Equipment management platform based on industrial internet

Abstract

The invention discloses an equipment management platform based on an industrial internet, and relates to the technical field of equipment control management. A temperature sensor, a humidity sensor and other sensors are arranged at an equipment end to collect environment parameters, and a correlation model is constructed to achieve dynamic environment perception and self-adaptive adjustment. A high-precision monitoring device and a precise synchronization mechanism are utilized to solve the problem of multi-device linkage synchronism, and the production efficiency is improved. The comprehensive energy consumption is evaluated by applying a system dynamics model, and an energy-saving strategy is formulated to reduce energy consumption. An intelligent scheduling algorithm based on reinforcement learning is adopted, edge and cloud computing resources are dynamically allocated, and the utilization rate of the computing resources is increased. And a data drift threshold is set and combined with a Kalman filtering algorithm, data drift is monitored and corrected, and data accuracy is guaranteed. The platform effectively solves many problems in industrial equipment management, improves the efficiency and quality of industrial production and the energy utilization level, and has remarkable economic benefits and practical values.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment control and management, and specifically to an equipment management platform based on the industrial Internet. Background Art

[0002] In the field of industrial production, with the increasing complexity of equipment and the continuous expansion of production scale, many problems have gradually emerged in traditional equipment management methods. On the one hand, the operation of equipment is greatly affected by environmental factors, but there is a lack of effective monitoring and analysis of environmental parameters, and it is impossible to adjust the equipment operation status in time to adapt to environmental changes, resulting in a decline in equipment performance and unstable product quality. On the other hand, when multiple devices are linked, due to differences in the response speed and operation cycle of different devices, there is a lack of a precise synchronization mechanism, resulting in product backlogs or waiting, and low production efficiency. At the same time, in terms of energy management, there are defects in the comprehensive energy consumption assessment across different types of equipment, and it is impossible to formulate a comprehensive and efficient energy-saving strategy, and the phenomenon of energy waste is serious. In addition, the equipment management platform also faces problems such as data drift and unreasonable allocation of computing resources, which seriously restrict the development of industrial production. Therefore, there is an urgent need for a new equipment management platform to solve these problems.

[0003] In view of this, the present application is specifically proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide an equipment management platform based on the industrial Internet to solve the problems mentioned in the above background art.

[0005] To solve the above technical problems, the equipment management platform based on the industrial Internet provided by the present invention includes:

[0006] Dynamic Environment Perception and Adaptive Module: It includes deploying temperature sensors, humidity sensors, electromagnetic interference sensors, and various environmental sensors at the equipment end to collect environmental parameters around the equipment in real time; using data analysis algorithms to construct an association model between environmental parameters and equipment performance, and adaptively adjusting the equipment operation parameters according to this model; in the stamping workshop of an automobile manufacturing factory, deploying PT100 platinum resistance temperature sensors, HIH-4000 humidity sensors, and Langer EMV3000 magnetic field sensors to collect environmental parameters; establishing an association model through multiple linear regression algorithms combined with regularization techniques, and adjusting the operation parameters of stamping equipment according to the model, such as adjusting the stamping speed and die cooling system parameters. Ensure the stable operation of equipment in a complex environment and improve product quality. For example, effectively control the accuracy deviation of stamped parts and reduce the scrap rate from 5% to 2%;

[0007] Multi-device linkage synchronization control module: It is used to solve the synchronization problem in the scenario of multi-device linkage, monitor the response speed and operation cycle of different devices, and ensure the efficient connection during the collaborative operation of multiple devices through a precise synchronization mechanism. In the chassis assembly line of an automobile final assembly workshop, tightening equipment, handling robots, and positioning jigs are equipped with high-precision monitoring devices. A GPS-based time synchronization device and a feedback control algorithm are adopted to adjust the speed according to the operation time deviation of the handling robot, and feedforward and predictive control technologies are introduced. The platform can also optimize the operation sequence and time interval of equipment, and adjust the equipment operation parameters according to the vehicle model. It solves the multi-device linkage synchronization problem, reduces product backlog or waiting phenomena, and improves production efficiency. For example, the automobile chassis assembly time is shortened from 35 minutes to 28 minutes, and the production efficiency is increased by more than 40%;

[0008] Comprehensive energy efficiency evaluation and energy-saving strategy formulation module: For the comprehensive energy consumption evaluation across different device types, it can evaluate the overall energy consumption of the collaborative operation of multiple devices under different production tasks, and then formulate a comprehensive and efficient energy-saving strategy. In the painting workshop of an automobile manufacturing factory, high-precision power monitoring instruments and other sensors are installed on painting robots, drying equipment, and ventilation systems to collect data. A system dynamics model is used to evaluate the comprehensive energy consumption, and the equipment operation parameters are adjusted according to the evaluation results, such as reducing the temperature of the drying equipment, adjusting the spray gun flow rate of the painting robot, and the fan speed of the ventilation system. It realizes the comprehensive energy consumption evaluation across different device types, formulates an energy-saving strategy, and reduces energy consumption. For example, the energy consumption in the painting workshop is reduced by about 16.9%, saving production costs;

[0009] Edge and cloud computing resource dynamic allocation module: It introduces an intelligent scheduling algorithm to monitor the data generation volume of devices, the resource usage of edge computing nodes and cloud servers in real time, including CPU utilization rate, memory occupancy, network bandwidth, etc., and dynamically adjusts the allocation ratio of edge computing and cloud computing resources according to the priority of production tasks and real-time computing requirements. The Zabbix tool is used to monitor the resources of the data center server, and the Advantech UNO-2482G industrial gateway is used to monitor the resources of the edge computing node. The Deep Q-Network (DQN) algorithm is adopted to dynamically allocate computing resources according to the data generation volume of devices, the priority of production tasks, and the resource usage. It optimizes the allocation of computing resources, improves the utilization rate of computing resources, and reduces data processing latency. The average utilization rate of computing resources is increased by 35%, and the average data processing latency is reduced by 45%, ensuring the efficient operation of the device management platform;

[0010] Data drift monitoring and correction module: It is used to monitor the data drift phenomenon caused by reasons such as sensor aging and the accumulation of subtle environmental factor changes during the long-term operation of the device; in the engine assembly workshop of an automobile manufacturing factory, the HBMT40B strain gauge torque sensor is used to collect data and transmit it through industrial Ethernet. Set the data drift threshold, use the Kalman filtering algorithm to process the data, fuse multi-source data to locate the cause of the drift, and establish a verification model to verify the data accuracy. Monitor and correct data drift, ensure data accuracy, improve the correctness of equipment status judgment and decision-making, and avoid misjudgment of equipment status and wrong decisions based on incorrect data. The accuracy rate of the engine assembly decision based on the corrected data is increased by more than 98%, improving the engine assembly quality and production efficiency.

