Energy-saving control method and system based on industrial internet and AI edge computing

Abstract

The invention discloses an energy-saving control method and system based on the industrial internet and AI edge computing, and the method comprises the steps: collecting equipment operation data in real time through the deployment of a sensor, and transmitting the data to an edge computing node through the industrial internet; the edge nodes preprocess the data, remove noise and process missing values; analyzing the data by using a deep learning or reinforcement learning algorithm, predicting an energy consumption trend and identifying an energy-saving opportunity; according to a prediction result, executing an energy-saving decision, such as adjusting equipment parameters and optimizing production scheduling; the energy-saving effect is monitored through a feedback mechanism, and the control strategy is continuously optimized; according to the energy-saving control method and system based on the industrial internet and the AI edge calculation, more efficient and intelligent energy-saving control is realized by combining the real-time data acquisition capability of the industrial internet and the intelligent decision-making capability of the AI edge calculation.

Description

Technical Field

[0001] This invention belongs to the field of energy-saving control based on industrial internet and AI edge computing, specifically relating to an energy-saving control method and system based on industrial internet and AI edge computing. Background Technology

[0002] Energy-saving control methods based on the Industrial Internet and AI edge computing are intelligent control approaches that combine Industrial Internet and AI edge computing technologies to improve industrial production efficiency and reduce energy consumption. By connecting all equipment, machines, and systems in a factory or industrial facility through sensors, devices, and intelligent systems, a vast data network is formed. This data includes equipment operating status, energy consumption, environmental factors, etc. Instead of transmitting all data to the cloud for computation, data processing and intelligent decision-making are performed on edge devices. This reduces latency, improves real-time performance, and alleviates the demand for network bandwidth. AI algorithms can analyze real-time data at these edge nodes and make energy-saving optimization decisions. This method utilizes real-time data acquired from the Industrial Internet, analyzes it through AI edge computing, and automatically adjusts the operating status or production parameters of equipment to achieve energy efficiency optimization. This includes dynamically adjusting machine operating modes or operating times to avoid ineffective or excessive energy consumption; predicting the energy efficiency status of equipment for proactive maintenance or optimization; and achieving optimal energy use strategies based on environmental and production conditions. The energy-saving control system typically adapts to changes in the production process to ensure that energy use throughout the entire production process is always optimized.

[0003] However, while existing energy-saving control methods based on the Industrial Internet and AI edge computing have great potential in improving industrial energy efficiency and reducing energy consumption, they also have some shortcomings and challenges. Existing energy-saving control methods still have some problems, including data processing lag, high system complexity, and poor equipment compatibility. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the purpose of this invention is to provide an energy-saving control method and system based on the Industrial Internet and AI edge computing. By combining the real-time data acquisition capabilities of the Industrial Internet with the intelligent decision-making capabilities of AI edge computing, more efficient and intelligent energy-saving control can be achieved.

[0005] The technical solution adopted by this invention to solve its technical problem is:

[0006] The energy-saving control method based on the Industrial Internet and AI edge computing includes the following steps:

[0007] The system collects real-time operating status data of the equipment by deploying sensors on the production equipment, including but not limited to temperature, humidity, pressure, and energy consumption parameters, and transmits the data to edge computing nodes through the Industrial Internet.

[0008] Edge computing nodes preprocess the received raw data, removing noise and handling missing values;

[0009] Deep learning or reinforcement learning algorithms are used to analyze the preprocessed data to predict the energy consumption trend of the equipment and identify potential energy-saving opportunities;

[0010] Based on the prediction results, the edge computing nodes execute energy-saving decisions, which include adjusting equipment operating parameters and optimizing production scheduling operations.

[0011] The energy-saving effect is monitored through a feedback mechanism, and the data is returned to the edge computing node to continuously optimize the control strategy.

[0012] As a preferred method, the operation status data of the production equipment is collected in real time by sensors deployed on the equipment, including but not limited to temperature, humidity, pressure, and energy consumption parameters, and the data is transmitted to the edge computing node via the Industrial Internet.

[0013] Temperature sensors, humidity sensors, pressure sensors, power meters, and flow meters are installed on production equipment to monitor the equipment's operating status in real time. The sensors are connected to edge computing nodes via wired or wireless means to periodically collect the equipment's operating data and convert the data into digital signals.

[0014] During the data acquisition phase, the sensor or edge computing node performs data preprocessing operations, which include:

[0015] Remove noise from sensor readings;

[0016] Imput or label missing values;

[0017] Filter or mark data indicating abnormal equipment status;

[0018] After data acquisition, the sensor will transmit the data to the edge computing node through a standardized industrial communication protocol.

[0019] As a preferred approach, the edge computing node preprocesses the received raw data by removing noise and handling missing values ​​as follows:

[0020] Data noise originates from sensor malfunctions, environmental interference, or data transmission errors. To ensure data reliability and accuracy, noise removal methods include:

[0021] Reduce short-term fluctuations in data by using weighted averages or sliding windows;

[0022] A filter is used to remove high-frequency noise while retaining low-frequency signals;

[0023] Outliers were identified and removed using Z-Score and IQR statistical methods.

