Distributed electromechanical equipment data control system based on Internet of Things

Abstract

The invention relates to the technical field of the Internet of Things, in particular to a distributed electromechanical equipment data control system based on the Internet of Things. Aiming at the problems of difficulty in heterogeneous equipment interconnection, low communication topology coordination efficiency, insufficient data security protection and the like of a traditional electromechanical system, the system realizes accurate acquisition and optimization of equipment parameters through an edge sensing and data preprocessing module, and a distributed self-organizing communication module adapts to a multi-interface protocol and supports dynamic routing; the hierarchical distributed control module constructs an edge-region-top layer three-level closed-loop control architecture, realizes global optimization under privacy protection in combination with the intelligent decision module, and guarantees data transmission and instruction execution security in cooperation with the multi-level dynamic security protection module. According to the invention, the bottleneck of cooperative control of a traditional system is solved, the data processing efficiency, control precision and operation safety of electromechanical equipment are improved, and the system is suitable for multi-equipment cooperative operation and maintenance requirements in an industrial scene.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things, and in particular to a distributed electromechanical equipment data control system based on Internet of Things. BACKGROUND

[0002] In the process of industrial intelligent transformation, the number of electromechanical equipment in manufacturing workshops, industrial parks and other scenarios has increased rapidly and is becoming increasingly heterogeneous, and the demand for real-time monitoring and accurate control of equipment data has significantly increased. Traditional electromechanical equipment control mostly uses centralized architecture, relying on a central server to manage the data collection and instruction issuance of all devices, connecting devices through a single interface protocol, relying on local servers to store data and execute control logic to meet basic operation and maintenance needs.

[0003] Traditional solutions mostly use centralized architecture to implement electromechanical equipment control, with the central server uniformly accessing sensors and controllers of all devices, collecting device operating parameters, and issuing control instructions after processing by the central computing power. In specific implementation, fixed workshop scenarios rely on industrial Ethernet to connect devices and central servers, and remote plants use simple wireless modules to transmit data, thereby achieving device monitoring and control.

[0004] However, the traditional centralized solution has obvious limitations: first, it has poor real-time performance, as the central server needs to process massive device data, resulting in high response delay of control instructions, which cannot meet the millisecond-level control requirements of motor speed regulation, valve switching, etc. Second, it has low reliability, as the central server or core communication link failure can easily cause the entire system to malfunction, affecting continuous operation of the equipment. Third, it is difficult to achieve heterogeneous compatibility, as different manufacturers' electromechanical equipment interfaces and protocols are not unified, requiring additional configuration of converters, which increases costs and reduces data transmission stability. Fourth, it has weak data security, as centralized storage is vulnerable to attacks, and there is a risk of tampering and leakage of device operation data and control instructions. Therefore, it is very important to develop a distributed electromechanical equipment data control system based on Internet of Things that can achieve multi-device distributed collaboration, low-latency control, heterogeneous compatibility, and data security control. SUMMARY

[0005] To overcome the above-mentioned defects of the prior art, embodiments of the present application provide a distributed electromechanical equipment data control system based on Internet of Things to solve the problems raised in the background art.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solution: a distributed electromechanical equipment data control system based on Internet of Things, comprising: a system management database for storing design data required during system operation and receiving real-time generated data, and constructing a distributed electromechanical equipment data control management database; Edge perception and data preprocessing module: used for collecting key operation parameters of electromechanical equipment, preprocessing collected data, and obtaining preprocessed equipment operation data; Distributed self-organizing communication module: used for realizing multi-link interconnection between the edge perception and data preprocessing module, the distributed trusted data storage module, the hierarchical distributed control module and the intelligent decision module; Distributed trusted data storage module: used for receiving data transmitted by the distributed self-organizing communication module; Hierarchical distributed control module: used for receiving preprocessed equipment operation data transmitted by the edge perception and data preprocessing module, and optimization control strategies transmitted by the intelligent decision module; Intelligent decision module: used for receiving historical data and real-time data transmitted by the distributed trusted data storage module, generating equipment fault prediction results, operation and maintenance optimization suggestions and global optimization control strategies; Multi-level dynamic security protection module: used for security protection of the operation process of the edge perception and data preprocessing module, the distributed self-organizing communication module, the distributed trusted data storage module, the hierarchical distributed control module and the intelligent decision module.

