Power supply intelligent management method and system, storage medium and program product

Abstract

The invention provides an intelligent power management method and system, a storage medium and a program product, and relates to the field of power management systems.The method comprises the steps that a time series data set is constructed according to historical power consumption data and corresponding electricity consumption in a first preset time period; training a machine learning model by using the time sequence data set to obtain a real-time load prediction model; inputting the historical power consumption data in a second preset time period into a real-time load prediction model by taking the current time as a starting point, and outputting a predicted power consumption demand in a third preset time period in the future; and adjusting a power distribution strategy according to the current power grid state and the predicted power consumption demand, so that the predicted power consumption demand is matched with the available power capacity. By implementing the method, the power distribution strategy can be dynamically adjusted according to the power consumption demand predicted in real time and the current power grid state, the power consumption demand and the available power capacity can be effectively matched, the utilization efficiency of power resources is improved, and energy waste is reduced.

Description

Technical Field

[0001] This application relates to the field of power management systems, and in particular, to an intelligent power management method, system, storage medium, and program product. Background Art

[0002] As electronic devices become increasingly popular, the advancement of power management systems is crucial for ensuring the stable operation of devices and achieving energy conservation. Currently, most traditional power management systems mainly operate based on preset thresholds or timing control strategies. These systems usually lack sufficient intelligence and adaptability, making them unable to cope effectively with the increasingly complex and variable power consumption demands.

[0003] In related technologies, power management usually predicts power demands through fixed power limits or basic algorithms, and then allocates and regulates power sources. This method can basically meet the power allocation requirements to a certain extent, but mainly relies on preset parameters for control.

[0004] However, this fixed control mode cannot accurately adapt to real-time load changes, and often has difficulty effectively predicting or coping with rapidly changing power demands. Especially in high-load data centers or complex network environments that need to manage multiple household appliances, it often leads to problems such as resource waste or insufficient power supply. Summary of the Invention

[0005] This application provides an intelligent power management method, system, storage medium, and program product, which are used to solve the problem of how to achieve more flexible and efficient power allocation in a power management system to adapt to the continuous changes in real-time power demands.

[0006] In a first aspect, this application provides an intelligent power management method, which is applied to an intelligent power management system. The method includes: Construct a time series data set based on historical power consumption data and corresponding power consumption within a first preset time period; Use the time series data set to train a machine learning model to obtain a real-time load prediction model; Taking the current time as the starting point, input the historical power consumption data within a second preset time period into the real-time load prediction model, and output the predicted power consumption demand within a future third preset time period; Adjust the power allocation strategy according to the current power grid state and the predicted power consumption demand, so that the predicted power consumption demand matches the available power capacity.

[0007] Through the above embodiments, the power intelligent management system constructs a time series data set based on historical power consumption data, and uses the data set to train a machine learning model to predict the real-time load, so that the system can dynamically adjust the power allocation strategy according to the real-time predicted power demand and the current power grid status, and can effectively match the power demand with the available power capacity, thereby improving the utilization efficiency of power resources, reducing energy waste, and optimizing the performance of the entire power system.

[0008] In some embodiments, the step of using the time series data set to train a machine learning model to obtain a real-time load forecasting model specifically includes: Extend the time series dataset with weather data, holiday data, and special event data; Use machine learning feature selection techniques to determine the impact weight corresponding to each input variable in the time series dataset; The machine learning model parameters are adjusted according to the impact weight to obtain a real-time load forecasting model.

[0009] Through the above embodiments, the power intelligent management system introduces additional weather data, holiday data, and special event data to expand the time series data set. By considering more factors affecting power demand, the accuracy and reliability of the prediction model are improved. The feature selection technology of machine learning helps determine the weight of each influencing factor and adjusts the model parameters accordingly, so that the prediction results are more in line with the actual power usage, enhancing the intelligence and responsiveness of power management.

[0010] In some embodiments, the step of adjusting the power distribution strategy according to the current power grid state and the predicted power demand to match the predicted power demand with the available power capacity specifically includes: Determine the task priority of each subsystem process and sort the priorities; In the priority ranking, the corresponding subsystem processes are suspended or closed in order from low to high, and the new power demand after each process adjustment is obtained; When it is detected that the new power demand matches the available power capacity, the process power consumption corresponding to the current subsystem process is maintained.

