Industrial user electrical load classification system and method

Abstract

The invention discloses an industrial user power utilization load classification system and method, and relates to the technical field of power utilization data processing. The industrial user electrical load classification system comprises a main system and a plurality of subsystems, and the plurality of subsystems are correspondingly carried on different industrial users respectively and are gathered in the main system for unified management; through cooperative work of the main system and the subsystems, comprehensive optimization from macroscopic to microscopic is realized. The main system is responsible for real-time acquisition and processing of global data and deep analysis of load characteristics, and provides accurate support for power grid dispatching; and the subsystems go deep into industrial users, finely collect device-level data, and encrypt and transmit the device-level data to the main system, so that the security and integrity of the data are ensured. The layout not only improves the intelligent level of power grid dispatching, but also enhances the refinement degree of internal power consumption management of industrial users.

Description

Technical Field

[0001] The present invention relates to the technical field of power consumption data processing, and in particular to an industrial user power consumption load classification system and method thereof. Background Art

[0002] Electricity load classification refers to the division of electrical equipment in the power system according to different standards and methods, so as to better carry out power system planning, peak management, time-of-use electricity prices, load forecasting and other tasks.

[0003] For example, a method and related device for classifying the electricity load of industrial users disclosed in publication number CN116561661A, for the same load vector, different load category numbers are used to train the FCM algorithm corresponding to the FCM model and generate the classification result corresponding to the load vector, and then the three types of evaluation indicators, namely the contour coefficient, CH index and Davidson-Botting index, are used to jointly evaluate the classification results corresponding to each load category number. The final classification result corresponding to the optimal load category number can be output according to the evaluation, which can ensure the accuracy of effectively classifying the electricity load of industrial users.

[0004] As shown in the above technology, the current electricity load classification system based on industrial users is a system for power plants or power systems to classify, manage and control the electricity loads of different industrial users (i.e. enterprises). This classification system divides and manages the electricity loads of industrial users in the entire power grid from a macro level. However, in actual use, due to factors such as equipment changes within the enterprise, the changes in electricity loads will be unpredictable. The use of a macro electricity load classification system is prone to errors in decision-making, which in turn affects the normal operation of some enterprises. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an industrial user power load classification system and method thereof, which solves the problems existing in the current power load classification system.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for classifying industrial user power loads, specifically comprising the following steps:

[0007] S1. Data collection and processing of the main system: Through the power grid dispatch center interface, smart meter network and subsystem upload interface, the original data of industrial users' electricity consumption is collected in real time and the data is pre-processed;

[0008] S2. Load characteristic analysis and classification: Calculate the load rate and peak-to-valley difference rate, cluster users using K-means, assign labels based on industry characteristics, and introduce user weights;

[0009] S3, Priority setting: Calculate the priority according to the load, peak-valley difference and user weight, sort and generate a list to assist in grid dispatching;

[0010] S4. Regional management and optimization: Divide the power grid into regions, analyze regional loads, use machine learning to optimize cross-regional power allocation, and formulate strategies;

[0011] S5, Subsystem data collection and encrypted transmission: The subsystem is installed on the industrial user and connected to the main system. The device is accessed through IoT, the collection frequency is configured, and data is securely transmitted to the main system in real time or periodically.

[0012] S6. Equipment load classification and label management: Analyze equipment load characteristics, calculate energy efficiency ratio, classify equipment and set labels, and manage them in combination with geographic location information;

[0013] S7. Load strategy formulation and optimization: Develop load control strategies based on classification and real-time data, and use intelligent algorithms to optimize operation plans.

[0014] Preferably, step S1 specifically includes:

[0015] S1.1 Data collection: Through the grid dispatch center interface, smart meter network and subsystem upload interface, the original data of industrial users’ electricity consumption E, voltage V, current I and power factor PF are collected in real time;

[0016] S1.2 Data storage: Distributed database technology is used to store data by timestamp and geographic location index;

[0017] S1.3 Data preprocessing: Clean the data, remove outliers, fill in missing values, and normalize them.