[0007] Technical effects and advantages of the present application: 1. The present application realizes heterogeneous device interconnection and hierarchical control through the collaborative design of "edge-area-top" hierarchical distributed control and self-organizing communication: the edge layer realizes real-time closed-loop control response quickly, the area layer guarantees the consistency of multi-device linkage, and the top layer optimizes global resource allocation, and the self-organizing routing dynamically repairs fault links, solving the problems of poor real-time performance and single-point failure paralysis of traditional centralized systems, and significantly improving the device control precision and data transmission reliability; 2. The present application realizes the "local training + global aggregation" mode driven by federated learning and intelligent decision: each regional node locally trains the model to protect data privacy, and the top layer aggregates to generate a global fault prediction model, and dynamically adjusts the operation and maintenance priority according to the production plan, solving the problems of data privacy leakage and poor regional adaptability of traditional centralized modeling, and improving the fault prediction ability and operation and maintenance efficiency; 3. The present application constructs a multi-level dynamic security protection system and realizes all-link security control: device end lightweight encryption, transmission end asymmetric encryption guarantee data security, command "one-time encryption" prevent replay attacks, edge node behavior baseline monitoring active interception of abnormalities, solving the problems of fragmentation of traditional system security protection and passive defense, and reducing the risk of data leakage and equipment misoperation in all directions. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed for the description of the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the basic field, other drawings can also be obtained without creative labor on the basis of these drawings.

[0009] Figure 1 The schematic diagram of the module connection of the present application.

[0010] Figure 2 The schematic diagram of the hierarchical distributed control module flow implementation of the present application. DETAILED DESCRIPTION

[0011] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.

[0012] Please refer to Figure 1 The present application provides a distributed electromechanical equipment data control system based on Internet of Things, which includes a system management database, an edge sensing and data preprocessing module, a distributed self-organizing communication module, a distributed trusted data storage module, a hierarchical distributed control module, an intelligent decision module and a multi-level dynamic security protection module.

[0013] This embodiment takes a manufacturing workshop as an application scenario, and m=50 electromechanical equipment (including 30 motors, 15 pump bodies and 5 valves) are deployed in the workshop. The system is deployed according to a three-level architecture of "edge-area-top", and specifically: the edge layer includes 1 edge node (including a sensor and an edge controller) matched with each electromechanical equipment, and there are 50 edge nodes in total; the area layer includes 2 area nodes (covering the east and west production areas) set in the workshop, and each area node manages 25 devices; the top layer includes 1 top platform node deployed in the factory operation and maintenance center, which is connected with the 2 area nodes in the workshop and the area nodes (q=5 in total) of other workshops in the factory. The modules are interconnected through the distributed self-organizing communication module, and the system management database is deployed on the top platform and realizes data intercommunication with all the modules.

[0014] The system management database is connected with the edge perception and data preprocessing module, the distributed self-organizing communication module, the distributed trusted data storage module, the hierarchical distributed control module, the intelligent decision module and the multi-level dynamic security protection module respectively, the edge perception and data preprocessing module is connected with the distributed trusted data storage module through the distributed self-organizing communication module, and the hierarchical distributed control module is connected with the intelligent decision module.

[0015] The system management database is connected with the edge perception and data preprocessing module, the distributed self-organizing communication module, the distributed trusted data storage module, the hierarchical distributed control module, the intelligent decision module and the multi-level dynamic security protection module respectively, the edge perception and data preprocessing module is connected with the distributed trusted data storage module through the distributed self-organizing communication module, and the hierarchical distributed control module is connected with the intelligent decision module. The system management database is connected with the edge perception and data preprocessing module, the distributed self-organizing communication module, the distributed trusted data storage module, the hierarchical distributed control module, the intelligent decision module and the multi-level dynamic security protection module respectively, the edge perception and data preprocessing module is connected with the distributed trusted data storage module through the distributed self-organizing communication module, and the hierarchical distributed control module is connected with the intelligent decision module. The system management database is connected with the edge perception and data preprocessing module, the distributed self-organizing communication module, the distributed trusted data storage module, the hierarchical distributed control module, the intelligent decision module and the multi-level dynamic security protection module respectively, the edge perception and data preprocessing module is connected with the distributed trusted data storage module through the distributed self-organizing communication module, and the hierarchical distributed control module is connected with the intelligent decision module. The system management database is connected with the edge perception and data preprocessing module, the distributed self-organizing communication module, the distributed trusted data storage module, the hierarchical distributed control module, the intelligent decision module and the multi-level dynamic security protection module respectively, the edge perception and data preprocessing module is connected with the distributed trusted data storage module through the distributed self-organizing communication module, and the hierarchical distributed control module is connected with the intelligent decision module.

[0016] The database adopts a distributed architecture and is backed up by multiple nodes to ensure data reliability, and a data correlation index is established based on a "time stamp + device ID" to ensure efficient and convenient subsequent queries and calls. In addition, the database establishes a real-time data interaction channel with the edge perception, control, decision and other modules, which not only provides design data support for the operation of each module, but also synchronously collects dynamic data, and finally builds a unified data management base supporting distributed mechanical and electrical equipment control.