[0024] Multi-scale decomposition is used to remove noise while preserving the main features of the signal;

[0025] Methods for handling missing values:

[0026] Directly delete samples containing missing values. If most of the data for a certain feature is missing, delete that feature.

[0027] Methods for imputing data by inferring the range of missing values ​​include:

[0028] Fill in the missing values ​​using the mean or median of the column;

[0029] Calculate missing values ​​based on the linear relationship between two adjacent known data points;

[0030] The accuracy of the filling can be further improved by combining KNN and regression analysis interpolation techniques.

[0031] As a preferred method, the preprocessed data is analyzed using deep learning or reinforcement learning algorithms to predict the energy consumption trend of the equipment and identify potential energy-saving opportunities.

[0032] A deep neural network is used for regression tasks to predict the energy consumption trend of equipment. The input data includes the historical energy consumption of the equipment, the operating environment and the equipment status characteristics. The model is trained to identify the patterns and trends of energy consumption and to predict future energy demand.

[0033] The Q-learning algorithm is adopted. Q-learning is a value function-based reinforcement learning method used in discrete state space. Q-learning learns the state-action value function to gradually optimize the control strategy of the device and dynamically adjust the working state of the device.

[0034] By using deep learning or reinforcement learning models, abnormal patterns in equipment operation can be identified. These abnormal patterns include sudden spikes in energy consumption or unexpected equipment behavior.

[0035] Clustering algorithms are used to perform cluster analysis on the operating status and energy consumption patterns of equipment, identify low-energy-consumption operating modes, and optimize the equipment's operating strategy based on different operating modes to achieve energy-saving goals.

[0036] As a preferred approach, based on the prediction results, the edge computing nodes perform energy-saving decisions. The methods for making energy-saving decisions, which involve adjusting equipment operating parameters and optimizing production scheduling operations, are as follows:

[0037] Based on predicted energy consumption trends and equipment operating status, edge computing nodes adjust equipment operating parameters in real time. By adjusting equipment power, operating speed, start-stop frequency, or temperature parameters, they ensure that the equipment operates in the most energy-efficient state.

[0038] The specific steps are as follows:

[0039] Adjust the load or operating intensity of equipment based on energy consumption forecasts;

[0040] By controlling the power consumption of the equipment, the motor speed and fan speed are automatically adjusted.

[0041] Schedule equipment start-up and shutdown based on predicted demand fluctuations;

[0042] Control the temperature of heaters or cooling equipment to ensure they only operate when demand is high;

[0043] Based on predicted energy consumption trends, edge computing nodes optimize production scheduling operations, which include task allocation for equipment, production sequence, and process arrangement.

[0044] The specific steps are as follows:

[0045] Based on energy consumption forecasts and equipment operating status, the priority of production tasks is dynamically adjusted, and high-energy-consuming tasks are assigned when equipment load is low.

[0046] In a multi-device environment, production tasks are allocated based on equipment energy efficiency and load forecasting to ensure balanced equipment load.

[0047] Adjust the pace or rate of production based on energy consumption forecasts;

[0048] Adjust production processes based on forecast results, and optimize the execution time and resource allocation for each process.

[0049] As a preferred approach, the method of monitoring energy-saving effects through a feedback mechanism and returning the data to the edge computing node to continuously optimize the control strategy is as follows:

[0050] During equipment operation, edge computing nodes continuously monitor real-time data of the equipment and return feedback information to the control system. Based on the feedback results, the system dynamically adjusts the equipment operating parameters or scheduling strategies.

[0051] The specific steps are as follows:

[0052] Real-time data collection of equipment operation is achieved through sensors and monitoring systems;

[0053] Monitoring data is transmitted back to the decision-making system through edge computing nodes, and feedback signals are generated to indicate the gap between the current energy efficiency status of the equipment and the target.

[0054] Based on feedback signals, adjust equipment operating parameters or optimize production scheduling strategies to ensure that energy-saving effects meet expectations.

[0055] Because there is a certain delay in the feedback signal, the control strategy is adjusted in advance based on historical data and prediction results;

[0056] Reinforcement learning algorithms optimize strategies by continuously acquiring environmental feedback. In energy-saving control, feedback information on the operating status of equipment is fed back to the learning model as a reward signal, and the model continuously adjusts its control strategy based on these reward signals.

[0057] The specific steps are as follows:

[0058] After each control action, the change in the device's operating state is fed back as a reward to Q-learning or other reinforcement learning algorithms to evaluate the effectiveness of the current strategy;

[0059] Reinforcement learning algorithms continuously adjust their policy functions based on feedback signals;

[0060] The model adapts to environmental changes over time and gradually improves the control strategy based on historical data and feedback signals;

[0061] Adaptive control strategies automatically adjust control parameters or strategies by analyzing feedback data in real time, and adjust the parameters of the model according to real-time feedback to maintain energy-saving optimization under different operating conditions.