[0011] Furthermore, in the dynamic environment perception and adaptive module, the data analysis algorithm includes at least one of the multiple linear regression algorithm and the neural network algorithm. By learning historical environmental parameters and equipment performance data, a more accurate correlation model is established; in the stamping workshop of an automobile manufacturing factory, the multiple linear regression algorithm combined with regularization technology is used to analyze 100 sets of environmental parameter and stamping part accuracy data collected every day in the past year, and establish a correlation model between environmental parameters and equipment performance. By learning historical data, a more accurate correlation model is established, providing a more reliable basis for the adaptive adjustment of equipment operation parameters, and further improving the adaptability of the equipment to environmental changes and product quality.

[0012] Furthermore, in the multi-device linkage and synchronous control module, the precise synchronization mechanism adopts high-precision clock synchronization technology and a dynamic adjustment algorithm based on feedback control to adjust the equipment operation rhythm in real time, ensuring that the synchronization accuracy of each device during linkage operation reaches the millisecond level; on the chassis assembly line in the general assembly workshop of an automobile, a high-precision time synchronization device based on GPS is used, and the synchronization accuracy can reach ±1μs. The dynamic adjustment algorithm based on feedback control is used to adjust the speed according to the running time deviation of the handling robot. Different adjustment coefficients are used when handling heavy parts and precision assembly. Ensure that the synchronization accuracy of each device during linkage operation reaches the millisecond level, improve the precision of multi-device collaborative operation, reduce product backlog or waiting, and enhance production efficiency.

[0013] Furthermore, in the comprehensive energy efficiency evaluation and energy-saving strategy formulation module, when evaluating the comprehensive energy consumption, multi-dimensional factors such as equipment load rate, operating time, and energy conversion efficiency are considered, and the system dynamics model is used to simulate and analyze energy consumption; in the paint shop of the automobile manufacturing plant, when evaluating the comprehensive energy consumption, the working time of the paint robot, the flow rate of the spray gun, the temperature and working time of the drying equipment, the speed and working time of the ventilation system fan, as well as the equipment load rate, energy conversion efficiency and other multi-dimensional factors are considered, and the system dynamics model is used to simulate and analyze energy consumption. Comprehensive and accurate evaluation of comprehensive energy consumption provides a basis for formulating more scientific and reasonable energy-saving strategies, effectively reducing energy consumption and improving energy utilization efficiency.

[0014] Furthermore, in the edge and cloud computing resource dynamic allocation module, the intelligent scheduling algorithm is an algorithm based on reinforcement learning. By continuously interacting with the system environment, it learns the optimal resource allocation strategy to improve the overall utilization of computing resources. The deep Q-network (DQN) algorithm based on reinforcement learning is adopted. Through 10,000 rounds of training, the intelligent agent and the environment interact to learn the optimal resource allocation strategy. Computing resources are dynamically allocated according to the resource conditions of edge computing nodes and cloud servers, the amount of device data generated, and the priority of production tasks. Learn the optimal resource allocation strategy, improve the overall utilization of computing resources, meet the complex and changeable production needs of automobile manufacturing plants, and improve the performance and response speed of the equipment management platform.

[0015] Furthermore, in the data drift monitoring and correction module, by setting a data drift threshold and using the Kalman filter algorithm to process the collected data in real time, effective monitoring and correction of data drift is achieved; in the engine assembly workshop of the automobile manufacturing plant, ±3% of the engine torque measurement value is set as the data drift threshold. The Kalman filter algorithm is used to process the sensor collected data in real time, and the drift cause is analyzed through multi-source data fusion and Bayesian network. A verification model is established to verify the data accuracy using RMSE and MAE indicators. Effectively monitor and correct data drift, accurately locate the drift cause by combining multi-source data, ensure data accuracy, and improve the reliability of data-based decision-making, thereby improving engine assembly quality and production efficiency.

[0016] Furthermore, it also includes an equipment status prediction module, which uses machine learning algorithms to analyze the equipment's historical operating data, real-time environmental parameters, and maintenance records, predict the equipment's future operating status, and warn of potential failures in advance; for example, it predicts the wear of stamping dies in high temperature environments and arranges equipment maintenance plans in advance. Early warning of potential failures allows companies to arrange maintenance in advance, reduce equipment downtime, ensure normal equipment operation, and reduce maintenance costs.

[0017] Furthermore, it also includes a user interaction interface module, which can intuitively display information such as the device operation status, environmental parameters, energy consumption data, and computing resource allocation, and receive instructions input by users to achieve convenient operation and control of the device management platform; it is convenient for operators to intuitively understand the device operation situation, realize convenient operation and control, and improve the convenience and efficiency of device management.

[0018] Furthermore, it also includes a data interaction module with external systems, including but not limited to at least one of the enterprise resource planning (ERP) system and the supply chain management (SCM) system, to achieve the interconnection and interoperability of production data and enterprise management data, improve the overall enterprise management efficiency, and promote the coordinated operation of various business links of the enterprise.

[0019] Furthermore, it also includes a storage and management module, which uses blockchain technology to store and manage key data in the device management platform to ensure the security, immutability, and traceability of the data.

[0020] Compared with the prior art, the beneficial effects of the present invention are:

[0021] 1. Improve the operation stability of the device and the product quality: The dynamic environment perception and adaptive module deploys a variety of high-precision environmental sensors at the device end, such as PT100 platinum resistance temperature sensors, HIH-4000 humidity sensors, etc., to collect environmental parameters in real time, and uses data analysis algorithms to construct an association model between environmental parameters and device performance. When environmental changes affect device performance, the platform automatically adjusts the device operation parameters according to the model. For example, in a stamping workshop, it can accurately control the precision deviation of stamping parts, reduce the scrap rate from 5% to 2%, greatly improve the product quality, and ensure the stable operation of the device in a complex environment.

[0022] 2. Improve production efficiency: The multi-device linkage synchronous control module solves the problem of synchronization in multi-device linkage with the help of high-precision monitoring devices and precise synchronization mechanisms. Taking the chassis assembly line in an automobile final assembly workshop as an example, by using GPS-based time synchronization devices and feedback control algorithms, the collaborative operation of handling robots and tightening devices is more accurate, and the product backlog or waiting phenomenon is reduced by more than 95%. The time required to assemble each automobile chassis is shortened from 35 minutes to 28 minutes, and the production efficiency is increased by more than 40%.

[0023] 3. Reduce energy consumption: The comprehensive energy efficiency evaluation and energy-saving strategy formulation module conducts a comprehensive energy consumption evaluation for cross-device types, considering multi-dimensional factors such as device load rate, operation duration, and energy conversion efficiency, and uses a system dynamics model for energy consumption simulation analysis. In the painting workshop of an automobile manufacturing factory, by optimizing device operation parameters and operation processes, such as reducing the temperature of drying equipment and adjusting the spray gun flow rate of painting robots, the energy consumption is reduced by about 16.9%, effectively saving production costs.