[0011] Through the above embodiments, the intelligent power management system can adjust the power demand to match the available power capacity while ensuring the continuous operation of key systems by determining the priority of subsystem processes and sorting them, and suspending or shutting down low-priority subsystem processes when necessary. This dynamic adjustment strategy helps to maintain stable operation of the system when the power supply is tight, and also maximizes the operating efficiency of the equipment when the power supply is sufficient, ensuring the reasonable allocation and efficient use of power resources.

[0012] In some embodiments, before the step of maintaining the process power consumption corresponding to the current subsystem process after detecting that the new power consumption demand matches the available power supply capacity, the method further includes: Training a machine learning model based on the historical power grid state data to obtain a power grid state prediction model; Inputting the current power grid state data into the power grid state prediction model to output the predicted power grid state within a third preset time period in the future; Determining the available power supply capacity according to the predicted power grid state.

[0013] Through the above embodiments, the power intelligent management system trains a power grid state prediction model and uses the model to predict the future power grid state, enabling the power management system to anticipate possible changes in the power grid in advance, timely adjust the power distribution strategy to cope with the future power grid state, thereby enhancing the system's adaptability to power grid fluctuations and response speed, and ensuring the continuity and security of power supply.

[0014] In some embodiments, before the step of adjusting the power distribution strategy according to the current power grid state and the predicted power consumption demand to make the predicted power consumption demand match the available power supply capacity, the method further includes: Judging whether the available power supply capacity is much greater than the predicted power consumption demand, where "much greater than" means exceeding a preset multiple of a numerical value; If so, sharing a preset proportion of the electric energy in the available power supply capacity with other system devices or power grid devices; If not, adjusting the power distribution strategy.

[0015] Through the above embodiments, when the power supply capacity is much greater than the demand, the power intelligent management system allows sharing the surplus electric energy with other system or power grid devices, which can not only increase the economic benefits of the power system, but also improve the energy efficiency and stability of the entire power grid.

[0016] In some embodiments, the step of taking the current time as the starting point, inputting the historical power consumption data within a second preset time period into the real-time load prediction model, and outputting the power consumption demand within a third preset time period in the future specifically includes: Dividing a plurality of power consumption modules according to the location information of each power consumption device; Inputting the historical power consumption data of each power consumption module within the second preset time period into the real-time load prediction model to output the predicted power consumption demand of the module within a third preset time period in the future; If it is detected that the change range of the module power consumption demand exceeds a preset threshold, an abnormal power consumption warning is given.

[0017] Through the above embodiments, the power intelligent management system divides multiple power consumption modules based on the location information of electrical equipment, and conducts independent power consumption data analysis and load prediction for each module, allowing the system to manage power resources more meticulously and accurately, as well as quickly and accurately warn of abnormal situations.

[0018] In some embodiments, after the step of adjusting the power distribution strategy according to the current power grid state and the predicted power consumption demand so that the predicted power consumption demand matches the available power capacity, the method further includes: Recording the actual power consumption demand within the third preset time period; If it is detected that the difference between the actual power consumption demand and the predicted power consumption demand exceeds a preset threshold, incremental training is performed on the real-time load prediction model.

[0019] Through the above embodiments, the power intelligent management system can continuously improve the accuracy of the real-time load prediction model by recording the actual power consumption demand, comparing it with the predicted value, and performing incremental training on the model when a large deviation is found. This enables the power intelligent management system to more precisely adapt to changes in the actual power consumption pattern, thereby continuously optimizing the configuration and use of power resources and enhancing the sustainability and intelligence level of the system.

[0020] In a second aspect, the present application provides a power intelligent management system, which includes: one or more processors and a memory; The memory is coupled to the one or more processors. The memory is used to store computer program code, and the computer program code includes computer instructions. The one or more processors call the computer instructions so that the power intelligent management system can implement a power intelligent management method provided in the above embodiments, which will not be elaborated here.

[0021] In a third aspect, the present application provides a computer-readable storage medium, including instructions, which, when running on a power intelligent management system, enable the power intelligent management system to implement a power intelligent management method provided in the above embodiments, which will not be elaborated here.

[0022] In a fourth aspect, the present application provides a computer program product, which, when running on a power intelligent management system, enables the power intelligent management system to implement a power intelligent management method provided in the above embodiments, which will not be elaborated here.

