Large-scale industrial fault diagnosis system and method based on Bluetooth MESH

Abstract

The invention relates to the technical field of industrial Internet of Things and wireless sensor networks, in particular to a large-scale industrial fault diagnosis system and method based on Bluetooth MESH, and efficient monitoring and fault early warning of industrial equipment are realized through low-power-consumption sensor nodes, edge computing, a Bluetooth Mesh network and a clustered Mesh tree architecture. The system comprises a client, a server, a gateway, a cluster head and a cluster node, the cluster node comprises a sensor node and a relay node, a lightweight neural network model is built in the sensor node, and edge reasoning and fault detection can be realized locally. Stable communication and real-time cooperation of large-scale nodes are ensured through a decentralized architecture of Bluetooth Mesh, establishment of a dynamic cluster head chain and optimization of an RSSI threshold value. And a low-power-consumption mechanism combining event driving and fixed polling is adopted, so that the energy consumption of the system is remarkably reduced. In-cluster communication loads are reduced through a clustering architecture, and the communication efficiency and the anti-interference capability of the system are improved in combination with channel separation of Mesh and BLE protocols.

Description

Technical Field

[0001] The present invention relates to the technical fields of industrial Internet of Things and wireless sensor networks, and particularly to a large-scale industrial fault diagnosis system and method based on Bluetooth MESH. Background Art

[0002] With the rapid development of the industrial Internet of Things, fault diagnosis technology has played an important role in the condition monitoring and maintenance of industrial equipment. However, traditional large-scale industrial fault diagnosis systems still face multiple challenges. First, traditional systems usually adopt a centralized architecture, where sensor nodes need to upload a large amount of raw data to the server for analysis, resulting in high network communication energy consumption and shortened node battery life, making it difficult to meet the low-power requirements of large-scale deployment. Second, the real-time monitoring of a large number of devices has high requirements for transmission delay and processing timeliness, while the centralized processing method is easily limited by network bottlenecks, leading to untimely fault detection and increasing industrial production risks. In addition, the centralized architecture requires a large amount of manual adjustment when facing dynamic addition and deletion of device nodes, with insufficient scalability and flexibility, and there is a risk of single-point failure. At the same time, the processing method relying on the central server is prone to becoming a bottleneck of system performance when dealing with large amounts of data and high complexity. Existing wireless communication protocols (such as Wi-Fi, ZigBee, etc.) have poor anti-interference and communication reliability in complex industrial environments, and it is difficult to meet the efficient cooperation requirements among a large number of nodes. Bluetooth Mesh, as a decentralized wireless communication protocol, has the advantages of multi-to-multi node communication, low power consumption, and broadcast message transmission, providing a new technical direction for solving the above problems. However, the application of Bluetooth Mesh in industrial fault diagnosis still lacks a mature solution, especially in combination with edge computing and dynamic node management.

[0003] Therefore, designing a large-scale industrial fault diagnosis system based on Bluetooth Mesh to meet the requirements of low power consumption, high real-time performance, and scalability has important research value and practical significance. Summary of the Invention

[0004] The purpose of the present invention is to overcome the problems of the above-mentioned prior art, and provide a large-scale industrial fault diagnosis system based on Bluetooth MESH to solve the technical problems such as high energy consumption, poor real-time performance, low diagnosis efficiency, and insufficient scalability existing in the existing large-scale industrial equipment fault diagnosis system.

[0005] The above purpose is achieved through the following technical solutions: A large-scale industrial fault diagnosis system based on Bluetooth MESH, comprising: A client: The client is used to interact with users, provide real-time monitoring, data display, and fault reporting functions, and communicate with the server or gateway through the network; Server side: The server side is deployed in the cloud or locally and is responsible for data storage, processing, and generating diagnostic reports; Gateway: The gateway bridges the Bluetooth Mesh network and the external network to achieve protocol conversion and data forwarding; Cluster head: The cluster head includes a station node and a BLE node. The station node communicates with the BLE node through SPI, is used to converge the fault information within the cluster and interact with the gateway through the BLE protocol, and at the same time supports peer-to-peer cooperation with other cluster head nodes; Cluster node: The cluster node includes a sensor node and a relay node; The sensor node is built-in with a low-power sensor and a lightweight neural network model, and is used to collect device operation data and perform local edge inference; The relay node is used to enhance the Mesh network coverage and data relay.

