Electric equipment energy-saving control method and system based on Internet of Things

Abstract

The invention discloses an electric equipment energy-saving control method and system based on the Internet of Things, and relates to the technical field of smart power grid and Internet of Things application, and the method comprises the steps: S1, obtaining the operation data and historical use records of user electric equipment, extracting the habit diversity and personalized preference analysis characteristics of a user, obtaining an equipment use preference model, and carrying out the operation of the user electric equipment; s2, extracting an operation rule of each device from the preference model, judging a priority adjustment parameter of the device, and obtaining a dynamic priority sequence, and S3, calculating an available power threshold value under current power resource limitation through real-time power grid load monitoring data, and determining a distribution power value required by important device guarantee; according to the electric equipment energy-saving control method and system based on the Internet of Things, the preference model and the priority sequence can be updated in time, and optimal distribution of electric power resources is achieved. The method can effectively balance the power demand of the user and the load of the power grid, improves the utilization efficiency of power resources, and provides a new solution for the management of the intelligent power grid.

Description

Technical Field

[0001] The present invention relates to the technical field of smart grid and Internet of Things applications, and particularly relates to an energy-saving control method and system for electrical equipment based on the Internet of Things. Background Art

[0002] In the development of modern smart grids, the demand response system, as an important technical means to optimize grid resource allocation and improve energy utilization efficiency, has received extensive attention. The core goal of this system is to achieve load peak shaving and valley filling by guiding users' electricity consumption behaviors during grid load fluctuations, thereby ensuring the stable operation of the grid. Existing demand response systems usually adjust power distribution through preset rules or simple user portraits. However, in actual applications, users' electricity consumption behaviors exhibit significant personalized characteristics and dynamic change trends, and traditional fixed models or static preference configurations are difficult to accurately reflect users' true intentions. This uncertainty makes it difficult for the system to effectively balance user satisfaction and grid operation efficiency when performing dynamic power distribution.

[0003] Firstly, users' electricity consumption habits have obvious diversity and personalized characteristics, and these behavioral characteristics are often hidden in a large amount of historical data. How to extract high-value electricity consumption preference characteristics from this massive data and construct a reliable user behavior model is one of the key challenges faced by the demand response system. Secondly, there are differences in the power supply priorities of electrical terminal devices, and the load contributions, user sensitivities, and criticalities among different devices are different. During the dynamic regulation process, how to reasonably rank the device priorities according to real-time demands and maximize the reduction of non-critical loads while ensuring the normal operation of critical devices involves an optimization decision problem under multi-variable constraints. Thirdly, the fluctuations of grid loads have strong randomness and real-time nature, and the system must respond quickly when adjusting power to prevent grid overload or imbalance. Although some current response strategies have certain flexibility, there is still a contradiction between the response speed and the stability of the power distribution scheme, which is likely to cause frequent fluctuations or adjustment lags and affect the system performance. In addition, users' long-term behavior patterns will evolve over time, and the original preference model may gradually become invalid. How to construct a preference recognition mechanism with adaptive update ability to adjust strategies in time before the model fails is the key to maintaining the long-term effectiveness of the system. Finally, in the context of limited power resources, how to formulate an optimal dynamic power distribution scheme under multi-objective optimization constraints to meet users' personalized electricity consumption needs and also take into account grid dispatching instructions urgently requires the introduction of efficient intelligent optimization algorithms to support decision-making. This poses higher requirements for algorithm performance, resource scheduling strategies, and the overall coordination ability of the system. Summary of the Invention

[0004] The purpose of the present invention is to provide an energy-saving control method and system for electrical equipment based on the Internet of Things to solve the problems existing in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solutions: An energy-saving control method for electrical equipment based on the Internet of Things, characterized by comprising: S1. Obtain the operation data and historical usage records of the user's electrical equipment, extract the characteristics of user habit diversity and personalized preferences for analysis, and obtain the equipment usage preference model; S2. Extract the operation rules of each device from the preference model, judge the device priority adjustment parameters, and obtain the dynamic priority sequence; S3. Calculate the available power threshold under the current power resource limit through real-time power grid load monitoring data, and determine the allocated power value required to ensure important devices; S4. Adopt an adaptive control strategy to adjust the power distribution ratio, obtain the reduction amplitude of the energy consumption of non-critical devices, and obtain the optimized power distribution scheme; S5. Extract the actual power supply parameters of each device, judge whether the power grid load fluctuation exceeds the preset threshold. If it exceeds, recalculate the multi-variable trade-off parameters to obtain the adjusted allocation result; S6. Obtain the adjusted allocation result and the operation status of the real-time electrical equipment, calculate the stability of key device guarantee and the reduction amount of non-critical device energy consumption, and determine the current resource allocation efficiency index; S7. Adopt an adaptive control strategy to update the device priority adjustment logic, judge whether the user habit diversity has changed. If it has changed, re-extract the preference model to obtain a new priority sequence; S8. Extract the device power distribution requirements, calculate the optimal distribution scheme under the power resource limit, and determine the final dynamic power distribution result.

[0006] Preferably, the S1 includes: Obtain the operation data and historical records of the electrical equipment, obtain the structured data set through database query, process the structured data set through time series analysis, extract the power consumption time and usage frequency characteristics, use the clustering algorithm to group according to the usage frequency and power change, obtain the classification result of the power change, calculate the relationship between the power change and the power consumption time according to the classification result, determine the habit diversity characteristics, extract the individual preferences from the habit diversity characteristics through regression analysis, obtain the user preference distribution, obtain the user preference distribution and analysis characteristics, use the preset threshold to judge the accuracy of the preference model, and obtain the final device usage preference model by comparing the updated preference model with the data extraction result.

[0007] Preferably, the S2 includes: Extract operation rules and importance indicators through a preference model, obtain the grid load by combining real-time data, get the device operation status parameters, process the operation rules and importance indicators using a multi-variable trade-off method. If the grid load exceeds the preset threshold, adjust the calculation method according to the load change, determine the device priority parameters, extract the load change trend from the real-time data, analyze the device parameters and importance indicators through the calculation method, judge the priority adjustment direction, generate a dynamic sequence for the priority adjustment result, obtain the device sorting after sequence update. If the load change trend continues to rise, recalculate the device parameters through multi-variable trade-off to get the adjusted dynamic sequence, verify the sequence update consistency by combining real-time data according to the dynamic sequence update result, determine the final device priority sequence, and optimize the dynamic sequence through the random forest algorithm to obtain a high-precision priority sequence.

[0008] Preferably, S3 includes: Obtain grid load data through real-time monitoring, calculate the load value at the current moment, get the available power range under resource constraints, extract the change trend from the load data, use a preset threshold to judge device dynamics, determine the adjustment direction of the priority sequence. For the device dynamics in the priority sequence, obtain the operation status of important devices, judge the specific range of guarantee requirements, calculate the power threshold according to the guarantee requirements and the available power range, determine the allocated power value of important devices, update the priority sequence through dynamic adjustment, extract the device status from the adjusted sequence to get the new guarantee requirement range, classify the load data and guarantee requirements using the support vector machine algorithm, judge whether the resource constraints meet the allocated power requirements. If the resource constraints are met, update the device operation status through the allocated power value to obtain the adjusted grid load data.

[0009] Preferably, S4 includes: According to the grid load data collected in real time, obtain its change trend, determine the dynamic priority sequence by analyzing the change trend, extract features from the change trend and the priority sequence, adjust the power distribution ratio using an adaptive control strategy to get a preliminary distribution plan. For the preliminary distribution plan, obtain the operation status and energy consumption data of non-critical devices, judge the reducible space, determine the adjusted power distribution ratio by calculating the reduction amplitude of the energy consumption of non-critical devices to get the intermediate result of the optimized plan. If the intermediate result of the optimized plan exceeds the preset threshold, update the change trend according to the real-time adjusted grid load data, recalculate the distribution ratio, verify the optimized plan using the support vector machine algorithm according to the updated distribution ratio and the energy consumption reduction amplitude to get the final power distribution plan, and obtain the actual reduction amplitude of non-critical devices by comparing the energy consumption data of the final power distribution plan with the initial plan to judge the implementation effect of the plan.

[0010] Preferably, S5 includes: Obtain the actual power supply parameters of each device from the optimized power distribution scheme, determine the device operating status, process the actual power supply parameters through real-time response technology, judge whether the response meets the speed requirements, obtain the grid load data, judge whether the load fluctuation exceeds the preset threshold. If the load fluctuation exceeds the preset threshold, calculate the adjustment parameters through a multi-variable trade-off model to obtain a preliminary adjustment plan, analyze the device parameters for the preliminary adjustment plan to determine the adjusted actual power supply status, process the adjusted actual power supply status through an optimization algorithm to obtain the final allocation result, update the power distribution scheme according to the final allocation result, and complete the grid load balance.