[0023] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. The power intelligent management system trains a machine learning model through a time series data set constructed based on historical power consumption data to achieve accurate prediction of real-time load. The system can timely adjust the power distribution strategy according to the prediction results, adapt to the current power grid state, and effectively match the power consumption demand with the available power capacity. This dynamic adjustment not only improves the utilization efficiency of power resources, reduces energy waste, but also optimizes the performance of the overall power system, enhancing the stability and response ability of the power grid.

[0024] 2. By expanding the time series data set and adding weather data, holiday data, and special event data, the load prediction model of the power intelligent management system considers more factors affecting power demand, greatly improving the accuracy and reliability of the prediction. Using feature selection techniques in machine learning, the system not only optimizes the model parameters but also ensures that the prediction results are highly consistent with the actual power usage situation, thereby further enhancing the intelligence and response ability of power management.

[0025] 3. The power intelligent management system records and compares the actual power consumption demand with the predicted value, and performs incremental training of the model for the deviation between the prediction and the actual. This continuous learning and optimization process enables the load prediction model to adapt to the continuous changes in the actual power consumption pattern, not only improving the prediction accuracy but also ensuring the adaptability and flexibility of the system in the face of dynamic changes in power demand. Description of the Drawings

[0026] Figure 1 is a flowchart of a power intelligent management method in an embodiment of the present application; Figure 2 is another flowchart of a power intelligent management method in an embodiment of the present application; Figure 3 is a schematic structural diagram of an entity device of a power intelligent management system in an embodiment of the present application. Detailed Embodiments

[0027] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification and appended claims of the present application, the singular forms "a", "an", "the", "above", "said", "this" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It should also be understood that the term "and / or" used in the present application refers to any and all possible combinations including one or more of the listed items.

[0028] Hereinafter, the terms "first" and "second" are only used for descriptive purposes and should not be construed as implying or suggesting relative importance or implicitly indicating the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0029] For ease of understanding, the method provided in this embodiment will be described in terms of its process below. Specifically, as Figure 1 shown, it is a schematic flowchart of a power intelligent management method in an embodiment of the present application.

[0030] S101. Construct a time series data set based on the historical power consumption data and the corresponding electricity consumption within a first preset time period.

[0031] The power intelligent management system collects the historical power consumption data and the corresponding actual electricity consumption within a first preset time period (such as the past year). Among them, the historical power consumption data may include information such as the power requirements and operating states of each electrical device or electrical unit at different time points. The corresponding electricity consumption refers to the total amount of electrical energy actually consumed at these time points.

[0032] The power intelligent management system arranges the collected data in chronological order to construct a time series data set. Each record in this data set contains the power consumption demand and the actual electricity consumption at a certain time point. In this way, the system establishes a mapping relationship between the power consumption demand and the actual electricity consumption.

[0033] S102. Use the time series data set to train a machine learning model to obtain a real-time load prediction model.

[0034] After constructing the time series data set, the power intelligent management system uses these historical data to train a real-time load prediction model through machine learning methods. It is used to estimate the power demand within a future period based on the historical power consumption data.

[0035] Specifically, the power intelligent management system divides the time series data set into a training set and a test set according to a certain ratio. Among them, the training set is used to train the model, and the test set is used to verify the performance of the model. In the training stage, the system selects appropriate machine learning algorithms, such as time series analysis, regression analysis, neural networks, etc., to construct the framework of the prediction model. Then, the training set data is input into the model, and through repeated iteration and parameter tuning, the fitting ability of the model to the historical data is continuously improved. Through this process, the model learns the key features affecting power consumption and the temporal law of power consumption changes from a large number of historical samples.

[0036] After training, the power intelligent management system uses the test set data to evaluate the model. By comparing the error between the model prediction results and the actual values, the system can evaluate the accuracy and generalization ability of the model. If the model performance meets the requirements, it will be used as the real-time load prediction model; if the performance is not ideal enough, it is necessary to further optimize the model structure or adjust the training data until a stable and reliable real-time load prediction model (for example, the prediction error is controlled within 5%) is obtained.

[0037] S103. Starting from the current time, input the historical power consumption data within the second preset time period into the real-time load prediction model, and output the predicted electricity demand within the next third preset time period.

[0038] When the power intelligent management system needs to predict the electricity demand for a period of time in the future, first, starting from the current time, collect the historical power consumption data within the second preset time period as the input data of the real-time load prediction model. Among them, usually the length of the second preset time period should be greater than or equal to the length of the third preset time period to ensure that there is enough historical information to support the prediction.