[0006] Furthermore, the system reduces interference during data transmission through channel separation technology, specifically including: Data of the Bluetooth Mesh and BLE protocols are carried on different communication channels respectively to avoid interference between protocols and improve data transmission efficiency; Among them, the Mesh communication part uses multiple hopping channels (signals 37, 38, 39), and Mesh nodes exchange data on multiple channels and complete data synchronization within a specific connection window to ensure efficient and low-power transmission; The BLE communication part uses an independent channel, which is different from the Mesh channel, to reduce signal conflicts between the two protocols and improve the reliability and stability of communication; The cluster head node supports a dual-mode communication mechanism, where the Mesh part processes multi-hop transmission within the cluster and forwards data to the upstream cluster head or relay, and the BLE part is used to communicate with the external gateway to achieve fast transmission of fault information.

[0007] Furthermore, the sensor node includes: Low-power sensor, used to collect device operation parameters, including but not limited to temperature, vibration, and pressure; Lightweight neural network inference unit, used to perform real-time fault diagnosis locally and generate fault information; Low-power communication module, used to transmit data to the cluster head node through the Bluetooth Mesh network.

[0008] A large-scale industrial fault diagnosis method based on Bluetooth MESH, including: A101. Classify the network based on device type or physical area; A102. Establish a cluster head chain, and a cluster head will be selected in each cluster to be responsible for coordination and information transfer between devices; A103. Establish a Mesh network within the cluster. Devices are interconnected through the Bluetooth Mesh protocol, including sensor nodes and relay nodes, and transmitted to the station node of the designated cluster head. A104. Edge computing and fault detection. The sensor nodes not only collect data but also perform edge computing tasks on the nodes. Edge computing allows data to be processed locally, reducing the need to transmit data to remote servers or the cloud. At this stage, the sensor nodes analyze and infer the collected data by running a neural network model for fault detection and early warning. A105. Management of adding or removing nodes. The network adds or removes device nodes as needed and automatically adjusts routing and cluster head allocation to ensure the stable operation of the network. A regular feedback mechanism helps optimize the network and ensures its efficient operation even in case of device replacement or failure.

[0009] Further, the establishment of the cluster head chain in step A102 includes: A201: The gateway detects the RSSI values of the BLE parts in each cluster head and sorts the RSSI values from high to low. A202: Set the RSSI threshold according to the actual scenario requirements. A203: Judge the detected RSSI values. A204: If the RSSI value is less than the RSSI threshold, it means that the signal quality of the cluster head node is low and is not suitable for direct connection with the gateway. The corresponding cluster head node enters the waiting pairing stage. A205: If the RSSI value is greater than or equal to the threshold, it means that the signal quality of the cluster head node is good and can be directly connected to the gateway to form a point-to-point connection. And allocate IDs to the connected clusters according to the size of the RSSI values.

[0010] A206: Each paired cluster head scans and detects the RSSI values of other unpaired cluster heads, summarizes and sorts all RSSI values from high to low. A207: Judge all the summarized RSSI values. If the RSSI value is less than the threshold, it means that the signal quality between the corresponding two cluster heads is low and is not suitable for point-to-point connection. Return to step A204. If the RSSI value is greater than the threshold, it means that the signal quality between the corresponding two cluster heads is good and is suitable for point-to-point connection. Jump to step A208. A208: Among all the sorted RSSI values, take the two corresponding cluster head nodes with the largest RSSI values for paired connection, and then take the two corresponding cluster head nodes with the second largest RSSI values for paired connection, and connect all the connectable cluster heads in turn. A209: Determine whether there are still cluster heads in the waiting pairing stage. If so, return to step A205; if not, jump to step A210.

[0011] A210: All cluster heads have completed pairing connections, and the cluster head chain is established.

[0012] Furthermore, the method for setting the RSSI threshold in step A202 includes: A401: Analyze requirements and scenario modeling, clarify network communication requirements, and establish a propagation model in combination with the node distribution, obstacles, and interference factors in the actual industrial scenario; A402: Estimate the distribution range of RSSI in combination with the propagation model, preliminarily determine the signal strength attenuation law between nodes based on the path loss model, and generate the theoretical RSSI value range; A403: Deploy a test environment, simulate the actual scenario to arrange nodes, and prepare test tools for collecting RSSI values and their corresponding link performances; A404: Collect RSSI data and evaluate link performance, record RSSI values, packet loss rates, and delay data by changing node distances and environmental conditions.