[0024] 4. Optimize computing resource allocation and improve data processing efficiency: The dynamic allocation module of edge and cloud computing resources adopts an intelligent scheduling algorithm based on reinforcement learning, which monitors the data generation volume of devices, the resource usage of edge computing nodes and cloud servers in real time, and dynamically allocates resources according to the production task priority and computing requirements. This increases the average utilization rate of computing resources by 35% and reduces the average data processing delay by 45%, ensuring the efficient operation of the device management platform and enabling it to quickly respond to production needs.

[0025] 5. Ensure data accuracy and decision reliability: The data drift monitoring and correction module processes the collected data in real time through setting data drift thresholds and Kalman filtering algorithms, effectively monitors and corrects data drift. In the engine assembly workshop, the processing of engine torque data controls the error within a very small range. Combining multi-source data fusion analysis, it accurately locates the cause of data drift. Through a strict data verification and feedback mechanism, the accuracy of engine assembly decisions based on the corrected data is increased by more than 98%, improving the equipment assembly quality and production efficiency.

[0026] 6. Strengthen the intelligence and integration of device management: The device status prediction module uses machine learning algorithms to predict potential device failures in advance based on the device's historical operation data, real-time environmental parameters and maintenance records, facilitating enterprises to arrange maintenance in advance and reducing device downtime. The user interface enables operators to intuitively understand information such as the device operation status and energy consumption data, and perform convenient operation and control; the data interaction function with external systems realizes the interconnection and interoperability of production data and enterprise management data, improving the overall management efficiency of enterprises. The application of blockchain technology ensures the security, immutability and traceability of key data, enhancing the credibility and stability of data management. Description of the Drawings

[0027] Figure 1 It is a schematic diagram of the composition of platform modules in the device management platform based on the industrial Internet;

[0028] Figure 2 It is a schematic diagram of the application of key technologies in the device management platform based on the industrial Internet;

[0029] Figure 3 It is a schematic diagram of the embodiment of platform advantages in the device management platform based on the industrial Internet;

[0030] Figure 4 It is a schematic diagram of the actual application scenario in the device management platform based on the industrial Internet. Detailed Embodiment

[0031] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0032] Please refer to Figures 1-4 , the present invention provides a technical solution: an equipment management platform based on the industrial Internet, including:

[0033] I. Dynamic Environment Perception and Adaptive System

[0034] Equipment Deployment: In the stamping workshop of an automobile manufacturing factory, the stable operation of large stamping equipment is crucial for the production quality of automobile parts. To accurately perceive the environmental parameters of equipment operation, a series of high-performance sensors are deployed. The selected PT100 platinum resistance temperature sensor adopts advanced thin-film platinum resistance technology, with a measurement accuracy of up to ±0.1°C and extremely high stability within a wide temperature range of -200°C to 850°C. The response time of this sensor is extremely short, only 0.2 seconds, and it can quickly capture the temperature changes of the stamping equipment caused by high-speed stamping, die friction, etc. For example, when stamping high-strength automotive steel plates, the temperature of the key parts of the stamping equipment may rapidly rise from normal temperature to 150°C in a short time, and the PT100 sensor can accurately and timely measure this temperature change.

[0035] The humidity sensor uses the HIH-4000 type, which is based on the principle of capacitive humidity sensing and has a measurement accuracy of up to ±2%RH. In the stamping workshop, changes in humidity will affect the surface characteristics of metal sheets and the lubrication effect of stamping dies, thereby affecting the accuracy of stamped parts. The measurement range of this sensor is 0%RH to 100%RH, and it can adapt to various humidity environments in the workshop. For example, during the plum rain season, the humidity in the workshop may rise from the normal 40%RH to 70%RH, and the HIH-4000 sensor can accurately monitor this humidity change.

[0036] In response to the complex electromagnetic environment in the workshop, the Langer EMV 3000 magnetic field sensor is introduced, with a measurement range of up to 1μA / m - 1000A / m, which can effectively detect electromagnetic interference generated by various electrical equipment. This sensor adopts a magnetic material with high magnetic permeability and an advanced induction coil design, and can also sensitively capture weak electromagnetic signals. In the stamping workshop, the electrical control systems of large stamping equipment, overhead cranes and other surrounding equipment will generate electromagnetic interference, and the Langer EMV3000 sensor can real-time monitor these interference signals, such as detecting that the electromagnetic interference intensity at a certain moment is 200A / m.

[0037] These sensors are connected to the device management platform through industrial-grade wired and wireless transmission modules. In areas where wiring is convenient and interference is low, an industrial-grade wired transmission module is used. Its transmission rate is as high as 10 Gbps, capable of quickly transmitting the environmental parameters collected by the sensors to the platform at a frequency of 500 times per second, with the transmission delay controlled within 1 millisecond. In some areas where devices move frequently or wiring is difficult, the wireless transmission module is enabled. This module uses the latest Wi-Fi6 technology, supports multi-user multiple-input multiple-output (MU-MIMO) function, can connect multiple sensors simultaneously, and can still maintain stable data transmission in a complex electromagnetic environment, with the transmission delay not exceeding 5 milliseconds.

[0038] Data analysis algorithms and correlation models:

[0039] The management platform uses the multiple linear regression algorithm and combines regularization techniques to establish a correlation model between environmental parameters and device performance. Let the relationship between the device performance index Y (such as the precision deviation of stamping parts, unit: mm) and temperature X1 (unit: °C), humidity X2 (unit: %RH), and electromagnetic interference intensity X3 (unit: A / m) be: Y = a0 + a1X1 + a2X2 + a3X3 + ∈;

[0040] Among them, a0, a1, a2, and a3 are regression coefficients, and ∈ is a random error term. To prevent the model from overfitting, an L2 regularization term is introduced, and the objective function becomes:

[0041]

[0042] Among them, N is the number of data samples, and λ is the regularization parameter. Through the cross-validation method, 100 groups of environmental parameters collected every day in the past year and the stamping part precision data on the same day are analyzed, and λ = 0.001 is optimized. The least squares method is used in combination with regularization techniques to estimate the regression coefficients, and finally a0 = 0.01, a1 = 0.002, a2 = 0.001, a3 = 0.0005 are obtained.