[0062] The specific steps are as follows:

[0063] Based on feedback data, the parameters of the equipment control model are dynamically updated;

[0064] When the external environment changes, the adaptive control strategy adjusts the control decisions according to the new environmental data;

[0065] Edge computing nodes adjust control algorithms based on real-time monitoring data to dynamically optimize the working status of the devices;

[0066] Edge computing nodes perform local decision-making and control operations, and through integration with cloud computing platforms, they use cloud computing capabilities to analyze and store data. The feedback mechanism enables data communication between edge nodes and cloud platforms, and continuously optimizes control strategies.

[0067] The specific steps are as follows:

[0068] Edge computing nodes collect device operation data in real time through sensors and perform preliminary analysis and decision-making based on local algorithms;

[0069] Edge computing nodes upload feedback data to the cloud platform, where more sophisticated machine learning models are used to further analyze energy-saving effects and generate optimization suggestions.

[0070] The cloud platform will return optimization suggestions to the edge nodes, and the edge nodes will then adjust the device's operating parameters or scheduling schemes according to the new strategy.

[0071] Energy-saving control systems based on the Industrial Internet and AI edge computing include:

[0072] Sensor network modules are used to deploy sensors on various production equipment to collect equipment operation data in real time.

[0073] Edge computing node modules are used to process collected data and execute energy-saving decisions. Edge computing nodes include at least one AI model to analyze and predict energy consumption trends.

[0074] The central control platform module is used to communicate with edge computing nodes through the industrial internet, collect data from the entire plant, and generate data to provide global monitoring and scheduling functions.

[0075] The feedback system module is used to track energy-saving effects in real time and transmit feedback information to edge computing nodes for decision optimization.

[0076] Another technical problem to be solved by the present invention is to provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the energy-saving control method and system based on industrial internet and AI edge computing as described above.

[0077] Another technical problem to be solved by the present invention is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements an energy-saving control method and system based on the Industrial Internet and AI edge computing.

[0078] The beneficial effects of this invention are:

[0079] By deploying sensors on production equipment and collecting data in real time via the Industrial Internet, the operating status of the equipment can be monitored in real time. Edge computing nodes preprocess the collected raw data, removing noise and handling missing values ​​to ensure high data quality. Through deep learning or reinforcement learning algorithms, the energy consumption trends of the equipment can be accurately predicted, and potential energy-saving opportunities can be identified. Based on the prediction results, edge computing nodes can adjust the operating parameters of the equipment and optimize production scheduling. Through a feedback mechanism, the energy-saving effect can be monitored in real time, and the results data can be fed back to the edge computing nodes to further optimize control strategies. Through automated prediction, decision-making, and feedback mechanisms, the need for manual intervention is reduced. By optimizing production scheduling and adjusting equipment operating parameters, not only is energy consumption reduced, but the operating efficiency and stability of the equipment are also improved. The combination of the Industrial Internet and AI edge computing has high scalability and can be flexibly applied to production environments of different scales. This solution provides decision support through a large amount of real-time data and advanced analytics, enabling enterprises to make more reasonable production and scheduling decisions based on accurate energy consumption prediction results. Attached Figure Description

[0080] Figure 1 This is a schematic diagram of the energy-saving control system based on industrial internet and AI edge computing of the present invention. Detailed Implementation

[0081] The principles and features of the present invention are described below. The examples given are for illustrative purposes only and are not intended to limit the scope of the invention. The invention is described more specifically by way of example in the following paragraphs. The advantages and features of the invention will become clearer from the following description and claims.

[0082] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0083] Example

[0084] The technical solution adopted by this invention to solve its technical problem is:

[0085] The energy-saving control method based on the Industrial Internet and AI edge computing includes the following steps:

[0086] The system collects real-time operating status data of the equipment by deploying sensors on the production equipment, including but not limited to temperature, humidity, pressure, and energy consumption parameters, and transmits the data to edge computing nodes through the Industrial Internet.

[0087] Edge computing nodes preprocess the received raw data, removing noise and handling missing values;

[0088] Deep learning or reinforcement learning algorithms are used to analyze the preprocessed data to predict the energy consumption trend of the equipment and identify potential energy-saving opportunities;

[0089] Based on the prediction results, the edge computing nodes execute energy-saving decisions, which include adjusting equipment operating parameters and optimizing production scheduling operations.

[0090] The energy-saving effect is monitored through a feedback mechanism, and the data is returned to the edge computing node to continuously optimize the control strategy.

[0091] By deploying sensors on production equipment, real-time data on equipment operation status is collected, enabling real-time monitoring of equipment status. Edge computing nodes preprocess the received data, removing noise and filling in missing values ​​to make the data more accurate and reliable. Deep learning or reinforcement learning algorithms analyze the preprocessed data to predict energy consumption trends based on historical data and current status, and identify potential energy-saving opportunities. Based on the prediction results, edge computing nodes automatically execute energy-saving decisions, adjusting equipment operating parameters and optimizing production scheduling. The decision-making process requires no manual intervention, and the system can adapt to different situations. The system monitors energy-saving effects through a feedback mechanism and returns data to the edge computing nodes, thereby continuously optimizing the control strategy. This reduces manual intervention and complex operations, with energy-saving control strategies based on automated and intelligent execution. By optimizing equipment operating parameters, excessive consumption or overload is avoided, improving equipment operating efficiency and reducing equipment failure rates.