[0013] A405: Classify data statistics, analyze the RSSI distribution characteristics, calculate its mean, variance, etc., and clarify the communication performance in different RSSI ranges; A406: Determine key performance indicators and thresholds, set the RSSI threshold that meets the requirements of packet loss rate and transmission success rate in combination with requirements and experimental results, and add a margin; A407: Calibrate comprehensively according to environmental parameters, correct the RSSI threshold according to the interference intensity and signal fluctuation, and verify the final threshold adaptability through simulation and testing.

[0014] Furthermore, the edge computing and fault detection in step A104 adopt a hierarchical response event-driven mode, and its hierarchical response mechanism is as follows: Station node: Issue diagnostic instructions within a specified time period, and at the same time receive fault information from sensor nodes; the station node generates summary data based on the received fault information and transmits it to the BLE part in the cluster head.

[0015] Relay node: Receive fault information from sensor nodes and forward it to the station node through the Bluetooth Mesh network; the relay node preferentially responds to the requests of sensor nodes according to the friendship mechanism to avoid data loss or delay.

[0016] Sensor node: According to the diagnostic instruction or wake-up request, collect the operation data of industrial equipment during the receiving window, transmit it through SPI and input it into the neural network for inference to generate a fault detection result.

[0017] Further, the event-driven working mode is specifically as follows: The sensor nodes are usually in a low-power sleep state and only perform edge computing when they wake up at specific time intervals and receive the diagnostic instructions sent by the station nodes; during the receiving window, the sensor nodes enter the active state and are ready to receive or send data.

[0018] Further, the edge computing processing flow is as follows: A301: The sensor nodes collect the device operation status data through the built-in low-power sensors; A302: The data collected by the sensors is transmitted to the computing unit through the FIFO mechanism of SPI; the FIFO buffer mechanism can temporarily store the transmitted data to ensure that even if the computing unit is momentarily busy or delayed during the data transmission process, the data collected by the sensors can still be completely received, avoiding data loss; A303: The lightweight neural network deployed on the sensor nodes performs inference calculations to detect whether there are faults or anomalies; A304: After the inference is completed, if a fault is detected, the node will generate a fault message and send it to the relay node or the station node through the Mesh network; A305: After completing the task, the sensor nodes enter the sleep state to reduce power consumption.

[0019] Further, the management of adding or removing nodes in step A105 is as follows: If the cluster head is removed, the gateway will check the routing table of the cluster head. When there is no subsequent connection to the cluster head removed, that is, when it is the last cluster head in the chain, the cluster head node will be deleted and other routes will remain unchanged; on the contrary, if there are other cluster heads subsequently, the other cluster heads will enter the unpaired stage and join the network as new cluster heads; when a new cluster head wants to join the network, the gateway or the existing cluster head receives the request and performs identity verification according to the preset verification mechanism; after the verification passes, the new cluster head is allowed to access the network and is assigned initial communication parameters, and then a cluster head chain is established for the new cluster head through the cluster head chain establishment process. Finally, the gateway will maintain and update the routing table of the cluster head; when a node joins, it sends a join request in a broadcast manner. After the corresponding cluster head receives the request and verifies the node's identity, it adds it to the network topology table, assigns a unique address and a routing path at the same time, and synchronizes the network configuration information; when a node needs to be removed, it sends an exit request. After the cluster head confirms, the relevant information is deleted from the topology table, the key is revoked, and the node's network configuration data is cleared; if the node fails to respond due to a fault, the cluster head marks it as invalid and removes it from the network after multiple failed connection attempts.

[0020] The present invention provides a large-scale industrial fault diagnosis system and method based on Bluetooth MESH, which ensures stable communication and real-time collaboration of large-scale nodes through the decentralized architecture of Bluetooth Mesh, the establishment of a dynamic cluster head chain and the optimization of RSSI thresholds. A low-power mechanism combining event-driven and fixed polling is adopted to significantly reduce the energy consumption of the system. The clustering architecture is used to reduce the communication load within the cluster, and the channel separation of Mesh and BLE protocols is combined to improve the communication efficiency and the anti-interference ability of the system. The present invention is particularly suitable for large-scale industrial equipment distribution scenarios, and has high real-time performance, low power consumption, high reliability and good scalability. It is particularly suitable for large-scale industrial equipment distribution scenarios, and has high real-time performance, low power consumption, high reliability and good scalability. The advantages are as follows: The adoption of Bluetooth Mesh network protocol and low-power design of sensor nodes, combined with a combination of event-driven and fixed polling mechanisms, significantly reduces the overall energy consumption of the system, especially in scenarios with large-scale node deployment.