[0011] Preferably, S6 includes: Obtain the real-time data of the electrical equipment, collect the operating status through sensors, determine the current load conditions of each device, extract the data of key devices and non-key devices from the operating status, adopt a dynamic power distribution algorithm to calculate the power distribution scheme, adjust the allocation result through the power distribution scheme to obtain the stability value of the key device and the energy consumption reduction of the non-key device. If the stability value is lower than the preset threshold, adjust the power distribution through priority sorting to determine the new allocation result, calculate the resource efficiency according to the allocation result to obtain the preliminary value of the efficiency index, perform iterative optimization on the efficiency index, adopt a linear regression algorithm to analyze the relationship between the real-time data and the allocation result, judge the optimized efficiency index, update the power distribution scheme through the optimized efficiency index, and determine the final device guarantee status.

[0012] Preferably, S7 includes: Collect the grid load and real-time data, use time series analysis to judge the load change trend to obtain the load fluctuation characteristics, calculate the efficiency index according to the load fluctuation characteristics and the resource allocation data to determine the current resource allocation status. For the efficiency index, adopt an adaptive control strategy to update the adjustment logic to obtain the device priority sequence. Analyze the diversity change characteristics through the device priority sequence and the user habit data to judge whether a significant change has occurred. If the diversity change characteristics exceed the preset threshold, re-extract the preference model through a clustering algorithm to obtain the updated preference sequence. Adjust the device priority logic according to the updated preference sequence to generate a new priority sequence. Verify the adaptability of the adjustment logic through the new priority sequence and the real-time data to determine the final priority sequence.

[0013] Preferably, S8 includes: Obtain the device power and allocation requirements through the priority sequence, use data extraction technology to obtain the initial allocation data set, extract multivariate trade-off parameters from the initial allocation data set, determine the trade-off coefficient through calculation and analysis, obtain the speed constraint conditions according to the trade-off coefficient and real-time response requirements, and get the adjusted allocation priority. For the adjusted allocation priority and power resources, analyze the resource limitations, determine the available power range, and use the linear programming algorithm to obtain a preliminary dynamic power allocation plan based on the available power range and allocation requirements. According to the preliminary dynamic power allocation plan and speed constraints, judge whether the real-time response is satisfied. If not, adjust the allocation parameters to obtain the optimized dynamic power allocation result, and determine the final allocation result through the optimized dynamic power allocation result and calculation analysis.

[0014] An energy-saving control system for electrical equipment based on the Internet of Things, which is used to implement the steps of the energy-saving control method for electrical equipment based on the Internet of Things. The system includes: A data acquisition module, which is used to obtain the operation data and historical usage records of user electrical equipment, extract the analysis features of user habit diversity and personalized preferences, and obtain a device usage preference model; A priority evaluation module, which is used to extract the operation rules of each device from the preference model, judge the device priority adjustment parameters, and obtain a dynamic priority sequence; A power calculation module, which is used to calculate the available power threshold under the current power resource limit according to the real-time power grid load monitoring data, and determine the allocated power value required for ensuring important equipment; A power regulation module, which is used to adjust the power allocation ratio by adopting an adaptive control strategy, obtain the reduction amplitude of the energy consumption of non-critical equipment, and obtain an optimized power allocation plan; A fluctuation response module, which is used to extract the actual power supply parameters of each device, judge whether the power grid load fluctuation exceeds the preset threshold. If it exceeds, recalculate the multivariate trade-off parameters to obtain an adjusted allocation result; An allocation evaluation module, which is used to obtain the adjusted allocation result and the operation status of real-time electrical equipment, calculate the stability of ensuring critical equipment and the reduction amount of energy consumption of non-critical equipment, and determine the current resource allocation efficiency index; A model update module, which is used to update the device priority adjustment logic based on the adaptive control strategy, judge whether the user habit diversity has changed. If it has changed, re-extract the preference model to obtain a new priority sequence; An optimal control module, which is used to extract the device power allocation requirements, calculate the optimal allocation plan under the power resource limit, and determine the final dynamic power allocation result.

[0015] As can be seen from the above technical solutions, the present invention has the following beneficial effects: The energy-saving control method and system for electrical equipment based on the Internet of Things analyze the operation data and historical records of users' electrical equipment, extract the characteristics of user habit diversity and personalized preferences, and establish a device usage preference model. Combining real-time power grid load data, a multi-variable trade-off calculation method is used to generate a dynamic priority sequence. According to the priority sequence and the change of power grid load, the present invention adopts an adaptive control strategy to dynamically adjust the power distribution ratio, ensuring the power supply stability of important equipment while reducing the energy consumption of non-critical equipment. When the user's habits change significantly, the present invention can update the preference model and priority sequence in a timely manner to achieve the optimal allocation of power resources. This method can effectively balance the user's power consumption demand and the power grid load, improve the utilization efficiency of power resources, and provide a new solution for intelligent power grid management. Brief Description of the Drawings

[0016] Figure 1 It is a flowchart of the method of the present invention; Figure 2 It is a connection diagram of system modules of the present invention. Detailed Embodiments

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0018] As Figure 1 shown, the present invention provides a technical solution: an energy-saving control method for electrical equipment based on the Internet of Things, including: S1. Obtain the operation data and historical usage records of the user's electrical equipment, extract the characteristics of user habit diversity and personalized preference analysis, and obtain a device usage preference model; S2. Extract the operation rules of each device from the preference model, judge the device priority adjustment parameters, and obtain a dynamic priority sequence; S3. Through the real-time power grid load monitoring data, calculate the available power threshold under the current power resource limit, and determine the allocated power value required to ensure important equipment; S4. Adopt an adaptive control strategy to adjust the power distribution ratio, obtain the reduction range of the energy consumption of non-critical equipment, and obtain an optimized power distribution plan; S5. Extract the actual power supply parameters of each device, judge whether the power grid load fluctuation exceeds the preset threshold. If it exceeds, recalculate the multi-variable trade-off parameters to obtain an adjusted allocation result; S6. Obtain the adjusted allocation result and the operating status of real-time electrical equipment, calculate the stability of critical equipment guarantee and the reduction of non-critical equipment energy consumption, and determine the current resource allocation efficiency index; S7. Update the device priority adjustment logic using an adaptive control strategy, determine whether the diversity of user habits has changed. If it has changed, re-extract the preference model to obtain a new priority sequence; S8. Extract the device power allocation requirements, calculate the optimal allocation plan under the power resource limit, and determine the final dynamic power allocation result.

[0019] This embodiment relies on the collaborative operation of the Internet of Things perception layer and the intelligent control center, and realizes the intelligent energy-saving regulation of electrical equipment through multi-source data fusion, behavior modeling, and real-time optimization. First, the system collects the operation data of equipment in user households or industrial sites through Internet of Things terminal nodes, combines the usage trajectories in the cloud historical record library, and uses feature extraction algorithms to analyze user operation habits, equipment start-stop cycles, frequently used time periods, etc., to construct a user preference feature vector set, and then uses clustering or deep learning models to generate a preference model. In this model, the system identifies the operation frequency, importance level, and controllability characteristics of each device in different scenarios, and forms the device dynamic priority for adjusting the control strategy. Subsequently, the system receives the grid load parameters from the power operator in real time (such as power capacity, fluctuation range, load peak prediction, etc.), combines the internal scheduling requirements, calculates the minimum guaranteed power demand of critical equipment, and designs the allocation boundary conditions based on this threshold. In the control layer, the system calls an adaptive controller based on multi-objective optimization to allocate device power in real time, compress the power consumption of non-critical equipment to reduce the total load, and trigger a backup calculation model when the load fluctuates abnormally, re-evaluate the power consumption weights of all equipment, and update the allocation plan. After each regulation cycle, a feedback evaluation is carried out on the execution result, including whether the operation of critical equipment is continuous and stable, whether the energy-saving degree of non-critical equipment meets the standard, etc., and the resource allocation efficiency index is output. At the same time, the system periodically re-models the user preference data to adapt to behavior changes, realizes continuous self-learning and parameter tuning, and ensures that the system is always in the optimal response state.

[0020] The present invention can implement customized energy-saving regulation based on individual behaviors and the real-time power grid environment, prioritize the guarantee of critical electrical equipment in the case of power shortage or frequent fluctuations, and effectively reduce the overall electricity cost. Through a continuously optimized adaptive control algorithm, the problem of lagging response of traditional fixed-value control can be avoided, and the response speed and control accuracy can be improved. At the same time, the present invention has strong scalability and can adapt to different scales and types of electricity usage environments, and has practical value and economic potential in the fields of smart home, industrial and commercial energy conservation, and urban distribution network optimization.

