Source-load interaction method based on electricity utilization information acquisition system

Abstract

The invention discloses a source-load interaction method based on an electricity utilization information acquisition system, and relates to the field of power distribution network operation and optimization. The method comprises the following four steps: S1, power consumption data acquisition: dynamically acquiring user power consumption, distributed energy output and power grid state data through intelligent equipment, and ensuring real-time accuracy; s2, load characteristic analysis: constructing a quantitative model to analyze characteristics such as a user load elastic coefficient and adjustable potential, establishing a supply and demand balance evaluation system in combination with distributed energy fluctuation, and determining peak and valley time periods and gaps; s3, generating an interactive strategy, and generating and issuing dynamic strategies such as electricity price excitation, load regulation and energy scheduling based on an analysis result; and S4, executing and feeding back the strategy, monitoring the execution effect in real time, and optimizing the strategy through a closed loop mechanism. According to the method, the problems of staticization, single mode and low efficiency of a traditional method are solved, the power grid supply and demand balance capability, the renewable energy consumption rate and the strategy accuracy are improved, and low-carbon and intelligent operation of the power distribution network is supported.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network operation and optimization, and particularly relates to a source-load interaction method based on an electricity consumption information collection system. Background Art

[0002] Existing source-load interaction methods have the following defects: static, single-mode, and low efficiency:

[0003] Static defects: Traditional methods mostly use fixed strategies (such as a single electricity price and fixed load regulation instructions), which cannot be dynamically adjusted according to the real-time operating status of the power grid (such as frequency fluctuations and voltage deviations) and fluctuations in distributed energy output (such as photovoltaic output changes with sunlight and wind power output changes with wind speed). This leads to poor strategy adaptability and difficulty in coping with complex power grid conditions.

[0004] Single-modal defect: Existing technologies mostly rely on a single data dimension (such as only considering user load or only focusing on energy output), and lack a comprehensive analysis of user load characteristics (elasticity coefficient, adjustable potential, electricity usage habits) and distributed energy volatility. As a result, strategy generation lacks comprehensive data support, and the peak-shaving and valley-filling effect is limited.

[0005] Low efficiency: Due to the lack of a closed-loop feedback mechanism, traditional methods are unable to optimize subsequent strategies based on the effectiveness of strategy execution, resulting in low user response rates, high renewable energy curtailment rates, and difficulty in improving grid operation efficiency.

[0006] Therefore, a source-load interaction method based on the electricity consumption information collection system is proposed. Summary of the Invention

[0007] In view of the above shortcomings in the existing technology, the purpose of the present invention is to address the shortcomings of the existing source-load interaction methods, such as static, single-modal, and inefficient. By constructing a dynamic knowledge graph, multi-modal demand analysis, intelligent matching optimization, and a closed-loop iterative mechanism, a complete technical solution of "power consumption data collection - load characteristic analysis - interactive strategy generation - strategy execution and feedback" is formed. This will achieve comprehensive management and control of user load characteristics and distributed energy volatility, accurately balance power grid supply and demand, improve renewable energy absorption rate and strategy adaptability, and provide a systematic solution for the low-carbon and intelligent operation of distribution networks.

[0008] In order to achieve the above objectives, the present invention adopts the following technical solutions:

[0009] A source-load interaction method based on an electricity consumption information collection system comprises the following steps:

[0010] S1. Electricity data collection: Real-time acquisition of user electricity usage data, as well as distributed energy output data and grid operation status information. Throughout the entire process, the data collection frequency is dynamically adjusted based on demand to ensure real-time and accurate data.

[0011] S2. Load characteristics analysis: Based on collected electricity consumption data, a quantitative model is constructed to analyze user load characteristics. Simultaneously, taking into account the output volatility of distributed energy resources, a grid supply and demand balance assessment system is established to ultimately identify grid peak and valley periods and supply and demand gaps.

[0012] S3. Interaction strategy generation: Based on the load characteristics analysis results and grid operation requirements, a dynamic source-load interaction strategy is generated. The generated dynamic source-load interaction strategy is distributed to user terminals and distributed energy control systems through the electricity consumption information collection system.

[0013] S4. Strategy execution and feedback: real-time monitoring of strategy execution results, collection of user response data and grid operation status, dynamic adjustment of interactive strategies; optimization of the accuracy and effectiveness of subsequent strategies through feedback mechanisms.

[0014] Furthermore, the electricity consumption data is obtained by real-time monitoring of operating data on the user side and the grid side through smart meters, sensors and other equipment. The electricity consumption data includes but is not limited to: user electricity load curve, real-time output data of distributed energy, grid frequency, voltage and other operating parameters; and the data is transmitted to the central processing system through the communication network.

[0015] Furthermore, the S1 includes:

[0016] S111. Collection equipment selection and deployment:

[0017] User side: Use smart meters that comply with the DL / T645-2007 protocol, supporting voltage, current, active power, and reactive power measurement. Each phase is equipped with an independent sampling channel to ensure data accuracy in three-phase imbalance scenarios.

[0018] Distributed energy side: PV inverters need to integrate RS485 interfaces to output active power and daily power generation in real time; wind turbine controllers need to collect wind speed, rotational speed, and actual output data, with the sampling frequency synchronized with the smart meter.

[0019] On the grid side: Wireless temperature sensors are installed on 10kV line towers, and synchronized phasor measurement units (PMUs) are deployed at the substation exit to collect grid frequency, voltage, and phase angle data. Throughout the entire process, the grid-side sampling frequency is 256 points / second.

[0020] S112. Dynamic acquisition frequency control:

[0021] Basic frequency: Under normal working conditions, the acquisition frequency is set to 5 minutes / time to meet routine monitoring needs;

[0022] High-frequency trigger conditions: When the grid frequency deviation exceeds ±0.2Hz, the voltage fluctuation exceeds ±5%, or the distributed energy output fluctuation rate exceeds 10%, it will automatically switch to 1-minute / time high-frequency collection and continue until 30 minutes after the parameters return to normal;

[0023] Customized scheduling: set up collection plans for special periods through the central system, and enforce collection every 2 minutes;

[0024] S113. Data transmission and preprocessing:

[0025] The transmission link adopts a dual-link architecture of "fiber-based and 4G / 5G backup". The fiber transmission delay is ≤50ms. The 4G / 5G backup link switches within 5 seconds after the main link is interrupted. Data is transmitted using the AES-128 encryption algorithm.