[0017] The edge perception and data preprocessing module is used for collecting key operating parameters of mechanical and electrical equipment and preprocessing the collected data to obtain preprocessed equipment operating data. S1.1: Collect vibration, current, speed and pressure parameters of mechanical and electrical equipment through industrial-grade sensors; S1.2: Perform noise filtering, outlier removal and data compression processing on the collected parameters: let a certain type of parameter sample set collected be , wherein n is the number of samples, is the i-th sample value; calculate the sample mean ; calculate the sample standard deviation ; set the statistical threshold range as wherein k is a confidence coefficient, and takes a value of 2.5 to 3.0, and a sample beyond the range is determined as an abnormal value; S1.3: The pre-processed equipment operation data is classified according to data types, and is respectively transmitted to a distributed self-organizing communication module and a hierarchical distributed control module.

[0018] It needs to be specifically explained in this embodiment that the collection logic and preprocessing process of the edge perception and data preprocessing module. Taking the mechanical and electrical equipment (motor, pump body, valve, etc.) in the manufacturing workshop as the monitoring object, the accurate parameter collection and data optimization processing are realized.

[0019] It needs to be specifically explained in this embodiment that each equipment is matched with an industrial-grade sensor. Through the MEMS vibration sensor, the Hall current sensor, etc., the core operation parameters of the equipment are collected, including the vibration parameters reflecting the equipment working condition, the current parameters, the rotating speed parameters reflecting the operation state, and the pressure parameters associated with the load, so as to ensure that the key monitoring dimensions of the equipment are covered.

[0020] It needs to be specifically explained in this embodiment that the data preprocessing first filters the noise (such as interference signals in equipment vibration) in the collected data through the algorithm combined with “sliding window + wavelet transform”; then the statistical threshold method is used to eliminate abnormal values. Based on a certain type of parameter sample set, the sample mean and standard deviation are calculated, the threshold range is set in combination with the confidence coefficient k of 2.5 to 3.0, and the sample beyond the range is determined as an abnormal value; finally, the data amount is reduced through the lightweight compression algorithm, and the subsequent transmission pressure is reduced.

[0021] It needs to be specifically explained in this embodiment that the data classification transmission divides the pre-processed equipment operation data according to types, and respectively transmits to a distributed self-organizing communication module (for subsequent storage and decision-making) and a hierarchical distributed control module (supporting real-time control instruction generation), so as to provide accurate and efficient data support for the subsequent distributed storage and hierarchical control of the system.

[0022] The distributed self-organizing communication module is used to realize the multi-link interconnection between the edge perception and data preprocessing module, the distributed trusted data storage module, the hierarchical distributed control module, and the intelligent decision-making module. The distributed self-organizing communication module realizes the following functions: Supporting distributed self-organizing routing, when a communication node fails, other nodes automatically re-plan the transmission path; Adopting a dynamic bandwidth allocation strategy, classifying data according to priority, and uniformly converting heterogeneous data into MQTT protocol or CoAP protocol.

[0023] The embodiment needs to be specifically explained is the actual landing logic of interface and protocol adaptation in the distributed self-organizing communication module. Combined with the different interface requirements of 50 heterogeneous mechanical and electrical equipment (30 motors, 15 pump bodies and 5 valves) in a medium-sized automobile parts manufacturing workshop, the communication module realizes the interconnection of all equipment through hardware integration and protocol compatibility design: for the requirement of high-speed real-time transmission of motor speed and torque data, the Profinet protocol interface module is built in the edge node communication unit, which is directly connected with the Profinet port of the motor controller, supports high-speed industrial Ethernet transmission rate, and meets the interactive requirement of millisecond-level control command of the motor; for the characteristics of pump body that needs to stably transmit periodic data such as flow and pressure, Modbus-RTU interface module is configured, which is connected with pump body controller through RS485 bus, and the baud rate is set according to the unified standard of workshop equipment to ensure the stability of data transmission; for the scene that the valve only needs to transmit discrete data such as switch state and action feedback, RS485 interface module is integrated and adapted with the valve controller to realize low-cost data acquisition. In the aspect of Internet of Things protocol adaptation, the edge node communication unit carries different protocol modules according to the data transmission distance and real-time requirement: LoRa module is used to adapt LoRaWAN protocol for short-distance, low-power conventional monitoring data (such as environmental temperature) in the workshop; 5G-Industrial module is used to adapt 5G-Industrial protocol for cross-regional, high real-time fault warning data; WiFi6E module is used to adapt WiFi6E protocol for high-bandwidth real-time control data (such as motor speed regulation instruction) in the workshop; NB-IoT module is used to adapt NB-IoT protocol for low-power pressure sensor data in remote corners. Through the integration of multi-protocol hardware modules and interface modules, the interconnection problem of “non-uniform interface and non-compatible protocol” of heterogeneous mechanical and electrical equipment in the workshop is completely solved, and the data of various equipment can be smoothly collected and transmitted.