[0021] Taking a smart home system as an example, the household contains typical electrical appliances such as air conditioners, refrigerators, water heaters, TVs, washing machines, etc. The system obtains the real-time power consumption data of each device through the ZigBee module connected by the home router, and records behaviors such as the user using the water heater at 7 am every day, turning on the air conditioner at 6 pm, and using the washing machine at 9 pm. During the peak summer power consumption period, when the power grid issues a load warning, the system automatically determines that the water heater and the air conditioner are key devices, while the TV and the washing machine are controllable devices. Based on the preference model and the adaptive control strategy, the system reduces the operating power of the washing machine and the TV, and allocates more power to the air conditioner and the water heater. During the continuous operation for one week, the system automatically updates the preference model according to the user's habit of gradually delaying the use of the washing machine, and improves the controllability level of the washing machine, thereby realizing dynamic energy-saving regulation. The test data shows that this solution saves about 18% of the power consumption on the premise of ensuring comfort, fully verifying its energy-saving effect and application feasibility.

[0022] S1 includes obtaining the operation data and historical records of electrical devices, obtaining a structured data set through database query, processing the structured data set through time series analysis, extracting the power consumption time and usage frequency features, using a clustering algorithm to group according to the usage frequency and power change, obtaining the classification result of the power change, calculating the relationship between the power change and the power consumption time according to the classification result, determining the habit diversity feature, extracting the individual preference from the habit diversity feature through regression analysis, obtaining the user preference distribution, obtaining the user preference distribution and analyzing the feature, using a preset threshold to judge the accuracy of the preference model, and obtaining the final device usage preference model by comparing the updated preference model with the data extraction result.

[0023] In step S1 described in this embodiment, the operation data and historical usage records of the user's electrical equipment are obtained through the Internet of Things terminal. The key lies in constructing a user equipment usage preference model for energy efficiency control. The specific process is as follows. First, the system collects the operation information of each device through sensor nodes, such as device number, current, voltage, power value, startup time, continuous usage time, etc., and uploads it to the central database. Through database query and retrieval operations, the original data is extracted into a structured data set with a unified format. The data set is indexed by timestamp and organized by device type, providing a clear data basis for subsequent analysis. After obtaining the structured data, the system performs time series analysis to identify the usage patterns of the devices within a day, a week, or a month cycle. Taking the startup time of each device as a reference, the daily startup period is statistically analyzed and defined as the "power consumption time characteristic"; at the same time, the number of startups of the device within a unit time is statistically analyzed and defined as the "usage frequency characteristic". For example, if a device starts up once at 7:00, 12:00, and 18:00 every day, its daily usage frequency is 3 times. Next, the system uses a clustering algorithm to group the devices with the usage frequency and average power change of each device as input features. For example, the K-means algorithm is used to classify the devices with large power changes but low frequencies into the "high-power low-frequency category", and the devices with small power changes but high usage frequencies into the "low-power high-frequency category". The result of this classification reflects the behavioral characteristics of different devices and provides a classification basis for subsequent modeling. Then, based on the device classification results, the system analyzes the corresponding relationship between the power change and the power consumption time. For example, the average power change values of different device categories are statistically analyzed between 7:00 and 9:00 every day, and the power consumption behavior curves for different periods of the day are drawn accordingly. By analyzing the fluctuation degree of such distribution curves, the system extracts the user's "habit diversity characteristic", that is, whether the user's behavior is regular and concentrated or scattered and diverse. The smaller the fluctuation, the more concentrated the habit, and the larger the fluctuation, the more diverse the behavior. Based on the above habit characteristics, the system further conducts regression analysis to fit the "habit diversity" with the changes in the user's individual power consumption process (such as daily active duration, power fluctuation range). The fitting result forms a "personalized preference parameter set", which reflects the individual preference tendency of the user in the power consumption behavior, such as whether the user prefers to use high-power devices at night, whether the user prefers to use a few key devices, etc. Finally, the system compares the currently extracted user preference characteristics with the existing preference model. If the difference between the two is within the preset threshold, for example, the average absolute difference in parameter changes is less than a certain fixed value (such as 0.1), it is determined that the preference model is still valid; if it exceeds the threshold, the system updates the preference model until the model output is stable and highly consistent with the actual behavior. This process ensures the accuracy and adaptability of the preference model, forms a complete device usage preference model, and supports subsequent dynamic power allocation and energy-saving control strategies.

[0024] This embodiment realizes high-precision user behavior modeling at the data processing level, constructs a scientific and iteratively optimizable preference extraction process, and improves the accuracy and dynamic adaptation ability of the preference model. Compared with traditional rule-of-thumb modeling, this method can adapt to diverse household or enterprise electricity consumption scenarios, fully explore personalized features in behaviors, and facilitate the precise deployment of subsequent energy consumption control strategies. At the same time, by introducing a model evaluation and update mechanism, the self-evolution ability of the model is realized to ensure that it always represents the latest user behavior characteristics.

[0025] Taking a commercial office building as an example, the office building is equipped with devices such as a lighting system, air conditioners, printers, electric water dispensers, and charging stations. The system records the operating power through the Internet of Things intelligent sockets connected to each device and uploads it to the data center daily. In the data of the past month, the lighting system is enabled 3 times a day, with an average duration of 8 hours each time, starting at 8 am, and the power change is stable; while the printer has a high enabling frequency but a drastic power change, and the usage time is scattered. The system uses K-means clustering to classify the lighting as "continuous and stable type", and the printer as "high-frequency fluctuation type". Further, through the time-power relationship fitting, it is found that the main usage of this office site is concentrated in the time period from 8:00 to 18:00, and the habit diversity index Dh = 0.12 is calculated, indicating that the behavior pattern is relatively concentrated. The personalized preference parameters extracted by regression analysis show high usage stability (preferences are concentrated on fixed devices). With the support of this model, subsequent operations of ensuring the lighting and air conditioners during power shortages are realized, thus achieving precise scheduling and energy-saving control.

[0026] S2 includes extracting operation rules and importance indicators through the preference model, obtaining the grid load by combining real-time data, getting the device operation status parameters, using a multi-variable trade-off method to process the operation rules and importance indicators. If the grid load exceeds the preset threshold, the calculation method is adjusted according to the load change to determine the device priority parameters, extracting the load change trend from the real-time data, analyzing the device parameters and importance indicators through the calculation method to judge the priority adjustment direction, generating a dynamic sequence for the priority adjustment result, obtaining the updated device sorting of the sequence. If the load change trend continues to rise, recalculate the device parameters through multi-variable trade-off to get the adjusted dynamic sequence, verifying the sequence update consistency by combining the real-time data according to the dynamic sequence update result, determining the final device priority sequence, and optimizing the dynamic sequence through the random forest algorithm to obtain a high-precision priority sequence.

[0027] In this embodiment, step S2 is mainly used to realize the dynamic judgment and optimized sorting of device priorities, ensuring that the normal operation of key devices can be reasonably guaranteed under the background of grid load changes, while taking into account the overall energy efficiency control goal. The implementation process is as follows: The system first extracts the operation rules and importance indicators of each device from the established user preference model. The operation rules refer to the average activation frequency, operation time distribution, power fluctuation characteristics, etc. of the device within a certain time period, which are used to describe the behavior pattern of the device in daily use. The importance indicator represents the importance degree of the device in meeting user needs. For example, devices related to safety, health, or basic services are given higher weights. At the same time, the system receives real-time power grid load data, such as parameters like total power consumption value, voltage fluctuation amplitude, power grid frequency, etc. By comparing with the preset safety load threshold, it is judged whether the current is in the peak load period. If the system detects that the current load value has exceeded this threshold, it will automatically start the priority evaluation program. In the priority evaluation, the system uses a weighted multi-variable algorithm to comprehensively process the operation rules and importance indicators of the devices. Each device will be assigned an initial priority score, which is calculated by weighting two parts: the first part is the weight value based on the device operation rules, reflecting its activity level in the current time period; the second part is the score based on the device importance, reflecting its key role in system stability or user experience. In the specific calculation, the system pre-sets two weight parameters respectively to balance the influence of "operation rules" and "importance indicators". For example, the weight of operation rules is set to 0.4, and the weight of importance indicators is set to 0.6. Then multiply these two values by their respective scores and sum them to obtain the comprehensive priority score of the device. Suppose a device has an operation rule score of 70 points and an importance indicator of 90 points, then the final priority score is: (0.4×70)+(0.6×90)=82 points. Sort the priority scores of all devices to obtain the current priority sequence. On this basis, the system also extracts the power grid load change trend from the real-time data, for example, by analyzing the load value changes in the past few minutes to judge whether it is in a continuous rising state. If it is found that the load is continuously rising, the system will recalculate the priority parameters for all devices and generate a new priority sequence. This sequence will be compared with the previous round sequence. If the sorting of the top few devices changes little, it is considered that the sequence update consistency is good and the sequence can be adopted; if the change is too large, further correction is required. Finally, in order to improve the accuracy and robustness of the sequence, the system inputs the constructed priority sequence into the random forest algorithm model. The model uses the historical usage data of the device, the frequency of power restriction, manual intervention records, etc. as feature variables to train and generate an optimized priority result. The model has good non-linear processing ability and can automatically adjust the feature weights, so as to output a more accurate and stable priority sorting result.