[0018] Preferably, step S2 specifically includes:

[0019] S2.1 Load characteristics analysis:

[0020] Load factor calculation: Where i represents the user number; T represents the timestamp; LF i Refers to the load rate of user i, which means the ratio of the maximum load of the user during the observation time to the total load; E i (T) refers to the electric load value of user i at timestamp T; max T Ei(T) refers to the maximum electric load value of user i in all observation timestamps T; ∑ T Ei(T) refers to the total electric load value of user i in all observation timestamps T;

[0021] Peak-to-valley rate calculation: Among them, min TEi(T) refers to the minimum electric load value of user i in all observation timestamps T;

[0022] S2.2 Cluster analysis: Using the K-means algorithm, users are divided into K categories (C1, C2, ..., C K );

[0023] S2.3 Classification criteria and labels: Assign labels to each category based on clustering results and industry characteristics;

[0024] Introducing user importance weight WF i , calculated based on user scale and economic contribution factors.

[0025] Preferably, step S3 specifically includes:

[0026] S3.1 Priority calculation: P i =α·LF i i+β·PVR i +γ·WF i , where α, β, and γ are weight coefficients, which are adjusted according to the grid strategy;

[0027] S3.2 Priority list: According to P i Sorting and generating priority lists for grid dispatching and load control.

[0028] Preferably, step S4 specifically includes:

[0029] S4.1 Regional division: According to the geographical location information, the power grid is divided into multiple regions and a set R = (R1, R2, ..., R N );

[0030] S4.2 Regional load analysis: Calculate the average load and peak-to-valley difference indicators of each region;

[0031] S4.3 Intelligent regional optimization: Use machine learning algorithms to analyze load complementarity and transmission efficiency between regions and formulate cross-regional power allocation strategies;

[0032] Strategy formulation: Based on the results of regional load analysis, formulate cross-regional power allocation plans or regional power restriction strategies.

[0033] Preferably, step S6 specifically includes:

[0034] S6.1 Equipment load characteristics analysis:

[0035] Time series modeling: Apply LSTM algorithm to the equipment load data E device (T) Modeling and identification of load patterns;

[0036] Calculate the energy efficiency ratio (EER) of the equipment device : Where P output is the device output power, P input Input power to the device;

[0037] S6.2 Equipment classification and labeling: Classify equipment into different categories and set labels based on load characteristics and energy efficiency ratio, and bind equipment labels to geographic locations;

[0038] S6.3 Dynamic classification adjustment: Monitor equipment load data in real time. When a significant change in equipment load pattern or energy efficiency ratio is detected, reclassify the equipment and update the label and geographic location binding.

[0039] Preferably, step S6.3 of dynamic classification adjustment specifically includes:

[0040] S6.3.1 Define change thresholds:

[0041] Load pattern change threshold Δ pattern : Used to quantify the degree of change in the equipment load pattern, calculated by comparing the difference between the current load pattern and the historical load pattern:

[0042] Δ pattern =D istance (E device (T current ), E device (T history ));

[0043] Among them, E device (T curren t) is the load data in the current time window, E device (T history ) is the load data in the historical time window, D istance () is the function that calculates the difference between two load patterns;

[0044] Energy efficiency ratio change threshold Δ EER : Used to quantify the degree of change in energy efficiency ratio, calculated by comparing the difference between the current energy efficiency ratio and the historical energy efficiency ratio:

[0045]

[0046] Among them, EER current is the current energy efficiency ratio, EER history is the historical energy efficiency ratio;

[0047] S6.3. Trigger conditions:

[0048] Reclassification is triggered when any of the following conditions are met:

[0049] (1) Load pattern changes significantly: Δ pattern >θ pattern , where θ pattern is the preset load pattern change threshold;

[0050] (2) Energy efficiency ratio changes significantly: Δ EER >θ EER , where θ EER is the preset energy efficiency ratio change threshold;

[0051] S6.3. Reclassification process:

[0052] Data collection: real-time collection of equipment load data and energy efficiency ratio data;

[0053] Change calculation: Calculate the difference between the current load pattern and the historical load pattern Δ pattern And the difference between the current energy efficiency ratio and the historical energy efficiency ratio Δ EER ;

[0054] Trigger judgment: Check whether any trigger condition is met;

[0055] Reclassification: If the trigger condition is met, the clustering algorithm or classification algorithm is reapplied to reclassify the equipment based on the current load pattern and energy efficiency ratio as features;

[0056] Update labels and bindings: Update the classification labels and geographic location bindings of the device based on the reclassification results.