[0092] The method of collecting real-time operational status data of production equipment through sensors deployed on the equipment, including but not limited to temperature, humidity, pressure, and energy consumption parameters, and transmitting the data to edge computing nodes via the Industrial Internet is as follows:

[0093] Temperature sensors, humidity sensors, pressure sensors, power meters, and flow meters are installed on production equipment to monitor the equipment's operating status in real time. The sensors are connected to edge computing nodes via wired or wireless means to periodically collect the equipment's operating data and convert the data into digital signals.

[0094] During the data acquisition phase, the sensor or edge computing node performs data preprocessing operations, which include:

[0095] Remove noise from sensor readings;

[0096] Imput or label missing values;

[0097] Filter or mark data indicating abnormal equipment status;

[0098] After data acquisition, the sensor will transmit the data to the edge computing node through a standardized industrial communication protocol.

[0099] By installing sensors for temperature, humidity, pressure, power, and flow on production equipment, the operating status of the equipment can be monitored in real time. Data preprocessing operations, including noise reduction, missing value imputation, and anomaly filtering, can significantly improve the accuracy and effectiveness of the data. Using standardized industrial communication protocols to transmit data to edge computing nodes ensures compatibility between devices and stable data transmission. After the data collected by sensors is transmitted to the edge computing nodes, it can be dynamically adjusted based on real-time data. By performing data preprocessing and local computation at the edge computing nodes, only important processing results or summary data are transmitted to the cloud or data center. Processing data locally at the edge computing nodes avoids the long-distance transmission of large amounts of sensitive data, reducing the risk of data leakage. By accumulating and analyzing long-term operating data of the equipment, potential energy-saving opportunities, production optimization points, and early signs of equipment failure can be identified, thereby helping enterprises formulate long-term optimization strategies.

[0100] The edge computing nodes preprocess the received raw data, removing noise and handling missing values ​​using the following methods:

[0101] Data noise originates from sensor malfunctions, environmental interference, or data transmission errors. To ensure data reliability and accuracy, noise removal methods include:

[0102] Reduce short-term fluctuations in data by using weighted averages or sliding windows;

[0103] A filter is used to remove high-frequency noise while retaining low-frequency signals;

[0104] Outliers were identified and removed using Z-Score and IQR statistical methods.

[0105] Multi-scale decomposition is used to remove noise while preserving the main features of the signal;

[0106] Methods for handling missing values:

[0107] Directly delete samples containing missing values. If most of the data for a certain feature is missing, delete that feature.

[0108] Methods for imputing data by inferring the range of missing values ​​include:

[0109] Fill in the missing values ​​using the mean or median of the column;

[0110] Calculate missing values ​​based on the linear relationship between two adjacent known data points;

[0111] The accuracy of the filling can be further improved by combining KNN and regression analysis interpolation techniques.

[0112] Noise removal and missing value imputation significantly improve the quality of raw data, ensuring its reliability and validity. Statistical methods such as Z-Score and IQR effectively remove outliers, preventing these outliers from distorting the analysis results. Multiple methods for handling missing values ​​are provided, including deleting data containing missing values, mean / median imputation, and interpolation techniques. Noise removal methods can remove short-term fluctuations and high-frequency noise while preserving the main characteristics of the signal. The solution combines multiple noise removal and missing value handling methods, allowing for flexible selection based on different data characteristics, ensuring both the accuracy of data preprocessing and the flexibility and adaptability of the system.

[0113] The method of using deep learning or reinforcement learning algorithms to analyze preprocessed data, predict equipment energy consumption trends, and identify potential energy-saving opportunities is as follows:

[0114] A deep neural network is used for regression tasks to predict the energy consumption trend of equipment. The input data includes the historical energy consumption of the equipment, the operating environment and the equipment status characteristics. The model is trained to identify the patterns and trends of energy consumption and to predict future energy demand.

[0115] The Q-learning algorithm is adopted. Q-learning is a value function-based reinforcement learning method used in discrete state space. Q-learning learns the state-action value function to gradually optimize the control strategy of the device and dynamically adjust the working state of the device.

[0116] By using deep learning or reinforcement learning models, abnormal patterns in equipment operation can be identified. These abnormal patterns include sudden spikes in energy consumption or unexpected equipment behavior.

[0117] Clustering algorithms are used to perform cluster analysis on the operating status and energy consumption patterns of equipment, identify low-energy-consumption operating modes, and optimize the equipment's operating strategy based on different operating modes to achieve energy-saving goals.

