, a node status table is generated for subsequent path construction and health analysis. For example, the A001 node of the bedside lamp in the 5F lighting subsystem records the following information: KNX protocol, address 1 / 0 / 15, and type is lighting actuator.

[0099] By reading the historical records of the control center's dispatch platform and extracting the scheduling process of the "Night Patrol Task," it was found that the task involved controlling the corridor lights and cameras in the 5th and 6th floor ward areas. Based on the existing topology, the system automatically calculated multiple feasible task control paths, such as:

[0100] Control path 1: Central host → 5F lighting controller → Corridor light L051

[0101] Control path 2: Central host → Security main control platform → Camera C061.

[0102] Step S200: By obtaining the node status table, calculate the path score of the historical records. Based on the average path score of each control path, obtain the main control path. At the same time, obtain the historical records of each control path, calculate the path stability index, communication latency index, and node health index of the historical records, and construct a path health feature set and fit it with the task execution results to generate a path health scoring function.

[0103] Step S200 includes:

[0104] Step S201: Obtain the historical record set of a certain task type executed in a certain functional area, extract the control path used to execute the task type in the historical record, classify the historical records of the same control path, collect the status table of each node involved in the historical record, extract the communication parameters in the status table, perform normalization processing, and calculate the path score according to the following formula:

[0105] ;

[0106] Where A represents the path score, B da C is represented as the normalized value of the a-th communication parameter in the d-th node. a Let D be the weight of the a-th communication parameter. d Let d represent the weight of the d-th node, b represent the total number of communication parameters, and e represent the total number of nodes.

[0107] Step S202: Obtain the path score corresponding to all historical records in a certain control path, calculate the average path score of the control path, sort the control paths used to execute a certain task type in a certain functional area from high to low according to the average path score, obtain a set of candidate control paths, and set the control path with the highest average path score as the main control path.

[0108] Step S203: Obtain the historical records of each control path, collect the task execution result, communication latency, and node communication parameters for each historical record, count the number of successfully executed tasks as H1 and the total number of historical records as H2, calculate the path stability index as P1 = H1 / H2, obtain the average communication latency of all communication links, where each communication link is the connection between every two nodes, and calculate the communication latency index as... , where t h Let be the average communication delay of the h-th communication link, and j be the total number of communication links. The node's communication parameters are normalized, and the node's health score is calculated using the following node health score formula:

[0109] ;

[0110] Where E represents the node's health score, B f C is represented as the normalized value of the f-th communication parameter. f The weight of the f-th communication parameter is represented by the sum of the health scores of all nodes. The average health score is calculated and set as the node health index.

[0111] Step S204: Normalize and summarize the path stability index, communication latency index, and node health index of the historical records to construct a path health feature set. Use the path health feature set of each historical record as input and the corresponding task execution result as label to construct a supervised learning sample set. Fit and train the following path health scoring function using the least squares method to train the weights of the path health scoring function. The path health scoring function is: K=Q1×P1+Q2×P2+Q3×P3, where K represents the path health score, and Q1, Q2, and Q3 represent the weights of the trained path stability index, communication latency index, and node health index, respectively.

[0112] For example, the collected communication parameter values ​​are shown in Table 1:

[0113]

[0114] Table 1

[0115] The normalized values ​​are shown in Table 2:

[0116]

[0117] Table 2

[0118] The path score obtained through weighted calculation is 0.281;

[0119] The average score of the historical score sets for control path 1 and control path 2 is taken.

[0120] The average path score for control path 1 is 0.285, and the average path score for control path 2 is 0.276. Therefore, control path 1 is the primary control path.

[0121] Assuming the system retrieves historical execution records for control path 1, totaling 30 tasks with 28 successes and 30 historical records, the path stability index is 0.933. Control path 1 includes two communication links: the average communication latency of the central host → 5F lighting controller communication link is 13, and the average communication latency of the 5F lighting controller → corridor light L051 communication link is 15, resulting in a communication latency index of 14. Based on the data in Table 2, the node health index is calculated to be 0.305.

[0122] Assume the historical task execution records are as shown in Table 3:

[0123]

[0124] Table 3

[0125] The Min-Max normalization formula is used to process the feature values ​​of each column. The above samples are trained using least squares regression, and the fitted weights are shown in the example: Q1=0.55, Q2=−0.30 (indicating that a high proportion of communication delay leads to a lower health score), and Q3=0.45.