[0026] Preprocessing rules:

[0027] Mark data with voltage greater than 420V or less than 180V and current greater than 120A as invalid values;

[0028] For single point missing values, the weighted average of the data before and after 10 minutes is used to fill in the missing values, where the weight coefficient of the first 5 minutes is 0.3 and the weight coefficient of the last 5 minutes is 0.7;

[0029] Standardize the raw data into JSON format, including device ID, collection timestamp, parameter name and value.

[0030] Furthermore, the user load characteristics include load elasticity coefficient, adjustable potential, power usage habits, etc.

[0031] Furthermore, the user load characteristic quantitative modeling method includes:

[0032] S21. Quantitative modeling of user load characteristics. Based on collected electricity consumption data, a quantitative model is constructed to analyze the response characteristics of user loads to electricity prices, their adjustable range, and electricity consumption patterns.

[0033] S22. Analyze the volatility of distributed energy output. Extract the distributed energy output data and corresponding meteorological data collected in S1, convert the data to a one-minute granularity, and eliminate data with zero or over-rated values ​​caused by equipment failures. Then, calculate the volatility index and, through curve fitting, identify the intraday output patterns of distributed energy and identify peak and trough output periods.

[0034] Calculate volatility indicators:

[0035] Standard deviation : ;

[0036] For the Minute output, is the average daily output, Minutes per day, =1440;

[0037] Coefficient of variation : ;

[0038] Photovoltaics , wind power , if it exceeds, it will be marked as abnormal fluctuation;

[0039] Climbing rate : , For the Minute output;

[0040] The maximum photovoltaic ramp rate is ≤ 15% of the rated capacity / 15min, and the maximum wind power photovoltaic ramp rate is ≤ 20% / 15min;

[0041] Curve fitting: PV uses a quadratic function to fit the intraday curve, and wind power converts wind power output based on the wind speed-power curve to identify the peak output phase;

[0042] The quadratic function fitting intraday curve is expressed as:

[0043] ;

[0044] for Intraday output plan; is the number of hours; is the quadratic term coefficient of the quadratic function, which is the key constant that determines the “convexity” of the distributed energy daily output curve; is the linear coefficient of the quadratic function, which is a constant that affects the “overall tilt trend” of the distributed energy daily output curve; is the constant of the quadratic function, which represents the “basic output value” of distributed energy;

[0045] The wind speed-power curve is expressed as:

[0046] ;

[0047] Indicates the output power and wind speed of wind power generation equipment (wind turbine) The corresponding relationship function; is the rated power of the fan, that is, the maximum power that the fan can continuously and stably output under design conditions (usually rated wind speed and other working conditions);

[0048] S23. Build a grid supply and demand balance assessment system that integrates user load and distributed energy data, calculates net load and divides peak and valley periods, quantifies the supply and demand gap, and provides clear targets for peak-shaving / valley-filling strategies.

[0049] Furthermore, the step of analyzing the response characteristics, adjustable space and electricity consumption patterns of user loads to electricity prices by constructing a quantitative model based on the collected electricity consumption data includes:

[0050] S211. Calculation of load elastic coefficient :

[0051] Extract samples containing electricity price adjustment events from the historical data collected by S1. The historical data covers at least five electricity price adjustment events, and each historical data covers 15-minute granular load records for seven days before and after the adjustment. Remove outliers from the data, and use the "weighted average of three points before and after" method to fill in missing values. In the weighted average method, the weight of the first point is 0.2, the weight of the first two points is 0.1, the weight of the last point is 0.4, and the weight of the last two points is 0.3.

[0052] Apply the elastic coefficient formula to calculate the load elastic coefficient , the formula is:

[0053] ;

[0054] is the load change, is the base load, is the change in electricity price, is the base electricity price;

[0055] Calculate the average elasticity coefficient by user type (residential, commercial, industrial), and the result must meet the following requirements: , business users , industrial users The significance of the results was verified by t-test, and abnormal samples that deviated from the group mean by 3 times the standard deviation were eliminated;

[0056] S212. Evaluate the potential for adjustment;

[0057] Load classification is based on the physical characteristics and operating rules of electrical equipment, combined with cross-validation of user equipment records and load curves collected by S1, to divide loads into three categories:

[0058] Rigid load, adjustment weight equals 0, verification standard is load curve fluctuation ≤ 5% / 24 hours;

[0059] Equipment with transferable load and time flexibility has an adjustment weight of 0.8. The verification standard is that the operating period can be delayed for ≥4 hours without affecting the use effect;

[0060] Equipment that can reduce load and temporarily reduce power has an adjustment weight of 0.6. The verification standard is that the power can be reduced by 30%-50% and the single reduction duration is ≤2 hours.

[0061] Potential calculation, total adjustable potential Calculation formula:

[0062] ;

[0063] For the The total rated power of the loads; For the The adjustment weight of the load that can be reduced reflects the "contribution coefficient" of this type of load during adjustment; Indicates the number of categories where load can be reduced;

[0064] At the same time, combined with the user compliance rate, calculate the actual callable capacity : ;

[0065] The K-means algorithm is used to cluster the transferable load curves, identify the peak power consumption period and the off-peak available period, and calculate the transfer rate. ,Require ≥50%.

[0066] ;

[0067] S213. Analyze electricity usage habits, group data for three consecutive months by "weekdays / weekends" and "seasons", eliminate outliers and fill in missing data, extract load peaks, time periods and other features, divide electricity usage patterns by clustering, and generate a portrait report for each user that includes "typical daily load curve", "peak hours", and "equipment type ratio".

[0068] Furthermore, the steps of integrating user load and distributed energy data, calculating net load and dividing peak and valley periods, and quantifying the supply and demand gap level include:

[0069] S231. Calculate Net Load : , is the total user load, The total output of distributed energy is plotted at a 15-minute granularity;

[0070] S232. Divide peak and valley periods based on the average net load of the past 30 days and standard deviation , define the peak period as: , lasting ≥30 minutes; trough period is: , lasting ≥30 minutes;

[0071] S233. Assessing the supply-demand gap. The supply-demand gap is being assessed.