[0118] By using deep neural networks for regression tasks, energy consumption trends of equipment can be accurately identified, and potential patterns can be mined from historical data. Q-learning, as a type of reinforcement learning, can automatically optimize control strategies based on the current state of the equipment. Through learning the state-action value function, the system can continuously improve the control method of the equipment in actual operation and reduce energy consumption. Deep learning and reinforcement learning models can not only predict the normal operating state of the equipment, but also help to discover abnormal patterns in the operation of the equipment. By clustering algorithms to perform cluster analysis on the operating state and energy consumption patterns of the equipment, efficient and low-energy-consumption operation modes can be identified. The solution combines a variety of advanced data-driven technologies such as deep learning, reinforcement learning, and clustering algorithms to provide a comprehensive and systematic energy-saving optimization strategy.

[0119] Based on the prediction results, the edge computing nodes execute energy-saving decisions, which involve adjusting equipment operating parameters and optimizing production scheduling operations.

[0120] Based on predicted energy consumption trends and equipment operating status, edge computing nodes adjust equipment operating parameters in real time. By adjusting equipment power, operating speed, start-stop frequency, or temperature parameters, they ensure that the equipment operates in the most energy-efficient state.

[0121] The specific steps are as follows:

[0122] Adjust the load or operating intensity of equipment based on energy consumption forecasts;

[0123] By controlling the power consumption of the equipment, the motor speed and fan speed are automatically adjusted.

[0124] Schedule equipment start-up and shutdown based on predicted demand fluctuations;

[0125] Control the temperature of heaters or cooling equipment to ensure they only operate when demand is high;

[0126] Based on predicted energy consumption trends, edge computing nodes optimize production scheduling operations, which include task allocation for equipment, production sequence, and process arrangement.

[0127] The specific steps are as follows:

[0128] Based on energy consumption forecasts and equipment operating status, the priority of production tasks is dynamically adjusted, and high-energy-consuming tasks are assigned when equipment load is low.

[0129] In a multi-device environment, production tasks are allocated based on equipment energy efficiency and load forecasting to ensure balanced equipment load.

[0130] Adjust the pace or rate of production based on energy consumption forecasts;

[0131] Adjust production processes based on forecast results, and optimize the execution time and resource allocation for each process.

[0132] Edge computing nodes can make rapid decisions based on real-time energy consumption predictions and equipment operating status, and dynamically adjust equipment operating parameters. By predicting energy consumption trends and equipment status, edge computing nodes can optimize production scheduling, rationally arrange equipment load, avoid equipment overwork, and ensure automatic adjustment of equipment start-up and shutdown during demand fluctuations to avoid energy waste. By optimizing production rhythm and processes based on energy consumption predictions, production operations during inefficient periods can be reduced, and the execution time of each process can be adjusted to avoid wasting resources and time. By adjusting equipment operating parameters in real time and optimizing production scheduling, equipment is prevented from running under high load for extended periods, reducing the risk of equipment failure and damage. Edge computing enables energy-saving decisions to respond quickly to real-time changes and has high adaptability. When production needs or the environment change, the system can adjust immediately, making the production process more flexible and adaptable.

[0133] The method of continuously optimizing the control strategy by monitoring energy-saving effects through a feedback mechanism and returning the data to the edge computing node is as follows:

[0134] During equipment operation, edge computing nodes continuously monitor real-time data of the equipment and return feedback information to the control system. Based on the feedback results, the system dynamically adjusts the equipment operating parameters or scheduling strategies.

[0135] The specific steps are as follows:

[0136] Real-time data collection of equipment operation is achieved through sensors and monitoring systems;

[0137] Monitoring data is transmitted back to the decision-making system through edge computing nodes, and feedback signals are generated to indicate the gap between the current energy efficiency status of the equipment and the target.

[0138] Based on feedback signals, adjust equipment operating parameters or optimize production scheduling strategies to ensure that energy-saving effects meet expectations.

[0139] Because there is a certain delay in the feedback signal, the control strategy is adjusted in advance based on historical data and prediction results;

[0140] Reinforcement learning algorithms optimize strategies by continuously acquiring environmental feedback. In energy-saving control, feedback information on the operating status of equipment is fed back to the learning model as a reward signal, and the model continuously adjusts its control strategy based on these reward signals.

[0141] The specific steps are as follows:

[0142] After each control action, the change in the device's operating state is fed back as a reward to Q-learning or other reinforcement learning algorithms to evaluate the effectiveness of the current strategy;

[0143] Reinforcement learning algorithms continuously adjust their policy functions based on feedback signals;

[0144] The model adapts to environmental changes over time and gradually improves the control strategy based on historical data and feedback signals;

[0145] Adaptive control strategies automatically adjust control parameters or strategies by analyzing feedback data in real time, and adjust the parameters of the model according to real-time feedback to maintain energy-saving optimization under different operating conditions.

[0146] The specific steps are as follows:

[0147] Based on feedback data, the parameters of the equipment control model are dynamically updated;

[0148] When the external environment changes, the adaptive control strategy adjusts the control decisions according to the new environmental data;

[0149] Edge computing nodes adjust control algorithms based on real-time monitoring data to dynamically optimize the working status of the devices;

[0150] Edge computing nodes perform local decision-making and control operations, and through integration with cloud computing platforms, they use cloud computing capabilities to analyze and store data. The feedback mechanism enables data communication between edge nodes and cloud platforms, and continuously optimizes control strategies.