[0021] By building a lightweight neural network model into the sensor node, edge computing can be implemented locally on the node, reducing the need to transmit data to the cloud or server. This not only improves the real-time performance of fault diagnosis, but also effectively reduces communication delays and data bandwidth requirements.

[0022] By coordinating the working hours and communication mechanisms of nodes at different levels through event-driven mode, efficient edge computing and fault detection in the Bluetooth Mesh network are achieved. The combination of fixed polling and event-driven strategies at the node level not only reduces overall energy consumption, but also ensures the real-time and reliability of data processing and transmission.

[0023] A clustered Mesh tree architecture is proposed to support dynamic adjustment of cluster head chains and the addition or removal of cluster nodes. When a new cluster head is added or an existing cluster head is removed, the system can ensure the self-healing ability and scalability of the network structure through an adaptive RSSI sorting mechanism and a dynamic cluster head chain update process.

[0024] For large-scale equipment deployment, the present invention effectively reduces the intra-cluster communication load through a clustered architecture, and separates the channels of the Mesh and BLE protocols, further reducing information collisions and improving communication efficiency. Finally, through the collaboration of the cluster head chain, the aggregation and sharing of data across the entire network are achieved, meeting the real-time and coverage requirements of large-scale industrial-grade monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a schematic diagram of the architecture of a large-scale industrial fault diagnosis system based on Bluetooth MESH according to the present invention; Figure 2 A schematic diagram for explaining cluster heads and cluster points in a large-scale industrial fault diagnosis system based on Bluetooth MESH according to the present invention; Figure 3 Schematic diagram of the cluster head chain establishment process in a large-scale industrial fault diagnosis method based on Bluetooth MESH according to the present invention; Figure 4 Schematic diagram of the cluster head chain establishment process in a large-scale industrial fault diagnosis method based on Bluetooth MESH according to the present invention; Figure 5 Schematic diagram of the RSSI threshold determination method in a large-scale industrial fault diagnosis method based on Bluetooth MESH according to the present invention; Figure 6 Schematic diagram of the event-driven mode of edge computing hierarchical response in a large-scale industrial fault diagnosis method based on Bluetooth MESH according to the present invention; Figure 7 Schematic diagram of the edge computing process in a large-scale industrial fault diagnosis method based on Bluetooth MESH according to the present invention. Detailed implementation manners

[0026] The present invention will be further described in detail below with reference to the drawings and embodiments. The described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0027] As Figure 1 shown, this solution provides a large-scale industrial fault diagnosis system based on Bluetooth MESH, aiming to achieve efficient and low-power fault detection and early warning of industrial equipment through low-power sensor nodes, edge computing, Bluetooth Mesh network and clustered Mesh tree architecture, and is particularly suitable for real-time monitoring and fault diagnosis scenarios of large-scale industrial equipment.

[0028] As Figure 1 and 2 shown, the system includes: Client: The client is the front-end user interaction part of the system, usually the terminal device used by the industrial control center, operators or maintenance personnel. The client provides functions such as real-time monitoring, data display, fault reporting and device status update through the user interface. The client can usually be a PC, mobile device or embedded device, with network connection capabilities, and can communicate with the server side or gateway to obtain and display the real-time data of the device.

[0029] Server side: The server side is the core computing and data processing unit of the entire fault diagnosis system, responsible for receiving requests from the client or other systems, analyzing data, generating diagnostic reports, and returning the results to the client. The main responsibilities of the server side include data storage, processing, and decision support. The server side can be deployed in the cloud, local data center, or industrial control system, with high computing power and storage capacity.

[0030] Gateway: The gateway is a bridging device between the BLE network and external networks (such as local area network, Internet, server side, etc.), responsible for transmitting data between the BLE network and external systems. The gateway plays the role of data forwarding, protocol conversion, and network connection.

[0031] Cluster head: The cluster head consists of two parts, a station node and a BLE node in the Bluetooth Mesh network. The two parts conduct wired data transmission through SPI. The station node is responsible for aggregating the fault diagnosis information of the corresponding cluster, and the BLE part is responsible for transmitting the aggregated fault information to the gateway through the data channel. The cluster head node not only aggregates the data of the station nodes it manages, but also collaboratively processes the data of other cluster head nodes through the point-to-point link of BLE to achieve fault diagnosis and data sharing across the entire network. BLE and Mesh protocols use different channels, thus reducing information collisions and ensuring efficient and stable communication.