[0043] To verify the accuracy of the model, the root mean square error (RMSE) and the coefficient of determination (R 2 ) are used for evaluation. After verifying 1000 groups of test data, the RMSE is 0.02 mm, and the R 2 value reaches 0.95, indicating that the model has high accuracy and goodness of fit. When the real-time collected temperature is 30 °C, the humidity is 50 %RH, and the electromagnetic interference intensity is 50 A / m, according to the model, the predicted precision deviation of the stamping parts is approximately

[0044] Adaptive adjustment and intelligent decision-making:

[0045] If the accuracy deviation of the prediction exceeds the allowable range (e.g., ±0.1 mm), the equipment management platform will automatically activate the adaptive adjustment mechanism. By establishing a high-speed data connection with the programmable logic controller (PLC) of the stamping equipment, the platform can adjust the operating parameters of the stamping equipment in real time and accurately. For example, when the deviation exceeds the range, if it is analyzed that the temperature factor has the greatest impact on the accuracy deviation, the platform will automatically reduce the stamping speed by 10% according to the pre-established temperature-stamping speed adjustment model. This model is established based on 500 stamping experiment data at different temperatures and simulation analysis. While reducing the stamping speed, by optimizing the water flow speed of the mold cooling system from 5 L / min to 7 L / min and reducing the temperature from 25 °C to 20 °C, the accuracy deviation of the stamped parts can be reduced to ±0.08 mm, meeting the accuracy requirements.

[0046] In addition, the platform has an intelligent decision-making function. When the accuracy deviation exceeds the range due to temperature factors for 3 consecutive times, the platform will start a deep analysis program. Using the convolutional neural network (CNN) in deep learning, it deeply mines the environmental data, equipment operation data, and accuracy deviation data in the past year. After training, the CNN model finds that after stamping 1000 times continuously in a high-temperature environment, the wear amount of the key parts of the stamping die will increase by 0.1 mm, which will affect the accuracy. Based on this, the platform automatically generates maintenance suggestions, arranges the equipment maintenance plan in advance, and replaces the die when the stamping times reach 800 times to ensure the long-term stable operation of the equipment, reducing the scrap rate of the stamped parts from the original 5% to 2%, and significantly improving the production efficiency and product quality.

[0047] II. Multi-device linkage synchronous control

[0048] Equipment and monitoring:

[0049] On the chassis assembly line in the automobile final assembly workshop, the accuracy of the coordinated operation of various equipment directly affects the assembly quality and production efficiency of the automobile. Each piece of equipment is equipped with a high-precision monitoring device to achieve real-time and accurate collection of the equipment operation status and position information.

[0050] As one of the key equipment for chassis assembly, the tightening equipment is installed with an advanced torque sensor. This sensor adopts the advanced strain gauge measurement principle, and the measurement accuracy can reach ±0.5% FS (full scale). For example, if the full-scale torque of the tightening equipment is 500 N·m, its measurement error can be controlled within ±2.5 N·m. At the same time, a high-precision encoder is installed on the rotating shaft of the tightening equipment, and the resolution can reach 0.01°. This means that during the process of tightening the bolt, the encoder can accurately measure the rotation of the rotating shaft every 0.01°, providing extremely critical motion parameters for the synchronous control of the equipment. In actual operation, the encoder can accurately judge whether the tightening equipment reaches the set tightening angle to ensure the accuracy of bolt tightening.

[0051] During the operation of the handling robot, the accuracy of its position and posture is crucial for the assembly quality. Therefore, a lidar and a vision recognition system are equipped for the handling robot. The lidar adopts high-precision TOF (Time of Flight) technology, and the ranging accuracy can reach ±1 mm. In the chassis assembly line, the handling robot needs to accurately grasp and place parts. The lidar can scan the surrounding environment in real time to obtain accurate position information. For example, when handling large parts of the car chassis, the lidar can accurately measure the distance between the robot and the target position, and the error is controlled within a very small range. The vision recognition system uses a high-resolution camera and advanced image processing algorithms to monitor and recognize the assembled parts and the working environment in real time. The camera resolution is as high as 5 million pixels, which can clearly identify the tiny features of the parts. It can not only accurately identify the position and posture of the parts, but also detect possible assembly defects, such as deformation and damage of the parts. Through the fusion of the lidar and the vision recognition system, the positioning accuracy of the handling robot can be controlled within ±2 mm, greatly improving the accuracy and reliability of the handling operation.

[0052] As a reference device for chassis assembly, the positioning tooling is equipped with high-precision displacement sensors and angle sensors. The displacement sensor adopts the magnetostrictive principle, and the accuracy can reach ±0.05 mm, which can monitor the displacement changes of the positioning tooling in the horizontal and vertical directions in real time. During the chassis assembly process, the positioning tooling needs to accurately adjust its position to ensure the accurate assembly of the chassis parts. For example, when adjusting the horizontal position of the positioning tooling, the displacement sensor can accurately measure the displacement changes to ensure the positioning accuracy. The angle sensor adopts fiber optic gyro technology, and the accuracy can reach ±0.001°, which is used to monitor the rotation angle changes of the positioning tooling. These sensors transmit the collected data to the equipment management platform through the industrial Ethernet at a frequency of 1000 times per second, ensuring that the platform can grasp the operating status of the equipment in real time.

[0053] Precision synchronization mechanism:

[0054] To achieve precise synchronization of multiple devices, a high-precision time synchronization device based on GPS is adopted. The synchronization accuracy of this device can reach ±1 μs, providing a unified and precise time reference for the entire assembly line. Each device is connected to this device through a time synchronization module to ensure the time consistency between devices.

[0055] At the same time, a dynamic adjustment algorithm based on feedback control is adopted to achieve device synchronization. In the precision synchronization mechanism, assume that the handling robot moves from one place to another, and the originally planned time is T 计 , and the actual time spent is T 实 , and the difference between the two is Δt = T 实 -T计 When ΔT exceeds the allowable range (such as ±30 milliseconds), the running speed v of the handling robot is adjusted according to this deviation, and the adjustment formula is:

[0056] where k is the adjustment coefficient. Through a large number of experiments and simulation analyses, an adaptive adjustment strategy is adopted to optimize the k value for different equipment types and operation scenarios. For example, when handling heavier components with a mass greater than 50 kilograms, the k value is increased to 0.7; when performing precision assembly operations with extremely high position accuracy requirements, the k value is decreased to 0.3.

[0057] For example, the planned running time of the handling robot is 8 seconds, and the actual running time is 8.05 seconds, that is, ΔT = 0.05 seconds, and the current speed v 现 = 1.2 m / s. If heavier components are being handled at this time and k = 0.7, then the adjusted speed In this way, it is ensured that the handling robot is precisely synchronized when collaborating with tightening equipment, positioning tooling, etc.

[0058] In addition, to further improve the accuracy and stability of synchronous control, feedforward control and predictive control technologies are introduced. Feedforward control predicts possible deviations in advance based on the operating status of the equipment and task requirements, and makes adjustments before the deviations occur. Predictive control uses the historical operating data and real-time monitoring data of the equipment to establish a prediction model of the equipment operating status, predicts the future operating trend of the equipment in advance, and provides a more accurate decision-making basis for synchronous control. Through the comprehensive application of these technologies, the phenomenon of product backlog or waiting is reduced by more than 90%, and the production efficiency is increased by 40%.