[0151] The specific steps are as follows:

[0152] Edge computing nodes collect device operation data in real time through sensors and perform preliminary analysis and decision-making based on local algorithms;

[0153] Edge computing nodes upload feedback data to the cloud platform, where more sophisticated machine learning models are used to further analyze energy-saving effects and generate optimization suggestions.

[0154] The cloud platform will return optimization suggestions to the edge nodes, and the edge nodes will then adjust the device's operating parameters or scheduling schemes according to the new strategy.

[0155] This solution utilizes feedback mechanisms and reinforcement learning algorithms to ensure continuous optimization of energy-saving effects. Edge computing nodes dynamically adjust equipment operating parameters or scheduling strategies based on real-time feedback data, promptly correcting deviations and ensuring equipment always operates in its most energy-efficient state. Adaptive control strategies and reinforcement learning algorithms can adjust control strategies in a timely manner according to changes in the external environment. Through real-time monitoring and feedback data transmission, the system can automatically adjust equipment operating parameters or production scheduling strategies, eliminating the delay of manual intervention and improving response speed. Reinforcement learning algorithms optimize control strategies based on feedback signals, continuously adjusting and improving strategy functions to enhance energy-saving effects. The combination of edge computing nodes and cloud computing platforms enables the system to quickly respond to changes in equipment operation locally while leveraging the powerful computing capabilities of the cloud for in-depth analysis and optimization. By dynamically adjusting the operating status of equipment, prolonged high-load operation or frequent start / stop operations are avoided, thereby reducing equipment wear and extending equipment lifespan. The system's feedback mechanism clearly reflects the execution effect of energy-saving strategies, allowing managers to view energy-saving data in real time, evaluate the effectiveness of control strategies, and make timely adjustments.

[0156] Energy-saving control systems based on the Industrial Internet and AI edge computing include:

[0157] Sensor network modules are used to deploy sensors on various production equipment to collect equipment operation data in real time.

[0158] Edge computing node modules are used to process collected data and execute energy-saving decisions. Edge computing nodes include at least one AI model to analyze and predict energy consumption trends.

[0159] The central control platform module is used to communicate with edge computing nodes through the industrial internet, collect data from the entire plant, and generate data to provide global monitoring and scheduling functions.

[0160] The feedback system module is used to track energy-saving effects in real time and transmit feedback information to edge computing nodes for decision optimization.

[0161] The sensor network module collects real-time operating data from production equipment, enabling timely understanding of the equipment's working status. Edge computing nodes process the collected data and execute energy-saving decisions in real time, avoiding data transmission delays to the cloud. AI models on edge computing nodes can analyze historical data to predict future energy consumption trends and provide optimization decisions. The central control platform module communicates with edge computing nodes through the Industrial Internet, collecting data from the entire plant and performing global monitoring and scheduling. The feedback system tracks energy-saving effects in real time and returns feedback information to the edge computing node module for further decision optimization. This system continuously optimizes energy-saving control strategies to avoid unnecessary energy waste and reduce energy consumption during equipment overload operation and idle states, thereby achieving energy-saving goals. The combination of AI models and edge computing nodes gives the system strong intelligent capabilities, enabling it to automatically adjust control strategies based on real-time data.

[0162] This embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the energy-saving control method and system based on the Industrial Internet and AI edge computing as described above.

[0163] This embodiment also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the energy-saving control method and system based on the Industrial Internet and AI edge computing as described above.

[0164] Kalman filtering can effectively reduce noise and uncertainty in sensor data, especially in dynamic systems. For missing values ​​in equipment operation data, mean imputation is used to fill the gaps when data is missing within a certain time period. Linear interpolation is used to estimate the missing values ​​based on the trend of changes in preceding and following data points. Standardization is applied to various types of equipment operation data to eliminate dimensional differences between different data dimensions. Regularization is used to appropriately scale the equipment operation data to prevent certain features from being too large and causing bias in subsequent analysis.

[0165] Historical operating data of the equipment is collected in real time through sensors, including but not limited to temperature, humidity, pressure, load, and energy consumption parameters. This data can be stored in a central database or cloud data storage. Based on time series data, a deep learning model suitable for predicting the energy consumption trend of the equipment is selected, and an appropriate loss function is selected to evaluate the prediction error, which is mean squared error or mean absolute error.

[0166] Reinforcement learning is a machine learning method that learns optimal decision-making strategies through interaction with the environment. In production scheduling optimization, the goal of reinforcement learning is to learn a scheduling strategy that balances energy consumption and production efficiency by having an agent continuously interact with the production environment.

[0167] In multiple equipment and processes, the production scheduling problem can be abstracted into a Markov decision process, which includes state space, action space, and reward function elements. In multiple equipment and processes, reinforcement learning agents need to coordinate the scheduling between equipment to ensure that while optimizing production efficiency, they adopt methods such as equipment collaborative scheduling and process optimization.