[0057] Preferably, step S7 specifically includes:

[0058] Strategy Development:

[0059] Based on equipment classification and real-time load data, formulate load control strategies, including setting equipment operation time window T window , adjust the equipment power P adjust ;

[0060] Introducing genetic algorithms to optimize equipment operation plans and reduce overall energy consumption;

[0061] Implementation and Feedback:

[0062] Execute load control strategies and monitor the execution results;

[0063] Collect the load data after execution E after , calculate the energy saving rate And fed back to the main system, where E before is the load data before execution.

[0064] The present invention also discloses an industrial user electricity load classification system, including a main system and a plurality of subsystems, wherein the plurality of subsystems are respectively mounted on different industrial users and are aggregated in the main system for unified management.

[0065] Preferably, the main system includes a data acquisition and processing module, a load characteristic analysis and classification module, a priority setting module, a regional management and optimization module and a subsystem management module. The data acquisition and processing module, the load characteristic analysis and classification module, the priority setting module, and the regional management and optimization module sequentially perform the operations of steps S1 to S4. The subsystem management module is used to connect and manage the subsystem;

[0066] The subsystem includes a data acquisition and encryption transmission module, an equipment load classification and label management module, and a load strategy formulation and optimization module. The data acquisition and encryption transmission module, the equipment load classification and label management module, and the load strategy formulation and optimization module perform operations from steps S5 to S7 in sequence, and the data acquisition and encryption transmission module is connected to the subsystem management module for data encryption upload.

[0067] The present invention provides an industrial user power load classification system and method. Compared with the prior art, it has the following beneficial effects:

[0068] 1. The industrial user electricity load classification system and method achieves comprehensive optimization from macro to micro through the collaborative work of the main system and subsystems. The main system is responsible for the real-time collection and processing of global data and the in-depth analysis of load characteristics, providing precise support for power grid dispatching; the subsystem goes deep into the industrial users, collects device-level data in detail, and transmits it to the main system in encrypted form, ensuring the security and integrity of the data. This layout not only improves the intelligence level of power grid dispatching, but also enhances the refinement of power management within industrial users. At the same time, machine learning and intelligent algorithms are used to optimize cross-regional power dispatch and load control strategies, effectively improving the overall efficiency and stability of the power system, and laying a solid foundation for the sustainable development of industrial electricity.

[0069] 2. The industrial user electricity load classification system and method provide a solid foundation for load characteristic analysis and accurate classification by real-time collection and preprocessing of multi-dimensional electricity consumption data. The calculation of load rate and peak-to-valley difference rate and K-means clustering analysis, combined with industry characteristics and user weights, realize the accurate identification and classification of user electricity consumption patterns. Furthermore, the priority list formulated based on the classification results directly assists power grid dispatch and optimizes resource allocation. The regional management and optimization strategy analyzes regional load characteristics through intelligent algorithms and formulates cross-regional deployment plans, which significantly improves the overall operation efficiency and stability of the power grid. This series of innovations not only promotes the rational use of energy, but also enhances the flexibility and resilience of the power grid, which is of great significance to the sustainable development of industrial electricity.

[0070] 3. The industrial user electricity load classification system and method, by introducing the LSTM algorithm to model the time series of equipment load characteristics, and combining energy efficiency ratio analysis, achieves deep insight and accurate classification of the operating status of industrial equipment. The dynamic classification adjustment mechanism monitors and responds to significant changes in equipment status in real time by setting load mode and energy efficiency ratio change thresholds, ensuring the timeliness and accuracy of classification labels. This not only improves the level of refinement of equipment management, but also provides a scientific basis for energy optimization and equipment maintenance. In addition, binding equipment tags to geographic locations further enhances the convenience and pertinence of management. Overall, the S6 step significantly improves the operating efficiency and energy utilization of industrial equipment, and promotes the intelligent transformation and sustainable development of the production process.