[0072] Total power supply of the power grid :

[0073] ;

[0074] The power supply of traditional power sources (such as thermal power, hydropower, nuclear power and other conventional power generation forms);

[0075] Supply and demand gap :

[0076] ;

[0077] Classification by gap ratio:

[0078] Severe power shortage: ;

[0079] General power shortage: ;

[0080] surplus: .

[0081] Furthermore, the source-load interaction strategy includes an electricity price incentive scheme, an adjustable load instruction scheme, a distributed energy scheduling scheme, etc.

[0082] Furthermore, the electricity price incentive scheme is based on the user load elasticity coefficient and the power grid supply and demand status. By dynamically adjusting electricity prices or providing demand response compensation, it encourages users to reduce load during peak hours and increase load during off-peak hours, thereby smoothing the peak-valley difference of the power grid. The steps for designing the electricity price incentive scheme include:

[0083] The load elasticity coefficient calculated based on S2, the larger the absolute value of the elasticity coefficient, the more significant the impact of electricity price adjustment on the load; combined with the real-time supply and demand ratio : ,in is the net load, is the available power supply capacity, The larger it is, the tighter the supply and demand is, and the more electricity prices need to be raised to curb the load;

[0084] Dynamic electricity price calculation formula:

[0085] Peak electricity price , ,in is the base electricity price, is the electricity price adjustment coefficient (value range is 0.3-0.5), is the absolute value of the elastic modulus;

[0086] Off-peak electricity prices , , encouraging users to increase electricity consumption by reducing prices;

[0087] Demand Response Compensation:

[0088] For users who actively participate in load shifting / reduction, compensation will be given according to the amount of reduction: ,in To actually reduce the load, is the compensation standard; the compensation amount increases with the increase of the supply-demand gap level;

[0089] Once the plan is issued, real-time electricity prices and compensation policies will be pushed through smart meters or user apps, with peak and off-peak period adjustments announced one hour in advance.

[0090] Furthermore, the adjustable load instruction scheme sends precise control instructions based on its regulation potential and grid demand to achieve load time shifting or power reduction. The method for generating the adjustable load instruction scheme includes:

[0091] Instruction object filtering:

[0092] Based on the S2 scalability potential assessment results, select users / devices with high scalability weights and a historical response rate ≥ 80%. Prioritize devices that have minimal impact on user experience.

[0093] Specific instruction content:

[0094] Transferable loads: Instructing equipment with time flexibility to transfer loads during specific periods, and providing supporting subsidies and incentives;

[0095] It can reduce the load by sending power adjustment range or operation duration instructions to the equipment that can temporarily reduce the power, and at the same time limit the power of a single air conditioner to no more than 70% of the rated value;

[0096] Calculation of instruction execution volume:

[0097] Total target for a single adjustment : ;

[0098] is the safety factor, ; For the supply-demand gap:

[0099] Allocate adjustment amount according to user adjustable capacity ratio :

[0100] ;

[0101] For the Adjustable capacity for each user;

[0102] Instruction issuance channel: Instructions are issued in the form of encrypted messages through the remote control interface of the electricity consumption information collection system to ensure that the instruction transmission delay is ≤10 seconds.

[0103] Furthermore, the distributed energy scheduling is used to coordinate the output of distributed energy, give priority to consuming renewable energy, reduce wind and solar power curtailment, and at the same time mitigate the impact of its output volatility on the power grid. The steps for formulating the distributed energy scheduling plan include:

[0104] Output plan formulation:

[0105] Formulate daily output plans based on S2 distributed energy output curve prediction , the planned photovoltaic output needs to match the changes in light intensity, and the planned wind power output needs to refer to the wind speed forecast;

[0106] The planned output must meet the grid constraints: , It is the grid acceptance capacity, and the ramp rate within 15 minutes is ≤ 10% of the rated capacity.

[0107] Real-time scheduling strategy:

[0108] Absorption priority, when When the amount is fully absorbed, no adjustment is required; Actual output of photovoltaic / wind power;

[0109] Power curtailment control, when the actual output of photovoltaic / wind power When the local energy storage system is started to charge: If the energy storage capacity is full, the output will be limited according to the principle of "wind power prioritizing photovoltaic power"; Charging power for the local energy storage system;

[0110] In conjunction with load regulation:

[0111] When the output of distributed energy resources drops sharply, an emergency load reduction command is immediately triggered to fill the output gap; when the output of distributed energy resources increases sharply, off-peak electricity price incentives are activated to attract users to increase their electricity consumption and absorb surplus electricity;

[0112] Dispatching instruction issuance: Output limit instructions are sent to photovoltaic inverters and wind turbine controllers via the RS485 interface, with accuracy controlled within ±5% of rated capacity.

[0113] Furthermore, the strategy execution and feedback method:

[0114] S41. Real-time monitoring of strategy execution results. By integrating data from the power consumption information collection system and the SCADA system, the monitoring screen is updated every 5 seconds, focusing on tracking user responses and grid status.

[0115] In terms of user response, calculate the electricity price incentive response rate:

[0116] ;

[0117] During the calculation process, it is required that the transferable load is ≥70% and the reducible load is ≥80%, and the user APP satisfaction score is collected;

[0118] In terms of grid status, monitor frequency, voltage and net load curves, and calculate the peak-to-valley difference changes after strategy execution :

[0119] ;

[0120] Before the strategy is executed, the maximum load of the power grid during the peak period (or the "peak value" in the peak-to-valley difference calculation) represents the peak pressure of the power grid before execution; The maximum load (or new “peak value”) of the power grid during peak hours after the strategy is executed, reflecting the effect of reducing peak load after demand response.