[0032] Cluster node: The cluster node contains edge devices in the Bluetooth Mesh network, responsible for collecting the operation data of industrial devices, performing edge computing, and then transmitting the data to the cluster head node through the Bluetooth Mesh network. The cluster node usually contains sensors and embedded neural network models for performing real-time fault detection tasks. The cluster node contains the following nodes: Sensor node: The sensor node has low power consumption characteristics, and each sensor node is built-in with sensors (such as temperature, vibration, pressure, etc.) and a lightweight neural network model. It collects the operation data of the device, performs inference tasks, and generates fault information. The sensor node transmits the industrial device fault information data to the station node of the corresponding cluster head through the Mesh network for further management.

[0033] Relay node: The relay node has a friendship function, forms a friendship relationship with low-power nodes, and is used to strengthen the propagation of network signals and data relay. The relay node helps to transmit the diagnostic information from the sensor node to ensure that the data can be stably transmitted to the cluster head node, avoiding information loss or delay.

[0034] As Figure 3 shown, this solution also provides a large-scale industrial fault diagnosis method based on Bluetooth MESH, including: A101: Classification and planning are carried out according to regions or device types. In an industrial environment, devices can be classified according to physical regions or device types. For example, they can be divided by workshop or floor, or clustered according to device functions (such as sensors, actuators). This helps with subsequent network design and management, ensuring that devices within each cluster work together.

[0035] A102: Establish a cluster head chain. One cluster head is selected within each cluster to be responsible for coordination and information transfer between devices; A103: Establish a Mesh network within the cluster. Devices are interconnected through the Bluetooth Mesh protocol, including sensor nodes and relay nodes, and finally transmitted to the station node of the designated cluster head. A104: Edge computing and fault detection. Sensor nodes not only collect data but also perform edge computing tasks on the nodes. Edge computing allows data to be processed locally, reducing the need to transmit data to remote servers or the cloud. At this stage, sensor nodes analyze and infer the collected data by running a neural network model for fault detection and early warning.

[0036] A105: Management of adding or removing nodes. The network can add or remove device nodes as needed and automatically adjust routing and cluster head allocation to ensure stable network operation.

[0037] As Figure 4 shown, the establishment of the Bluetooth Mesh cluster head chain in step A102 includes the following steps: A201: The gateway detects the RSSI values of the BLE parts in each cluster head and sorts the RSSI values from high to low.

[0038] A202: To pursue good transmission performance, an RSSI threshold is set according to actual scenario requirements.

[0039] A203: Judge the detected RSSI values.

[0040] A204: If the RSSI value is less than the threshold, it means that the signal quality of the cluster head node is low and it is not suitable for the gateway to make a direct connection. The corresponding cluster head node enters the waiting pairing stage; A205: If the RSSI value is greater than or equal to the threshold, it means that the signal quality of the cluster head node is good and it can make a direct connection with the gateway to form a point-to-point connection. And IDs are assigned to the connected clusters according to the size of the RSSI values (for example, the cluster head with the highest RSSI has an ID of 01, followed by 02, 03...).

[0041] A206: Each paired cluster head scans and detects the RSSI values of other unpaired cluster heads, summarizes and sorts all RSSI values from high to low (it can be sent to the gateway for sorting and then distributed). A207: Judge all the aggregated RSSI values. If the RSSI value is less than the threshold, it means that the signal quality between the corresponding two cluster heads is low and not suitable for point-to-point connection. Return to step A204, that is, the unpaired cluster head nodes are still in the waiting pairing stage. If the RSSI value is greater than the threshold, it means that the signal quality between the corresponding two cluster heads is good and suitable for point-to-point connection. Jump to step A208; A208: Among all the sorted RSSI values, select the two corresponding cluster head nodes with the largest RSSI value for paired connection, and then select the two corresponding cluster head nodes with the second largest RSSI value for paired connection, and connect all the connectable cluster heads in turn.

[0042] A209: Judge whether there are still cluster heads in the waiting pairing stage. If there are, return to step A205. If not, jump to step A210.

[0043] A210: All cluster heads have been paired and connected, and the cluster head chain is established.

[0044] As Figure 5 shown, in step A202, in order to pursue good transmission performance, the RSSI threshold is set according to the actual scenario requirements. The method for determining the RSSI threshold is as follows: A401: Analyze the requirements and scenario modeling, clarify the network communication requirements (such as packet loss rate, latency, coverage range), and establish a propagation model in combination with the node distribution, obstacles, and interference factors in the actual industrial scenario.