[0071] 4. The industrial user electricity load classification system and method realizes refined control of industrial equipment load through intelligent strategy formulation and execution feedback mechanism. Combining equipment classification with real-time data, setting optimized operation time and power adjustment effectively reduces overall energy consumption. The introduction of genetic algorithm further optimizes the equipment operation plan and improves energy utilization efficiency. Post-execution effect monitoring and energy saving rate calculation not only ensure the effectiveness of the strategy, but also provide data support for subsequent optimization. Overall, the S7 step significantly enhances the energy-saving and consumption-reducing capabilities of industrial production and promotes the development of green manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 It is a schematic diagram of the steps of the present invention;

[0073] Figure 2 A schematic diagram of the determination process for reclassification of the present invention;

[0074] Figure 3 It is a system principle block diagram of the present invention. DETAILED DESCRIPTION

[0075] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0076] See also Figure 1-Figure 2 The present invention discloses a method for classifying the power load of industrial users and provides the following four technical solutions:

[0077] The first implementation method includes the following steps:

[0078] S1. Data collection and processing of the main system: Through the power grid dispatch center interface, smart meter network and subsystem upload interface, the original data of industrial users' electricity consumption is collected in real time and the data is pre-processed;

[0079] S2. Load characteristic analysis and classification: Calculate the load rate and peak-to-valley difference rate, cluster users using K-means, assign labels based on industry characteristics, and introduce user weights;

[0080] S3, Priority setting: Calculate the priority according to the load, peak-valley difference and user weight, sort and generate a list to assist in grid dispatching;

[0081] S4. Regional management and optimization: Divide the power grid into regions, analyze regional loads, use machine learning to optimize cross-regional power allocation, and formulate strategies;

[0082] S5, Subsystem data collection and encrypted transmission: The subsystem is installed on the industrial user and connected to the main system. The device is accessed through IoT, the collection frequency is configured, and data is securely transmitted to the main system in real time or periodically.

[0083] S6. Equipment load classification and label management: Analyze equipment load characteristics, calculate energy efficiency ratio, classify equipment and set labels, and manage them in combination with geographic location information;

[0084] S7. Load strategy formulation and optimization: Develop load control strategies based on classification and real-time data, and use intelligent algorithms to optimize operation plans.

[0085] Through the collaborative work of the main system and the subsystem, comprehensive optimization from macro to micro is achieved. The main system is responsible for the real-time collection and processing of global data, as well as the in-depth analysis of load characteristics, to provide precise support for power grid dispatching; the subsystem goes deep into the internal industrial users, collects device-level data in detail, and transmits it to the main system in encrypted form, ensuring the security and integrity of the data. This layout not only improves the intelligence level of power grid dispatching, but also enhances the refinement of power management within industrial users. At the same time, machine learning and intelligent algorithms are used to optimize cross-regional power dispatching and load control strategies, effectively improving the overall efficiency and stability of the power system, and laying a solid foundation for the sustainable development of industrial electricity consumption.

[0086] The second implementation mode is mainly different from the first implementation mode in that step S1 specifically includes:

[0087] S1.1 Data collection: Through the grid dispatch center interface, smart meter network and subsystem upload interface, the original data of industrial users’ electricity consumption E, voltage V, current I and power factor PF are collected in real time;

[0088] S1.2 Data storage: Distributed database technology is used to store data by timestamp and geographic location index to ensure fast retrieval and efficient analysis;

[0089] S1.3 Data preprocessing: Clean the data, remove outliers, fill in missing values, and normalize them.

[0090] Step S2 specifically includes:

[0091] S2.1 Load characteristics analysis:

[0092] Load factor calculation: Where i represents the user number; T represents the timestamp; LF i Refers to the load rate of user i, which means the ratio of the maximum load of the user during the observation time to the total load; E i (T) refers to the electric load value of user i at timestamp T; max T Ei(T) refers to the maximum electric load value of user i in all observation timestamps T; ∑ T Ei(T) refers to the total electric load value of user i in all observation timestamps T;

[0093] Peak-to-valley rate calculation: Among them, min T Ei(T) refers to the minimum electric load value of user i in all observation timestamps T;

[0094] S2.2 Cluster analysis: Using the K-means algorithm, users are divided into K categories (C1, C2, ..., CK );

[0095] The K-means algorithm is a widely used clustering algorithm that aims to assign n observations (or data points) to k clusters so that each point belongs to the cluster represented by the mean (cluster center) closest to it, and the mean (i.e. centroid) of all points that are cluster centers defines this cluster. The K-means algorithm is an iterative process that works by trying to optimize the sum of the squared distances from data points to their cluster centers (often called the inertia or within-cluster sum of squares).