[0059] In another implementation scenario, as shown in Tables 1, 2, and 3 below:

[0060]

[0061] Table 1: Diagram of the change in the running time deviation of the handling robot

[0062]

[0063] Table 2: Comparison diagram of the handling robot's speed before and after adjustment

[0064] Item Before Optimization After Optimization Time (min) for Assembling Each Automobile Chassis 35 28 Incidence Rate of Product Backlog or Waiting Phenomenon High Reduced by More than 95% Proportion of Improved Production Efficiency - 50％

[0065] Table 3: Comparison diagram of the production efficiency of the equipment linkage synchronous control before and after optimization

[0066] As can be seen from the above:

[0067] Real-time feedback and optimization adjustment of the system:

[0068] The device management platform receives the operation data transmitted by each device in real time and processes and analyzes the data in real time through the built-in data analysis module. When a synchronization deviation occurs between devices, the platform can not only adjust the operation parameters of the devices in time, but also optimize the entire assembly process flow.

[0069] For example, through in-depth analysis of the assembly data in the past six months, the platform found that after the handling robot transported the chassis parts to the tightening station, the tightening device needed to wait for an average of 2.5 seconds before starting to work, and this waiting time resulted in a loss of production efficiency. The platform used an optimization algorithm to adjust the operation sequence and time interval of the handling robot and the tightening device, shortening the waiting time to within 0.5 seconds.

[0070] At the same time, the platform has self-learning and adaptive optimization functions. Through learning and analysis of a large amount of historical data, the platform can automatically identify the best operation parameters and synchronization strategies of different devices in different operation scenarios, and make dynamic adjustments according to the actual production situation. For example, when producing different models of car chassis, the platform can automatically adjust the operation parameters and synchronization strategies of the devices according to the vehicle model characteristics and assembly requirements. When producing the chassis of an SUV model, due to the larger size and heavier weight of the parts, the running speed of the handling robot is reduced by 15%, and at the same time, the torque parameter of the tightening device is increased by 20% to ensure efficient and precise multi-device linkage synchronization control under different production tasks.

[0071] III. Comprehensive Energy Efficiency Evaluation and Energy Saving Strategy Formulation

[0072] 1. Equipment and Energy Consumption Monitoring

[0073] In the painting workshop of an automobile manufacturing plant, multiple devices such as painting robots, drying equipment, and ventilation systems operate together, and their energy consumption directly affects the production cost and energy utilization efficiency of the plant. To achieve accurate energy consumption monitoring, high-precision power monitoring instruments are installed at the power input end of each device, and the specific information is as follows:

[0074]

[0075] These power monitoring instruments can collect the energy consumption data of the devices at a frequency of once per minute and transmit the data to the device management platform in real time through the industrial Ethernet. At the same time, other types of sensors are also installed on the devices to monitor the operation status of the devices. For example, a spray gun flow sensor with a measurement accuracy of ±0.1 L / min is installed on the painting robot to monitor the paint flow of the spray gun in real time; a temperature sensor with a measurement accuracy of ±0.5 °C is installed on the drying equipment to monitor the drying temperature in real time; a fan speed sensor with a measurement accuracy of ±10 r / min is installed in the ventilation system to obtain the fan speed in real time.

[0076] 2. Comprehensive Energy Consumption Assessment Model and Energy Saving Strategies

[0077] Comprehensive Energy Consumption Assessment Model

[0078] Use the system dynamics model to evaluate the comprehensive energy consumption. Taking the painting workshop as an example, assume that the energy consumption E1 of the painting robot is related to the working hours t1 (unit: h) and the spray gun flow rate q1 (unit: L / min), the energy consumption E2 of the drying equipment is related to the temperature T (unit: °C) and the working hours t2 (unit: h), and the energy consumption E3 of the ventilation system is related to the fan speed n (unit: r / min) and the working hours t3 (unit: h). Through the analysis and fitting of a large amount of historical data, the following comprehensive energy consumption model is established:

[0079]

[0080] Example of Energy Consumption Analysis

[0081] When the daily production task of the workshop is to spray 1000 cars, the operating parameters of each device are as follows: the working hours of the painting robot t1 = 9h, the spray gun flow rate q1 = 6L / min; the temperature of the drying equipment T = 160°C, the working hours t2 = 7h; the fan speed of the ventilation system n = 1600r / min, the working hours t3 = 11h.

[0082] According to the above comprehensive energy consumption model, calculate the energy consumption of each device and the total energy consumption:

[0083] Energy consumption of the painting robot: E1 = 0.6×9×6 = 32.4kWh;

[0084] Energy consumption of the drying equipment: E2 = 0.012×160×7 = 13.44kWh;

[0085] Energy consumption of the ventilation system: E3 = 0.004×1600×11 = 70.4kWh;

[0086] Comprehensive energy consumption: E 总 = E1 + E2 + E3 = 32.4 + 13.44 + 70.4 = 116.24kWh.

[0087] Formulation and Effect Analysis of Energy Saving Strategies

[0088] Through in-depth analysis of the comprehensive energy consumption model, it is found that the energy consumption of the drying equipment has a greater impact on the total energy consumption, and within a certain range, reducing the temperature of the drying equipment can effectively reduce the energy consumption without affecting the painting quality. After multiple experiments and data analysis, it is determined to reduce the temperature of the drying equipment to 140°C, and at the same time dynamically adjust the operating hours of the equipment according to the production task. The specific adjustments are as follows:

[0089] Paint spraying robot: Keep the working duration \(t1 = 9h\) unchanged, optimize the spray gun flow control algorithm, and reduce the spray gun flow rate \(q1\) to \(5.5L / min\).

[0090] Drying equipment: Reduce the temperature \(T\) to \(140^{\circ}C\), and adjust the working duration \(t2\) to \(6.5h\).

[0091] Ventilation system: According to the reduction of the drying equipment temperature, reduce the fan speed \(n\) to \(1400r / min\), and adjust the working duration \(t3\) to \(10h\).

[0092] Recalculate the energy consumption of each device and the total energy consumption after adjustment:

[0093] Energy consumption of paint spraying robot: \(E′1 = 0.6×9×5.5 = 29.7kWh\);

[0094] Energy consumption of drying equipment: \(E′2 = 0.012×140×6.5 = 10.92kWh\);

[0095] Energy consumption of ventilation system: \(E′3 = 0.004×1400×10 = 56kWh\);

[0096] Total energy consumption after adjustment: \(E′\) 总 \(= E′1 + E′2 + E′3 = 29.7 + 10.92 + 56 = 96.62kWh\).

[0097] The energy consumption reduction ratio is:

[0098] By formulating and implementing the above energy-saving strategies, on the premise of ensuring the painting quality, the goal of reducing energy consumption by about 16.9% has been achieved. At the same time, through continuous monitoring and optimization, the energy-saving potential can be further explored and the energy utilization efficiency can be improved.