[0028] Through this embodiment, the system can quickly and reasonably adjust the priority sequence of electrical equipment when the power grid load changes, achieving a balance between ensuring critical equipment and overall energy-saving control. The multi-variable trade-off model enhances the objectivity and flexibility of priority assessment, avoiding the influence of a single factor on the control decision. The introduction of the random forest algorithm further improves the model's ability to automatically learn and optimize historical behaviors and feature weights, achieving high-precision and high-reliability equipment scheduling and ranking. The system can handle a rapidly changing load environment and take into account the personalized needs of users.

[0029] Taking a smart office park as an example, the park is equipped with a conference room lighting system, a central air-conditioning system, elevators, printing terminals, advertising screens, etc. First, the system determines that the central air-conditioning and elevators are high-importance equipment based on historical usage data, and their preference models show high usage frequencies at 8:00 am and 6:00 pm on weekdays. The power grid load threshold is set at 90%. When the real-time load reaches 93%, the system starts dynamic priority adjustment. Through the weight parameters α = 0.4 (emphasizing operation frequency) and β = 0.6 (emphasizing equipment function guarantee), it is calculated that the air-conditioning and elevators have high priorities, while the advertising screens and printers have low priorities. Under the continuous upward trend of the load, the system adjusts the priority parameters again and uses the random forest model to optimize the ranking, finally turning off the power of the advertising screens and some printers to relieve the load. The response time of the whole process is less than 2 seconds, and the user experience of high-priority equipment is not significantly affected, verifying the effectiveness of this dynamic priority control mechanism.

[0030] S3 includes obtaining power grid load data through real-time monitoring, calculating the load value at the current moment, obtaining the available power range under resource constraints, extracting the change trend from the load data, using a preset threshold to judge equipment dynamics, determining the adjustment direction of the priority sequence, for the equipment dynamics in the priority sequence, obtaining the operating status of important equipment, judging the specific range of guarantee requirements, according to the guarantee requirements and the available power range, calculating the power threshold, determining the allocated power value of important equipment, updating the priority sequence through dynamic adjustment, extracting the equipment status from the adjusted sequence, obtaining the new guarantee requirement range, using the support vector machine algorithm to classify the load data and the guarantee requirements, judging whether the resource constraints meet the allocated power requirements, if the resource constraints meet the requirements, then updating the equipment operating status through the allocated power value, and obtaining the adjusted power grid load data.

[0031] Step S3 described in this embodiment is aimed at the problem of dynamic configuration of power resources. Through real-time load analysis of the power grid and key equipment protection strategy, dynamic allocation control of power resources is realized. This process is based on the power grid operation data continuously monitored by the system, combined with the equipment operation status and protection requirements, and gradually optimizes and makes strategic decisions. First, the system collects the power grid load value at the current moment from the monitoring node, including key parameters such as total load power, voltage fluctuation amplitude, and frequency change. The upper limit of the power that can be allocated by the system under the current power grid state is determined by the calculation formula "current available power = set maximum power supply capacity - current total load". For example, if the maximum power supply capacity is 100kW and the current total load is 82kW, the available power is 18kW. Then, the system analyzes the change trend of the load data. By calculating the load change value per minute in the past 5 minutes, the growth rate is obtained. For example, if the load rises from 80kW to 85kW within 5 minutes, the average growth rate is 1kW / minute. If the rate exceeds the set threshold (such as 0.8kW / minute), it is judged that the system is in a load increase state and the current equipment priority sequence needs to be adjusted. At this time, the system focuses on the part marked as "important equipment" in the current priority sequence and reads its real-time operating status, including whether it is running, current power demand, operating duration, etc. Based on this, the minimum guarantee demand of each key equipment is judged. For example, a medical oxygen supply device needs to guarantee 3kW, and a data center server needs to guarantee a minimum of 8kW. The system adds up the guarantee demand of all important equipment as the minimum guaranteed power demand value. Subsequently, the system compares the minimum guarantee demand with the current available power to calculate whether the demand is met. If it is met, power is allocated to key equipment one by one according to the importance, and the equipment operating status is dynamically updated. If it is not met, the priority sequence is readjusted, the power request of some controllable equipment is reduced, and the released power is used for key equipment protection. In this process, in order to improve the judgment accuracy, the system introduces the support vector machine (SVM) algorithm. SVM uses the load history change curve, the equipment operation load curve and the guaranteed power demand as feature vectors to construct a distribution model and classify the current state into the category of "satisfactory allocation" or "unable to meet allocation". If the classification result is "satisfactory", the equipment operating status is updated according to the current allocation plan, and the system then obtains the new load data after allocation to complete the dynamic feedback control closed loop.

[0032] This solution uses a dynamic power allocation strategy based on real-time data to ensure the continuous operation of important equipment when power resources are tight, and improve the system's ability to respond to sudden grid pressure. Compared with static control strategies, this method can flexibly respond to load changes, dynamically adjust equipment priorities and redistribute strategies, greatly improving system robustness and energy efficiency. At the same time, the introduction of support vector machines enhances the system's judgment ability, enabling more accurate load classification and strategy adaptation, and has good practicality and scalability in a data-driven environment.

[0033] Taking a certain intelligent elderly care center as an example, the center is equipped with electrical equipment such as intelligent bed monitors, heart rate monitors, air conditioners, TVs, lighting, and water heaters. The system sets the heart rate monitor and the intelligent bed monitor as important equipment and requires all-weather power supply. Real-time monitoring shows that the high temperature weather on that day led to a sharp increase in the electricity consumption of the air conditioner, the grid load reached 96%, and the available power was only 4 kW left. The system judges that the load continues to rise, automatically updates the equipment priority sequence, partially shuts down the air conditioner, and retains the power supply for the heart rate monitor. The system calculates that the total power to be guaranteed for important equipment is 3.5 kW, and the SVM model judges that the guarantee demand is "satisfiable" at the current load level. Finally, the system distributes power according to the guarantee level sequence and updates the equipment status. Subsequent data shows that the grid load drops to 92%, and the heart rate monitoring runs without interruption, achieving the goals of energy-saving control and stable guarantee simultaneously.

[0034] S4 includes obtaining the change trend according to the real-time collected grid load data, determining the dynamic priority sequence by analyzing the change trend, extracting features from the change trend and the priority sequence, adopting an adaptive control strategy to adjust the power distribution ratio to obtain a preliminary distribution plan. For the preliminary distribution plan, obtain the operating status and energy consumption data of non-critical equipment, judge the reducible space, determine the adjusted power distribution ratio by calculating the reduction amplitude of the energy consumption of non-critical equipment, and obtain the intermediate result of the optimized plan. If the intermediate result of the optimized plan exceeds the preset threshold, update the change trend according to the real-time adjusted grid load data, recalculate the distribution ratio, verify the optimized plan using the support vector machine algorithm according to the updated distribution ratio and the energy consumption reduction amplitude, obtain the final power distribution plan, and judge the implementation effect of the plan by comparing the energy consumption data of the final power distribution plan and the initial plan to obtain the actual reduction amplitude of non-critical equipment.

[0035] Step S4 described in this embodiment aims to optimize and regulate the energy consumption of non-critical devices through an adaptive control strategy, thereby improving the overall power consumption efficiency and ensuring the stable operation of the power grid. First, the system collects real-time power grid load data, including the total load value and the load change amount per unit time. By performing trend analysis on this data, such as using the data of the past 5 minutes or 15 minutes for moving window averaging, the system can identify whether the current load is in an increasing, decreasing, or stable state. Based on this trend judgment, combined with the device dynamic priority sequence generated in the previous steps, features such as device response delay, energy consumption level, and control sensitivity are extracted to support the design of subsequent allocation schemes. Then, based on the current power resource limitations, the system uses an adaptive allocation model to perform preliminary power allocation, and the system sets an initial allocation ratio. For example, key devices (such as medical equipment, elevators, etc.) are given priority, and the remaining power is allocated to non-critical devices. This allocation ratio is the "preliminary plan". Next, the system obtains the real-time operating power and historical energy consumption data of non-critical devices, and combines with the power compression model to judge the energy consumption reduction space of each device under the current operating task. For example, if an air conditioner currently has an operating power of 2 kW, but only 1.2 kW is required to maintain basic operation, then its reduction amplitude can be 0.8 kW. Summing up the reducible amplitudes of all non-critical devices, the system readjusts the allocation ratio and redistributes the freed-up power resources to guarantee-type or more efficient devices. This is the "intermediate result of the optimized plan". If the reduction power of any device in this optimized result exceeds the preset threshold (for example, 20%), it is considered that this plan may have an unstable impact on the power grid load, and the system will re-collect the current load data and re-perform trend identification and allocation strategy calculation. Based on the new allocation ratio and the reducible energy consumption data, the system uses the support vector machine (SVM) algorithm to verify the feasibility of the optimized plan. This algorithm uses the optimized power value of each device, the total power grid load, device type, device response sensitivity, etc. as input variables to classify and judge whether the current power allocation is stable and reasonable, and outputs two results: "feasible" or "needs correction". If the classification result is "feasible", then this optimized plan is executed as the final power allocation strategy. The system records the energy consumption difference of non-critical devices before and after execution, that is, the "actual reduction amplitude", and compares it with the estimated value to judge the actual energy-saving effect of the plan execution for subsequent strategy learning and model optimization.