[0039] Specifically, the power intelligent management system arranges the historical power consumption data within the second preset time period in chronological order to form an input sequence. Each item of this sequence contains the power consumption characteristics at the corresponding moment, such as the total power demand, the power demand of sub-devices, etc. The system inputs the input sequence into the trained real-time load prediction model, and the model gradually scans these historical data to capture the electricity consumption patterns and trend characteristics contained therein. With the help of the electricity consumption patterns learned by the model, the system can extrapolate the electricity demand within the next third preset time period. Among them, the output form of the model is usually a time series, and each item represents the estimated power demand at a certain future moment.

[0040] S104. Adjust the power distribution strategy according to the current power grid state and the predicted electricity demand to make the predicted electricity demand match the available power capacity.

[0041] After determining the predicted electricity demand for a period of time in the future, the power intelligent management system dynamically optimizes and adjusts the power distribution strategy in combination with the actual operating state of the current power grid to balance the power supply and demand and improve the power utilization rate.

[0042] Specifically, the power intelligent management system obtains the operating parameters of the current power grid in real time, including the output of the generating units, the power flow of the transmission lines, the state of the electrical equipment, etc. Among them, the operating parameters of the power grid can be obtained by docking data with intelligent meters, monitoring devices, etc. The available power capacity comes from the system's own power generation capacity, the power supply capacity of the power grid, and the power reserve of the energy storage devices.

[0043] Based on the predicted electricity demand and available power capacity, the power intelligent management system formulates an optimal power distribution strategy. Specifically, if the predicted demand is less than the available capacity, the surplus power is reasonably distributed, such as being retransmitted to the power grid, etc.; if the predicted demand is greater than the available capacity, the system takes measures to cut or shift the peak electricity demand to avoid power supply shortages. Specific optimization strategies can include adjusting the output power of generating units, controlling the power consumption time periods of interruptible loads, using time-of-use electricity prices to encourage users to change their electricity consumption habits, coordinating the grid connection operation of distributed power sources, etc. The formulation of the strategy needs to consider multiple factors such as equipment capacity, environmental protection restrictions, economic benefits, and user experience, which are not limited here.

[0044] The following further describes the more specific process of the method provided in this embodiment. Specifically, as Figure 2 shown, it is another process schematic diagram of a power intelligent management method in an embodiment of the present application.

[0045] S201. Use weather data, holiday data, and special event data to expand the time series data set.

[0046] When constructing the time series data set, the power intelligent management system not only considers historical power consumption data and corresponding electricity consumption, but also introduces additional dimensional data such as weather, holidays, and special events. For example, during the peak electricity consumption period in the hot summer, the use of cooling equipment such as air conditioners will increase significantly, resulting in a sharp increase in power demand. During major holidays such as the Spring Festival and National Day, enterprises shut down and residents take vacations, and the electricity consumption may drop significantly. Similarly, special events such as major social activities and sports events will also have a significant impact on the electricity consumption in local areas.

[0047] Therefore, the power intelligent management system combines multi-dimensional data such as historical weather conditions (such as temperature, humidity, wind force, etc.), holiday arrangements (such as legal holidays, weekends, etc.), and special events (such as large-scale exhibitions, sports events, etc.) with the original time series data set to form a richer and more three-dimensional feature space. By capturing the deep driving factors behind the load changes, it helps to improve the accuracy of prediction.

[0048] S202. Use machine learning feature selection techniques to determine the influence weights corresponding to each input variable in the time series data set.

[0049] After introducing external data such as weather and holidays to enrich the time series data set, the power intelligent management system evaluates the importance of each input feature for load prediction. Specifically, the power intelligent management system uses feature selection techniques in the field of machine learning, such as feature importance evaluation based on tree models, regularization methods (such as L1 regularization), mutual information, etc., to quantitatively describe the dependence relationship between each input feature and the target variable, and accordingly assigns them different weight coefficients.

[0050] S203. Adjust the parameters of the machine learning model according to the influence weights to obtain a real-time load prediction model.

[0051] Based on the feature weight information obtained in step S202, the power intelligent management system performs targeted optimization on the key parameters of the real-time load prediction model. Among them, the real-time load prediction model can adopt mature machine learning algorithms such as support vector regression, random forest, LSTM, etc., which are not limited here.