[0099] IV. Dynamic Allocation of Edge and Cloud Computing Resources

[0100] 1. Resource Monitoring Equipment

[0101] In an automotive manufacturing plant, in order to achieve the dynamic allocation of edge and cloud computing resources, it is necessary to accurately monitor the resources of the data center and edge computing nodes.

[0102] Data Center Monitoring

[0103] The Zabbix server performance monitoring tool is deployed in the data center. This tool has powerful monitoring capabilities and can obtain multiple key metrics of the server in real time. It monitors the CPU utilization rate, memory occupancy rate, network bandwidth, etc. of the server at a frequency of once per second. Its monitoring accuracy is extremely high. The monitoring error of the CPU utilization rate is controlled within ±0.1%, which means that even a slight change in the CPU utilization rate can be accurately captured; the monitoring accuracy of the memory occupancy rate reaches ±0.05%, which can accurately reflect the memory usage; the monitoring accuracy of the network bandwidth is ±1 Mbps, which can detect network bandwidth fluctuations in a timely manner.

[0104] The Zabbix tool collects various performance data of the system by installing an agent program on the server. These data are regularly sent to the Zabbix server for analysis and storage. The server side processes these data and generates detailed reports and visual charts so that administrators can understand the running status of the server in real time.

[0105] Edge computing node monitoring

[0106] The edge computing node adopts the Advantech UNO-2482G industrial gateway, and its built-in monitoring module has the ability to monitor the computing resources of the edge computing node in real time. This module collects data at an ultra-high frequency of once every 0.5 seconds to ensure that it can timely reflect the resource changes of the edge computing node. For the statistics of the number of CPU cores, the error is controlled within ±0.1 core, which makes the evaluation of CPU resources very accurate; the monitoring error of the memory capacity is within ±0.1 GB, which can accurately master the memory usage status.

[0107] The monitoring module obtains the usage information of the CPU and memory by interacting with the operating system of the industrial gateway. These information are encapsulated into data packets and sent to the device management platform through industrial Ethernet or wireless communication.

[0108] 2. Intelligent scheduling algorithm and resource allocation

[0109] Definition of state space and action space

[0110] To achieve dynamic allocation of computing resources, it is first necessary to clarify the state space and action space. Let the computing resources of the edge computing node be R 边缘 , represented by the number of CPU cores and the memory capacity (unit: GB); the computing resources of the cloud server are R 云端 , also measured by the number of CPU cores and the memory capacity (unit: GB); the device data generation amount is D (unit: MB / s), which reflects the amount of data generated by the device per unit time; the production task priority is P, divided into levels 1-5, and level 5 represents the highest priority.

[0111] The state space S is composed of these four factors, i.e., S = (R 边缘 , R 云端 , D, P). The action space A is a series of resource allocation strategies, specifically manifested as allocating x% of the computing tasks to edge computing nodes and (100 - x)% to cloud servers, where the value range of x is from 0 to 100.

[0112] Reinforcement Learning Algorithm Selection and Training

[0113] The Deep Q-Network (DQN) algorithm is adopted to achieve intelligent scheduling. The DQN algorithm combines the ideas of deep learning and Q-learning and can learn the optimal action strategy in a high-dimensional state space.

[0114] Randomly initialize the parameters of the DQN network, including the weights and biases of the neural network. At the same time, create an experience replay buffer to store the experience data of the interaction between the agent and the environment. Initially, the buffer is empty.

[0115] The agent selects an action a t from the action space A t . In the initial stage of training, to explore more action spaces, the ε-greedy strategy is adopted, that is, randomly select an action with a probability of ε and select the action with the largest Q value with a probability of 1 - ε. As the training progresses, the value of ε gradually decreases, making the agent more inclined to select the optimal action.

[0116] After executing the action a t , the system will transfer to the next state S t+1 and obtain the corresponding reward r t . The design of the reward function r comprehensively considers factors such as computing resource utilization, data processing latency, and task completion time. Its calculation formula is:

[0117] r = α×U - β×D l - γ×T c ;

[0118] where U is the computing resource utilization, which reflects the usage efficiency of computing resources in the system; D l is the data processing latency, that is, the time spent from data generation to processing completion; T c is the task completion time, which measures the time required for a production task from start to end. α = 0.6, β = 0.3, γ = 0.1 are weight coefficients, and these weight coefficients can be adjusted according to different production requirements and priorities.

[0119] Put the state S t , the action a t , and the reward r tand the next state S t+1 It is stored as an empirical data in the experience replay buffer.

[0120] Experience replay: Randomly sample a batch of empirical data from the experience replay buffer for training the DQN network. By minimizing the prediction error of the Q value, update the parameters of the DQN network. Specifically, calculate the mean square error between the target Q value and the current Q value, and then use the gradient descent algorithm to update the network parameters to make the predicted Q value closer to the target Q value.

[0121] Training process: A total of 10,000 episodes of training are carried out, and each episode simulates different production scenarios and task requirements. In each episode, the agent continuously interacts with the environment to learn the optimal resource allocation strategy. As the training progresses, the performance of the DQN network gradually improves, and the agent can better select the optimal action according to the current state.

[0122] Suppose at a certain moment, the CPU utilization rate of the edge computing node is 65%, the memory occupancy rate is 55%, and the remaining network bandwidth is 90 Mbps. The computing resource R of the edge computing node 边缘 is 4 CPU cores and 8 GB of memory; the computing resource R of the cloud server 云端 is 20 CPU cores and 64 GB of memory. The device data generation rate is 55 MB / s, and the production task priority is level 4.

[0123] At this time, the agent, according to the current state S=(R 边缘 , R 云端 , D, P), calculates the Q values of each action through the trained DQN network. After calculation, it is selected to allocate 75% of the computing tasks to the edge computing node and 25% to the cloud server.

[0124] Under this allocation strategy, the computing resource utilization rate is increased from the original 50% to 65%. This is because when there is a certain amount of remaining computing resources in the edge computing node, it undertakes most of the computing tasks and makes full use of its idle resources; at the same time, the cloud server also shares part of the tasks to ensure the balance of the overall computing power. The data processing delay is reduced from the original 100 ms to 60 ms, which is because the edge computing node is close to the data source and can process data faster, reducing the data transmission time.

[0125] Long-term effect evaluation

[0126] After a period of verification, under different production tasks and data generation volumes, the intelligent scheduling algorithm based on reinforcement learning has shown significant advantages. The average utilization rate of computing resources has increased by 35%, which means that the system can utilize computing resources more efficiently and reduce resource waste. The average data processing delay has been reduced by 45%, enabling the device management platform to respond more quickly to production demands and improve production efficiency. At the same time, through real-time monitoring and dynamic adjustment, the efficient utilization of computing resources and the timely processing of tasks are ensured, meeting the complex and changing production demands of the automobile manufacturing plant.

