A regional load regulation method and system

Abstract

This invention discloses a regional load regulation method and system. By constructing a dynamic tag library, it quantifies the adjustable load potential at the user level, dynamically groups regional users, generates and executes regional load regulation strategies, evaluates the regulation effect, and continuously optimizes the tags. Through this technical solution, by introducing multi-dimensional electricity consumption tags covering internal and external factors, it achieves refined tapping of user load potential. Based on refined user grouping, it realizes differentiated and flexible regulation with "one policy per category" or even "one policy per household," minimizing the impact of regulation on users' production and lives while ensuring grid security and increasing user participation.

Description

Technical Field

[0001] This invention belongs to the field of power system operation and control technology, and particularly relates to a regional load regulation method and system. Background Technology

[0002] With the increasing proportion of renewable energy and the growing complexity of power load characteristics, the peak-valley difference in the power grid is widening, posing a severe challenge to its regulation capacity. The traditional "source follows load" model is no longer adequate for the requirements of the new power system, making it crucial to tap the regulation potential on the load side to ensure the safe, stable, and economical operation of the power grid.

[0003] Currently, assessments of adjustable loads are mostly focused on the macro level, such as simple estimations based on industry type experience. These assessments fail to consider the micro-level characteristics of users, such as equipment configuration, electricity consumption habits, and production processes, resulting in crude potential assessments that deviate significantly from actual adjustable capacity. Therefore, there is an urgent need for a method that can refine and dynamically assess user load potential and generate personalized, adaptive control strategies accordingly. Summary of the Invention

[0004] To address the problems existing in the background art, the present invention provides a regional load regulation method.

[0005] A regional load regulation method, characterized by comprising the following steps:

[0006] Step 1) Construct a dynamic tag library, which includes an internal feature tag set and an external constraint tag set;

[0007] Step 2) Real-time data collection and dynamic updating of various tags;

[0008] Step 3) Quantify the user's adjustable load potential; obtain the user load adjustment capability feature vector through weighted calculation;

[0009] Step 4) Dynamically segment users by region;

[0010] Step 5) Generate and execute regional load control strategies for different groups respectively;

[0011] Step 6) After the load control strategy is implemented, if the system determines that there is still a gap, it will initiate temporary bidding to reduce peak loads for specific groups.

[0012] Step 7) Evaluate the effect of the load control strategy and optimize the labels.

[0013] Furthermore, the internal feature tag set is used to describe the user's inherent load regulation capabilities, electricity consumption characteristics, and user attribute characteristics. The internal feature tag set includes: load elasticity feature tags, which include the duration of interruptible power supply, the capacity of the transferable power shield, and the power that can be reduced under the current load of the unit; electricity consumption behavior feature tags, which include the electricity consumption regularity of the current unit, the load curve shape (peak power, flat power, valley point characteristics), and the equipment usage habits of the equipment belonging to the current unit; and user attribute feature tags, which include industry type, daily production shift, and energy consumption priority of various energy-consuming equipment within the current unit, distinguishing between critical and non-critical energy loads.

[0014] Furthermore, the external constraint label set is used to describe external factors that influence a user's willingness and ability to adjust. This external constraint label set includes time factor labels, environmental factor labels, and market and policy factor labels. Time factor labels include season type, weekday / weekend, statutory holidays, temporary work schedule adjustments, and power control measures due to force majeure (natural disasters, war, etc.). Environmental factor labels include temperature, humidity, and weather conditions. Market and policy factor labels include current regional real-time electricity prices, incentive compensation standards, orderly electricity consumption levels, and power supply guarantee requirements. The user's labels are dynamically updated using real-time data collected from the internal feature label set and the external constraint label set.

[0015] Furthermore, the quantification of user adjustable load potential is achieved by constructing a user adjustable load potential quantification model through the internal feature label set and the external constraint label set. The user adjustable load potential quantification model comprehensively calculates the potential values ​​of interruptibility, transferability, and load reduction of users under external constraint factors, forming potential feature vectors for different users.