[0121] When the command execution rate is less than 60% or the frequency deviation exceeds ±0.3Hz, an audible and visual alarm is triggered to notify the dispatcher to intervene;

[0122] S42. Feedback data collection and analysis. Feedback data collection covers the user side, energy side, and grid side: the user side records the actual load curve through smart meters , calculate the adjustment amount : , The energy side records the deviation between the actual output and the planned output of distributed energy. : , To formulate a daily output plan, Actual output of photovoltaic / wind power; frequency fluctuation and voltage fluctuation collected on the grid side; during analysis, calculate the strategic benefit index : , is the duration, Effective; Evaluate elastic modulus deviation , is the actual elasticity coefficient calculated after the strategy is executed, To predict the elastic coefficient; The model needs to be modified; statistical response rate differences are calculated by user type to identify high-response and low-response user groups;

[0123] S43. Dynamic adjustment and optimization of strategies, dynamic optimization of strategy parameters based on feedback analysis results: electricity price adjustment coefficient The adjustment formula is:

[0124] ;

[0125] The updated electricity price adjustment coefficient (such as the peak-valley price surcharge coefficient) affects the intensity of user-side electricity price incentives; is the electricity price adjustment coefficient before the update, which is the basic value for adjustment;

[0126] If the actual elastic coefficient is lower than the model prediction, it will increase , instruction allocation is tilted towards high-response users, reducing the quota for low-response users; distributed energy plan output is based on Correction; in the long-term iteration, update the user elasticity coefficient and adjustable potential model weekly, review the strategy effect monthly, and retain the strategy benefit index The Q-learning algorithm is used to train the strategy selection model to automatically match the optimal strategy with the grid status. High-response users are awarded "Green Electricity Certification" and given additional discounts, while low-response industrial users are given higher compensation standards to continuously improve the adaptability of the strategy.

[0127] In summary, due to the adoption of the above technical solution, the beneficial technical effects of the invention are:

[0128] This invention achieves systematic optimization of the source-load interaction in the distribution network through the overall technical solution of "electricity data collection - load characteristics analysis - interactive strategy generation - strategy execution and feedback". The comprehensive beneficial effects are as follows:

[0129] The overall solution, through its dynamic, multi-dimensional, closed-loop design, completely breaks through the limitations of traditional methods. Dynamic data collection ensures real-time and accuracy, providing a reliable foundation for subsequent analysis. Load characteristic analysis quantifies user load characteristics and distributed energy volatility to build a comprehensive supply and demand balance assessment system, providing an accurate basis for strategy generation. Interactive strategy generation combines user and energy characteristics to generate personalized, multi-type collaborative strategies, achieving precise matching of sources and loads. Strategy execution and feedback continuously improve strategy effectiveness through real-time monitoring and iterative optimization.

[0130] Overall, the solution significantly improves the grid's ability to balance supply and demand. By accurately identifying peak and valley periods and supply and demand gaps, combined with dynamic electricity prices, load regulation and distributed energy scheduling, it effectively smooths out the peak and valley differences in the grid and reduces grid operating pressure. It significantly increases the renewable energy absorption rate, reduces wind and solar power curtailment, and promotes the low-carbon transformation of distribution networks. It enhances the accuracy of strategies and user response rates, improves user participation through personalized strategies and feedback optimization, and ensures efficient implementation of strategies. At the same time, it ensures the safe and economical operation of the grid, avoids operational risks through real-time monitoring and alarm mechanisms, reduces operating costs, and improves power supply reliability.

[0131] The present invention provides a systematic solution for the low-carbon and intelligent operation of the distribution network, and comprehensively improves the economy, safety and sustainability of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0132] Figure 1 The present invention is a flow chart of a source-load interaction method based on an electricity consumption information collection system. DETAILED DESCRIPTION

[0133] In order to make the purpose, technical solutions and advantages of the invention more clear, the invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described here are only used to explain the invention and are not used to limit the invention.

[0134] like Figure 1 As shown in FIG, a source-load interaction method based on the power consumption information collection system:

[0135] A source-load interaction method based on an electricity consumption information collection system comprises the following steps:

[0136] S1. Electricity data collection: Real-time acquisition of user electricity data (voltage, current, power, and electricity consumption, etc.), as well as output data of distributed energy resources (such as photovoltaic and wind power) and grid operating status information (such as frequency and voltage fluctuations). Throughout the entire process, the data collection frequency is dynamically adjusted based on demand to ensure real-time and accurate data.

[0137] The electricity consumption data is obtained by real-time monitoring of user-side and grid-side operating data using smart meters, sensors, and other equipment. The electricity consumption data includes, but is not limited to, user load curves, real-time output data of distributed energy resources, grid frequency, voltage, and other operating parameters. The data is transmitted to a central processing system via a communication network to provide a basis for subsequent analysis.

[0138] S111. Collection equipment selection and deployment:

[0139] User side: Use smart meters that comply with the DL / T645-2007 protocol, supporting voltage (0-400V), current (0-100A), active power (0.2-level accuracy), and reactive power (0.5-level accuracy) measurement. Each phase is equipped with an independent sampling channel to ensure data accuracy in three-phase imbalance scenarios.

[0140] On the distributed energy side: The photovoltaic inverter must be integrated with an RS485 interface to output real-time active power (resolution 0.1kW) and daily power generation (resolution 0.01kWh). The wind turbine controller must collect wind speed (0-30m / s), rotation speed (0-1500r / min) and actual output data, and the sampling frequency must be synchronized with the smart meter.

[0141] Grid side: Wireless temperature sensors (temperature measurement range -40°C to +125°C, accuracy ±1°C) are installed on 10kV line towers, and synchronized phasor measurement units (PMUs) are deployed at substation exits to collect grid frequency (50Hz±0.5Hz, measurement accuracy ±0.001Hz), voltage (10kV±10%, accuracy 0.2), and phase angle data. The grid-side sampling frequency is 256 points / second.

[0142] S112. Dynamic acquisition frequency control:

[0143] Basic frequency: Under normal operating conditions (grid frequency 50±0.1Hz, voltage deviation ≤±2%), the acquisition frequency is set to 5 minutes / time to meet routine monitoring needs;

[0144] High-frequency trigger conditions: When the grid frequency deviation exceeds ±0.2Hz, the voltage fluctuation exceeds ±5%, or the distributed energy output fluctuation rate (change in 1 minute / rated capacity) exceeds 10%, it will automatically switch to 1-minute / time high-frequency collection and continue until 30 minutes after the parameters return to normal;

[0145] Customized scheduling: Set data collection plans for special periods (such as holidays and peak electricity consumption) through the central system. For example, during the summer period of 2:00 PM to 4:00 PM (peak air conditioning load), data collection is enforced every 2 minutes.