[0127] V. Data Drift Monitoring and Correction

[0128] Data Acquisition Equipment and Monitoring:

[0129] In the engine assembly workshop of the automobile manufacturing plant, precise monitoring of engine torque is crucial. An advanced strain gauge torque sensor, model HBM T40B, is selected. This sensor uses special alloy strain gauges with exquisite bonding technology to ensure a measurement accuracy of up to ±0.1% FS (Full Scale). Assuming the full-scale torque is 500 N·m, its measurement error can be strictly controlled within ±0.5 N·m. The sensor has excellent stability, with a built-in temperature compensation circuit that can effectively resist the influence of temperature fluctuations in the workshop on the measurement accuracy. Within the temperature range of -20°C to 80°C, the measurement accuracy drift does not exceed ±0.05% FS.

[0130] The sensor transmits data to the device management platform through industrial Ethernet. To ensure high-speed and stable data transmission, an industrial-grade network switch from Huawei, model NetEngine 8000 M14, is used. It has 40GE high-speed ports and supports port aggregation technology, which can bundle multiple physical ports into a logical port to achieve a transmission bandwidth of up to 400 Gbps. At the same time, it supports redundant power modules. When the main power supply fails, the backup power supply can switch seamlessly within milliseconds to ensure uninterrupted data transmission. During the data transmission process, the AES-256 encryption algorithm is used to encrypt the data to ensure the security and integrity of the data and prevent the data from being stolen or tampered with during transmission.

[0131] Data Drift Monitoring and Correction Mechanism:

[0132] Data Drift Monitoring: By setting a data drift threshold to determine whether data has drifted. Taking the engine torque data as an example, based on the statistical analysis of historical data and the requirements of the engine assembly process, ±3% of the torque measurement value is set as the data drift threshold. If the torque sensor measurement value is 400 N·m, the normal range is 388 N·m to 412 N·m. When the measurement value exceeds this range, it is initially judged that data drift may have occurred.

[0133] Implementation of the Kalman Filter Algorithm: The Kalman filter algorithm is used to process the collected data in real time. Let the sensor measurement value be z k , the predicted value of the system state be The state update value is The Kalman gain is K k . The prediction equation is:

[0134]

[0135] P k|k-1 = AP k-1|k-1 A T + Q;

[0136] The update equation is:

[0137] K k = P k|k-1 H T (HP k|k-1 H T + R) -1 ;

[0138]

[0139] P k|k = (I - K k H)P k|k-1 ;

[0140] Among them, A is the state transition matrix. In the engine torque monitoring scenario, assuming that the engine operating state is relatively stable, A is set as the identity matrix I; B is the control input matrix. Since there is no direct influence of external control input on torque measurement in this scenario, B is set as the zero matrix; H is the observation matrix, with a value of 1 because the torque value is directly observed; Q is the process noise covariance matrix. According to the actual noise level of engine operation and through statistical analysis of a large amount of experimental data, Q is set as 0.01; R is the observation noise covariance matrix. Considering the measurement accuracy and stability of the sensor, R is set as 0.005; I is the identity matrix.

[0141] Example of the correction effect: Let the current time be k and the previous time be k - 1. Let the sensor measurement value be z, the predicted value of the system state be The state update value is The Kalman gain is K.

[0142] The prediction equation is:

[0143]

[0144] P 预 = AP k-1 A T + Q;

[0145] The update equation is as follows:

[0146]

[0147] P 新 =(I - KH)P 预 ;

[0148] Among them, in the engine torque monitoring scenario, assuming that the engine operating state is relatively stable, A is set as the identity matrix I; since there is no direct influence of external control input on torque measurement in this scenario, B is set as the zero matrix; H takes the value of 1 because the torque value is directly observed; according to the actual noise level of engine operation, through a large number of experimental data statistics, Q is set as 0.01; considering the measurement accuracy and stability of the sensor, R is set as 0.005; I is the identity matrix.

[0149] For example, when the engine torque value measured by the sensor is z = 420 N·m, which exceeds the normal range. Assume that the state estimate value at the previous moment Covariance P k-1 = 0.05. After calculation, the predicted value (because A = I, B = 0), covariance P 预 = 0.05 + 0.01 = 0.06. The Kalman gain The updated state estimate value Covariance P 新 =(1 - 0.923)×0.06 = 0.00462. After processing, a torque estimate closer to the true value is obtained, ensuring quality control during the engine assembly process.

[0150] Multi-source data fusion and in-depth analysis:

[0151] To more accurately determine the cause of data drift, the engine torque data is fused and analyzed with other multi-source data. In the engine assembly workshop, the operating state data such as the engine speed, temperature, and oil pressure, as well as the environmental parameters such as the workshop temperature, humidity, and electromagnetic interference, are collected simultaneously. By establishing a correlation model between the data and using a Bayesian network to construct the causal relationship between the engine torque and other operating state parameters and environmental parameters.

[0152] For example, after training on 1000 sets of historical data, the Bayesian network model finds that when the engine speed suddenly increases and the electromagnetic interference in the workshop increases simultaneously, the probability of data drift in the engine torque data increases by 80%. Through multi-source data fusion and in-depth analysis, the root cause of data drift can be located more quickly and accurately, providing strong support for taking targeted corrective measures.

[0153] Verification and feedback mechanism after data correction:

[0154] The data corrected by the Kalman filtering algorithm enters a strict data verification process. By establishing a verification model, the corrected data is compared and verified with historical normal data and the operating data of other related devices. For example, the corrected torque data is compared with the historical torque data of engines in the same batch under the same assembly process, and cross-verification is carried out in combination with the data of other devices on the assembly line (such as the torque data of bolt tightening devices and the position data of engine positioning devices).

[0155] Verification metrics are set, such as Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). In the verification of 500 sets of corrected data, the RMSE is controlled within 1.5 N·m and the MAE is controlled within 1 N·m, indicating that the corrected data has high accuracy. If the verification passes, the data is applied to the quality control and decision-making in the engine assembly process; if the verification fails, the data drift monitoring and correction process is rechecked to analyze whether there are problems such as improper algorithm parameter settings and sensor failures, and the problems are fed back to the equipment management platform. Through this strict data verification and feedback mechanism, it is ensured that the accuracy rate of engine assembly decisions based on the corrected data is increased by more than 98%, effectively improving the engine assembly quality and production efficiency.