[0126] Step S300: Set a monitoring cycle, calculate the health score of each stage when executing the current task, and determine whether the path switching determination process is triggered. If the path switching determination process is not triggered, continue to collect data. If the path switching determination process is triggered, filter the candidate control paths and synchronize the candidate control paths that meet the conditions to the control center.

[0127] Step S300 includes:

[0128] Step S301: Set a preset monitoring period, obtain the current task type and functional area, determine the current main control path, collect the communication parameters of all nodes on the main control path according to the preset monitoring period, perform normalization processing, and calculate the node health score according to the node health score formula.

[0129] Step S302: Preset a health score threshold, compare the health score of each node with the health score threshold, if the health score of a node is lower than the health score threshold, then set the node as an unhealthy node. If any node in the main control path is an unhealthy node, the system triggers a path switching determination process. If there are no unhealthy nodes in the main control path, the system continues to collect data according to the preset monitoring cycle.

[0130] Step S303: After the system triggers the path switching determination process, it traverses the candidate control path set, obtains the communication parameters of each node in the candidate control path, calculates the health score of each node, removes candidate control paths with unhealthy nodes, calculates the path stability index, communication latency index, and node health index of the remaining candidate control paths, and inputs them into the path health scoring function to calculate the path health score of the remaining candidate control paths. The remaining candidate control paths are then sorted from high to low according to the path health score, and the sorted set of remaining candidate control paths is synchronized to the control center.

[0131] For example, preset cycle: once every 10 minutes, current task type: night patrol task, current functional area: 5F ward area, current main control path: control path 1;

[0132] The communication parameters collected during this period, after normalization, are shown in Table 4:

[0133]

[0134] Table 4

[0135] The calculated health score of the control center is 0.34, the health score of the 5F lighting controller is 0.685, and the health score of the corridor light L051 is 0.89. The preset health score threshold is 0.7. Therefore, the control center and the 5F lighting controller are unhealthy nodes, and the system triggers the path switching judgment process.

[0136] The candidate control path set is as follows: Control path 2: Control center → Security main control platform → Camera C061; Control path 3: Central host → Environmental controller → Temperature and humidity detector TMP501;

[0137] The normalized data for control path 2 is shown in Table 5:

[0138]

[0139] Table 5

[0140] The calculated health score of the control center is 0.34, the health score of the security main control platform is 0.42, and the health score of camera C061 is 0.57. There are unhealthy nodes, so this control path is removed.

[0141] Control path 3 calculates the health score of the central host to be 0.71, the health score of the environmental controller to be 0.78, and the health score of the temperature and humidity detector TMP501 to be 0.82. Therefore, control path 3 is retained.

[0142] After normalization, the path stability index, communication latency index, and node health index in control path 3 are 1, 0, and 0.833, respectively, and the calculated path health score is 0.925.

[0143] Step S400: In the control center, determine the new primary control path, activate the policy copy of the new primary control path, set the initial duration of the switchover, and determine whether the new primary control path meets the preset consistency conditions. If it does, continue to execute the new primary control path; otherwise, trigger the fault handling mechanism.

[0144] Step S400 includes:

[0145] Step S401: In the control center, select the control path with the highest path health score from the remaining candidate control path set and set it as the new main control path. Call the policy template of the current task type through the control center, generate a control policy copy based on the target control node in the new main control path, and distribute the policy copy to the target control node through the rapid deployment mechanism. Mark the policy copy as pending activation.

[0146] Step S402: Extract task context information from the node executing the current task in the original main control path. The task context information includes task status variables, sensor feedback status, and the output of the last action. Write the extracted task context information into the task context middleware for caching, and transmit the task context information synchronously to the target control node in the new main control path that deploys the strategy copy through the control channel.

[0147] Step S403: Set the policy copy in the new master control path to the active state, resume task execution, preset the initial switching duration, and perform output consistency verification by comparing the control commands output by the new master control path with the output of the original master control path in the corresponding state. If the output consistency verification meets the preset consistency conditions, the task execution of the new master control path will continue to be maintained. If the output consistency verification does not meet the preset consistency conditions, the fault handling mechanism will be triggered.

[0148] Step S404: When the fault handling mechanism is triggered, determine whether the original primary control path is available. If the original primary control path is available, switch back to the original primary control path and reactivate the original control policy copy to resume task execution. If the original primary control path is unavailable, select the path with the second highest path health score as the new primary control path based on the remaining candidate control path set, and repeat the steps of deploying control policy copy, synchronizing task context, and activating policy copy until task recovery is completed.