[0036] This embodiment realizes the intelligent energy consumption compression of non-critical devices through a closed-loop optimization mechanism, effectively releasing resources for key load guarantee, and improving the energy efficiency utilization rate on the premise of ensuring operation safety. The adaptive control combined with the SVM classification mechanism provides a fast and accurate response path, making the power allocation still have stability and forward-looking in dynamic changes. The monitoring and feedback of the energy consumption amplitude improve the evaluability and traceability of the strategy, providing practical technical support for the construction of smart grids and energy-saving systems.

[0037] Taking the experimental building of a certain university as an example, this building is equipped with central air conditioning, a projection system, a humidifier, an electric kettle, and a teaching lighting system. During a certain high-temperature period in summer, the system detected a sharp increase in the power grid load. The system identified the load trend as continuous increase, and took lighting and air conditioning as guarantee equipment, initially allocating 5 kW to the air conditioning and 3 kW to the lighting, and the rest to other equipment. The system analysis found that the electric kettle and the projection system were in a non-critical state, and the current total power was 3.2 kW, and it was estimated that it could be reduced to 1.8 kW under the minimum operation. So the system readjusted the allocation ratio, reduced the power of non-critical equipment to 2 kW, and redistributed the saved 1.2 kW to the lighting guarantee equipment. The SVM model judged that the new plan was stable and reliable, and finally implemented this optimization strategy, actually reducing the energy consumption by about 1.1 kW. The system verified that the adjustment and implementation were successful, the energy-saving effect was obvious, and at the same time, the lighting and air conditioning were maintained to run continuously, ensuring the normal progress of teaching tasks.

[0038] S5 includes obtaining the actual power supply parameters of each device from the optimized power distribution plan, determining the device operation status, processing the actual power supply parameters through real-time response technology, judging whether the response meets the speed requirement, obtaining the power grid load data, judging whether the load fluctuation exceeds the preset threshold. If the load fluctuation exceeds the preset threshold, then calculate the adjustment parameters through a multivariable trade-off model to obtain a preliminary adjustment plan, analyze the device parameters for the preliminary adjustment plan to determine the actual power supply status after adjustment, process the actual power supply status after adjustment through an optimization algorithm to obtain the final allocation result, and update the power distribution plan according to the final allocation result to complete the power grid load balance.

[0039] The S5 step described in this embodiment is mainly used to achieve dynamic adjustment during the execution of the optimized allocation scheme. By detecting the power supply status of devices, analyzing the grid load fluctuations, and calculating the adjustment parameters, the power distribution is finally corrected to ensure that the grid load is within a stable range. First, the system extracts the power supply parameters of each electrical device from the generated optimized power distribution scheme, including voltage value, current value, the current power supply of the device (i.e., the product of voltage and current), and the response time. Based on these parameters, it is determined whether the device is operating normally and whether its adjustment response is timely. The system sets a threshold for the response speed, such as five seconds, as the evaluation criterion. If the actual response time of a certain device exceeds this threshold, the system will determine that the device fails to execute the power adjustment in a timely manner and needs to enter the adjustment and optimization process. Subsequently, the system detects the total grid load value at the current moment and compares it with the load value at the previous moment to obtain the difference between the two. If the absolute value of this difference exceeds the pre-set load fluctuation threshold (such as three kilowatts), it is considered that the current grid fluctuates greatly and the system needs to re-allocate power resources. At this time, the system uses a multi-variable trade-off model to determine which devices should be adjusted first. For each device, the system comprehensively considers its response ability, importance, and power change trend, and assigns corresponding weight scores. This score is calculated from three factors: the first is the device response time score (the faster the response, the higher the score), the second is the device priority score (for example, medical devices have a higher score, and lighting devices have a lower score), and the third is the rate score of the device power change. The system will set different weight ratios for these three factors, for example, the response ability accounts for 40%, the priority accounts for 40%, and the power change rate accounts for 20%, and sum them up weighted to obtain a comprehensive score. Devices with higher scores are more likely to be adjusted to optimize the overall power configuration. Based on the above scoring results, the system formulates a preliminary adjustment plan, that is, to increase or decrease the power supply of some devices. For example, if the original power supply of a certain device is one kilowatt and 200 watts are cut according to the adjustment result, its new power supply is 800 watts. The new power supply of all devices constitutes a new power supply state. For this adjusted state, the system applies optimization algorithms, such as genetic algorithms, simulated annealing algorithms, or particle swarm optimization, etc., to search for the optimal solution from multiple possible power distribution combinations. Its optimization goals usually include two aspects: one is to reduce the total sum of the differences before and after the adjustment of all devices, and the other is to ensure that the total new power of all devices is as close as possible to the current grid load requirements. The system sets a balance factor for these two goals to achieve the overall optimum. The finally obtained optimal power supply combination is the final result of this adjustment. The system updates the power distribution strategy accordingly and issues a new scheduling instruction to make each device enter a new power supply state, so as to achieve the dynamic balance of the grid load.

[0040] This embodiment has a very strong adaptive regulation ability, can automatically identify risks when the power grid load changes rapidly, and quickly adjust the power distribution strategy to ensure the stable operation of the system. The response speed control mechanism improves the protection ability for critical mission equipment, and the introduction of the optimization algorithm enables the system to obtain better regulation effects in a multi-objective and multi-constraint environment. In addition, this solution has good scalability and can be widely applied to scenarios such as smart home, industrial automation, and smart city energy management.

[0041] Taking a large data center as an example, its main electrical equipment includes servers, cooling systems, uninterruptible power supplies (UPS), lighting, and auxiliary office equipment. During the peak operation period, the system monitors that the power grid load fluctuation reaches 5.2 kW, exceeding the preset threshold of 3.5 kW, triggering the dynamic power adjustment mechanism. The system first evaluates the response speed of the servers and the cooling system and finds that the response time of the cooling system is 3 seconds, meeting the 5-second threshold, while the lighting system responds slowly. After evaluation by the multi-variable trade-off model, it is decided to reduce the power of the lighting system by 1.2 kW and slightly increase the power of the cooling system to ensure. Subsequently, the genetic algorithm is used to optimize the overall power configuration, and finally the power grid load drops back to the safe range (less than 80 kW), the servers remain stable, the lighting is dimmed moderately but still meets the basic lighting requirements, the system balance and service quality are taken into account, and the dual effects of energy conservation and stability are achieved.

[0042] S6 includes obtaining the real-time data of the electrical equipment, collecting the operating status through sensors, determining the current load conditions of each device, extracting the data of critical and non-critical devices from the operating status, using the dynamic power distribution algorithm to calculate the power distribution plan, adjusting the distribution result through the power distribution plan to obtain the stability value of the critical device and the energy consumption reduction of the non-critical device. If the stability value is lower than the preset threshold, the power distribution is adjusted through priority sorting to determine the new distribution result. Calculate the resource efficiency according to the distribution result to obtain the preliminary value of the efficiency index, perform iterative optimization on the efficiency index, use the linear regression algorithm to analyze the relationship between the real-time data and the distribution result, judge the optimized efficiency index, and update the power distribution plan through the optimized efficiency index to determine the final device guarantee status.