[0146] S113. Data transmission and preprocessing:

[0147] The transmission link adopts a dual-link architecture of "fiber-based and 4G / 5G backup". The fiber transmission delay is ≤50ms. The 4G / 5G backup link switches within 5 seconds after the main link is interrupted. Data is transmitted using the AES-128 encryption algorithm.

[0148] Preprocessing rules:

[0149] Abnormal value elimination: data with voltage > 420V or < 180V, and current > 120A are marked as invalid values;

[0150] Data completion: For single point missing values, the weighted average of the data before and after 10 minutes is used to fill the missing value (weight coefficient: 0.3 for the first 5 minutes and 0.7 for the last 5 minutes);

[0151] Format conversion: Standardize the raw data into JSON format, including the device ID, collection timestamp (accurate to milliseconds), parameter name and value.

[0152] S2. Load characteristics analysis: Based on collected electricity consumption data, a quantitative model is constructed to analyze user load characteristics. Furthermore, considering the output volatility of distributed energy resources, a grid supply and demand balance assessment system is established to ultimately identify grid peak and valley periods and supply and demand gaps.

[0153] S21. Quantitative modeling of user load characteristics. Based on collected electricity usage data, a quantitative model is constructed to analyze the response characteristics of user loads to electricity prices, their adjustable range, and their electricity usage patterns, providing a user-side basis for strategy generation. This includes:

[0154] S211. Calculation of load elastic coefficient :

[0155] Extract samples containing electricity price adjustment events from the historical data collected by S1. The historical data must cover at least five electricity price adjustment events, and each time, the historical data must cover 15-minute granularity load records for seven days before and after the adjustment. Remove outliers from the data (such as sudden drops / surges in load due to power outages or equipment failures). Fill missing values ​​using the "weighted average of three points before and after" method. In this method, the weight of the first point is 0.2, the weight of the first two points is 0.1, the weight of the last point is 0.4, and the weight of the last two points is 0.3.

[0156] Apply the elastic coefficient formula to calculate the load elastic coefficient , the formula is:

[0157] ;

[0158] is the load change, is the base load, is the change in electricity price, is the base electricity price;

[0159] Calculate the average elasticity coefficient by user type (residential, commercial, industrial), and the result must meet the following requirements: , business users , industrial users The significance of the results was verified by t-test (confidence level ≥ 95%), and abnormal samples that deviated from the group mean by 3 times the standard deviation were eliminated;

[0160] S212. Evaluating Adjustable Potential:

[0161] Load classification is based on the physical characteristics and operating rules of electrical equipment, combined with cross-validation of user equipment records and load curves collected by S1, to divide loads into three categories:

[0162] For rigid loads, such as medical monitors and refrigerators that require 24-hour continuous power supply, the adjustment weight is equal to 0, and the verification standard is that the load curve fluctuation is ≤5% / 24 hours;

[0163] Transferable loads, such as electric vehicle charging stations and storage-type electric water heaters, have time-flexible equipment. The adjustment weight is 0.8, and the verification standard is that the operating period can be delayed by 4 hours or more without affecting the performance.

[0164] Loads that can be reduced, such as air conditioners and non-core lighting, have a temporary power reduction. The adjustment weight is 0.6. The verification standard is that the power can be reduced by 30%-50% and the single reduction duration is ≤2 hours.

[0165] Potential calculation, total adjustable potential calculation formula:

[0166] ;

[0167] For the The total rated power of the loads; For the The adjustment weight of the load that can be reduced reflects the "contribution coefficient" of this type of load during adjustment; Indicates the number of categories where load can be reduced;

[0168] At the same time, combined with the user compliance rate, calculate the actual callable capacity : ;

[0169] The K-means algorithm is used to cluster the transferable load curves, identify the peak power consumption period and the off-peak available period, and calculate the transfer rate. ,Require ≥50%.

[0170] ;

[0171] S213. Analyze electricity usage habits. Group three consecutive months of data by "weekday / weekend" and "season." Remove outliers and fill in missing data. Extract characteristics such as load peaks and time periods. Use clustering to categorize electricity usage patterns. Generate a profile report for each user, including "typical daily load curve," "peak hours," and "device type share."

[0172] S22. Analyze the volatility of distributed energy output. Extract the distributed energy output data collected in S1 (1 minute / point for photovoltaic power and 10 seconds / point for wind power) and corresponding meteorological data (light intensity and wind speed). Convert the data to a 1-minute granularity, eliminating data with zero or over-rated values ​​caused by equipment failures. Then, calculate volatility indicators and use curve fitting to identify the intraday output patterns of distributed energy resources and identify peak and trough output periods.

[0173] Calculate volatility indicators:

[0174] Standard deviation :

[0175] For the Minute output, is the average daily output, Minutes per day, =1440;

[0176] Coefficient of variation : ,

[0177] Photovoltaics , wind power , if it exceeds, it will be marked as abnormal fluctuation;

[0178] Climbing rate : , For the Minute output;

[0179] The maximum photovoltaic ramp rate is ≤ 15% of the rated capacity / 15min, and the maximum wind power photovoltaic ramp rate is ≤ 20% / 15min;

[0180] Curve fitting: PV uses a quadratic function to fit the intraday curve, and wind power converts wind power output based on the wind speed-power curve to identify the peak output phase;

[0181] The quadratic function fitting intraday curve is expressed as:

[0182] ;

[0183] for Intraday output plan; is the number of hours; is the quadratic term coefficient of the quadratic function, which is the key constant that determines the “convexity” of the distributed energy daily output curve; is the linear coefficient of the quadratic function, which is a constant that affects the “overall tilt trend” of the distributed energy daily output curve; is the constant of the quadratic function, which represents the “basic output value” of distributed energy;

[0184] The wind speed-power curve is expressed as:

[0185] ;

[0186] Indicates the output power and wind speed of wind power generation equipment The corresponding relationship function; is the rated power of the fan, that is, the maximum power that the fan can continuously and stably output under the design conditions;

[0187] S23. Build a grid supply and demand balance assessment system that integrates user load and distributed energy data, calculates net load, divides peak and valley periods, quantifies the supply and demand gap, and provides clear targets for peak-shaving / valley-filling strategies.