[0096] S2.3 Classification criteria and labels: Assign labels to each category based on clustering results and industry characteristics;

[0097] Introducing user importance weight WF i , calculated based on factors such as user scale and economic contribution.

[0098] Step S3 specifically includes:

[0099] S3.1 Priority calculation: P i =α·LF i i+β·PVR i +γ·WF i , where α, β, and γ are weight coefficients, which are adjusted according to the grid strategy;

[0100] S3.2 Priority list: According to P i Sorting and generating priority lists for grid dispatching and load control.

[0101] Step S4 specifically includes:

[0102] S4.1 Regional division: According to the geographical location information, the power grid is divided into multiple regions and a set R = (R1, R2, ..., R N );

[0103] S4.2 Regional load analysis: Calculate the average load and peak-to-valley difference indicators of each region;

[0104] S4.3 Intelligent regional optimization: Use machine learning algorithms, such as deep reinforcement learning or graph neural networks, to analyze the load complementarity and transmission efficiency between regions, formulate cross-regional power allocation strategies, and optimize the configuration of power grid resources;

[0105] Deep reinforcement learning (DRL) is a machine learning method that combines deep learning (DL) and reinforcement learning (RL). It solves complex decision-making problems by using deep neural networks to approximate the policy or value function in reinforcement learning;

[0106] Graph Neural Network (GNN) is a deep learning model specifically designed to process graph structured data. Graph structured data consists of nodes and edges, where nodes represent entities and edges represent the relationship between entities. Graph Neural Network uses this complex network structure to extract and discover features and patterns in graph structured data to meet the needs of graph learning tasks such as clustering, classification, prediction, segmentation, and generation.

[0107] Strategy formulation: Based on the results of regional load analysis, formulate cross-regional power allocation plans or regional power restriction strategies.

[0108] By collecting and preprocessing multi-dimensional electricity consumption data in real time, a solid foundation is provided for load characteristic analysis and accurate classification. The calculation of load rate and peak-to-valley difference rate and K-means clustering analysis, combined with industry characteristics and user weights, realize the accurate identification and classification of user electricity consumption patterns. Furthermore, the priority list formulated based on the classification results directly assists power grid dispatch and optimizes resource allocation. The regional management and optimization strategy analyzes regional load characteristics through intelligent algorithms and formulates cross-regional deployment plans, which significantly improves the overall operation efficiency and stability of the power grid. This series of innovations not only promotes the rational use of energy, but also enhances the flexibility and resilience of the power grid, which is of great significance to the sustainable development of industrial electricity consumption.

[0109] The third implementation mode is mainly different from the second implementation mode in that step S6 specifically includes:

[0110] S6.1 Equipment load characteristics analysis:

[0111] Time series modeling: Apply LSTM algorithm to the equipment load data E device (T) Modeling and identification of load patterns;

[0112] LSTM (Long Short-Term Memory) is a special recurrent neural network (RNN). It improves the internal structure of RNN and introduces a structure called "gate" to control the flow of information, so that it can effectively retain information in long intervals. It is suitable for tasks such as processing and predicting sequence data.

[0113] Calculate the energy efficiency ratio (EER) of the equipment device : Where P output is the device output power, P input Input power to the device;

[0114] S6.2 Equipment classification and labeling: According to load characteristics and energy efficiency ratio, equipment is classified into different categories (such as high-efficiency production equipment, low-efficiency auxiliary equipment, etc.), and labels are set (such as Tagdevice = "high-efficiency motor"), and the equipment label is bound to the geographical location (Locdevice), such as workshop number, production line number, etc.;

[0115] S6.3 Dynamic classification adjustment: Monitor equipment load data in real time. When a significant change in equipment load pattern or energy efficiency ratio is detected, reclassify the equipment and update the label and geographic location binding.