[0149] For example, the system calls the task strategy template function, which includes patrol execution logic, lighting judgment, environmental feedback collection, and feedback forwarding. According to the new path structure, the target control node is: temperature and humidity sensor TMP501. The strategy copy will be deployed on the TMP501 node side and will be activated after context synchronization.

[0150] The system extracts the task execution context from the original main path as the patrol stage variables: 2 Patrol to the west section of the corridor; the last device output was that the light was on for 5 seconds; the sensor feedback was that the light was turned on successfully, and the reporting delay was 120ms; the above data is written into the task context middleware and synchronously transmitted to the node where TMP501 is located through the control channel. The environment controller reads the context data and prepares for policy activation.

[0151] Initial switching phase: Set to a 60-second observation window.

[0152] Table 6 shows the consistency between the system's output of control path 3 and the historical output of control path 1 under the same task state:

[0153]

[0154] Table 6

[0155] Although the commands are different—one is a light-on command and the other is an environment-aware command—they both meet the mission objective of "visualizing the patrol status + confirming the status," and are therefore considered to have passed the functional consistency verification.

[0156] Step S500: During path switching and task recovery, nodes marked as abnormal to be confirmed are identified. By monitoring the running status data of the nodes marked as abnormal to be confirmed and drawing a fluctuation feature map, it is determined whether the nodes marked as abnormal to be confirmed are high-risk nodes. If they are high-risk nodes, maintenance personnel are reminded to handle them.

[0157] Step S500 includes:

[0158] Step S501: During path switching and task recovery, if several output consistency verification failures, repeated rollbacks of control tasks, and path selection failures are detected, the system obtains the association relationship between the abnormal triggering node and its upstream and downstream nodes based on path switching records, abnormal logs, and control path topology, constructs a set of suspected faulty nodes, and marks the abnormal triggering node as an abnormal node to be confirmed.

[0159] Step S502: Monitor the abnormal node to be confirmed, continuously collect the running status data of the abnormal node to be confirmed, the running status data includes communication delay, instruction response time, control output offset, upstream and downstream communication success rate, and draw the fluctuation feature map of the running status data within the preset judgment period.

[0160] Step S503: Based on the fluctuation characteristic map of the running status data, calculate the fluctuation slope of the running status data, preset the fluctuation slope threshold, and when the fluctuation slope exceeds the fluctuation slope threshold, update the abnormal node to be confirmed as a high-risk node and perform logical isolation operation. At the same time, push the high-risk node to the operation and maintenance management platform to prompt maintenance personnel to intervene for diagnosis and handling.

[0161] For example, the environmental controller and the security main control platform are abnormal nodes to be confirmed. The data collection period is set to 5 minutes, and data is collected once every 5 seconds to draw a fluctuation characteristic map of the operating status data.

[0162] The system automatically performs fluctuation characteristic analysis on the operating data and extracts the fluctuation slope as shown in Table 7:

[0163]

[0164] Table 7

[0165] Set the latency fluctuation slope threshold to 30ms / min, the response time fluctuation slope threshold to 20ms / min, the output offset slope threshold to 0.1 offset / min, and the communication success rate change slope threshold to 5% communication success rate / min;

[0166] The system automatically performs the following operations: removes the security control platform from the set of all path candidates and stops issuing control policies to that node; pushes alarms to the operation and maintenance platform.

[0167] To better implement the above methods, an intelligent control system for low-voltage equipment based on the Internet of Things is also proposed. The system includes a topology modeling module, a control path analysis module, a task monitoring module, a switching judgment module, and an abnormal node identification module.

[0168] Topology modeling module: Sets low-voltage equipment as nodes, summarizes the physical connection relationship and control logic channel of each node, generates a topology diagram of the low-voltage control network, and generates a set of control paths corresponding to different task types in different functional areas through historical records.

[0169] The topology modeling module includes topology structure units and control path set units:

[0170] Topology Unit: All low-voltage electrical equipment deployed in the hospital building is set as nodes. By parsing the wiring diagram, control system configuration table and communication link log, the physical connection relationship and control logic channel of each low-voltage electrical equipment are obtained. Equipment with the same superior control node is classified into the same functional group, the logical connection path between each node is constructed, and the topology diagram of the low-voltage control network is generated.