[0045] A402: Estimate the distribution range of RSSI in combination with the propagation model, preliminarily determine the signal strength attenuation law between nodes based on the path loss model, and generate the theoretical RSSI value range.

[0046] A403: Deploy the test environment, simulate the actual scenario to arrange nodes, and prepare test tools for collecting RSSI values and their corresponding link performance.

[0047] A404: Collect RSSI data and evaluate the link performance, and record the RSSI value, packet loss rate, latency and other performance data by changing the node distance and environmental conditions.

[0048] A405: Classify data statistics, analyze the RSSI distribution characteristics, calculate its mean, variance, etc., and clarify the communication performance in different RSSI ranges.

[0049] A406: Determine the key performance indicators and thresholds, and set the RSSI threshold that meets the requirements of packet loss rate and transmission success rate in combination with the requirements and experimental results, and add a margin.

[0050] A407: Comprehensive environmental parameter calibration, correct the RSSI threshold according to the interference intensity and signal fluctuation, and verify the final threshold adaptability through simulation and testing.

[0051] As Figure 6 shown, in step A104, the edge computing and fault detection are in an event-driven mode with hierarchical response, and its hierarchical response mechanism is as follows: Station node: Issue diagnostic instructions within a specified time period, and at the same time receive fault information from sensor nodes. The station node generates summary data based on the received fault information and transmits it to the BLE part in the cluster head.

[0052] Relay node: Receive fault information from sensor nodes and forward it to the station node through the Bluetooth Mesh network. The relay node preferentially responds to the requests of sensor nodes according to the friendship mechanism to avoid data loss or delay.

[0053] Sensor node: According to the diagnostic instruction or wake-up request, collect the operation data of industrial equipment during the receiving window, transmit it through SPI and input it into the neural network for inference to generate a fault detection result.

[0054] Its event-driven working mode is as follows: Sensor nodes are usually in a low-power sleep state and only perform edge computing when they wake up at specific time intervals and receive the diagnostic instructions issued by the station node. During the receiving window, the sensor nodes enter the active state to prepare to receive or send data.

[0055] As Figure 7 shown, the edge computing processing flow is as follows: A301: The sensor node collects the device operation status data through built-in low-power sensors (such as temperature, vibration or pressure).

[0056] A302: The data collected by the sensor is transmitted to the computing unit through SPI.

[0057] A303: The lightweight neural network deployed on the sensor node performs inference calculations to detect whether there are faults or abnormalities.

[0058] A304: After the inference is completed, if a fault is detected, the node will generate a fault message and send it to the relay node or the station node through the Mesh network.

[0059] A305: After completing the task, the sensor node enters the sleep state to reduce power consumption.

[0060] The management of adding or removing nodes in step A105 of this embodiment is as follows: If the cluster head is removed, the gateway will check the routing table of the cluster head. When there is no subsequent cluster head connected to the removed cluster head, that is, when it is the last cluster head in the chain, the cluster head node will be deleted and other routes will remain unchanged. On the contrary, if there are other cluster heads subsequently, these other cluster heads will enter the unpaired stage and join the network as new cluster heads. When a new cluster head wants to join the network, the gateway or an existing cluster head receives the request and performs authentication according to a preset verification mechanism. After successful verification, the new cluster head is allowed to access the network and is assigned initial communication parameters. Then, through the cluster head chain establishment process, a cluster head chain is established for the new cluster head. Finally, the gateway will maintain and update the routing table of the cluster head.

[0061] When a node needs to be removed, it sends an exit request. After the cluster head confirms, the relevant information is deleted from the topology table, the key is revoked, and the network configuration data of the node is cleared. When a node joins, it sends a join request via broadcast. After the corresponding cluster head receives the request and verifies the node's identity, it adds the node to the network topology table, assigns a unique address and a routing path, and synchronizes the network configuration information. If the node fails to respond due to a fault, the cluster head marks it as invalid and removes it from the network after multiple failed connection attempts. The above is only to illustrate the embodiments of the present invention and is not intended to limit the present invention. For those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A large-scale industrial fault diagnosis system based on Bluetooth MESH, characterized in that Including: Client: The client is used to interact with users, providing functions such as real-time monitoring, data display, and fault reporting, and communicating with the server or gateway through the network; Server: The server is deployed in the cloud or locally, responsible for data storage, processing, and generating diagnostic reports; Gateway: The gateway bridges the Bluetooth Mesh network and the external network to achieve protocol conversion and data forwarding; Cluster head: The cluster head includes a station node and a BLE node. The station node communicates with the BLE node through SPI, is used to converge the fault information within the cluster and interact with the gateway through the BLE protocol, and at the same time supports peer-to-peer cooperation with other cluster head nodes; Cluster node: The cluster node includes a sensor node and a relay node; The sensor node is built-in with a low-power sensor and a lightweight neural network model, and is used to collect device operation data and perform local edge inference; The relay node is used to enhance the Mesh network coverage and data relay.