[0188] First, calculate the net load : , is the total user load, The total output of distributed energy is plotted at a 15-minute granularity;

[0189] Secondly, the peak and valley periods are divided based on the average net load of the past 30 days. and standard deviation , define the peak period as: , lasting ≥30 minutes; trough period is: , lasting ≥30 minutes;

[0190] Finally, assess the supply and demand gap. The supply and demand gap is being assessed;

[0191] Total power supply of the power grid :

[0192] ;

[0193] The power supply of traditional power sources (such as thermal power, hydropower, nuclear power and other conventional power generation forms);

[0194] Supply and demand gap :

[0195] ;

[0196] Classification by gap ratio:

[0197] Severe power shortage: ;

[0198] General power shortage: ;

[0199] surplus: .

[0200] S3. Interaction strategy generation: Based on the load characteristics analysis results of S2 (user elasticity coefficient, adjustable potential, peak and valley periods, supply and demand gap, etc.) and the real-time operating status of the power grid, dynamic source-load interaction strategies are generated. These strategies include three categories: electricity price incentive schemes, adjustable load instruction schemes, and distributed energy resource scheduling schemes. These strategies are distributed to user terminals and distributed energy resource control systems through the electricity consumption information collection system. They aim to balance grid supply and demand, optimize operational efficiency, prioritize renewable energy consumption, and reduce wind and solar power curtailment. Strategy generation must take into account user sensitivity to electricity prices, load adjustability, and the output characteristics of distributed energy resources to ensure targetedness and effectiveness.

[0201] The electricity price incentive scheme is based on the user load elasticity coefficient and the grid supply and demand status. By dynamically adjusting electricity prices or providing demand response compensation, it encourages users to reduce load during peak hours and increase load during off-peak hours, thereby smoothing the peak-valley difference of the grid. The steps for designing the electricity price incentive scheme include:

[0202] The load elasticity coefficient calculated based on S2, the larger the absolute value of the elasticity coefficient (such as residential users ∈[−0.5,−0.2]), the more significant the impact of electricity price adjustment on load; combined with the real-time supply and demand ratio : ,in is the net load, is the available power supply capacity, The larger it is, the tighter the supply and demand is, and the more electricity prices need to be raised to curb the load;

[0203] Dynamic electricity price calculation formula:

[0204] Peak electricity price , ,in is the base electricity price, is the adjustment coefficient (value range is 0.3-0.5), is the absolute value of the elastic coefficient;

[0205] Off-peak electricity prices , , encouraging users to increase electricity consumption by reducing prices;

[0206] Demand Response Compensation:

[0207] For users who actively participate in load shifting / reduction, compensation will be given according to the amount of reduction: ,in To actually reduce the load, The compensation standard; the compensation amount increases with the level of supply and demand gap (when there is severe power shortage Increase by 50%);

[0208] Once the plan is issued, real-time electricity prices and compensation policies will be pushed through smart meters or user apps, with peak and off-peak period adjustments announced one hour in advance.

[0209] For adjustable load (transferable, curtailable) instruction schemes, precise control instructions are sent based on their regulation potential and grid demand to achieve load time shifting or power curtailment. The adjustable load instruction generation method includes:

[0210] Instruction object filtering:

[0211] Based on the S2 scalability potential assessment results, select users / devices with high scalability weights (load transferability and load reduction) and a historical response rate ≥ 80%. Prioritize devices with minimal impact on user experience (e.g., electric vehicle charging stations and air conditioners).

[0212] Specific instruction content:

[0213] Shiftable loads: Instructing time-flexible devices (electric vehicle users, storage water heaters) to shift loads to specific time periods (e.g., from peak loads of 18:00-20:00 to off-peak loads of 23:00-6:00 the following day), along with supporting subsidies (e.g., 0.1 yuan / kWh).

[0214] It can reduce loads by sending limited power adjustment range instructions or operation duration instructions (single ≤ 2 hours) to devices that can temporarily reduce power. For example, "Raise the air conditioner temperature from 26°C to 28°C, with a compensation of 0.5 yuan per unit." At the same time, it limits the power of a single air conditioner to no more than 70% of the rated value.

[0215] Calculation of instruction execution volume:

[0216] Total target for a single adjustment : ;

[0217] is the safety factor, ; The gap between supply and demand;

[0218] Allocate adjustment amount according to user adjustable capacity ratio :

[0219] ;

[0220] For the Adjustable capacity for each user;

[0221] Instruction issuance channel: Instructions are issued in the form of encrypted messages through the remote control interface of the electricity consumption information collection system to ensure that the instruction transmission delay is ≤10 seconds.

[0222] The distributed energy scheduling plan is used to coordinate the output of distributed energy, give priority to consuming renewable energy, reduce wind and solar power curtailment, and mitigate the impact of its output fluctuations on the power grid. The steps for formulating the distributed energy scheduling plan include:

[0223] Output plan formulation:

[0224] Formulate daily output plans based on S2 distributed energy output curve prediction , the planned photovoltaic output needs to match the changes in light intensity, and the planned wind power output needs to refer to the wind speed forecast;

[0225] The planned output must meet the grid constraints: , It is the grid acceptance capacity, and the ramp rate within 15 minutes is ≤ 10% of the rated capacity.

[0226] Real-time scheduling strategy:

[0227] Absorption is prioritized when photovoltaic / wind power actually outputs When the amount is fully absorbed, no adjustment is required;

[0228] Power curtailment control, when When the local energy storage system is started to charge: If the energy storage capacity is full, the output will be limited according to the principle of "wind power prioritizing photovoltaic power"; Charging power for the local energy storage system;

[0229] In conjunction with load regulation:

[0230] When the output of distributed energy resources drops suddenly (drops by more than 20% of the rated capacity within 15 minutes), an emergency load reduction command is immediately triggered to fill the output gap;

[0231] When the output of distributed energy resources increases suddenly, the off-peak electricity price incentive is activated to attract users to increase their electricity consumption and absorb surplus electricity;

[0232] Dispatching instruction issuance: Output limit instructions are sent to photovoltaic inverters and wind turbine controllers via the RS485 interface, with accuracy controlled within ±5% of rated capacity.