[0016] Furthermore, the dynamic grouping of users within the region employs a k-means clustering algorithm to dynamically divide users into groups with similar regulatory characteristics. The clustering algorithm uses the "potential feature vector" of each user unit as its primary input. This vector is input by normalizing the basic potential value and combining it with external constraint factors to synthesize real-time potential values, ultimately constructing a complete potential feature vector. This vector, combined with the user's geographical location and transformer substation information, dynamically groups users within the region. Based on the grouping results, users are divided into groups with similar regulatory characteristics, which are broadly categorized as "high interruptibility users," "high transferability users," "temperature-sensitive users," and "rigid load users."

[0017] The basic potential value normalization refers to transforming the qualitative descriptions and raw values ​​in the user's internal feature labels into a unified, dimensionless potential index, which ranges from 0 to 1.

[0018] The external constraint factor refers to the quantified value of the inhibitory or enhancing effect of the external constraint label on each basic potential.

[0019] Furthermore, the generation and execution of regional load regulation strategies for different groups involves issuing different suggestions and instructions to different groups based on the real-time situation of the power grid. These include interruptible load incentive instructions, load shifting suggestions, and power reduction instructions. Based on the real-time regulation needs of the power grid (such as peak shaving, valley filling, and providing reserves), differentiated regulation strategies are matched and generated for different user groups: for the "highly interruptible industrial user" group, interruptible load incentive instructions are prioritized during peak periods; for the "highly mobile commercial user" group, load shifting suggestions are issued under electricity price incentives to guide them to shift their electricity consumption from peak to off-peak or valley periods; for the "temperature-sensitive building" group, instructions for setting temperature fine-tuning or power reduction are issued to utilize their thermal inertia for flexible adjustment.

[0020] After load regulation is implemented, if the system determines that a gap still exists, it can initiate temporary bidding to reduce peak loads for specific groups and make up for the gap.

[0021] Furthermore, the effectiveness of the load control strategy is evaluated, including actual load reduction, load transfer amount, response speed, and user satisfaction; and the user's electricity consumption label is updated based on the evaluation results.

[0022] While adopting the above technical solutions, the method of the present invention is implemented through the following system, which is constructed based on a computer, storage medium and network. The system's working content includes: a data acquisition and processing module, an electricity tag management module, a load potential quantification module, a user dynamic grouping module, a load control strategy generation module, an instruction execution and communication module and an effect evaluation and optimization module.

[0023] The data acquisition and processing module is used to collect users' electricity consumption data, equipment operation data, environmental data, and market policy data.

[0024] Electricity Tag Management Module: Used to store, update, and manage the multi-dimensional user electricity tag system.

[0025] Load potential quantification module: Used to run the user-level potential quantification model and calculate and generate the "potential feature vector" for each user.

[0026] User dynamic grouping module: Used to execute clustering algorithms to complete multi-dimensional dynamic grouping of users in a region.

[0027] Load control strategy generation module: used to generate differentiated control strategy instructions based on grid demand and user grouping results.

[0028] Command execution and communication module: used to send control commands to the user-side smart terminal and receive execution feedback.

[0029] Effect Evaluation and Optimization Module: Used to evaluate the control effect and drive the optimization and updating of the electricity labeling system and quantitative model based on the evaluation results.

[0030] The beneficial effects of this invention are that it provides a regional load regulation method and system. By introducing multi-dimensional electricity consumption tags that cover internal and external factors, it achieves refined tapping of user load potential, significantly improving the accuracy of potential assessment. Based on refined user grouping, it realizes differentiated and flexible regulation with "one policy per category" or even "one policy per household". While ensuring power grid safety, it minimizes the impact of regulation on users' production and life and improves user participation. Attached Figure Description

[0031] Figure 1 This is a flowchart of a regional load regulation method according to Embodiment 1 of the present invention;

[0032] Figure 2 This is a structural diagram of a regional load control system according to Embodiment 2 of the present invention. Detailed Implementation

[0033] The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.

[0034] Example 1

[0035] The regional load control system involved in this embodiment (see system structure block diagram) Figure 2 ).

[0036] S1: Building a dynamic tag library

[0037] Before load regulation begins, a dynamic tag library has been built in advance. The tag library includes internal feature tag sets and external constraint tag sets.