2. The large-scale industrial fault diagnosis system based on Bluetooth MESH according to claim 1, wherein The system reduces interference during data transmission through channel separation technology, specifically including: Data of the Bluetooth Mesh and BLE protocols are carried on different communication channels respectively to avoid interference between protocols and improve data transmission efficiency; Among them, the Mesh communication part uses multiple hopping channels (signals 37, 38, 39), and Mesh nodes exchange data on multiple channels and complete data synchronization within a specific connection window to ensure efficient and low-power transmission; The BLE communication part uses an independent channel, different from the Mesh channel, to reduce signal conflicts between the two protocols and improve the reliability and stability of communication; The cluster head node supports a dual-mode communication mechanism, where the Mesh part processes multi-hop transmission within the cluster and forwards data to the upstream cluster head or relay, and the BLE part is used to communicate with the external gateway to achieve fast transmission of fault information.

3. A large-scale industrial fault diagnosis system based on Bluetooth MESH according to claim 1, characterized in that, The sensor node includes: Low-power sensor, used to collect device operation parameters, including but not limited to temperature, vibration, and pressure; Lightweight neural network inference unit, used to perform real-time fault diagnosis locally and generate fault information; Low-power communication module, used to transmit data to the cluster head node through the Bluetooth Mesh network.

4. A large-scale industrial fault diagnosis method based on Bluetooth MESH, characterized in that, Including: A101. Classify the network based on device type or physical area; A102. Establish a cluster head chain, and a cluster head will be selected in each cluster to be responsible for coordination and information transfer between devices; A103. Establish a Mesh network within the cluster, and devices are interconnected through the Bluetooth Mesh protocol, including sensor nodes and relay nodes, and are transmitted to the station node of the designated cluster head; A104. Edge computing and fault detection. The sensor node not only collects data but also executes edge computing tasks on the node; Edge computing allows data to be processed locally, reducing the need to transmit data to a remote server or the cloud; At this stage, the sensor node analyzes and infers the collected data by running a neural network model to perform fault detection and early warning; A105. Management of adding or removing nodes. The network adds or removes device nodes as needed and automatically adjusts routing and cluster head allocation to ensure the stable operation of the network. A regular feedback mechanism helps optimize the network and ensures its efficient operation even in the case of device replacement or failure.

5. A large-scale industrial fault diagnosis method based on Bluetooth MESH according to claim 4, characterized in that, The establishment of the cluster head chain described in step A102 includes: A201: The gateway detects the RSSI values of the BLE parts in each cluster head and sorts the RSSI values from high to low. A202: Set the RSSI threshold according to the actual scenario requirements. A203: Judge the detected RSSI values. A204: If the RSSI value is less than the RSSI threshold, it means that the signal quality of the cluster head node is low and it is not suitable for direct connection with the gateway. The corresponding cluster head node enters the waiting pairing stage. A205: If the RSSI value is greater than or equal to the threshold, it means that the signal quality of the cluster head node is good and it can be directly connected to the gateway to form a point-to-point connection. And assign IDs to the connected clusters according to the size of the RSSI values. A206: Each paired cluster head scans and detects the RSSI values of other unpaired cluster heads, summarizes and sorts all the RSSI values from high to low. A207: Judge all the summarized RSSI values. If the RSSI value is less than the threshold, it means that the signal quality between the corresponding two cluster heads is low and it is not suitable for point-to-point connection. Return to step A204. If the RSSI value is greater than the threshold, it means that the signal quality between the corresponding two cluster heads is good and it is suitable for point-to-point connection. Jump to step A208. A208: Among all the sorted RSSI values, select the two corresponding cluster head nodes with the largest RSSI values for paired connection, and then select the two corresponding cluster head nodes with the second largest RSSI values for paired connection, and connect all the connectable cluster heads in turn. A209: Judge whether there are still cluster heads in the waiting pairing stage. If so, return to step A205. If not, jump to step A210. A210: All cluster heads have been paired and connected, and the cluster head chain is established.