[0233] S4. Strategy execution and feedback: real-time monitoring of strategy execution results, collection of user response data and grid operation status, dynamic adjustment of interactive strategies; optimization of the accuracy and effectiveness of subsequent strategies through feedback mechanisms.

[0234] S41. Real-time monitoring of strategy execution results. By integrating data from the power consumption information collection system and the SCADA system, the monitoring screen is updated every 5 seconds, focusing on tracking user responses and grid status.

[0235] In terms of user response, calculate the electricity price incentive response rate:

[0236] ;

[0237] During the calculation process, it is required that the transferable load is ≥70% and the reducible load is ≥80%, and the user APP satisfaction score is collected;

[0238] In terms of grid status, monitor frequency, voltage and net load curves, and calculate the peak-to-valley difference changes after strategy execution :

[0239] ;

[0240] is the maximum load of the power grid during the peak period before the strategy is implemented, representing the peak pressure of the power grid before implementation; It is the maximum load of the power grid during peak hours after the strategy is executed, reflecting the effect of reducing peak load after demand response;

[0241] When the command execution rate is less than 60% or the frequency deviation exceeds ±0.3Hz, an audible and visual alarm is triggered to notify the dispatcher to intervene;

[0242] S42. Feedback data collection and analysis. Feedback data collection covers the user side, energy side, and grid side: the user side records the actual load curve through smart meters , calculate the adjustment amount : , The energy side records the deviation between the actual output and the planned output of distributed energy. : To formulate a daily output plan, Actual output of photovoltaic / wind power; frequency fluctuation and voltage fluctuation collected on the grid side; during analysis, calculate the strategic benefit index : , is the duration, Effective; Evaluate elastic modulus deviation , is the actual elasticity coefficient calculated after the strategy is executed, To predict the elastic coefficient; The model needs to be modified; statistical response rate differences are calculated by user type to identify high-response and low-response user groups;

[0243] S43. Dynamic adjustment and optimization of strategies, dynamic optimization of strategy parameters based on feedback analysis results: electricity price adjustment coefficient The adjustment formula is:

[0244] ;

[0245] is the updated electricity price adjustment coefficient, which affects the incentive intensity of the user-side electricity price; is the electricity price adjustment coefficient before the update, which is the basic value for adjustment;

[0246] If the actual elastic coefficient is lower than the model prediction, it will increase ; Instruction allocation is tilted towards high-response users, reducing the quota for low-response users; Distributed energy plan output is based on Correction; in the long-term iteration, update the user elasticity coefficient and adjustable potential model weekly, review the strategy effect monthly, and retain the strategy benefit index The Q-learning algorithm is used to train the strategy selection model to automatically match the optimal strategy with the grid status. High-response users are awarded "Green Electricity Certification" and given additional discounts, while low-response industrial users are given higher compensation standards to continuously improve the adaptability of the strategy.

[0247] The above description is a preferred embodiment of the invention and is not intended to limit the invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the invention should be included in the scope of protection of the invention.

Claims

1. A source-load interaction method based on an electricity consumption information collection system, characterized in that: The following steps are involved: S1. Electricity data collection: Real-time acquisition of user electricity usage data, as well as distributed energy output data and grid operation status information. The data collection frequency is dynamically adjusted based on demand to ensure real-time and accurate data. S2. Load characteristics analysis: Based on collected electricity consumption data, a quantitative model is constructed to analyze user load characteristics. Furthermore, considering the output volatility of distributed energy resources, a grid supply and demand balance assessment system is established to ultimately identify grid peak and valley periods and supply and demand gaps. S3. Interaction strategy generation: Based on the load characteristics analysis results and grid operation requirements, a dynamic source-load interaction strategy is generated. The generated dynamic source-load interaction strategy is distributed to user terminals and distributed energy control systems through the electricity consumption information collection system. S4. Strategy execution and feedback: real-time monitoring of strategy execution results, collection of user response data and grid operation status, dynamic adjustment of interactive strategies; optimization of the accuracy and effectiveness of subsequent strategies through feedback mechanisms.

2. The source-load interaction method based on the power consumption information collection system according to claim 1, characterized in that: The electricity consumption data is obtained by real-time monitoring of the operating data on the user side and the grid side through smart meters, sensors and other equipment. The electricity consumption data includes but is not limited to: user electricity load curve, real-time output data of distributed energy, grid frequency, and voltage operating parameters; and the data is transmitted to the central processing system through the communication network.

3. The source-load interaction method based on the power consumption information collection system according to claim 1, characterized in that: Said S1 comprises: S111. Collection equipment selection and deployment: User side: Use smart meters that comply with the DL / T645-2007 protocol, supporting voltage, current, active power, and reactive power measurement. Each phase is equipped with an independent sampling channel to ensure data accuracy in three-phase imbalance scenarios. Distributed energy side: PV inverters need to integrate RS485 interfaces to output active power and daily power generation in real time; wind turbine controllers need to collect wind speed, rotational speed, and actual output data, with the sampling frequency synchronized with the smart meter. On the grid side: Wireless temperature sensors are installed on 10kV line towers, and synchronized phasor measurement units (PMUs) are deployed at the substation exit to collect grid frequency, voltage, and phase angle data. Throughout the entire process, the grid-side sampling frequency is 256 points / second. S112. Dynamic acquisition frequency control: Basic frequency: Under normal working conditions, the acquisition frequency is set to 5 minutes / time to meet routine monitoring needs; High-frequency trigger conditions: When the grid frequency deviation exceeds ±0.2Hz, the voltage fluctuation exceeds ±5%, or the distributed energy output fluctuation rate exceeds 10%, it will automatically switch to 1-minute / time high-frequency collection and continue until 30 minutes after the parameters return to normal; Customized scheduling: set up collection plans for special periods through the central system, and enforce collection every 2 minutes; S113. Data transmission and preprocessing: The transmission link adopts a dual-link architecture of "fiber-based, 4G / 5G backup". The fiber transmission delay is ≤50ms. The 4G / 5G backup link switches within 5 seconds after the main link is interrupted. Data is transmitted using the AES-128 encryption algorithm. Preprocessing rules: Mark data with voltage greater than 420V or less than 180V and current greater than 120A as invalid values; For single point missing values, the weighted average of the data before and after 10 minutes is used to fill in the missing values, where the weight coefficient of the first 5 minutes is 0.3 and the weight coefficient of the last 5 minutes is 0.7; Standardize the raw data into JSON format, including device ID, collection timestamp, parameter name and value.