[0024] The embodiment needs to be specifically explained that the fault detection and path re-planning process of distributed self-organizing routing, which is based on the practical application of improved AODV algorithm in inter-vehicle communication network: 2 regional nodes in the vehicle manage 25 devices respectively, and multiple LoRa gateways are deployed under each regional node. The gateways are distributed in a "grid-like" topology, each gateway periodically sends "hello packet" to adjacent gateways, which contains gateway ID, current load, communication state, and adjacent gateways need to reply "acknowledgement packet" within a short time after receiving. If a gateway does not receive the "acknowledgement packet" of the target gateway for many times in a row, it is determined that the gateway is faulty (for example, a LoRa gateway in the east region stops working due to power failure). After fault determination, the gateway initiating the detection immediately broadcasts "fault information" (including fault gateway ID, fault time, and current network topology) to all other gateways in the region, and other gateways start path re-planning after receiving: taking "the nearest distance and the lowest load" as the principle, the gateway built-in routing calculation chip quickly filters and selects the replacement path, for example, a gateway in the east region originally transmits device data through the faulty gateway, and automatically switches to transmit through other normal gateways after the fault, and the path re-planning time is short. At the same time, the regional node receives the topology update information of each gateway in real time, synchronously updates the regional communication routing table, and synchronously updates the update result to the top node, ensuring the dynamic transparency of the entire communication network topology. Through the logic of "active detection-quick response-collaborative re-planning", the mechanism avoids the interruption of device data transmission in the region due to the failure of a single gateway, and ensures the continuity and reliability of the data transmission of 50 mechanical and electrical equipment in the workshop.

[0025] The embodiment needs to be specifically explained that the actual execution process of dynamic bandwidth allocation and protocol conversion, the core is to realize the accurate allocation of communication resources and the unified compatibility of heterogeneous data based on data priority: first, data priority determination, the priority identification chip built-in the communication module will determine the priority according to the "tag field" (such as fault warning tag, control instruction tag, monitoring data tag) carried by the data, combined with the running rules of the workshop equipment : when the motor current and vibration acceleration exceed the rated range of the equipment, the data is determined as (fault warning data); when the data is motor speed regulation instruction and valve switch control instruction, it is determined as (real-time control data); when the data is non-critical data such as workshop environment temperature and equipment surface temperature, it is determined as (conventional monitoring data). Then dynamic bandwidth allocation: the bandwidth management module of the regional node communication unit allocates link resources according to priority, data occupies 5G-Industrial link main bandwidth, ensuring that fault warning information is transmitted to the regional node with low delay; data occupies a certain bandwidth of industrial Ethernet link, meeting the high-speed interaction demand of control instruction; data occupies a small amount of bandwidth of the LoRa link, and realizes the conventional data transmission at a low bandwidth cost. Finally, the protocol conversion, the protocol conversion chip (supporting multi-protocol analysis) built in the communication module converts the collected heterogeneous data: the current analog data of the motor Profinet protocol, the pressure digital data of the pump Modbus-RTU protocol, and the switch state data of the valve RS485 protocol into the MQTT protocol (used for the real-time data of 、 ) or the CoAP protocol (used for the low-power consumption data of ). In the conversion process, the data integrity is ensured through the data format verification (such as checksum calculation), and the conversion delay is low. Through the process, the real-time and reliability of the key data are ensured, the unified access of the heterogeneous data is realized, and the standardized data format is provided for the subsequent distributed storage, hierarchical control module data processing.

[0026] The hierarchical distributed control module is used for receiving the preprocessed equipment operation data transmitted by the edge perception and data preprocessing module and the optimization control strategy transmitted by the intelligent decision module. The distributed trusted data storage module includes the following aspects: The hierarchical storage architecture is adopted, the edge node stores the real-time operation data of the mechanical and electrical equipment in the last one hour, the regional node stores the historical operation data of the mechanical and electrical equipment in the last thirty days, and the top node stores the global operation and maintenance data across regions; The distributed index is established based on the timestamp and the equipment ID; The real-time operation data and the historical operation data are stored in the distributed time series database, and the equipment fault record and the control instruction log are stored in the industrial alliance chain.