[0038] An internal feature tag set is used to describe the user's inherent load regulation capabilities and power consumption characteristics.

[0039] The internal feature label set includes:

[0040] Load resilience features include interruptible duration, transferable capacity, and power reduction capability.

[0041] Electricity consumption behavior characteristics tags include electricity consumption regularity, load curve shape (peak, flat, and valley characteristics), and equipment usage habits.

[0042] User attribute feature tags, including industry type, production shift, and energy priority (critical / non-critical load).

[0043] Equipment refers to energy-consuming equipment that can participate in load regulation in various industries. For example, commercial users include air conditioning systems, lighting systems, and power systems, while industrial users include office air conditioners, lighting equipment, auxiliary production equipment, power equipment, and production equipment.

[0044] Equipment usage habits include the number of times the equipment is turned on and off per unit time, the equipment's operating period, basic parameter settings, and whether the equipment's usage fluctuates regularly with the seasons (such as air conditioning and heating), weekdays / holidays, or production cycles.

[0045] External constraint label set: Used to describe external factors that affect a user's willingness and ability to adjust.

[0046] The external constraint label set includes:

[0047] Time factor tags include season type, weekday / weekend, and holiday.

[0048] Environmental factors labels include temperature, humidity, and weather conditions.

[0049] Market and policy factors are tagged, including real-time electricity prices, incentive compensation standards, orderly electricity consumption levels, and power supply guarantee requirements.

[0050] S2: Complete data collection and dynamic tag updates

[0051] At 10:00 AM on a summer workday, the system initiated a new round of data collection and tag update cycle:

[0052] Data collection:

[0053] Internal data: Obtain load data for all users at 96 points (every 15 minutes) from the acquisition system for the current day and the same period in history.

[0054] Environmental data: Real-time weather information obtained from the meteorological platform: "Sunny, temperature 35℃, expected to rise to 35℃ at 13:00 in the afternoon, humidity 70%".

[0055] Market policy data: The dispatch center released information that the "peak electricity price" will be launched today, from 13:00 to 15:00, with an electricity price of 3 yuan / kWh. The response requirement is to respond within the day.

[0056] Update tags: The system dynamically updates the tags of some users based on newly collected data.

[0057] Taking user A (a data center) as an example, their tag is updated as follows:

[0058] Internal feature tags: {Industry type: Digital services, Load curve shape: Stationary, Interruptionability: Very low, Portability: Low, Reduceability: Medium (IT load cooling)}

[0059] Taking user B (an electric vehicle charging station) as an example:

[0060] Internal characteristic tags: {Industry type: Service industry, Load curve shape: Bimodal (midday, evening), Interruptibility: Medium, Transferability: High, Reduceability: Medium}

[0061] Take user C (a precision electronics manufacturing company) as an example:

[0062] Internal feature tags: {Industry type: Manufacturing, Production shift system: Three shifts, Energy priority: Production line (extremely high), Comfort air conditioning (high), Office power consumption (medium), Interruptionability: Low (Production line), Portability: Medium (Some testing processes), Reduction potential: Medium (Air conditioning in non-constant temperature and humidity workshops)}

[0063] External constraint labels (for all users): {Season: Summer, Date type: Weekday, Temperature: 32℃ (forecast 35℃), Electricity pricing policy: Peak pricing, Orderly electricity consumption level: Orange}

[0064] S3: Quantify user adjustable load potential

[0065] At 10:30 AM, the system invoked the "Load Potential Quantification Module" to calculate the real-time potential of all users under the current external constraints.

[0066] The user's adjustment potential is calculated using the internal feature labels and external constraint labels.

[0067] The user's adjustment potential is calculated using a weighted fuzzy logic algorithm. This algorithm comprehensively calculates the user's interruptible, transferable, and reduceable load potential under specific external constraints, forming a "potential feature vector" for each user. The structured "potential feature vector" consists of: {Interruptible potential (kW), Transferable potential (kW), Reduceable potential (kW), Response speed, Durable duration (h)}.