[0052] The power intelligent management system adaptively adjusts the model hyperparameters according to the feature weights, so as to maximize the predictive power of the key features. For example, for continuous variables with higher weights such as temperature, the system will correspondingly increase the maximum depth of the decision tree to capture finer-grained non-linear relationships; for discrete variables with higher weights such as holidays, the system will appropriately increase the number of trees to reflect the differences in load under different date types.

[0053] By combining domain knowledge (feature weights) with data-driven learning algorithms (parameter tuning), the power intelligent management system can automatically extract a real-time load prediction model with high robustness and strong generalization ability from a large amount of historical operation data. This model can dynamically adapt to changing influencing factors, timely capture the fluctuation trend of power demand, and provide a reliable basis for grid dispatching optimization decisions.

[0054] In the above embodiment, the power intelligent management system introduces additional weather data, holiday data, and special event data to expand the time series data set. By considering more factors affecting power demand, the accuracy and reliability of the prediction model are improved. The feature selection technology of machine learning helps to determine the weights of each influencing factor and adjusts the model parameters accordingly, so that the prediction results are more in line with the actual power usage situation, enhancing the intelligence and response ability of power management.

[0055] S204. Divide multiple power consumption modules according to the location information of each power consumption device.

[0056] In the actual operation of the power grid, the power demand patterns of users in different regions and of different types often have significant differences. For example, the peak power demand in the commercial area usually appears during the day, while the peak power consumption in the residential area is mostly at night. To achieve more refined load management, the power intelligent management system divides the power consumption devices into multiple independent power consumption modules according to their geographical location information.

[0057] Specifically, the system comprehensively collects geographical information of electrical equipment (such as substations, distribution rooms, gateway meters, etc.) within its jurisdiction and marks the longitude and latitude coordinates where they are located. Then, combining prior information such as administrative division maps and user type distribution maps, data mining algorithms such as clustering are used to automatically classify equipment with adjacent locations and similar electricity consumption characteristics into the same module. Among them, the electricity consumption demands within each module are relatively convergent, while there are obvious differences between different modules.

[0058] S205. Input the historical power consumption data of each electricity consumption module within the second preset time period into the real-time load prediction model, and output the predicted electricity consumption demand of the module within the future third preset time period.

[0059] Based on the result of the electricity consumption module division, the power intelligent management system independently predicts the power demands of each module. First, the system takes the current moment as a benchmark and extracts the historical power consumption data of each module within the past second preset time period (such as the previous 3 hours) as the model input. At the same time, the system collects data on external influencing factors such as the real-time weather conditions in the areas where each module is located and the current date type (weekday / holiday), and inputs them together with the historical load data into the real-time load prediction model to output the predicted value of the electricity consumption demand of the module within the future third preset time period (such as the next 1 hour), reflecting its short-term load change trend. If it is detected that the change range of the electricity consumption demand of the module exceeds the preset threshold, an abnormal electricity consumption warning is issued.

[0060] Through the above embodiments, the power intelligent management system divides multiple electricity consumption modules based on the location information of electrical equipment, and conducts independent power consumption data analysis and load prediction for each module, allowing the system to manage power resources more meticulously and accurately, as well as quickly and accurately warn of abnormal situations.

[0061] S206. Train a machine learning model based on historical power grid state data to obtain a power grid state prediction model.

[0062] The power intelligent management system constructs a historical power grid state data set reflecting the operation status of the power grid by collecting historical power grid state data, including records of key parameters such as the voltage, current, power factor, and frequency of the power grid at different time points, as well as relevant environmental factors such as weather conditions and load distributions.

[0063] The power intelligent management system trains a machine learning model based on this historical power grid state dataset. Optional models include Recurrent Neural Network (RNN), Long Short-Term Memory Network (LSTM), etc., which are good at processing time series data and capturing long-term dependencies therein. The system continuously optimizes the prediction performance of the model by adjusting the hyperparameters of the model (such as the number of hidden layers, the number of neurons) and the training algorithm (such as the Adam optimizer). Through repeated iterative optimization, a power grid state prediction model that can accurately predict the operating state of the power grid for a period of time in the future is obtained.

[0064] S207. Input the current power grid state data into the power grid state prediction model, and output the estimated power grid state within the third preset time period in the future.

[0065] After completing the training of the power grid state prediction model, the power intelligent management system can use this model to estimate the future power grid state. Specifically, the system collects the operating parameters of the current power grid in real time, such as voltage, current, etc., and organizes them into an input in the same format as the training data. Then, the system inputs the data representing the current power grid state into the prediction model, and the model calculates and outputs the estimated value of the power grid state within a period of time (the third preset time period) in the future.