[0027] It is necessary for the embodiment to specifically describe the landing logic of the hierarchical storage architecture of the distributed trusted data storage module, and the data hierarchical management is realized in combination with the three-level node layout of the workshop "edge-region-top": the edge node corresponds to the local storage unit of each mechanical and electrical equipment, and only stores the real-time operation data of the equipment in the last one hour (such as the real-time current of the motor and the real-time pressure of the pump body). The cyclic coverage strategy is adopted, when the storage capacity reaches the upper limit, the earliest data is automatically covered, and the demand of the edge layer for "real-time data fast calling and low storage cost" is met. The regional node (one in the east region and one in the west region) stores the historical operation data of 25 devices in the last thirty days (such as the daily average speed and the periodic pressure change) in the jurisdiction, and the data is classified and archived according to "equipment type-date", so as to facilitate the regional operation and maintenance personnel to trace the recent equipment operation trend. The top node stores the cross-regional global operation and maintenance data of the whole workshop and other regions of the factory (such as multi-workshop fault statistics and overall energy consumption data), and provides data support for factory-level resource optimization. Through hierarchical storage, the storage pressure of the edge node is avoided, and the redundant data transmission between the regional and top nodes is reduced, realizing "data storage nearby and calling on demand".

[0028] The embodiment needs to be specifically explained that the distributed data index design based on timestamp and device ID, the core is to quickly locate the target data: allocate "device ID-time stamp" double identification for each stored data, the device ID is set according to the workshop device number rule (such as motor number "M001-M030", pump body number "P001-P015"), the time stamp is accurate to millisecond level, and the two form a unique index key. The index structure adopts distributed B+ tree, the edge node index is only associated with local device data, the regional node index integrates the jurisdiction device index and is associated with the edge node index address, and the top node index summarizes all regional node indexes. When the operation and maintenance personnel query the data of a certain time period of a certain device (such as M012 motor), the system first locates the corresponding regional node through the device ID, then filters the target period according to the time stamp, and finally directly calls the original data of the edge or regional node, realizes millisecond level query response, and greatly improves the data retrieval efficiency.

[0029] The embodiment needs to be specifically explained that the fusion application logic of time series database and alliance chain: for the real-time running data of the device (such as the current value collected every second), the historical running data (such as daily energy consumption statistics), because it has the characteristics of "high frequency writing, time sequence association", it is stored in a distributed time series database, which supports high concurrency writing and time sequence dimension query, and adapts to the read-write requirements of machine and electrical equipment data; for the device fault record (such as motor overheating alarm details), control instruction log (such as valve opening and closing instruction content) and other key data, because it needs to be "tamper-proof, traceable", it is stored in an industrial alliance chain, the alliance chain nodes cover the workshop regional nodes and the factory top nodes, and a unique hash value is generated when each key data is written. If the data of a node is tampered with, the hash value does not match and cannot be verified. Through the fusion mode of "storing regular data in time series database and storing key data in alliance chain", the read-write efficiency of regular data is guaranteed, and the credibility of key data is ensured.

[0030] Hierarchical distributed control module: used for receiving the preprocessed device running data transmitted by the edge sensing and data preprocessing module, and the optimization control strategy transmitted by the intelligent decision module, the hierarchical distributed control module realizes: The edge control level receives the preprocessed device running data, and realizes real-time closed-loop control by using adaptive PID algorithm; The regional control level coordinates the linkage of multiple machine and electrical equipment, ensures that the control targets of multiple devices are consistent, and the control error is ≤0.5%; The control error is calculated as follows: assuming that the target control value of multiple devices is T, the actual control value of the jth device is , the control error is , wherein m is the number of devices; The top control level dynamically allocates power of the power grid based on global data optimization of resource allocation; The control instruction is double-checked before execution, and the authorized signature of the regional node and the rationality of the local data are verified.

[0031] Further, the adaptive PID algorithm optimizes parameters through reinforcement learning, and defines a reward function to evaluate the parameter optimization effect: let the current PID parameters be the proportional coefficient Kp, the integral coefficient Ki, and the differential coefficient Kd, the actual output value of the device be y(t), and the target output downlink value be r(t), then the error e(t) = |r(t)-y(t)|; the reward function wherein: a, b, d, f are weight coefficients (the value range is 0.1-0.4, and a+b+d+f=1), and c is a very small positive number (the value is 0.001, to avoid e(t)=0 denominator is 0); ΔKp, ΔKi, ΔKd are the differences between the current parameters and the parameters of the last period, respectively; The larger the R value, the better the parameter optimization effect.