[0068] The process of potential feature vector quantization includes the following steps:

[0069] S101: Base Potential Value Normalization

[0070] First, the qualitative descriptions and raw values ​​in the user's internal feature labels are transformed into a unified, dimensionless potential index (range 0-1).

[0071] Input: User's internal feature labels (e.g., interruptibility: medium, transferable capacity: 300kW).

[0072] Processing: The system has a built-in "label-value mapping table" and a normalization algorithm.

[0073] For qualitative labels: fuzzy membership degrees are used for mapping. For example:

[0074] Interruptibility: {Very Low: 0.1, Low: 0.3, Medium: 0.6, High: 0.8, Very High: 0.95}

[0075] User C's interruptibility: Low, quantized as 0.3.

[0076] For quantitative labels: use Min-Max normalization or S-curve normalization based on industry benchmarks. For example, for a medium-sized manufacturing company, a transferable capacity of 300kW might be normalized to 0.65.

[0077] Output: A set of normalized base potential indices {P_int_base, P_shift_base, P_curt_base} are obtained, representing the inherent potential of interruption, transfer, and reduction, respectively.

[0078] S102: Calculation of External Constraint Factors

[0079] Secondly, quantify the inhibitory or enhancing effect of external constraint labels on each basic potential, and calculate the constraint factor C (usually C ≤ 1, indicating inhibition; under strong incentive policies, C > 1, indicating enhancement).

[0080] Input: External constraint labels (e.g., temperature: 35℃, electricity pricing policy: peak electricity price, orderly electricity consumption level: orange).

[0081] Processing: The system has a built-in "constraint rule library" which contains a series of IF-THEN rules or influence coefficient matrices.

[0082] Example Rule 1 (Environmental Constraint): IF Temperature > 33℃ AND User Type = "Temperature-Sensitive Building" THENC_curt = 0.7. In hot weather, users have higher requirements for air conditioning comfort and are less willing to reduce air conditioning usage; therefore, the constraint factor for the reduction potential is 0.7.

[0083] Rule Example 2 (Policy Constraint): IF Orderly Electricity Level = "Orange" THEN C_int = 1.2. Under higher-level orderly electricity requirements, users' cooperation obligations are enhanced, interruptibility potential is amplified, and the constraint factor is 1.2.

[0084] Example Rule 3 (Market Incentives): IF Electricity Price Policy = "Peak Price" AND User Electricity Price Sensitivity = "High" THEN C_shift = 1.5. High electricity prices are a strong incentive for price-sensitive users, significantly increasing their willingness to switch.

[0085] Output: The dynamic constraint factors {C_int, C_shift, C_curt} of the three potentials are obtained.

[0086] S103: Real-time potential value synthesis calculation

[0087] By combining the basic potential index with constraint factors and the user's baseline load, the actual, dimensional potential value (unit: kW) is calculated.

[0088] Real-time potential value (kW) = Basic potential index × Dynamic constraint factor × Baseline load ratio factor × Typical user load value (kW)

[0089] Baseline load proportion factor: This refers to the proportion of the potential load in its typical load, which can be learned from historical data. For example, if user C's air conditioning load accounts for about 20% of its total load, then its reduction potential proportion factor can be set to 0.2.

[0090] Typical load value for a user: The average or predicted load value for that user during the current time period (e.g., midday peak). This is a summer load curve for a shopping mall. The mall's daily electricity load is relatively regular, varying with the season and the mall's operating hours. Therefore, the typical load value for a user can be largely derived from the mall's load curve.

[0091] Calculation example (user C's "cut-off potential"):

[0092] The basic potential index P_curt_base = 0.6;

[0093] The dynamic constraint factor C_curt = 0.7;

[0094] The baseline load scaling factor F_curt = 0.2;

[0095] User typical load value Load_typical = 2000 kW;

[0096] The potential reduction is 0.6 × 0.7 × 0.2 × 2000 kW = 168 kW.

[0097] The system is rounded to 170 kW based on its equipment accuracy.

[0098] S104: Construct the complete potential feature vector

[0099] Finally, all the calculated potential values, along with other key attributes, are encapsulated into a structured vector.