[0171] Control path set unit: In the topology diagram, the basic attributes of each node are collected. The basic attributes include node type, unique identification information, communication parameters, and control load type. The basic attributes are written into the node's status table to generate a status table for all nodes. In the control center, the historical records of task scheduling are obtained, and the task types and functional areas corresponding to the weak current equipment executing the scheduling are extracted from the historical records. The task path correspondence is mapped, and combined with the current topology, a feasible control path set is generated.

[0172] Control path analysis module: By acquiring the status table of the nodes, the path score of the historical records is calculated. Based on the average path score of each control path, the main control path is obtained. At the same time, the historical records of each control path are acquired, and the path stability index, communication latency index, and node health index of the historical records are calculated. The path health feature set is constructed and fitted with the task execution results for training to generate the path health scoring function.

[0173] Task monitoring module: It has a preset monitoring cycle, calculates the health score of each stage when the current task is executed, and determines whether the path switching determination process is triggered. If the path switching determination process is not triggered, it continues to collect data. If the path switching determination process is triggered, it filters the candidate control paths and synchronizes the candidate control paths that meet the conditions to the control center.

[0174] The task monitoring module includes a path switching determination unit and a unit for determining remaining candidate control paths.

[0175] Path switching determination unit: It has a preset monitoring period, acquires the currently executing task type and functional area, and determines the current main control path. According to the preset monitoring period, it collects the communication parameters of all nodes on the main control path and performs normalization processing. It calculates the node's health score according to the node health score formula, presets a health score threshold, and compares the health score of each node with the health score threshold. If the health score of a node is lower than the health score threshold, the node is set as an unhealthy node. If any node in the main control path is an unhealthy node, the system triggers the path switching determination process. If there are no unhealthy nodes in the main control path, the system continues to collect data according to the preset monitoring period.

[0176] Determine the remaining candidate control path units: After the system triggers the path switching determination process, it traverses the candidate control path set, obtains the communication parameters of each node in the candidate control path, calculates the health score of each node, eliminates candidate control paths with unhealthy nodes, calculates the path stability index, communication latency index, and node health index of the remaining candidate control paths, and inputs them into the path health scoring function to calculate the path health score of the remaining candidate control paths. The remaining candidate control paths are then sorted from high to low according to the path health score, and the sorted set of remaining candidate control paths is synchronized to the control center.

[0177] Switching judgment module: In the control center, determine the new primary control path, activate the policy copy of the new primary control path, set the initial duration of the switch, and determine whether the new primary control path meets the preset consistency conditions. If it does, continue to execute the new primary control path; if it does not, trigger the fault handling mechanism.

[0178] Anomaly Node Identification Module: During path switching and task recovery, nodes marked as anomalies to be confirmed are identified. By monitoring the operational status data of these marked nodes and plotting fluctuation characteristic graphs, the module determines whether the node is a high-risk node. If it is a high-risk node, maintenance personnel are alerted to take action.

[0179] The abnormal node identification module includes units for marking abnormal nodes to be confirmed and high-risk node units:

[0180] Marking unconfirmed abnormal node units: During path switching and task recovery, if several output consistency verification failures, repeated rollbacks of control tasks, and path selection failures are detected, the system obtains the association relationship between the abnormal trigger node and its upstream and downstream nodes based on path switching records, abnormal logs, and control path topology, constructs a set of suspected fault nodes, and marks the abnormal trigger node as an unconfirmed abnormal node.

[0181] High-risk node unit: Monitors nodes awaiting confirmation of abnormality, continuously collects operational status data of these nodes, including communication latency, command response time, control output offset, and upstream / downstream communication success rate. Within a preset judgment period, it plots a fluctuation characteristic map of the operational status data, calculates the fluctuation slope based on the map, and sets a preset fluctuation slope threshold. When the fluctuation slope exceeds the threshold, the node awaiting confirmation of abnormality is updated to a high-risk node, and logical isolation is performed. Simultaneously, the high-risk node is pushed to the operation and maintenance management platform, prompting maintenance personnel to intervene for diagnosis and handling.