[0032] The embodiment needs to be specifically explained that the real-time closed-loop control logic of the edge control level, the core revolves around the landing application of the adaptive PID algorithm: after the edge node receives the preprocessed data (such as motor current, speed) transmitted by the edge perception module, it immediately starts the real-time closed-loop control process. Taking motor speed control as an example, the module first obtains the target output value (such as the speed required by production) and the actual output value of the device, calculates the error between the two, and calls the adaptive PID algorithm to generate a control instruction. The algorithm dynamically optimizes the proportional coefficient, integral coefficient, and differential coefficient through reinforcement learning, and the optimization basis is the reward function: the function comprehensively considers the error size and parameter change amplitude, finally selects the parameter combination with the maximum reward value to execute control, ensures fast control response, and meets the millisecond-level real-time adjustment requirements of mechanical and electrical equipment.

[0033] The embodiment needs to be specifically explained that the multi-device coordination logic of the regional control level, which realizes device linkage control based on the consensus algorithm: taking the linkage of multiple motors of a production line in a workshop as an example, the regional node first receives the unified target control value (such as the synchronous speed of the production line) issued by the top layer or the decision module, and then collects the actual control values of each motor. The overall control error is calculated through the consensus algorithm, that is, the deviation average of the actual values of multiple devices and the target value, if the error exceeds the allowed range, the regional node immediately issues parameter adjustment instructions to the edge node corresponding to the motor with larger deviation, until the actual values of all motors are consistent with the target value, ensuring the synchronization and stability of the production line device linkage.

[0034] The embodiment needs to be specifically explained that the global resource optimization logic of the top control level focuses on the dynamic allocation of power grid power: the top node receives the mechanical and electrical equipment load data uploaded by each regional node in real time (such as the total power of each regional motor and the running power consumption of the pump body), combines the overall power supply capacity of the power grid, and allocates the power quota of each regional power grid according to the principle of "load matching and high efficiency and energy saving", appropriately increases the quota of the region with high load, and reasonably reduces the quota of the region with low load, so as to avoid that the power overload of a region leads to equipment downtime, and at the same time, maximize the utilization rate of power grid resources and support stable operation of multi-region equipment.

[0035] The embodiment needs to be specifically explained that the control instruction double-checking process builds a strong defense for safe operation of the equipment: in the instruction issuing stage, the hierarchical distributed control module first verifies the regional node authorization signature carried by the instruction, and confirms by the encryption algorithm that the instruction source is a legal regional node; after the signature verification passes, it is further verified whether the equipment operation parameters required by the instruction match the rated range of the equipment. If the parameter exceeds the rated value, the module immediately terminates the instruction execution, synchronously generates alarm information and uploads it to the system management database and the operation and maintenance end, so as to avoid the damage risk of the equipment caused by misoperation or malicious instruction from the source.

[0036] Intelligent decision module: used for receiving historical data and real-time data transmitted by the distributed trusted data storage module, generating equipment fault prediction results, operation and maintenance optimization suggestions and global optimization control strategies, the intelligent decision module realizes the following functions: Each regional node transmits the trained model parameters to the top node, and the top node aggregates the model parameters to generate a global fault prediction model; The model parameter aggregation adopts a weighted average algorithm: let the model parameters of the pth regional node be , the sample quantity of the node be , the total quantity of global samples be , and q be the number of regional nodes, then the aggregated global parameters ; Based on the global fault prediction model, the mechanical and electrical equipment is fault predicted, and a dynamic operation and maintenance priority is generated combining with the production plan, and the equipment with low health degree and affecting the urgent production task is preferentially included in the maintenance plan.

[0037] The embodiment needs to be specifically explained that the intelligent decision module realizes the logic, and the core is developed around "data privacy protection + global model optimization": first, data reception and local training, the module obtains device running data, environment data and operation and maintenance historical data from the distributed trusted data storage module, and each regional node carries out model training based on the federal learning framework; second, parameter aggregation, each regional node transmits the model parameters of local training to the top node, and the top node aggregates the global parameters according to the "sample size weighted average" strategy, that is, according to the proportion of the sample size of each regional node, the global fault prediction model is calculated to ensure that the model takes into account the running characteristics of the equipment in each region; finally, prediction and strategy output, the module calls the global fault prediction model to calculate the fault probability of the equipment, generates a dynamic operation and maintenance scheme combining the production plan priority, and the equipment with low health degree and urgent production influence is maintained preferentially, and the global optimization control strategy (such as regional power allocation scheme) is transmitted to the hierarchical distributed control module to provide intelligent basis for equipment control.