[0116] Step S6.3 dynamic classification adjustment specifically includes:

[0117] S6.3.1 Define change thresholds:

[0118] Load pattern change threshold Δ pattern : Used to quantify the degree of change in the equipment load pattern, calculated by comparing the difference between the current load pattern and the historical load pattern, such as using methods such as Euclidean distance, cosine similarity, or dynamic time warping (DTW):

[0119] Δ pattern =D istance (E device (T current ), E device (T history ));

[0120] Among them, E device (T curren t) is the load data in the current time window, E device (T history ) is the load data in the historical time window, D istance () is the function that calculates the difference between two load patterns;

[0121] Energy efficiency ratio change threshold Δ EER : Used to quantify the degree of change in energy efficiency ratio, calculated by comparing the difference between the current energy efficiency ratio and the historical energy efficiency ratio:

[0122]

[0123] Among them, EER current is the current energy efficiency ratio, EER historyIt is the historical energy efficiency ratio (it can be the energy efficiency ratio at the time of the most recent classification or the average energy efficiency ratio over a period of time);

[0124] S6.3. Trigger conditions:

[0125] Reclassification is triggered when any of the following conditions are met:

[0126] (1) Load pattern changes significantly: Δ pattern >θ pattern , where θ pattern is the preset load pattern change threshold;

[0127] (2) Energy efficiency ratio changes significantly: Δ EER >θ EER , where θ EER is the preset energy efficiency ratio change threshold;

[0128] S6.3. Reclassification process:

[0129] Data collection: real-time collection of equipment load data and energy efficiency ratio data;

[0130] Change calculation: Calculate the difference between the current load pattern and the historical load pattern Δ pattern And the difference between the current energy efficiency ratio and the historical energy efficiency ratio Δ EER ;

[0131] Trigger judgment: Check whether any of the above trigger conditions is met;

[0132] Reclassification: If the trigger condition is met, reapply the clustering algorithm (such as K-means, which is also called K-means clustering algorithm, a widely used unsupervised clustering algorithm. This algorithm divides the data set into K clusters in an iterative manner, and each cluster is represented by the mean (i.e., "center of mass") of all samples within it) or classification algorithm (such as decision tree, support vector machine, etc. Decision tree is a graphical method that intuitively uses probability analysis. It solves the expected value of net present value by building a tree structure, thereby evaluating project risks and judging its feasibility. In machine learning, decision trees are recursively applied to Select the best features for division and build a tree structure from the root node to the leaf node. Each leaf node represents a category or prediction result. Support vector machine is a generalized linear classifier that performs binary classification of data in a supervised learning manner. The core idea is to find an optimal hyperplane to separate sample points of different categories and maximize the margin, thereby improving the accuracy of classification. This hyperplane is called the maximum margin hyperplane, and the data point closest to the hyperplane is called the support vector). The current load mode and energy efficiency ratio are used as features to reclassify the equipment.

[0133] Update labels and bindings: Update the classification labels and geographic location bindings of the device based on the reclassification results.

[0134] By introducing the LSTM algorithm to model the time series of equipment load characteristics and combining it with energy efficiency ratio analysis, we can achieve deep insight and accurate classification of the operating status of industrial equipment. The dynamic classification adjustment mechanism monitors and responds to significant changes in equipment status in real time by setting load mode and energy efficiency ratio change thresholds, ensuring the timeliness and accuracy of classification labels. This not only improves the level of refinement of equipment management, but also provides a scientific basis for energy optimization and equipment maintenance. In addition, binding equipment tags to geographic locations further enhances the convenience and pertinence of management. Overall, the S6 step significantly improves the operating efficiency and energy utilization of industrial equipment, and promotes the intelligent transformation and sustainable development of the production process.

[0135] The fourth implementation mode is mainly different from the third implementation mode in that step S7 specifically includes:

[0136] Strategy Development:

[0137] Based on equipment classification and real-time load data, formulate load control strategies, including setting equipment operation time window T window , adjust the equipment power P adjust ;

[0138] Introducing genetic algorithms to optimize equipment operation plans and reduce overall energy consumption;

[0139] Implementation and Feedback:

[0140] Execute load control strategies and monitor the execution results;

[0141] Collect the load data after execution E after , calculate the energy saving rate And fed back to the main system, where E before is the load data before execution.

[0142] Through intelligent strategy formulation and execution feedback mechanism, refined control of industrial equipment load is achieved. Combining equipment classification and real-time data, setting optimized operation time and power adjustment effectively reduces overall energy consumption. The introduction of genetic algorithm further optimizes equipment operation plan and improves energy efficiency. Post-execution effect monitoring and energy saving rate calculation not only ensure the effectiveness of the strategy, but also provide data support for subsequent optimization. Overall, the S7 step significantly enhances the energy-saving and consumption-reducing capabilities of industrial production and promotes the development of green manufacturing.