[0043] The step S6 described in this embodiment is mainly used to implement the performance evaluation and intelligent optimization of the device power distribution scheme, ensure the stability of the operation of key devices, and improve the power consumption efficiency of the entire system. This process combines multiple links such as device status collection, dynamic power distribution, stability calculation, efficiency modeling, and regression analysis to form a closed-loop feedback optimization mechanism. First, the system collects the operating status of each power-consuming device through sensors deployed on them to obtain real-time power data. Based on this, the system calculates the current load value of each device, that is, the ratio of the current power of the device to its rated power. This value is used to determine whether the device is operating at a high load or a low load state. Devices are classified into key devices and non-key devices according to their functions. For key devices, the system focuses on monitoring the fluctuation range of their power output, which is defined as the "stability value". This value can be calculated by the following formula: Stability value of key device = Standard deviation of device operating power ÷ Rated power of device. Among them, the standard deviation of the operating power is the degree of fluctuation of the power sampling values of the device within a certain time window, which is used to measure whether its operation is stable. If the stability value is lower than the set stability threshold (for example, ten percent), it is considered that the device is operating unstably, and the system needs to re-adjust the power distribution scheme. After the system performs the preliminary power distribution, the energy consumption reduction of non-key devices is calculated. It is defined as: Energy consumption reduction of non-key device = Original power supply power - Current power supply power. The reduction amounts of multiple non-key devices are summed up to evaluate the current power transfer space. Based on the stability value of key devices and the energy consumption reduction of non-key devices, the system constructs a resource efficiency index, and the preliminary calculation method is as follows: Resource efficiency index = Sum of stable powers of key devices ÷ Total system power supply power. Among them, the stable power is the sum of the power supply powers of key devices that meet the stability conditions, and the total system power supply power is the sum of the current powers of all devices. If the resource efficiency index is lower than the expected value (for example, less than seventy-five percent), the system will enter the optimization stage. At this time, the system uses the linear regression algorithm to analyze the relationship between multiple input features and the efficiency index. The main input features include: device power change range, operating duration, response speed, device type, etc. The linear regression model is expressed as: Efficiency index = Coefficient one × Power change range + Coefficient two × Operating duration + Coefficient three × Response time +... + Constant term. The regression coefficients of each input variable are fitted according to historical data, and the current real-time data is used to substitute into the model to calculate the optimized predicted efficiency index. If the predicted value is better than the current efficiency index (for example, increased by more than five percentage points), the system updates the power distribution result using the adjusted scheme after regression analysis, and finally generates a new device guarantee status.

[0044] In this embodiment, by introducing a dual-index system of stability and energy consumption, the evaluation and iteration of the power distribution strategy are realized, and the capabilities of accurate identification, adjustment, and verification are possessed. The combination of dynamic power distribution and regression analysis significantly improves the scientificity and response ability of the control strategy. Compared with the traditional static threshold control, this method can adapt to complex and changeable electricity consumption scenarios, and while ensuring the continuity of critical loads, the overall energy-saving goal is achieved.

[0045] In an intelligent shopping mall, the system deploys equipment such as central air conditioners, elevators, advertising screens, lighting systems, and background music. Air conditioners and elevators are defined as critical equipment, and the rest are non-critical equipment. During the peak customer period at noon on a certain day, the system detects that the grid load is approaching the upper limit. Through real-time sensors, the system identifies that the current load of the air conditioner is 80% and that of the advertising screen is 40%. The system transfers some power from the advertising screen to the air conditioner according to the power distribution algorithm to improve its operation guarantee. At the same time, it is recorded that the power of the advertising screen drops by 300 watts, and the total system energy consumption drops by about 1000 watts. Subsequently, the system calculates that the stability index of critical equipment is 95% (i.e., the operating power fluctuation is extremely small), and the reduction ratio of non-critical equipment is 47%. The initial efficiency index is 0.79. After the system adopts linear regression optimization, the index is improved to 0.85, and the new distribution plan is adopted and implemented, finally achieving the control goal of ensuring the comfort of the passenger flow while optimizing the overall power consumption efficiency of the system.

[0046] S7 includes collecting grid load and real-time data, using time series analysis to judge the load change trend, obtaining the load fluctuation characteristics, calculating the efficiency index according to the load fluctuation characteristics and resource allocation data, determining the current resource allocation state, adopting an adaptive control strategy to update and adjust the logic for the efficiency index, obtaining the equipment priority sequence, analyzing the diversity change characteristics through the equipment priority sequence and user habit data, judging whether a significant change occurs. If the diversity change characteristics exceed the preset threshold, then re-extract the preference model through the clustering algorithm to obtain the updated preference sequence, adjust the equipment priority logic according to the updated preference sequence to generate a new priority sequence, and verify the adaptability of the adjustment logic through the new priority sequence and real-time data to determine the final priority sequence.

[0047] In this embodiment, by periodically updating the device priority logic, the system's dynamic adaptability to user behavior changes and grid status is ensured. The entire process takes load trend analysis and diversity feature detection as the core, and combines the adaptive update mechanism of the preference model. First, the system collects the grid load value and the real-time operation data of each device, and uses time series analysis methods (such as moving average, exponentially weighted moving average, etc.) to judge the load change trend. On this basis, load fluctuation features are extracted, such as average fluctuation amplitude, peak-valley period, mutation frequency, etc. Then, the system combines the current power resource allocation data and calculates the efficiency index according to the following formula: Efficiency index = Total stable power supply of key devices ÷ Total allocated power. Among them, the total stable power supply of key devices refers to the total power obtained by the key loads that can operate stably and continuously under the current priority strategy. The efficiency index is used to measure the effectiveness of the current resource usage configuration. If the efficiency index is low, the system triggers the adjustment logic optimization process. Through an adaptive control strategy (such as a priority adjustment logic with feedback), the current device priorities are dynamically rearranged to generate a new priority sequence. After that, the system extracts behavior features from the current priority sequence and the user's historical preference model, and compares to find out whether the user's usage pattern has changed. The judgment of the diversity change feature is based on the following formula: Diversity change index = Difference value between the current device preference distribution and the historical model preference distribution. The difference value can be measured by Euclidean distance, Kullback-Leibler divergence, etc. If this index exceeds the set threshold (such as 0.3), it is determined that the user's behavior has changed significantly. The system will trigger a clustering algorithm (such as K-means or DBSCAN), and based on features such as device usage frequency, usage time period, and power demand in the recent period, re-extract a new user preference model and generate an updated preference sequence. Combining the new preference sequence with the current grid status, the system adjusts the priority allocation logic and generates a new priority sequence again. This sequence will be used for power allocation decisions in future regulation cycles. To verify whether the new priority logic is adaptable, the system executes the new strategy in a small range and compares it with the device response data under the original strategy, such as whether the stability of key devices has improved and whether the energy consumption of non-key devices has decreased, and finally confirms the feasibility of this new sequence.

[0048] In this embodiment, by introducing a behavior change detection and preference model reconstruction mechanism, the intelligent and dynamic evolution ability of the device priority control strategy is realized. The system can not only sense the grid pressure, but also adapt to the change of the user's behavior pattern, improve the flexibility of device management and power supply efficiency, especially suitable for commercial or industrial scenarios where user loads change frequently. At the same time, the use of clustering algorithms improves the accuracy of preference modeling and enhances the real-time adjustment ability of the system.

[0049] In an energy-saving management system for an office park, the system updates the power distribution strategy every hour. One day at noon, the system found that the grid load fluctuated continuously and frequently, and the efficiency index dropped to 68%. The system immediately invoked the adaptive strategy to adjust the device priority sequence. The system also detected that multiple office users had temporarily changed their electricity consumption habits. For example, printers were used frequently, and air conditioners were turned on intermittently. The difference index of the preference model compared with the previous day reached 0.42. The system automatically invoked the clustering algorithm to generate a new preference model based on the new behavior data and re-order the device priorities. After the adjustment, the efficiency index rose to 84%. The system verified that the new sequence was feasible and successfully adapted to the dual challenges of office behavior changes and load fluctuations.

[0050] S8 includes obtaining device power and distribution requirements through the priority sequence, using data extraction technology to obtain the initial distribution data set, extracting multi-variable trade-off parameters from the initial distribution data set, determining the trade-off coefficient through calculation and analysis, obtaining the speed constraint conditions according to the trade-off coefficient and real-time response requirements, obtaining the adjusted distribution priority, analyzing the resource limitations for the adjusted distribution priority and power resources, determining the available power range, using the linear programming algorithm to obtain a preliminary dynamic power distribution plan based on the available power range and distribution requirements, judging whether the real-time response is satisfied according to the preliminary dynamic power distribution plan and speed constraint. If not, adjust the distribution parameters to obtain the optimized dynamic power distribution result, and determine the final distribution result through the optimized dynamic power distribution result and calculation analysis.