6. The large-scale industrial fault diagnosis method based on Bluetooth MESH according to claim 5, characterized in that The method for setting the RSSI threshold described in step A202 includes: A401: Analyze requirements and scenario modeling, clarify the network communication requirements and establish a propagation model in combination with the node distribution, obstacles and interference factors in the actual industrial scenario. A402: Estimate the distribution range of RSSI in combination with the propagation model, preliminarily determine the signal strength attenuation law between nodes based on the path loss model, and generate the theoretical RSSI value range. A403: Deploy the test environment, arrange nodes to simulate the actual scenario, and prepare test tools for collecting RSSI values and their corresponding link performances. A404: Collect RSSI data and evaluate the link performance, record RSSI values, packet loss rates and delay data by changing the node distance and environmental conditions. A405: Classify data statistics, analyze the RSSI distribution characteristics, calculate its mean, variance, etc., and clarify the communication performance in different RSSI ranges. A406: Determine key performance indicators and thresholds, set the RSSI threshold that meets the requirements of packet loss rate and transmission success rate in combination with requirements and experimental results, and add a margin; A407: Conduct comprehensive environmental parameter calibration, correct the RSSI threshold according to the interference intensity and signal fluctuation, and verify the adaptability of the final threshold through simulation and testing.

7. A large-scale industrial fault diagnosis method based on Bluetooth MESH according to claim 4, characterized in that The edge computing and fault detection described in step A104 adopt an event-driven mode with hierarchical response, and its hierarchical response mechanism is as follows: Station node: Issue diagnostic instructions within a specified time period and simultaneously receive fault information from sensor nodes; the station node generates summary data based on the received fault information and transmits it to the BLE part in the cluster head; Relay node: Receive fault information from sensor nodes and forward it to the station node through the Bluetooth Mesh network; the relay node preferentially responds to the requests of sensor nodes according to the friendship mechanism to avoid data loss or delay; Sensor node: Collect the operation data of industrial equipment during the reception window according to the diagnostic instruction or wake-up request, transmit it through SPI and input it into the neural network for inference to generate a fault detection result.

8. A large-scale industrial fault diagnosis method based on Bluetooth MESH according to claim 7, characterized in that, The specific event-driven working mode is as follows: Sensor nodes are usually in a low-power sleep state and only perform edge computing when they wake up at specific time intervals and receive the diagnostic instructions issued by the station node; during the reception window, sensor nodes enter the active state and are ready to receive or send data.

9. The method for large-scale industrial fault diagnosis based on Bluetooth MESH according to claim 8, wherein, The edge computing processing flow is as follows: A301: The sensor node collects the device operation status data through the built-in low-power sensor; A302: The data collected by the sensor is transmitted to the computing unit through the FIFO mechanism of SPI; the FIFO buffer mechanism can temporarily store the transmitted data to ensure that even if the computing unit is momentarily busy or delayed during the data transmission process, the data collected by the sensor can still be completely received, avoiding data loss; A303: The lightweight neural network deployed on the sensor node performs inference calculations to detect whether there are faults or abnormalities; A304: After the inference is completed, if a fault is detected, the node will generate a fault message and send it to the relay node or the station node through the Mesh network; A305: After completing the task, the sensor node enters the sleep state to reduce power consumption.

10. A large-scale industrial fault diagnosis method based on Bluetooth MESH according to claim 4, characterized in that, The management of adding or removing nodes described in step A105 is as follows: If the cluster head is removed, the gateway will check the routing table of the cluster head. When there is no subsequent cluster head connected to the removed cluster head, that is, when it is the last cluster head in the chain, the cluster head node will be deleted and other routes will remain unchanged; conversely, if there are other cluster heads subsequently, the other cluster heads will enter the unpaired stage and join the network as the new cluster head; when a new cluster head wants to join the network, the gateway or the existing cluster head receives the request and performs identity verification according to the preset verification mechanism; after the verification passes, the new cluster head is allowed to access the network and is assigned initial communication parameters, and then a cluster head chain is established for the new cluster head through the cluster head chain establishment process. Finally, the gateway will maintain and update the routing table of the cluster head; When a node joins, it sends a join request via broadcast. After the corresponding cluster head receives the request and verifies the node's identity, it adds the node to the network topology table, assigns a unique address and routing path, and synchronizes the network configuration information; when a node needs to be removed, it sends a leave request, and after the cluster head confirms, it deletes the relevant information from the topology table, revokes the key, and clears the node's network configuration data; if a node fails to respond due to a fault, the cluster head marks it as invalid and removes it from the network after multiple failed connection attempts.

Images