[0066] It should be noted that this estimation process can be carried out in a rolling manner. For example, if the third preset time period is 24 hours in the future, the system can obtain the latest power grid state data every 1 hour, input it into the model for prediction, and then collect the actual state data 1 hour later, compare it with the predicted value, and use it to evaluate and optimize the model performance. At the same time, the newly obtained state data will also be updated to the training dataset in a timely manner for the next round of incremental learning, so that the model can continue to improve and adapt to the changes in the operating characteristics of the power grid.

[0067] S208. Determine the available power capacity based on the estimated power grid state.

[0068] After obtaining the estimated power grid state within a period of time in the future, the power intelligent management system further analyzes and judges how much power supply capacity the power grid can provide under these estimated states, that is, determines the upper limit of the available power capacity.

[0069] Specifically, the system comprehensively considers various key indicators in the power grid state prediction results, such as the predicted power grid voltage level, line power flow distribution, equipment load rate, etc., and combines static information such as the physical topology structure of the power grid and equipment parameters to analyze and calculate the power transmission capacity and power source dispatching space of the power grid. For example, in one embodiment, the power source intelligent management system predicts that the voltage of the 220 kV backbone line of the power grid will be maintained within ±2% of the nominal voltage within the next 1 hour, the average line load rate is 60%, and there is no significant power flow overload phenomenon. At the same time, the power outputs of the grid-connected wind farms and photovoltaic power plants are expected to be 80 MW and 120 MW respectively, and a ±20% power output adjustment can be achieved through dispatching. The system calculates that in this predicted state, the power grid can provide a maximum power supply capacity of 500 MW, of which the capacity of conventional power sources is 380 MW and the adjustable capacity of new energy is 120 MW.

[0070] In the above embodiment, the power source intelligent management system trains the power grid state prediction model and uses the model to predict the future power grid state, enabling the power management system to anticipate possible changes in the power grid in advance, timely adjust the power distribution strategy to cope with the future power grid state, thereby enhancing the system's adaptability and response speed to power grid fluctuations and ensuring the continuity and security of power supply.

[0071] S209. Determine whether the available power source capacity is much larger than the predicted power consumption demand.

[0072] After determining the available power source capacity, the power source intelligent management system compares it with the future power consumption demand predicted in step S205 to determine whether the power supply capacity is sufficient.

[0073] Specifically, the system calculates the difference or ratio between the available power source capacity and the predicted power consumption demand. If this difference is greater than a preset threshold (such as 100 MW), or the ratio is greater than a preset multiple (such as 1.5 times), it is determined that the available power source capacity is "much larger than" the predicted power consumption demand. The setting of the threshold or multiple here can be adjusted according to engineering experience based on factors such as the operating characteristics of the power grid and the requirements for stability margin, and is not limited here.

[0074] S210. Determine the task priorities of each subsystem process and perform priority sorting.

[0075] The power source intelligent management system first evaluates and sorts the task importance of each subsystem process. The system can comprehensively consider factors such as the function of each process, the degree of influence on the overall system performance, and historical operation data, and use machine learning algorithms such as AHP (Analytic Hierarchy Process) to quantitatively calculate and sort the priorities of the processes.

[0076] For example, in an intelligent manufacturing system, core processes such as the motion control process and workpiece quality inspection process of a CNC machining center, which are directly related to product quality and production efficiency, are usually assigned the highest priority. While auxiliary processes such as equipment status monitoring and environmental data collection can be ranked at relatively lower priorities. The system sorts each process according to the calculated priority to form a priority queue, providing a decision basis for subsequent dynamic adjustments.

[0077] S211. Pause or shut down the corresponding subsystem processes in ascending order from low to high in the priority sorting to obtain the new power consumption demand after each process adjustment.

[0078] When the power intelligent management system predicts that there may be a power supply shortage in the next period of time, it takes proactive measures to reduce the power load to ensure the operation of critical tasks. The system will start from the process with the lowest priority according to the process priority queue obtained in step S210, and gradually pause or shut down the corresponding subsystem processes. After each process adjustment, the system recalculates the current total power consumption demand until it matches the available power capacity.

[0079] S212. When it is detected that the new power consumption demand matches the available power capacity, maintain the process power consumption corresponding to the current subsystem processes.