[0038] Multi-level dynamic security protection module: used for security protection of the running process of the edge perception and data preprocessing module, the distributed self-organizing communication module, the distributed trusted data storage module, the hierarchical distributed control module and the intelligent decision module, the multi-level dynamic security protection module realizes the following: A normal behavior baseline is established for each edge node, and an alarm is triggered and the node is isolated when the node behavior deviates from the baseline, and the behavior baseline includes the node regular communication frequency and data transmission amount range; Communication frequency deviation calculation: let the node regular communication frequency be , the current communication frequency be , and the deviation , when β exceeds the preset threshold, an alarm is triggered; Data transmission amount deviation calculation: let the node regular unit time transmission amount be , the current unit time transmission amount be , and the deviation , when λ exceeds the preset threshold, an alarm is triggered.

[0039] The embodiment needs to be specifically explained that the multi-level dynamic security protection module has a hierarchical encryption logic, and the encryption scheme is adapted to different scenes: the device end adopts a lightweight national encryption algorithm to encrypt the collected raw data due to limited computing power, and ensures the safety of the data source in low computing power consumption; the global data is transmitted between the regional node and the top node, and is encrypted by using an asymmetric national encryption algorithm, so that the data is not eavesdropped or tampered with in cross-level transmission through the "public key encryption-private key decryption" mechanism, and the security and transmission efficiency are considered.

[0040] The embodiment needs to be specifically explained that the control instruction anti-replay attack mechanism: each time the control instruction (such as motor speed regulation, valve opening and closing instruction) is generated, the module synchronously generates a random key, and the key is bound and associated with the instruction; the key needs to be verified by the instruction receiving end, and the key is only valid in the current instruction period, and subsequent attacks cannot be initiated by repeatedly using old instructions and old keys, so as to block the replay risk from the source.

[0041] The embodiment needs to be specifically explained that the edge node behavior baseline monitoring process: first, establish a normal behavior baseline for each edge node, including the normal communication frequency and the data transmission amount range per unit time; the current communication frequency deviation and the data transmission amount deviation of the node are calculated in real time, and when any deviation exceeds the preset threshold, an alarm is triggered immediately and the node is isolated to prevent the spread of abnormal nodes.

[0042] The embodiment needs to be specifically explained that the whole-process operation audit mechanism: the module records all key operations (such as control instruction issuing, PID parameter modification, model parameter updating), including the operator, operation time and operation content; the operation record is stored in the distributed trusted data storage module in real time, relying on the tamper-proof feature of the storage module to ensure that the operation is traceable and convenient for subsequent audit and investigation of abnormal operations.

[0043] Secondly: in the drawings of the disclosed embodiment, only the structures related to the disclosed embodiment are involved, other structures can refer to the usual design, and in the case of no conflict, the same embodiment and different embodiments of the present application can be combined with each other; Finally: the above only describes the preferred embodiments of the present application and is not used to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An Internet of Things based distributed electromechanical equipment data control system, characterized in that, Comprise: System management database: for storing design data required during system operation and receiving real-time generated data, building a distributed mechanical and electrical equipment data control and management database; Edge perception and data preprocessing module: for collecting key operating parameters of mechanical and electrical equipment, preprocessing the collected data, and obtaining preprocessed equipment operating data; Distributed self-organizing communication module: for realizing multi-link interconnection between edge perception and data preprocessing module, distributed trusted data storage module, hierarchical distributed control module and intelligent decision module; Distributed trusted data storage module: for receiving data transmitted by the distributed self-organizing communication module; Hierarchical distributed control module: for receiving preprocessed equipment operating data transmitted by the edge perception and data preprocessing module, and optimization control strategy transmitted by the intelligent decision module; Intelligent decision module: for receiving historical data and real-time data transmitted by the distributed trusted data storage module, generating equipment fault prediction results, operation and maintenance optimization suggestions and global optimization control strategy; Multi-level dynamic security protection module: for security protection of the operation process of edge perception and data preprocessing module, distributed self-organizing communication module, distributed trusted data storage module, hierarchical distributed control module and intelligent decision module.

2. The distributed electromechanical equipment data control system based on the Internet of Things according to claim 1, characterized in that: The edge perception and data preprocessing module comprises: S1.1: collecting vibration, current, speed and pressure parameters of mechanical and electrical equipment through industrial-grade sensors; S1.2: Noise filtering, outlier rejection and data compression processing are performed on the collected parameters: let a certain type of parameter sample set collected be , where n is the number of samples, is the i-th sample value; the sample mean is calculated ; the sample standard deviation is calculated ; the statistical threshold range is set to , where k is the confidence coefficient, and the value is 2.5 to 3.0, and samples outside this range are determined as outliers; S1.3: classifying the preprocessed equipment operating data according to data types, and transmitting them to the distributed self-organizing communication module and the hierarchical distributed control module respectively.