[0100] Vector format: {Interruptible potential (kW), Transferable potential (kW), Cuttable potential (kW), Response time (minutes), Durability (hours)}

[0101] "Response speed" and "duration" are also derived through label mapping and rule base. For example, interruptibility: high users, whose response speed is usually mapped to "fast" (quantified as <15 minutes); productivity: continuous users, whose duration may be mapped to "short" (quantified as <1 hour).

[0102] Example of quantification results:

[0103] User A (Data Center): Due to its extremely low interruptibility but moderate scalability, the model calculates its potential vector under the "orange" warning constraint as: {0, 0, 150, moderate, 2}. This means: no interruptibility / transfer capability, but under the premise of ensuring the safety of core equipment, the cooling load can be reduced by 150kW for 2 hours by appropriately increasing the chilled water temperature.

[0104] User B (charging station): Considering peak electricity prices and high portability, the model calculates its potential vector as: {50, 200, 30, fast, 1}. This means: the 50kW fast charging load can be immediately interrupted, and the dispatch system can guide vehicle owners to transfer the 200kW charging demand to after 15:00, while appropriately reducing the charging power by 30kW.

[0105] User C (manufacturing company): Its production line load interruptibility is low, heavily constrained by policy. The model calculates its potential vector as: {0, 100, 80, slow, 3}. This means: the production line cannot be interrupted, but the 100kW product aging test process can be moved to the night, and the air conditioning in non-core areas can be turned off, reducing the load by 80kW.

[0106] S4: Dynamically segment users by region

[0107] At 10:45 AM, the "User Dynamic Segmentation Module" receives the real-time potential feature vectors of all users and executes the K-means clustering algorithm.

[0108] Grouping Results: The 50 users were divided into 4 groups, as shown in Table 1:

[0109] Table 1 Example of User Dynamic Grouping Results

[0110]

[0111] S5: Generate and execute regional load regulation strategies

[0112] At 13:00, the power grid entered its peak period, and the measured load had reached the warning line. The dispatcher issued an order to "reduce peak load by 5000kW".

[0113] Strategy Generation: The "Load Control Strategy Generation Module" automatically generates and recommends the following differentiated strategy combinations based on the clustering results and control objectives:

[0114] For Group 1: Immediately implement the interruptible load incentive procedure, issuing instructions to 5 users, requiring them to interrupt their contracted loads within 15 minutes. Expected contribution: 800kW.

[0115] For Group 2: Implement unified air conditioning load control instructions, and through the building automation system, uniformly increase the central air conditioning temperature setpoints of all users in the group by 1℃. Expected contribution: 2500kW.

[0116] For Group 3: Implement load shifting incentives, sending SMS / APP push notifications to users to encourage them to move transferable operations (such as user C's testing operations) to off-peak hours, and providing additional subsidies. Expected contribution: 1000kW.

[0117] For group 4: No active control will be performed; it will only be used as a monitoring object for the system's operating status.

[0118] Total expected contribution: 800 + 2500 + 1000 = 4300kW. The system determines that 700kW is still needed, so temporary bidding can be initiated for some high-willing users in Group 1 and Group Rental 2 to smooth out the peak and make up for the shortfall.

[0119] Command execution: The "Command Execution and Communication Module" compiles the above strategies into standardized commands that can be recognized by each user's energy management system or smart terminal through interfaces and communication gateways, and then issues them. User-side devices execute them automatically or after confirmation.

[0120] S6: Evaluate the regulatory effect and optimize the labels.

[0121] The peak period ends at 3:00 PM.

[0122] Effect Evaluation: The "Effect Evaluation and Optimization Module" initiates analysis. It compares the actual load curve during the regulation period with the predicted baseline load curve.

[0123] Actual total peak reduction: 5200kW, exceeding the target.

[0124] Group contribution analysis: Group 1 actually interrupted 750kW (achievement rate 94%), Group 2 actually reduced 2300kW (achievement rate 92%), and Group 3 transferred 900kW (achievement rate 90%).

[0125] Closed-loop optimization:

[0126] The system discovered that the actual interruption response speed of user B (charging station) exceeded the model's expectations. Therefore, it corrected the response speed label from fast to extremely fast and added a positive label of "high reliability" to the user profile. In similar scenarios in the future, it will be prioritized for use.