[0143] See also Figure 3The present invention also discloses an industrial user electricity load classification system, including a main system and several subsystems. The several subsystems are respectively installed in different industrial users and are aggregated in the main system for unified management.

[0144] The main system includes a data acquisition and processing module, a load characteristic analysis and classification module, a priority setting module, a regional management and optimization module and a subsystem management module. The data acquisition and processing module, the load characteristic analysis and classification module, the priority setting module, and the regional management and optimization module sequentially perform the operations of steps S1 to S4. The subsystem management module is used to connect and manage the subsystem;

[0145] The subsystem includes a data acquisition and encryption transmission module, an equipment load classification and label management module, and a load strategy formulation and optimization module. The data acquisition and encryption transmission module, the equipment load classification and label management module, and the load strategy formulation and optimization module perform steps S5 to S7 in sequence, and the data acquisition and encryption transmission module connects to the subsystem management module for data encryption upload.

[0146] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0147] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0148] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for classifying power loads of industrial users, characterized in that: The specific steps include: S1. Data collection and processing of the main system: Through the power grid dispatch center interface, smart meter network and subsystem upload interface, the original data of industrial users' electricity consumption is collected in real time and the data is pre-processed; S2. Load characteristic analysis and classification: Calculate the load rate and peak-to-valley difference rate, cluster users using K-means, assign labels based on industry characteristics, and introduce user weights; S3, Priority setting: Calculate the priority according to the load, peak-valley difference and user weight, sort and generate a list to assist in grid dispatching; S4. Regional management and optimization: Divide the power grid into regions, analyze regional loads, use machine learning to optimize cross-regional power allocation, and formulate strategies; S5, Subsystem data collection and encrypted transmission: The subsystem is installed on the industrial user and connected to the main system. The device is accessed through IoT, the collection frequency is configured, and data is securely transmitted to the main system in real time or periodically. S6. Equipment load classification and label management: Analyze equipment load characteristics, calculate energy efficiency ratio, classify equipment and set labels, and manage them in combination with geographic location information; S7. Load strategy formulation and optimization: Develop load control strategies based on classification and real-time data, and use intelligent algorithms to optimize operation plans.

2. A method for classifying power loads of industrial users according to claim 1, characterized in that: Step S1 specifically includes: S1.1 Data collection: Through the grid dispatch center interface, smart meter network and subsystem upload interface, the original data of industrial users’ electricity consumption E, voltage V, current I and power factor PF are collected in real time; S1.2 Data storage: Distributed database technology is used to store data by timestamp and geographic location index; S1.3 Data preprocessing: Clean the data, remove outliers, fill in missing values, and normalize them.

3. A method for classifying power loads of industrial users according to claim 1, characterized in that: Step S2 specifically includes: S2.1 Load characteristics analysis: Load factor calculation: Where i represents the user number; T represents the timestamp; LF i Refers to the load rate of user i, which means the ratio of the maximum load of the user during the observation time to the total load; E i (T) refers to the electric load value of user i at timestamp T; max T Ei(T) refers to the maximum electric load value of user i in all observation timestamps T; ∑ T Ei(T) refers to the total electric load value of user i in all observation timestamps T; Peak-to-valley rate calculation: Among them, min T Ei(T) refers to the minimum electric load value of user i in all observation timestamps T; S2.2 Cluster analysis: Using the K-means algorithm, users are divided into K categories (C1, C2, ..., C K ); S2.3 Classification criteria and labels: Assign labels to each category based on clustering results and industry characteristics; Introducing user importance weight WF i , calculated based on user scale and economic contribution factors.

4. A method for classifying power loads of industrial users according to claim 1, characterized in that: Step S3 specifically includes: S3.1 Priority calculation: P i =α·LF i i+β·PVR i +γ·WF i , where α, β, and γ are weight coefficients, which are adjusted according to the grid strategy; S3.2 Priority list: According to P i Sorting and generating priority lists for grid dispatching and load control.

5. A method for classifying power loads of industrial users according to claim 1, characterized in that: Step S4 specifically includes: S4.1 Regional division: According to the geographical location information, the power grid is divided into multiple regions and a set R = (R1, R2, ..., R N ); S4.2 Regional load analysis: Calculate the average load and peak-to-valley difference indicators of each region; S4.3 Intelligent regional optimization: Use machine learning algorithms to analyze load complementarity and transmission efficiency between regions and formulate cross-regional power allocation strategies; Strategy formulation: Based on the results of regional load analysis, formulate cross-regional power allocation plans or regional power restriction strategies.