[0156] In summary: In the present invention:

[0157] Under the synergistic effect of the dynamic environment perception and adaptive module and the multi-device linkage synchronous control module: The dynamic environment perception and adaptive module real-time monitors the environmental parameters around the equipment, providing a decision-making basis for the multi-device linkage synchronous control module. When the change of environmental parameters affects the equipment performance, the former can timely adjust the equipment operation parameters to ensure the stable operation of the equipment in different environments, thereby ensuring the precise synchronization of multi-device linkage. For example, in a stamping workshop, the increase in temperature may affect the accuracy of stamping equipment. After this module adjusts the equipment operation parameters, it can avoid the problem of out-of-sync multi-device linkage caused by equipment performance fluctuations, improving production efficiency and product quality.

[0158] Under the synergistic effect of the multi-device linkage synchronous control module and the comprehensive energy efficiency evaluation and energy-saving strategy formulation module: The multi-device linkage synchronous control module ensures the high efficiency of equipment collaborative operation, reduces equipment waiting and idling time, and directly reduces energy consumption. The comprehensive energy efficiency evaluation and energy-saving strategy formulation module optimizes the equipment operation parameters and operation processes according to the equipment operation status and energy consumption data, further improving the energy utilization efficiency. For example, in an automobile general assembly workshop, by optimizing the equipment linkage sequence and time interval, unnecessary operation time of the equipment is reduced, and at the same time, the equipment power is adjusted according to the energy efficiency evaluation results to achieve the energy-saving goal.

[0159] Under the collaborative action of the comprehensive energy efficiency evaluation and energy-saving strategy formulation module and the edge and cloud computing resource dynamic allocation module: The comprehensive energy efficiency evaluation and energy-saving strategy formulation module generates a large amount of energy consumption data. The edge and cloud computing resource dynamic allocation module reasonably allocates computing resources according to the characteristics and processing requirements of these data. For example, when processing large-scale energy consumption data analysis tasks, part of the tasks are allocated to the cloud for centralized computing to improve computing efficiency; for the processing of energy consumption monitoring data with high real-time requirements, it is allocated to the edge computing nodes to reduce data transmission latency. This resource allocation method not only meets the computing requirements of energy efficiency evaluation and energy-saving strategy formulation but also improves the utilization rate of computing resources.

[0160] Under the collaborative action of the edge and cloud computing resource dynamic allocation module and the data drift monitoring and correction module: The edge and cloud computing resource dynamic allocation module provides computing resource support for the data drift monitoring and correction module. The data drift monitoring and correction module processes a large amount of sensor data and has high requirements for computing power. When the data volume is large, cloud computing resources can be used for complex data analysis and algorithm processing, such as using deep learning algorithms to analyze the reasons for data drift; the edge computing nodes are responsible for the preliminary processing and monitoring of real-time data to detect data drift signs in a timely manner. The two work together to ensure the accuracy and reliability of the data.

[0161] Under the collaborative action of the data drift monitoring and correction module and the dynamic environment perception and adaptation module: The data drift monitoring and correction module ensures the accuracy and reliability of the environmental parameters and device performance data obtained by the dynamic environment perception and adaptation module. If the sensor data drifts, it will affect the accuracy of the associated model, which will in turn lead to incorrect adjustment of device operation parameters. This module corrects data drift in a timely manner, ensuring the normal operation of the dynamic environment perception and adaptation module, enabling the device to make adaptive adjustments based on real environmental and performance data, and improving the stability and reliability of device operation.

Claims

1. The equipment management platform based on the Industrial Internet is characterized by: include: Dynamic environment perception and adaptive module: including deploying temperature sensors, humidity sensors, electromagnetic interference sensors, and environmental sensors on the device side to collect the environmental parameters around the device in real time; using data analysis algorithms to build a correlation model between environmental parameters and device performance, and adaptively adjust the device operating parameters based on the correlation model; Multi-device linkage synchronization control module: used to solve the synchronization problem in multi-device linkage scenarios, monitor the response speed and operation cycle of different devices, and ensure efficient connection when multiple devices work together through precise synchronization mechanism; Comprehensive energy efficiency evaluation and energy-saving strategy formulation module: Comprehensive energy consumption evaluation across equipment types, evaluates the overall energy consumption of multiple equipment operating in coordination under different production tasks, and then formulates energy-saving strategies; Dynamic allocation module of edge and cloud computing resources: Introduces intelligent scheduling algorithms to monitor the amount of device data generated, the resource usage of edge computing nodes and cloud servers in real time, including CPU utilization, memory occupancy, and network bandwidth, and dynamically adjusts the allocation ratio of edge computing and cloud computing resources according to the priority of production tasks and real-time computing needs; Data drift monitoring and correction module: used to monitor data drift caused by sensor aging and accumulation of subtle changes in environmental factors during long-term operation of the equipment.

2. The industrial Internet-based equipment management platform according to claim 1, characterized in that: In the dynamic environment perception and adaptation module, the data analysis algorithm includes at least one of a multivariate linear regression algorithm and a neural network algorithm, and an association model is established by learning historical environmental parameters and equipment performance data.

3. The industrial Internet-based equipment management platform according to claim 1, characterized in that: In the multi-device linkage synchronization control module, the precise synchronization mechanism adopts high-precision clock synchronization technology and a dynamic adjustment algorithm based on feedback control to adjust the equipment operation rhythm in real time.

4. The industrial Internet-based equipment management platform according to claim 1, characterized in that: In the comprehensive energy efficiency evaluation and energy-saving strategy formulation module, when evaluating the comprehensive energy consumption, multiple dimensional factors such as equipment load rate, operating time, and energy conversion efficiency are considered, and the system dynamics model is used to simulate and analyze energy consumption.

5. The industrial Internet-based equipment management platform according to claim 1, characterized in that: In the edge and cloud computing resource dynamic allocation module, the intelligent scheduling algorithm is an algorithm based on reinforcement learning, which learns the optimal resource allocation strategy by continuously interacting with the system environment.

6. The industrial Internet-based equipment management platform according to claim 1, characterized in that: In the data drift monitoring and correction module, by setting a data drift threshold, the Kalman filter algorithm is used to process the collected data in real time to effectively monitor and correct the data drift.

7. The industrial Internet-based equipment management platform according to claim 1, characterized in that: It also includes an equipment status prediction module, which uses machine learning algorithms to analyze the equipment's historical operating data, real-time environmental parameters, and maintenance records to predict the equipment's future operating status and provide early warning of potential failures.

8. The industrial Internet-based equipment management platform according to claim 1, characterized in that: It also includes a user interaction interface module that intuitively displays equipment operating status, environmental parameters, energy consumption data, and computing resource allocation information.

9. The industrial Internet-based equipment management platform according to claim 1, characterized in that: It also includes a data interaction module with external systems, including at least one of an enterprise resource planning system and a supply chain management system, to achieve interconnection between production data and enterprise management data.

10. The industrial Internet-based equipment management platform according to claim 1, characterized in that: It also includes a storage and management module, which uses blockchain technology to store and manage key data to ensure the security, non-tamperability and traceability of the data.

Images