[0080] Through the dynamic adjustment in step S211, the power intelligent management system finally finds a combination of process configurations that can maintain the stable operation of the system under the current power supply conditions. At this time, the system will suspend further adjustments and control the power consumption of each subsystem process to remain at the current level.

[0081] In the above embodiment, the power intelligent management system can adjust the power consumption demand to match the available power capacity while ensuring the continuous operation of critical systems by determining and sorting the priorities of subsystem processes and pausing or shutting down low-priority subsystem processes when necessary. This dynamic adjustment strategy helps to maintain the stable operation of the system during power shortages, and also maximizes the operation efficiency of the equipment when the power supply is sufficient, ensuring the reasonable allocation and efficient use of power resources.

[0082] S213. Record the actual power consumption demand within the third preset time period.

[0083] The power intelligent management system will continuously record the actual power consumption of each subsystem and device in the next period of time (the third preset time period). The system obtains real-time data such as the power, current, and voltage of each load unit through intelligent electricity meters, power consumption information collection devices, etc., and summarizes and calculates to obtain the actual total power consumption demand of the entire system.

[0084] S214. If the difference between the detected actual power consumption demand and the predicted power consumption demand exceeds a preset threshold, perform incremental training on the real-time load prediction model.

[0085] Due to the fact that actual power consumption behaviors may be affected by many uncertain factors, it is inevitable that there will be deviations in the load prediction model trained by the power intelligent management system based on historical data. Therefore, during the operation of the system, it continuously monitors the difference between the prediction result and the actual power consumption demand. When the accumulated deviation exceeds a certain threshold, it promptly triggers the incremental training and optimization of the prediction model.

[0086] In the above embodiment, the power intelligent management system records the actual power consumption demand and compares it with the predicted value. When a large deviation is found, it performs incremental training on the model, which can continuously improve the accuracy of the real-time load prediction model, enabling the power intelligent management system to more precisely adapt to the changes in the actual power consumption pattern, thereby continuously optimizing the allocation and use of power resources and enhancing the sustainability and intelligence level of the system.

[0087] S215. Share a preset proportion of the electrical energy in the available power capacity with other system devices or grid devices.

[0088] When the power supply is relatively sufficient, the power intelligent management system can share the surplus electrical energy with other systems to improve energy utilization efficiency and economic benefits. Specifically, the system evaluates the difference between its available power capacity and power consumption demand. When this difference exceeds a certain proportion (such as 20%), it establishes a sharing mechanism with other power demand parties.

[0089] In the above embodiment, when the power capacity is much larger than the demand, the power intelligent management system allows the surplus electrical energy to be shared with other systems or grid devices, which can not only increase the economic benefits of the power system but also improve the energy efficiency and stability of the entire power grid.

[0090] The power intelligent management system of the embodiment of the present invention is applied to an electronic device. Figure 3 The schematic diagram of the architecture of the electronic device suitable for implementing the embodiment of the present invention is shown.

[0091] It should be noted that Figure 3 The shown electronic device is only an example and should not bring any limitations to the functions and usage scope of the embodiment of the present invention.

[0092] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions (computer programs), or by controlling relevant hardware through instructions (computer programs). These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor. The electronic device of this embodiment includes a storage medium and a processor. Among them, multiple instructions are stored in the storage medium, and these instructions can be loaded by the processor to execute any step of the method provided by the embodiments of the present invention.

[0093] Specifically, the storage medium and the processor are directly or indirectly electrically connected to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more signal lines. The computer execution instructions for implementing the data access control method are stored in the storage medium, including at least one software function module that can be stored in the storage medium in the form of software or firmware. The processor executes various functional applications and data processing by running the software programs and modules stored in the storage medium. The storage medium can be, but is not limited to, a random access storage medium (Random Access Memory, abbreviated as RAM), a read-only storage medium (Read Only Memory, abbreviated as ROM), a programmable read-only storage medium (Programmable Read-Only Memory, abbreviated as PROM), an erasable read-only storage medium (Erasable Programmable Read-Only Memory, abbreviated as EPROM), an electrically erasable read-only storage medium (Electric Erasable Programmable Read-Only Memory, abbreviated as EEPROM), etc. Among them, the storage medium is used to store programs, and the processor executes the programs after receiving the execution instructions.