[0127] It was discovered that the actual reduction in office space for a certain building was only 80% of the expected amount. Analysis revealed that this was because an important client visited that day, and the control order was manually overridden. The system lowered the confidence level of its reduceability label from 0.9 to 0.7 and will apply this discount in the next calculation.

[0128] Through the implementation of this embodiment, the system has successfully achieved precise, flexible, and efficient regulation of regional loads. Furthermore, through a closed-loop optimization mechanism, the system model has become increasingly accurate, providing strong load-side support for the safe, economical, and stable operation of the power grid.

[0129] A regional load control system, comprising:

[0130] Data acquisition and processing module 101: Used to collect users' electricity consumption data, equipment operation data, environmental data and market policy data.

[0131] Electricity Tag Management Module 102: Used to store, update and manage the multi-dimensional user electricity tag system.

[0132] Load potential quantification module 103: used to run the user-level potential quantification model and calculate and generate the "potential feature vector" for each user.

[0133] User dynamic grouping module 104: Used to execute clustering algorithms to complete multi-dimensional dynamic grouping of users in a region.

[0134] Load control strategy generation module 105: used to generate differentiated control strategy instructions based on grid demand and user grouping results.

[0135] Instruction execution and communication module 106: used to send control instructions to the user-side smart terminal and receive execution feedback.

[0136] Effect evaluation and optimization module 107: Used to evaluate the control effect and drive the optimization and updating of the electricity labeling system and quantitative model based on the evaluation results.

[0137] The above specific embodiments are used to explain and illustrate the present invention, and are only preferred embodiments of the present invention, not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made to the present invention within the spirit and scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A regional load regulation method, characterized in that: This is achieved through the following steps: Step 1) Construct a dynamic tag library, which includes an internal feature tag set and an external constraint tag set; Step 2) Real-time data collection and dynamic updating of various tags; Step 3) Quantify the user's adjustable load potential; obtain the user load adjustment capability feature vector through weighted calculation; Step 4) Dynamically segment users by region; Step 5) Generate and execute regional load control strategies for different groups respectively; Step 6) After the load control strategy is implemented, if the system determines that there is still a gap, it will initiate temporary bidding to reduce peak loads for specific groups. Step 7) Evaluate the effect of the load control strategy and optimize the labels.

2. The regional load regulation method as described in claim 1, characterized in that: The internal feature tag set is used to describe the user's inherent load regulation capability, power consumption characteristics, and user attribute characteristics.

3. The regional load regulation method as described in claim 1, characterized in that: The external constraint label set is used to describe external factors that affect a user's willingness and ability to adjust.

4. The regional load regulation method as described in claim 1, characterized in that: The quantification of user adjustable load potential is achieved by constructing a user-level potential quantification model using the internal feature label set and the external constraint label set. The user-level potential quantification model comprehensively calculates the user's interruptibility, transferability, and load reduction potential values ​​under external constraints, forming potential feature vectors for different users.

5. The regional load regulation method as described in claim 1, characterized in that: The method involves dynamically grouping regional users using a clustering algorithm to divide users into groups with similar regulatory characteristics.

6. The regional load regulation method as described in claim 1, characterized in that: The aforementioned generation and execution of regional load regulation strategies for different groups involves issuing different suggestions and instructions to different groups based on the real-time situation of the power grid, including interruptible load incentive instructions, load shifting suggestions, and power reduction instructions.

7. The regional load regulation method as described in claim 1, characterized in that: The evaluation of the effects of the load control strategy includes actual load reduction, load transfer amount, response speed, and user satisfaction. And based on the evaluation results, we will optimize and update the user's electricity usage tags.

8. The regional load regulation method as described in claim 1 is implemented through the following system, which is constructed based on a computer, storage medium, and network, and its working components include: The system includes a data acquisition and processing module, an electricity tag management module, a load potential quantification module, a user dynamic grouping module, a load control strategy generation module, an instruction execution and communication module, and an effect evaluation and optimization module.

Images