[0182] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for intelligent control of low-voltage equipment based on the Internet of Things, characterized in that, The methods include: Step S100: Set the low-voltage equipment as nodes, summarize the physical connection relationship and control logic channel of each node, generate the topology diagram of the low-voltage control network, and generate the control path set corresponding to different task types in different functional areas through historical records. Step S200: By obtaining the node status table, calculate the path score of the historical records. Based on the average path score of each control path, obtain the main control path. At the same time, obtain the historical records of each control path, calculate the path stability index, communication latency index, and node health index of the historical records, and construct a path health feature set and fit it with the task execution results to generate a path health scoring function. Step S300: Set a monitoring cycle, calculate the health score of each stage when executing the current task, and determine whether the path switching determination process is triggered. If the path switching determination process is not triggered, continue to collect data. If the path switching determination process is triggered, filter the candidate control paths and synchronize the candidate control paths that meet the conditions to the control center. Step S400: In the control center, determine the new primary control path, activate the policy copy of the new primary control path, set the initial duration of the switchover, and determine whether the new primary control path meets the preset consistency conditions. If it does, continue to execute the new primary control path; otherwise, trigger the fault handling mechanism. Step S500: During path switching and task recovery, nodes marked as abnormal to be confirmed are identified. By monitoring the running status data of the nodes marked as abnormal to be confirmed and drawing a fluctuation feature map, it is determined whether the nodes marked as abnormal to be confirmed are high-risk nodes. If they are high-risk nodes, maintenance personnel are reminded to handle them. Step S500 includes the following steps: Step S501: During path switching and task recovery, if several output consistency verification failures, repeated rollbacks of control tasks, and path selection failures are detected, the system obtains the association relationship between the abnormal triggering node and its upstream and downstream nodes based on path switching records, abnormal logs, and control path topology, constructs a set of suspected faulty nodes, and marks the abnormal triggering node as an abnormal node to be confirmed. Step S502: Monitor the abnormal node to be confirmed, continuously collect the running status data of the abnormal node to be confirmed, the running status data includes communication delay, instruction response time, control output offset, upstream and downstream communication success rate, and draw the fluctuation feature map of the running status data within the preset judgment period. Step S503: Calculate the fluctuation slope of the running status data based on the fluctuation characteristic map of the running status data, preset the fluctuation slope threshold, and when the fluctuation slope exceeds the fluctuation slope threshold, update the abnormal node to be confirmed as a high-risk node and perform logical isolation operation. At the same time, push the high-risk node to the operation and maintenance management platform to prompt maintenance personnel to intervene for diagnosis and handling.

2. The intelligent control method for low-voltage equipment based on the Internet of Things according to claim 1, characterized in that, Step S100 includes the following steps: Step S101: Set all low-voltage electrical equipment deployed in the hospital building as nodes, and obtain the physical connection relationship and control logic channel of each low-voltage electrical equipment by parsing the wiring diagram, control system configuration table and communication link log; Step S102: Group devices with the same superior control node into the same functional group, construct the logical connection path between each node, and generate the topology diagram of the weak current control network. Step S103: In the topology diagram, collect the basic attributes of each node, including node type, unique identification information, communication parameters, and control load type. Write the basic attributes into the node's status table to generate the status table of all nodes. Step S104: In the control center, obtain the historical records of task scheduling, extract the task types and functional areas corresponding to the weak current equipment executing the scheduling from the historical records, map the corresponding relationship of task paths, and generate a set of feasible control paths in combination with the current topology.

3. The intelligent control method for low-voltage equipment based on the Internet of Things according to claim 2, characterized in that, Step S200 includes the following steps: Step S201: Obtain the historical record set of a certain task type executed in a certain functional area, extract the control path used to execute the task type in the historical record, classify the historical records of the same control path, collect the status table of each node involved in the historical record, extract the communication parameters in the status table, perform normalization processing, and calculate the path score according to the following formula: ； Where A represents the path score, B da C is represented as the normalized value of the a-th communication parameter in the d-th node. a Let D be the weight of the a-th communication parameter. d Let d represent the weight of the d-th node, b represent the total number of communication parameters, and e represent the total number of nodes. Step S202: Obtain the path score corresponding to all historical records in a certain control path, calculate the average path score of the control path, sort the control paths used to execute a certain task type in a certain functional area from high to low according to the average path score, obtain a set of candidate control paths, and set the control path with the highest average path score as the main control path. Step S203: Obtain the historical records of each control path, collect the task execution result, communication latency, and node communication parameters for each historical record, count the number of successfully executed tasks as H1 and the total number of historical records as H2, calculate the path stability index as P1 = H1 / H2, obtain the average communication latency of all communication links, where each communication link is the connection between every two nodes, and calculate the communication latency index as... , where t h Let be the average communication delay of the h-th communication link, and j be the total number of communication links. The node's communication parameters are normalized, and the node's health score is calculated using the following node health score formula: ； Where E represents the node's health score, B f C is represented as the normalized value of the f-th communication parameter. f The weight of the f-th communication parameter is represented by the sum of the health scores of all nodes. The average health score is calculated and set as the node health index. Step S204: Normalize and summarize the path stability index, communication latency index, and node health index of the historical records to construct a path health feature set. Use the path health feature set of each historical record as input and the corresponding task execution result as label to construct a supervised learning sample set. Fit and train the following path health scoring function using the least squares method to train the weights of the path health scoring function. The path health scoring function is: K=Q1×P1+Q2×P2+Q3×P3, where K represents the path health score, P3 represents the node health index, and Q1, Q2, and Q3 represent the weights of the trained path stability index, communication latency index, and node health index, respectively.