[0094] Furthermore, the software programs and modules in the above storage medium may further include an operating system, which may include various software components and / or drivers for managing system tasks (such as memory management, storage device control, power management, etc.), and can communicate with various hardware or software components to provide a running environment for other software components. The processor can be an integrated circuit chip with signal processing capabilities. The above-mentioned processor can be a general-purpose processor, including a central processing unit (Central Processing Unit, abbreviated as CPU), a network processor (Network Processor, abbreviated as NP), etc., which can implement or execute the various methods, steps and logic flow block diagrams disclosed in this embodiment. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0095] Since the instructions stored in the storage medium can execute the steps in any of the methods provided in the embodiments of the present invention, the beneficial effects of any of the methods provided in the embodiments of the present invention can be achieved. For details, please refer to the previous embodiments and will not be repeated here.

[0096] The above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A power supply intelligent management method, applied to a power supply intelligent management system, characterized in that: The method comprises: Constructing a time series data set based on historical power consumption data and corresponding power consumption data within a first preset time period; Using the time series data set to train a machine learning model to obtain a real-time load forecasting model; Taking the current time as the starting point, inputting the historical power consumption data within the second preset time period into the real-time load forecasting model, and outputting the predicted power demand within the third preset time period in the future; The power distribution strategy is adjusted according to the current power grid state and the predicted power demand, so that the predicted power demand matches the available power capacity.

2. The method according to claim 1, characterized in that The step of using the time series data set to train a machine learning model to obtain a real-time load forecasting model specifically includes: Extend the time series dataset with weather data, holiday data, and special event data; Using machine learning feature selection techniques to determine the impact weight corresponding to each input variable in the time series data set; The machine learning model parameters are adjusted according to the impact weights to obtain a real-time load forecasting model.

3. The method according to claim 1, characterized in that: The step of adjusting the power distribution strategy according to the current power grid state and the predicted power demand to match the predicted power demand with the available power capacity specifically includes: Determine the task priority of each subsystem process and sort the priorities; In the priority ranking, pausing or shutting down corresponding subsystem processes in order from low to high, and obtaining a new power demand after each process adjustment; When it is detected that the new power demand matches the available power capacity, the process power consumption corresponding to the current subsystem process is maintained.

4. The method according to claim 3, characterized in that Before the step of maintaining the process power consumption corresponding to the current subsystem process after detecting that the new power demand matches the available power capacity, the method further includes: Training a machine learning model based on the historical power grid state data to obtain a power grid state prediction model; Inputting current grid state data into the grid state prediction model, and outputting an estimated grid state within a third preset time period in the future; The available power capacity is determined according to the estimated power grid status.

5. The method according to claim 1, characterized in that Before the step of adjusting the power distribution strategy according to the current power grid state and the predicted power demand to match the predicted power demand with the available power capacity, the step further includes: Determine whether the available power capacity is much greater than the predicted power demand, where much greater means exceeding a preset value by a multiple; If yes, then sharing a preset proportion of the electric energy in the available power capacity with other system devices or power grid devices; If not, adjust the power distribution strategy.

6. The method according to claim 1, characterized in that The step of taking the current time as the starting point, inputting the historical power consumption data within the second preset time period into the real-time load prediction model, and outputting the power demand within the future third preset time period specifically includes: Divide multiple power consumption modules according to the location information of each power consumption device; Inputting the historical power consumption data of each power consumption module in the second preset time period into the real-time load prediction model, and outputting the predicted power demand of the module in the future third preset time period; If the variation of the power demand of the module detected exceeds a preset threshold, an abnormal power consumption warning is issued.

7. The method according to claim 1, characterized in that After the step of adjusting the power distribution strategy according to the current power grid state and the predicted power demand so that the predicted power demand matches the available power capacity, the method further includes: Recording the actual power demand within the third preset time period; If it is detected that the difference between the actual power demand and the predicted power demand exceeds a preset threshold, incremental training is performed on the real-time load prediction model.

8. A power supply intelligent management system, characterized in that: The power intelligent management system includes: one or more processors and memory; The memory is coupled to the one or more processors, and the memory is used to store computer program codes, wherein the computer program codes include computer instructions, and the one or more processors call the computer instructions to enable the power intelligent management system to execute the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instruction is executed on the intelligent power management system, the intelligent power management system executes the method as claimed in any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product runs on an intelligent power management system, the intelligent power management system is enabled to execute the method according to any one of claims 1 to 7.

Images