[0051] This embodiment is used to complete the final output link of the entire power distribution strategy. Through priority constraints, multi-variable trade-off analysis, and linear programming modeling, dynamic allocation decisions are realized. Considering the requirements of the device for response timeliness, speed regulation and algorithm optimization are carried out on the initial allocation results to ensure timely system response and reasonable resource allocation. First, the system obtains the power demand and expected power supply of each device according to the determined device priority sequence, and constructs an initial allocation data set. The content of the data set includes information such as device identification, type, power demand value, maximum tolerable power, priority number, etc. From this data set, the system extracts several variable parameters that affect the allocation strategy, such as device response speed, energy consumption level, regulation sensitivity, historical stability, etc. Each variable is assigned a relative weight, and a "multi-variable trade-off coefficient" is calculated comprehensively based on these weights. The trade-off coefficient can be expressed as: trade-off coefficient = response speed weight × response ability + energy consumption weight × power consumption per unit time + stability weight × fluctuation frequency. This trade-off coefficient is used to sort and adjust the device priorities, and combined with the system's requirements for real-time response, speed constraints are set. For example, if a device needs to complete power adjustment within three seconds, but its actual adjustment time is five seconds, it is considered not to meet the speed constraint and its priority needs to be reduced. Next, the system evaluates the available power resources of the power grid at the current moment according to the adjusted priority sequence. For example, the current total available power is 100 kilowatts, and the system takes this value as the "power limit constraint". Under the joint constraints of power limit and device allocation requirements, the system constructs a linear programming model with the goal of maximizing the total power supply of key devices while taking into account response timeliness and resource efficiency. The objective function of the linear programming is described as follows: Objective: Maximize the sum of the actual allocated powers of all devices, and give priority to satisfying the devices with high priority, fast response, and low energy consumption. The constraint conditions include: the allocated power of each device shall not exceed its maximum rated value; the sum of the allocated powers of all devices shall not exceed the total available power; the response time shall meet the real-time control threshold. By solving this linear programming model, the system obtains a "preliminary dynamic power allocation plan". Subsequently, the system verifies whether this plan is feasible under the response limitations of each device. If there are unsatisfied response situations, the parameters of some devices are readjusted, their target powers are reduced or the speed buffer time limit is reset to generate an optimized power allocation plan. Finally, based on the optimized results, the system comprehensively analyzes the allocation effect in combination with the operation data, including the power supply guarantee rate of key devices, the reduction of energy consumption of non-key devices, the overall load stability, etc. If the analysis results meet the expected goals, the result is determined as the final power distribution plan and pushed to the dispatching system for execution.

[0052] This implementation method realizes the optimal allocation of power resources under complex constraints by integrating a multi-factor trade-off model with a linear programming solution method. Meanwhile, considering the real-time response requirements, it effectively guarantees critical loads, reduces non-critical energy consumption, and improves the execution efficiency of the allocation strategy. Introducing response time constraints and a feedback mechanism enhances the practicality and engineering deployability of the system's control strategy, making it particularly suitable for power consumption scenarios with large-scale and concurrent operation of multiple types of devices.

[0053] In the energy management system of an airport, it involves devices such as runway lighting, terminal building air conditioning, charging piles, elevators, and self-service boarding machines. During a certain peak operation stage, the grid load is close to the upper limit, and only 80 kilowatts of available power remains. The system extracts initial allocation data based on the priority and actual demand of the devices. After calculating the trade-off coefficient and adjusting the speed constraint, it is found that the response time of some self-service boarding machines is too slow, and their priorities are adjusted. The system constructs a linear programming model based on the current load status. In the preliminary plan, the power supply for the elevators is insufficient, resulting in response delays. The system automatically adjusts the parameters and optimizes them. Finally, the generated allocation result is: 20 kilowatts for runway lighting, 25 kilowatts for air conditioning, 15 kilowatts for elevators, 10 kilowatts for charging piles, and 10 kilowatts for self-service devices. After execution, the overall energy consumption of the system is reduced by about 12%, the key devices operate stably, and the energy-saving effect is significant.

[0054] As Figure 2 shown, there is also provided an energy-saving control system for power-consuming devices based on the Internet of Things, which is used to implement the steps of the energy-saving control method for power-consuming devices based on the Internet of Things. The system includes: A data acquisition module, which is used to obtain the operation data and historical usage records of user power-consuming devices, extract the analysis features of user habit diversity and personalized preferences, and obtain a device usage preference model; A priority evaluation module, which is used to extract the operation rules of each device from the preference model, judge the device priority adjustment parameters, and obtain a dynamic priority sequence; A power calculation module, which is used to calculate the available power threshold under the current power resource limit according to the real-time grid load monitoring data, and determine the allocated power value required to guarantee important devices; A power regulation module, which is used to adjust the power allocation ratio by adopting an adaptive control strategy, obtain the reduction amplitude of the energy consumption of non-critical devices, and obtain an optimized power allocation plan; A fluctuation response module, which is used to extract the actual power supply parameters of each device, judge whether the grid load fluctuation exceeds a preset threshold. If it exceeds, recalculate the multi-variable trade-off parameters to obtain an adjusted allocation result; An allocation evaluation module, which is used to obtain the adjusted allocation result and the operation status of real-time power-consuming devices, calculate the stability of critical device guarantee and the reduction amount of non-critical device energy consumption, and determine the current resource allocation efficiency index; A model update module, which is used to update the device priority adjustment logic based on an adaptive control strategy, determine whether the diversity of user habits has changed, and if so, re-extract the preference model to obtain a new priority sequence; An optimal control module, which is used to extract the device power allocation requirements, calculate the optimal allocation scheme under the power resource limit, and determine the final dynamic power allocation result.

[0055] The system described in this embodiment works collaboratively through eight modules to build a complete closed-loop mechanism for power consumption energy-saving control. First, the data acquisition module obtains the operation data and historical usage records of the user's electrical equipment, extracts the diversity of user habits and personalized preference features based on time series analysis, frequency statistics, and clustering methods, and constructs a device usage preference model. Subsequently, the priority evaluation module calculates the priority adjustment parameters and generates a dynamic priority sequence based on the operation rules, importance levels, and behavior preference scores of the devices. The power calculation module analyzes the current power resource limit state according to the real-time collected grid load data, calculates the available power threshold, and pre-allocates the guaranteed power for key devices in combination with the priority information. The power regulation module adopts an adaptive control strategy to compress the power of non-critical devices while ensuring key loads, thereby obtaining an optimized power allocation ratio. The fluctuation response module monitors the grid load fluctuation in real time. If it exceeds the set threshold, it recalculates and adjusts the allocation result through a multi-variable trade-off mechanism (based on device response speed, energy consumption level, and priority weight). The allocation evaluation module calculates the power supply stability of key devices and the energy-saving ratio of non-critical devices based on the adjusted operation state data, and generates a resource allocation efficiency index accordingly. If the user's habits change significantly, the model update module will reconstruct the preference model based on real-time data and output a new device priority sequence to correct the control logic. Finally, the optimal control module comprehensively considers the device power requirements, grid power limit, response speed requirements, and priority weight, adopts a linear programming algorithm to generate an optimal dynamic power allocation scheme, and verifies its adaptability in real time to ensure that the system achieves stable and efficient power consumption scheduling control under the background of changing grid pressure and diverse user behaviors.

[0056] This system integrates core technologies such as Internet of Things real-time monitoring, edge computing modeling, and AI optimization scheduling, and has high intelligence and self-adaptability. Through a modular structure, it realizes full-process closed-loop control such as device identification, behavior modeling, scheduling optimization, and fluctuation response, significantly improves energy efficiency, reduces unnecessary energy consumption, and strengthens the operation guarantee ability of key services.

[0057] In a smart office building, the system is deployed on the main power lines and devices such as air conditioners, lighting, elevators, and projection systems on each floor. During the peak load at noon, the system monitors an increase in grid pressure and detects significant fluctuations in the power of the air conditioners. The priority assessment module identifies the air conditioners and elevators as critical devices, and the power regulation module reduces the lighting power in some meeting rooms. After the optimization scheme, the energy consumption of non-critical devices is reduced by 22%. The resource efficiency index transmitted back by the system is increased from 76% to 88%, and the stability of critical loads is maintained above 95%, ensuring reliable scheduling of the system on the basis of energy conservation.

[0058] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An energy-saving control method for electrical equipment based on the Internet of Things, characterized in that, Including: S1. Obtain the operation data and historical usage records of the user's electrical equipment, extract the analysis features of user habit diversity and personalized preferences, and obtain the equipment usage preference model; S2. Extract the operation rules of each equipment from the preference model, judge the equipment priority adjustment parameters, and obtain the dynamic priority sequence; S3. Calculate the available power threshold under the current power resource limit through the real-time power grid load monitoring data, and determine the allocated power value required for ensuring important equipment; S4. Adopt an adaptive control strategy to adjust the power distribution ratio, obtain the reduction range of the energy consumption of non-critical equipment, and obtain the optimized power distribution scheme; S5. Extract the actual power supply parameters of each equipment, judge whether the power grid load fluctuation exceeds the preset threshold. If it exceeds, recalculate the multi-variable trade-off parameters to obtain the adjusted allocation result; S6. Obtain the adjusted allocation result and the operation status of the real-time electrical equipment, calculate the stability of ensuring critical equipment and the reduction amount of the energy consumption of non-critical equipment, and determine the current resource allocation efficiency index; S7. Adopt an adaptive control strategy to update the equipment priority adjustment logic, judge whether the user habit diversity has changed. If it has changed, re-extract the preference model to obtain a new priority sequence; S8. Extract the equipment power distribution requirements, calculate the optimal distribution scheme under the power resource limit, and determine the final dynamic power distribution result.