[0168] The closed-loop control system improves production scheduling efficiency and energy saving through real-time data acquisition and transmission, real-time feedback and target setting, production scheduling adjustment, control strategy optimization, energy saving effect evaluation and improvement steps. The combination of real-time data stream transmission and the closed-loop control system is the core of improving production scheduling efficiency and energy saving effect. The specific process is as follows:

[0169] Through industrial sensors, smart devices, and cloud platforms, information on equipment operation, energy consumption, and production progress is collected and transmitted in real time.

[0170] The control system analyzes the current energy consumption and production efficiency based on real-time data and feeds the information back to the production scheduling system. This feedback includes not only suggestions for adjusting production efficiency, but also revisions to energy-saving targets.

[0171] The system adjusts equipment scheduling, production processes, and process parameters in real time based on feedback results;

[0172] By connecting different devices in the factory through the Industrial Internet, real-time operating data of the devices can be collected. This data may include temperature, pressure, vibration, current, voltage, and power consumption information. Through the Industrial Internet, the data of all devices is transmitted to the central platform in real time.

[0173] The central platform aggregates data from various devices and sensors, and generates the overall operating status of the plant through analysis. This data includes power consumption, production progress, load, and failure rate information of each production line, machine, and equipment. Based on this information, the system has a comprehensive understanding of the energy usage of the entire plant.

[0174] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0175] Those skilled in the art will understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0176] The above embodiments of the present invention are not intended to limit the scope of protection of the present invention. The implementation of the present invention is not limited thereto. All other modifications, substitutions or alterations made to the above structure of the present invention based on the above content of the present invention, in accordance with ordinary technical knowledge and common practice in the field, without departing from the basic technical idea of ​​the present invention, shall fall within the scope of protection of the present invention.