3. The distributed electromechanical equipment data control system based on the Internet of Things according to claim 1, characterized in that: The distributed self-organizing communication module comprises: Supporting distributed self-organizing routing, when a communication node fails, other nodes automatically re-plan the transmission path; Adopting dynamic bandwidth allocation strategy, classifying data according to priority, and uniformly converting heterogeneous data into MQTT protocol or CoAP protocol.

4. The distributed electromechanical equipment data control system based on the Internet of Things according to claim 1, characterized in that: The distributed trusted data storage module comprises: Adopting hierarchical storage architecture, edge nodes store real-time operating data of mechanical and electrical equipment in the last one hour, regional nodes store historical operating data of mechanical and electrical equipment in the last thirty days, and top nodes store global operation and maintenance data across regions; Based on timestamp + device ID to establish distributed index; Real-time operating data and historical operating data are stored in distributed time series database, and device fault records and control instruction logs are stored in industrial alliance chain.

5. The distributed electromechanical equipment data control system based on the Internet of Things according to claim 1, characterized in that: The hierarchical distributed control module comprises: The edge control level receives preprocessed equipment operating data and realizes real-time closed-loop control by using adaptive PID algorithm; The regional control level coordinates the linkage of multiple mechanical and electrical equipment to ensure that the control targets of multiple equipment are consistent, and the control error is ≤0.5%; The control error is calculated as follows: let the target control value of a plurality of devices be T, the actual control value of the jth device be , and the control error be , where m is the number of devices. The top control level optimizes resource allocation based on global data to dynamically allocate power grid power; Double-checking is performed before executing the control instruction to verify the authorization signature of the regional node and the rationality of the local data.

6. The distributed electromechanical equipment data control system based on the Internet of Things according to claim 1, characterized in that: The intelligent decision module comprises: Receiving equipment operating data, environmental data and operation and maintenance history data transmitted by the distributed trusted data storage module, and training local models in each regional node; Each regional node transmits the trained model parameters to the top node, and the top node aggregates the model parameters to generate a global fault prediction model; The model parameter aggregation adopts a weighted average algorithm: let the model parameter of the pth regional node be , the sample quantity of the node be , the total quantity of global samples be , and q be the quantity of regional nodes, then the aggregated global parameter is ; Based on the global fault prediction model, the mechanical and electrical equipment is predicted for failure, and the dynamic operation and maintenance priority is generated combined with the production plan. The equipment with low health degree and urgent impact on production tasks is preferentially included in the maintenance plan.

7. The distributed electromechanical equipment data control system based on the Internet of Things according to claim 1, characterized in that: The multi-level dynamic security protection module realizes: A normal behavior baseline is established for each edge node, and an alarm is triggered and the node is isolated when the node behavior deviates from the baseline. The behavior baseline includes the node's regular communication frequency and data transmission range. Communication frequency deviation degree calculation: assuming the node's regular communication frequency is , the current communication frequency is , then the deviation degree is , and an alarm is triggered when β exceeds a preset threshold value; Data transmission quantity deviation degree calculation: assuming that the regular unit time transmission quantity of a node is , the current unit time transmission quantity is , then the deviation degree is , and an alarm is triggered when λ exceeds a preset threshold.

8. The distributed electromechanical equipment data control system based on the Internet of Things according to claim 5, characterized in that: In the edge control level of the hierarchical distributed control module, the adaptive PID algorithm optimizes the parameters through reinforcement learning, and the parameter optimization period is adjusted according to the change of device load. When the device load fluctuation amplitude exceeds the preset value, the optimization period is shortened. The device load fluctuation amplitude is calculated by continuously collecting the device current parameters, and the calculation period is one minute. The specific calculation is as follows: Let the current parameter sequence collected continuously in one minute be , where t is the collection times, is the current value collected for the kth time; calculate the maximum value of the current and the minimum value ; the load fluctuation amplitude , where ; when ΔI> the preset threshold, output a first-level signal, and when ΔI≤ the preset threshold, output a second-level signal.

9. The distributed electromechanical equipment data control system based on the Internet of Things according to claim 5, characterized in that: The adaptive PID algorithm optimizes parameters through reinforcement learning, and defines a reward function to evaluate the parameter optimization effect: let the current PID parameters be the proportional coefficient Kp, the integral coefficient Ki, and the differential coefficient Kd, the actual output value of the equipment be y(t), the target output downlink value be r(t), and the error e(t) = |r(t)-y(t)|; the reward function wherein: a, b, d, f are weight coefficients (the value range is 0.1-0.4, and a+b+d+f=1), and c is a very small positive number (value 0.001, to avoid e(t)=0 denominator 0); ΔKp, ΔKi, ΔKd are the differences between the current parameters and the parameters of the last period, respectively.

Images