6. A method for classifying power loads of industrial users according to claim 1, characterized in that: Step S6 specifically includes: S6.1 Equipment load characteristics analysis: Time series modeling: Apply LSTM algorithm to the equipment load data E device (T) Modeling and identification of load patterns; Calculate the energy efficiency ratio (EER) of the equipment device : Where P output is the device output power, P input Input power to the device; S6.2 Equipment classification and labeling: Classify equipment into different categories and set labels based on load characteristics and energy efficiency ratio, and bind equipment labels to geographic locations; S6.3 Dynamic classification adjustment: Monitor equipment load data in real time. When a significant change in equipment load pattern or energy efficiency ratio is detected, reclassify the equipment and update the label and geographic location binding.

7. A method for classifying industrial user power loads according to claim 6, characterized in that: Step S6.3 dynamic classification adjustment specifically includes: S6.3.1 Define change thresholds: Load pattern change threshold Δ pattern : Used to quantify the degree of change in the equipment load pattern, calculated by comparing the difference between the current load pattern and the historical load pattern: Δ pattern =D istance (E device (T current ),E device (T history )); Among them, E device (T curren t) is the load data in the current time window, E device (T history ) is the load data in the historical time window, D istance () is the function that calculates the difference between two load patterns; Energy efficiency ratio change threshold Δ EER : Used to quantify the degree of change in energy efficiency ratio, calculated by comparing the difference between the current energy efficiency ratio and the historical energy efficiency ratio: Among them, EER current is the current energy efficiency ratio, EER history is the historical energy efficiency ratio; S6.

3. Trigger conditions: Reclassification is triggered when any of the following conditions are met: (1) Load pattern changes significantly: Δ pattern >θ pattern , where θ pattern is the preset load pattern change threshold; (2) Energy efficiency ratio changes significantly: Δ EER >θ EER , where θ EER is the preset energy efficiency ratio change threshold; S6.

3. Reclassification process: Data collection: real-time collection of equipment load data and energy efficiency ratio data; Change calculation: Calculate the difference between the current load pattern and the historical load pattern Δ pattern And the difference between the current energy efficiency ratio and the historical energy efficiency ratio Δ EER ; Trigger judgment: Check whether any trigger condition is met; Reclassification: If the trigger condition is met, the clustering algorithm or classification algorithm is reapplied to reclassify the equipment based on the current load pattern and energy efficiency ratio as features; Update labels and bindings: Update the classification labels and geographic location bindings of the device based on the reclassification results.

8. A method for classifying power loads of industrial users according to claim 1, characterized in that: Step S7 specifically includes: Strategy Development: Based on equipment classification and real-time load data, formulate load control strategies, including setting equipment operation time window T window , adjust the equipment power P adjust ; Introducing genetic algorithms to optimize equipment operation plans and reduce overall energy consumption; Implementation and Feedback: Execute load control strategies and monitor the execution results; Collect the load data after execution E after , calculate the energy saving rate And fed back to the main system, where E before is the load data before execution.

9. An industrial user power load classification system, using the industrial user power load classification method according to any one of claims 1 to 8 for classification, characterized in that: It includes a main system and several subsystems. The subsystems are respectively installed in different industrial users and are aggregated into the main system for unified management.

10. The industrial user power load classification system according to claim 9 is characterized in that: The main system includes a data acquisition and processing module, a load characteristic analysis and classification module, a priority setting module, a regional management and optimization module and a subsystem management module. The data acquisition and processing module, the load characteristic analysis and classification module, the priority setting module, and the regional management and optimization module sequentially perform the operations of steps S1 to S4. The subsystem management module is used to connect and manage the subsystem; The subsystem includes a data acquisition and encryption transmission module, an equipment load classification and label management module, and a load strategy formulation and optimization module. The data acquisition and encryption transmission module, the equipment load classification and label management module, and the load strategy formulation and optimization module perform operations from steps S5 to S7 in sequence, and the data acquisition and encryption transmission module is connected to the subsystem management module for data encryption upload.

Images