Claims

1. An energy-saving control method based on industrial internet and AI edge computing, characterized in that, Comprising the following steps: Real-time collection of equipment operating state data through sensors deployed on production equipment, including but not limited to temperature, humidity, pressure, energy consumption parameters, transmission of data to edge computing nodes through industrial internet; The edge computing node pre-processes the received raw data, eliminating noise and handling missing values; Using deep learning or reinforcement learning algorithms to analyze the pre-processed data, predict the energy consumption trend of the equipment, and identify potential energy-saving opportunities; Based on the prediction results, the edge computing node executes energy-saving decisions, which are adjustments to equipment operating parameters and optimization of production scheduling operations; Through the feedback mechanism, monitor the energy-saving effect, and return the data to the edge computing node to continuously optimize the control strategy. 2.The energy-saving control method based on industrial internet and AI edge computing according to claim 1, wherein, The method for collecting real-time operating state data of the equipment through sensors deployed on the production equipment, including but not limited to temperature, humidity, pressure, energy consumption parameters, and transmitting the data to the edge computing node through the industrial internet is: Install temperature sensors, humidity sensors, pressure sensors, power meters, and flow meter sensors on the production equipment to monitor the operating state of the equipment in real time. The sensors are connected to the edge computing node through wired or wireless methods, periodically collect the operating data of the equipment, and convert the data into digital signals; During the data collection stage, the sensor end or the edge computing node performs data preprocessing operations, which include: Remove noise from sensor readings; Interpolate or label missing values; Filter or mark data with abnormal equipment states; After data collection, the sensor transmits the data to the edge computing node through standardized industrial communication protocols. 3.The energy-saving control method based on industrial internet and AI edge computing according to claim 2, characterized in that, The method for the edge computing node to pre-process the received raw data, eliminate noise, and handle missing values is: Data noise comes from sensor failure, environmental interference, or data transmission errors. To ensure data reliability and accuracy, noise elimination methods include: Reduce short-term fluctuations in data through weighted averaging or sliding window; Use a filter to remove high-frequency noise and retain low-frequency signals; Identify and eliminate outliers using Z-Score and IQR statistical methods; Use multi-scale decomposition to remove noise and retain the main features of the signal; Methods for handling missing values: Directly delete samples containing missing values. If most of the data for a particular feature is missing, delete that feature; Methods for filling in missing values include: Fill in missing values with the mean or median of the column; Estimate missing values based on the linear relationship between the two adjacent known data points; Further improve the filling accuracy using KNN and regression analysis interpolation techniques. 4.The energy-saving control method based on industrial internet and AI edge computing according to claim 3, characterized in that, The method for using deep learning or reinforcement learning algorithms to analyze the pre-processed data, predict the energy consumption trend of the equipment, and identify potential energy-saving opportunities is: Use deep neural networks for regression tasks to predict the energy consumption trend of the equipment. The input data includes the historical energy consumption, operating environment, and equipment state features of the equipment. Through training the model to identify the pattern and trend of energy consumption, predict future energy demand. Q-learning algorithm is used, which is a value function-based reinforcement learning method for discrete state space. Q-learning optimizes the control strategy of the device step by step by learning the state-action value function, and dynamically adjusts the working state of the device. Through deep learning or reinforcement learning model, abnormal patterns in device operation are identified, which are sudden energy consumption surges or device behaviors inconsistent with expectations. Clustering algorithm is used to cluster the operating state and energy consumption pattern of the device, identify the low-energy operation mode, and optimize the working strategy of the device according to different operation modes to achieve energy-saving goals. 5.The energy-saving control method based on industrial internet and AI edge computing according to claim 4, characterized in that, Based on the prediction results, the edge computing node executes energy-saving decisions, which are methods for adjusting device operating parameters and optimizing production scheduling operations: Based on the predicted energy consumption trend and the operating state of the device, the edge computing node adjusts the operating parameters of the device in real time, such as adjusting the power, speed, start-stop frequency, or temperature parameters of the device to ensure that the device operates in the most energy-efficient state. Specific operations are as follows: Adjust the load or operating intensity of the device according to the energy consumption prediction; Automatically adjust the motor speed and fan speed by controlling the power consumption of the device; Arrange the start and stop of the device according to the predicted demand fluctuations; Control the temperature of the heater or cooling device to ensure it only works when demand is high; Based on the predicted energy consumption trend, the edge computing node optimizes production scheduling operations, including task allocation, production sequence, and process arrangement of the device. Specific operations are as follows: According to the energy consumption prediction and the operating state of the device, dynamically adjust the priority of production tasks, and arrange high-energy-consuming tasks when the device load is low; In a multi-device environment, allocate production tasks according to the energy efficiency and load prediction of the device to ensure load balancing of the device; Adjust the rhythm or speed of production according to the energy consumption prediction; Adjust the production process according to the prediction results to optimize the execution time and resource allocation of each process. 6.The energy-saving control method based on industrial internet and AI edge computing according to claim 5, characterized in that, Through the feedback mechanism, monitor the energy-saving effect, and return the data to the edge computing node to continuously optimize the control strategy. The method is as follows: During the operation of the device, the edge computing node returns feedback information to the control system by continuously monitoring the real-time data of the device, and dynamically adjusts the operating parameters or scheduling strategy of the device according to the feedback results. Specific operations are as follows: Collect real-time data of the device through sensors and monitoring systems; The monitoring data is returned to the decision system through the edge computing node, and a feedback signal is generated to indicate the gap between the current energy efficiency of the device and the target; According to the feedback signal, adjust the operating parameters of the device or optimize the production scheduling strategy to ensure that the energy-saving effect meets the expectations; Based on the feedback signal, there is a certain delay, and the control strategy is based on historical data and prediction results to make adjustments in advance; The reinforcement learning algorithm optimizes the strategy by continuously obtaining environmental feedback. In energy-saving control, the feedback information of the device operating state is fed back to the learning model as a reward signal, and the model continuously adjusts its control strategy based on these reward signals. Specific operations are as follows: After each control action, the change in the operating state of the device is fed back as a reward to the Q-learning or other reinforcement learning algorithm to evaluate the effectiveness of the current strategy. The reinforcement learning algorithm continuously adjusts the policy function based on feedback signals; The model adapts to environmental changes over time and gradually improves the control strategy based on historical data and feedback signals; The adaptive control strategy automatically adjusts control parameters or strategies by analyzing feedback data in real time, adjusting the model's parameters based on real-time feedback to maintain energy optimization under different operating conditions; The specific operation is as follows: According to the feedback data, dynamically update the parameters of the device control model; When the external environment changes, the adaptive control strategy adjusts the control decision according to the new environmental data; The edge computing node adjusts the control algorithm based on real-time monitoring data to dynamically optimize the working state of the device; The edge computing node executes local decision and control operation, and through the combination with the cloud computing platform, uses the computing power of the cloud to analyze and store data, and the feedback mechanism realizes the data intercommunication between the edge node and the cloud platform, continuously optimizes the control strategy; The specific operation is as follows: The edge computing node collects device operation data in real time through sensors and performs preliminary analysis and decision based on local algorithms; The edge computing node uploads feedback data to the cloud platform, and the cloud uses more complex machine learning models to further analyze energy saving effect and generate optimization suggestions; The cloud platform returns the optimization suggestions to the edge node, and the edge node adjusts the working parameters or scheduling scheme of the device according to the new strategy.

7. An energy saving control system based on industrial internet and AI edge computing, characterized in that, It includes: Sensor network module, for deploying sensors on each production equipment to collect real-time operation data of the equipment; Edge computing node module, for processing collected data and executing energy saving decisions, the edge computing node includes at least one AI model to analyze and predict energy consumption trend; Central control platform module, for communicating with edge computing nodes through industrial internet, collecting plant-wide data and generating global monitoring and scheduling functions; Feedback system module, for tracking energy saving effect in real time and transmitting feedback information to edge computing nodes for decision optimization.

8. An electronic device, comprising: A computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the energy saving control method based on industrial internet and AI edge computing according to any one of claims 1-6.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the energy saving control method based on industrial internet and AI edge computing according to any one of claims 1-6.

Images