4. The intelligent control method for low-voltage equipment based on the Internet of Things according to claim 3, characterized in that, Step S300 includes the following steps: Step S301: Set a preset monitoring period, obtain the current task type and functional area, determine the current main control path, collect the communication parameters of all nodes on the main control path according to the preset monitoring period, perform normalization processing, and calculate the node health score according to the node health score formula. Step S302: Preset a health score threshold, compare the health score of each node with the health score threshold, if the health score of a node is lower than the health score threshold, then set the node as an unhealthy node. If any node in the main control path is an unhealthy node, the system triggers a path switching determination process. If there are no unhealthy nodes in the main control path, the system continues to collect data according to the preset monitoring cycle. Step S303: After the system triggers the path switching determination process, it traverses the candidate control path set, obtains the communication parameters of each node in the candidate control path, calculates the health score of each node, removes candidate control paths with unhealthy nodes, calculates the path stability index, communication latency index, and node health index of the remaining candidate control paths, and inputs them into the path health scoring function to calculate the path health score of the remaining candidate control paths. The remaining candidate control paths are then sorted from high to low according to the path health score, and the sorted set of remaining candidate control paths is synchronized to the control center.

5. The intelligent control method for low-voltage equipment based on the Internet of Things according to claim 4, characterized in that, Step S400 includes the following steps: Step S401: In the control center, select the control path with the highest path health score from the remaining candidate control path set and set it as the new main control path. Call the policy template of the current task type through the control center, generate a control policy copy based on the target control node in the new main control path, and distribute the policy copy to the target control node through the rapid deployment mechanism. Mark the policy copy as pending activation. Step S402: Extract task context information from the node executing the current task in the original main control path. The task context information includes task status variables, sensor feedback status, and the output of the last action. Write the extracted task context information into the task context middleware for caching, and transmit the task context information synchronously to the target control node in the new main control path that deploys the strategy copy through the control channel. Step S403: Set the policy copy in the new master control path to the active state, resume task execution, preset the initial switching duration, and perform output consistency verification by comparing the control commands output by the new master control path with the output of the original master control path in the corresponding state. If the output consistency verification meets the preset consistency conditions, the task execution of the new master control path will continue to be maintained. If the output consistency verification does not meet the preset consistency conditions, the fault handling mechanism will be triggered. Step S404: When the fault handling mechanism is triggered, determine whether the original primary control path is available. If the original primary control path is available, switch back to the original primary control path and reactivate the original control policy copy to resume task execution. If the original primary control path is unavailable, select the path with the second highest path health score as the new primary control path based on the remaining candidate control path set, and repeat the steps of deploying the control policy copy, synchronizing the task context, and activating the policy copy until the task recovery is completed.