2. The energy-saving control method for electrical equipment based on the Internet of Things according to claim 1, characterized in that: The said S1 includes: Obtain the operation data and historical records of the electrical equipment, obtain the structured data set through database query, process the structured data set through time series analysis, extract the power consumption time and usage frequency features, use the clustering algorithm to group according to the usage frequency and power change, obtain the classification result of the power change, calculate the relationship between the power change and the power consumption time according to the classification result, determine the habit diversity feature, extract the individual preference from the habit diversity feature through regression analysis, obtain the user preference distribution, obtain the user preference distribution and analysis features, use the preset threshold to judge the accuracy of the preference model, and obtain the final equipment usage preference model by comparing the updated preference model with the data extraction result.

3. The energy-saving control method for electrical equipment based on the Internet of Things according to claim 1, characterized in that: The said S2 includes: Extract the operation rules and importance indicators through the preference model, combine the real-time data to obtain the power grid load, and obtain the equipment operation status parameters. Use the multi-variable trade-off method to process the operation rules and importance indicators. If the power grid load exceeds the preset threshold, adjust the calculation method according to the load change to determine the equipment priority parameters. Extract the load change trend from the real-time data, analyze the equipment parameters and importance indicators through the calculation method, judge the priority adjustment direction, generate a dynamic sequence for the priority adjustment result, obtain the equipment sorting after the sequence update. If the load change trend continues to rise, recalculate the equipment parameters through the multi-variable trade-off to obtain the adjusted dynamic sequence. According to the dynamic sequence update result, verify the sequence update consistency by combining the real-time data to determine the final equipment priority sequence, and optimize the dynamic sequence through the random forest algorithm to obtain a high-precision priority sequence.

4. A method for energy-saving control of electrical equipment based on the Internet of Things according to claim 1, characterized in that: The said S3 includes: Obtain power grid load data through real-time monitoring, calculate the load value at the current moment, obtain the available power range under resource constraints, extract the change trend from the load data, use a preset threshold to judge the device dynamics, determine the adjustment direction of the priority sequence, for the device dynamics in the priority sequence, obtain the operating status of important devices, judge the specific range of the guarantee requirements, calculate the power threshold according to the guarantee requirements and the available power range, determine the allocated power value of important devices, update the priority sequence through dynamic adjustment, extract the device status from the adjusted sequence, obtain the new guarantee requirement range, use the support vector machine algorithm to classify the load data and the guarantee requirements, judge whether the resource constraints meet the allocated power requirements, if the resource constraints meet the requirements, update the device operating status through the allocated power value, and obtain the adjusted power grid load data.

5. The energy-saving control method for electrical equipment based on the Internet of Things according to claim 1, wherein: The S4 includes: According to the power grid load data collected in real time, obtain its change trend, determine the dynamic priority sequence by analyzing the change trend, extract features from the change trend and the priority sequence, use an adaptive control strategy to adjust the power distribution ratio, obtain a preliminary distribution plan, for the preliminary distribution plan, obtain the operating status and energy consumption data of non-critical devices, judge the reducible space, determine the adjusted power distribution ratio by calculating the reduction amplitude of the energy consumption of non-critical devices, obtain the intermediate result of the optimized plan, if the intermediate result of the optimized plan exceeds the preset threshold, update the change trend according to the real-time adjusted power grid load data, recalculate the distribution ratio, verify the optimized plan using the support vector machine algorithm according to the updated distribution ratio and the energy consumption reduction amplitude, obtain the final power distribution plan, and judge the execution effect of the plan by comparing the energy consumption data of the final power distribution plan with the initial plan.

6. The energy-saving control method for electrical equipment based on the Internet of Things according to claim 1, wherein: The S5 includes: Obtain the actual power supply parameters of each device from the optimized power distribution plan, determine the device operating status, process the actual power supply parameters through real-time response technology, judge whether the response meets the speed requirements, obtain the power grid load data, judge whether the load fluctuation exceeds the preset threshold, if the load fluctuation exceeds the preset threshold, calculate the adjustment parameters through a multivariable trade-off model, obtain a preliminary adjustment plan, analyze the device parameters for the preliminary adjustment plan, determine the adjusted actual power supply status, process the adjusted actual power supply status through an optimization algorithm, obtain the final allocation result, update the power distribution plan according to the final allocation result, and complete the power grid load balance.

7. A method for energy-saving control of electrical equipment based on the Internet of Things according to claim 1, characterized in that: The S6 includes: Obtain the real-time data of the electrical equipment, collect the operating status through sensors, determine the current load conditions of each device, extract the data of key devices and non-key devices from the operating status, adopt a dynamic power distribution algorithm, calculate the power distribution plan, adjust the distribution result through the power distribution plan, obtain the stability value of key devices and the energy consumption reduction of non-key devices. If the stability value is lower than the preset threshold, adjust the power distribution through priority sorting to determine the new distribution result. Calculate the resource efficiency according to the distribution result to obtain the preliminary value of the efficiency index. Conduct iterative optimization for the efficiency index, adopt a linear regression algorithm to analyze the relationship between the real-time data and the distribution result, judge the optimized efficiency index, and update the power distribution plan through the optimized efficiency index to determine the final device guarantee status.

8. The energy-saving control method for electrical equipment based on the Internet of Things according to claim 1, characterized in that: The S7 includes: By collecting the grid load and real-time data, using time series analysis to judge the load change trend, obtaining the load fluctuation characteristics, calculating the efficiency index according to the load fluctuation characteristics and resource allocation data, determining the current resource allocation status, adopting an adaptive control strategy to update the adjustment logic for the efficiency index to obtain the device priority sequence, analyzing the diversity change characteristics through the device priority sequence and user habit data, and judging whether a significant change has occurred. If the diversity change characteristics exceed the preset threshold, re-extract the preference model through a clustering algorithm to obtain the updated preference sequence, adjust the device priority logic according to the updated preference sequence to generate a new priority sequence, and verify the adaptability of the adjustment logic through the new priority sequence and real-time data to determine the final priority sequence.

9. The energy-saving control method for electrical equipment based on the Internet of Things according to claim 1, wherein: The S8 includes: Obtain the device power and allocation requirements through the priority sequence, use data extraction technology to obtain the initial allocation data set, extract the multi-variable trade-off parameters from the initial allocation data set, determine the trade-off coefficient through calculation and analysis, obtain the speed constraint conditions according to the trade-off coefficient and real-time response requirements, and get the adjusted allocation priority. Analyze the resource limitations for the adjusted allocation priority and power resources to determine the available power range. Obtain the preliminary dynamic power distribution plan through the available power range and allocation requirements using a linear programming algorithm. Judge whether the real-time response is satisfied according to the preliminary dynamic power distribution plan and speed constraint. If not, adjust the allocation parameters to obtain the optimized dynamic power distribution result, and determine the final distribution result through the optimized dynamic power distribution result and calculation analysis.

10. An energy-saving control system for power-consuming devices based on the Internet of Things, which is used to implement the steps of the energy-saving control method for power-consuming devices based on the Internet of Things according to any one of claims 1-9, characterized in that, The system includes: A data collection module, used to obtain the operation data and historical usage records of the user's electrical equipment, extract the user habit diversity and personalized preference analysis features, and obtain the device usage preference model; A priority evaluation module, used to extract the operation rules of each device from the preference model, judge the device priority adjustment parameters, and obtain the dynamic priority sequence; A power calculation module, used to calculate the available power threshold under the current power resource limit according to the real-time grid load monitoring data, and determine the allocated power value required for the guarantee of important devices; The power regulation module is used to adjust the power distribution ratio by adopting an adaptive control strategy, obtain the reduction amplitude of the energy consumption of non-critical devices, and obtain an optimized power distribution scheme; The fluctuation response module is used to extract the actual power supply parameters of each device, judge whether the power grid load fluctuation exceeds a preset threshold. If it exceeds, recalculate the multi-variable trade-off parameters to obtain an adjusted allocation result; The allocation evaluation module is used to obtain the adjusted allocation result and the operating status of real-time electrical equipment, calculate the stability of key equipment guarantee and the reduction of non-critical equipment energy consumption, and determine the current resource allocation efficiency index; The model update module is used to update the device priority adjustment logic based on the adaptive control strategy, judge whether the user habit diversity has changed. If it has changed, re-extract the preference model to obtain a new priority sequence; The optimal control module is used to extract the device power distribution requirements, calculate the optimal distribution scheme under the power resource limit, and determine the final dynamic power distribution result.

Images