6. An intelligent control system for low-voltage electrical equipment based on the Internet of Things (IoT), used to implement the intelligent control method for low-voltage electrical equipment based on the IoT as described in any one of claims 1-5, characterized in that, The system includes a topology modeling module, a control path analysis module, a task monitoring module, a switching judgment module, and an abnormal node identification module. The topology modeling module sets low-voltage equipment as nodes, summarizes the physical connection relationship and control logic channel of each node, generates a topology diagram of the low-voltage control network, and generates a set of control paths corresponding to different task types in different functional areas through historical records. The control path analysis module: obtains the node status table, calculates the path score of the historical records, obtains the main control path based on the average path score of each control path, obtains the historical records of each control path, calculates the path stability index, communication latency index, and node health index of the historical records, and constructs a path health feature set and performs fitting training with the task execution results to generate a path health scoring function. The task monitoring module has a preset monitoring period, calculates the health score of each stage when executing the current task, and determines whether the path switching determination process is triggered. If the path switching determination process is not triggered, it continues to collect data. If the path switching determination process is triggered, it filters the candidate control paths and synchronizes the candidate control paths that meet the conditions to the control center. The switching judgment module: In the control center, it determines the new primary control path, activates the policy copy of the new primary control path, sets the initial switching duration, and judges whether the new primary control path meets the preset consistency conditions. If it does, it continues to execute the new primary control path; if it does not, it triggers the fault handling mechanism. The abnormal node identification module: during path switching and task recovery, marks nodes as abnormal nodes to be confirmed. By monitoring the running status data of the nodes marked as abnormal nodes to be confirmed and drawing fluctuation feature maps, it determines whether the nodes to be confirmed are high-risk nodes. If they are high-risk nodes, it alerts maintenance personnel to handle them.

7. The intelligent control system for low-voltage equipment based on the Internet of Things according to claim 6, characterized in that, The topology modeling module includes topology structure units and control path set units: The topology unit: all low-voltage electrical equipment deployed in the hospital building is set as nodes. By parsing the wiring diagram, control system configuration table and communication link log, the physical connection relationship and control logic channel of each low-voltage electrical equipment are obtained. Equipment with the same superior control node is classified into the same functional group, the logical connection path between each node is constructed, and the topology diagram of the low-voltage control network is generated. The control path set unit: In the topology diagram, it collects the basic attributes of each node, including node type, unique identification information, communication parameters, and control load type. It writes the basic attributes into the node's status table to generate a status table for all nodes. In the control center, it obtains the historical records of task scheduling, extracts the task types and functional areas corresponding to the weak current equipment executing the scheduling from the historical records, maps the task path correspondence, and generates a feasible control path set in combination with the current topology.

8. The intelligent control system for low-voltage equipment based on the Internet of Things according to claim 6, characterized in that, The task monitoring module includes a path switching determination unit and a unit for determining remaining candidate control paths: The path switching determination unit: It has a preset monitoring period, acquires the currently executed task type and functional area, and determines the current main control path. According to the preset monitoring period, it collects the communication parameters of all nodes on the main control path and performs normalization processing. It calculates the node's health score according to the node health score formula, presets a health score threshold, and compares the health score of each node with the health score threshold. If the health score of a node is lower than the health score threshold, the node is set as an unhealthy node. If any node in the main control path is an unhealthy node, the system triggers the path switching determination process. If there are no unhealthy nodes in the main control path, the system continues to collect data according to the preset monitoring period. The remaining candidate control path determination unit: After the system triggers the path switching determination process, it traverses the candidate control path set, obtains the communication parameters of each node in the candidate control path, calculates the health score of each node, eliminates candidate control paths with unhealthy nodes, calculates the path stability index, communication latency index, and node health index of the remaining candidate control paths, and inputs them into the path health scoring function to calculate the path health score of the remaining candidate control paths, sorts the remaining candidate control paths from high to low according to the path health score, and synchronizes the sorted set of remaining candidate control paths to the control center.

9. The intelligent control system for low-voltage equipment based on the Internet of Things according to claim 6, characterized in that, The abnormal node identification module includes units for marking abnormal nodes to be confirmed and high-risk node units: The unit for marking abnormal nodes to be confirmed: During path switching and task recovery, if several output consistency verification failures, repeated rollbacks of control tasks, and path selection failures are detected, the system obtains the association relationship between the abnormal triggering node and its upstream and downstream nodes based on path switching records, abnormal logs, and control path topology, constructs a set of suspected faulty nodes, and marks the abnormal triggering node as an abnormal node to be confirmed. The high-risk node unit monitors unconfirmed abnormal nodes, continuously collects operational status data of these nodes, including communication latency, command response time, control output offset, and upstream / downstream communication success rate. It also plots a fluctuation characteristic map of the operational status data within a preset judgment period, calculates the fluctuation slope based on the map, and sets a preset fluctuation slope threshold. When the fluctuation slope exceeds the threshold, the unconfirmed abnormal node is updated to a high-risk node, and a logical isolation operation is performed. Simultaneously, the high-risk node is pushed to the operation and maintenance management platform, prompting maintenance personnel to intervene for diagnosis and handling.

Images