4. The source-load interaction method based on the power consumption information collection system according to claim 1, characterized in that: The user load characteristics include load elasticity coefficient, adjustable potential, and electricity usage habits.

5. The source-load interaction method based on the power consumption information collection system according to claim 1, characterized in that: The user load characteristic quantitative modeling method includes: S21. Quantitative modeling of user load characteristics. Based on collected electricity consumption data, a quantitative model is constructed to analyze the response characteristics of user loads to electricity prices, their adjustable range, and electricity consumption patterns. S22. Analyze the volatility of distributed energy output. Extract the distributed energy output data and corresponding meteorological data collected in S1, convert the data to a one-minute granularity, and eliminate data with zero or over-rated values ​​caused by equipment failures. Then, calculate the volatility index and, through curve fitting, identify the intraday output patterns of distributed energy and identify peak and trough output periods. Calculate volatility indicators: Standard deviation : ; For the Minute output, is the average daily output, Minutes per day, =1440; Coefficient of variation : ; Photovoltaics , wind power , if it exceeds, it will be marked as abnormal fluctuation; Climbing rate : , For the Minute output; The maximum photovoltaic ramp rate is ≤ 15% of the rated capacity / 15min, and the maximum wind power photovoltaic ramp rate is ≤ 20% / 15min; Curve fitting: PV uses a quadratic function to fit the intraday curve, and wind power converts wind power output based on the wind speed-power curve to identify the peak output phase; The quadratic function fitting intraday curve is expressed as: ； for Intraday output plan; is the number of hours; is the coefficient of the quadratic term of the quadratic function; is the coefficient of the linear term of the quadratic function; is the constant of the quadratic function; The wind speed-power curve is expressed as: ； Indicates the output power and wind speed of wind power generation equipment The corresponding relationship function; is the rated power of the fan, that is, the maximum power that the fan can continuously and stably output under the design conditions; S23. Build a grid supply and demand balance assessment system that integrates user load and distributed energy data, calculates net load and divides peak and valley periods, quantifies the supply and demand gap, and provides clear targets for peak-shaving / valley-filling strategies.

6. The source-load interaction method based on the power consumption information collection system according to claim 5, characterized in that: The steps of analyzing the response characteristics of user loads to electricity prices, adjustable space, and electricity usage patterns by constructing a quantitative model based on the collected electricity usage data include: S211. Calculation of load elastic coefficient : Extract samples containing electricity price adjustment events from the historical data collected by S1. The historical data must cover at least five electricity price adjustment events, and each historical data covers 15-minute granular load records for seven days before and after the adjustment. Remove outliers from the data, and fill missing values ​​using the "weighted average of three points before and after" method. In this method, the weight of the first point is 0.2, the weight of the first two points is 0.1, the weight of the last point is 0.4, and the weight of the last two points is 0.

3. Apply the elastic coefficient formula to calculate the load elastic coefficient , the formula is: ； is the load change, is the base load, is the change in electricity price, is the base electricity price; Calculate the average elasticity coefficient by user type group, and the result must meet the following requirements: Residential users , business users , industrial users The significance of the results was verified by t-test, and abnormal samples that deviated from the group mean by 3 times the standard deviation were eliminated; S212. Evaluate the potential for adjustment; Load classification is based on the physical characteristics and operating rules of electrical equipment, combined with cross-validation of user equipment records and load curves collected by S1, to divide loads into three categories: Rigid load, adjustment weight equals 0, verification standard is load curve fluctuation ≤ 5% / 24 hours; Equipment with transferable load and time flexibility has an adjustment weight of 0.

8. The verification standard is that the operating period can be delayed for ≥4 hours without affecting the use effect; Equipment that can reduce load and temporarily reduce power has an adjustment weight of 0.

6. The verification standard is that the power can be reduced by 30%-50% and the single reduction duration is ≤2 hours. Potential calculation, total adjustable potential Calculation formula: ； For the The total rated power of the loads; For the The adjustment weight of the load that can be reduced reflects the "contribution coefficient" of this type of load during adjustment; Indicates the number of categories where load can be reduced; At the same time, combined with the user compliance rate, calculate the actual callable capacity : ; The K-means algorithm is used to cluster the transferable load curves, identify the peak power consumption period and the off-peak available period, and calculate the transfer rate. , ≥50%; ； S213. Analyze electricity usage habits. Group three consecutive months of data by "weekdays / weekends" and "seasons." Remove outliers and complete missing data. Extract characteristics such as load peaks and time periods. Cluster electricity usage patterns to generate a profile report for each user, including "typical daily load curve," "peak hours," and "equipment type ratio." 7. The source-load interaction method based on the power consumption information collection system according to claim 5, characterized in that: The steps of integrating user load and distributed energy data, calculating net load, dividing peak and valley periods, and quantifying the supply and demand gap level include: S231. Calculate Net Load : , is the total user load, The total output of distributed energy is plotted at a 15-minute granularity; S232. Divide peak and valley periods based on the average net load of the past 30 days and standard deviation , define the peak period as: , lasting ≥30 minutes; trough period is: , lasting ≥30 minutes; S233. Assessing the supply-demand gap. The supply-demand gap is being assessed. Total power supply of the power grid : ； The power supply of the traditional power supply; Supply and demand gap : ； Classification by gap ratio: Severe power shortage: ; General power shortage: ; surplus: .

8. The source-load interaction method based on the power consumption information collection system according to claim 1, characterized in that: The dynamic source-load interaction strategy includes three categories: electricity price incentive scheme, adjustable load instruction scheme, and distributed energy scheduling scheme.

9. The source-load interaction method based on the power consumption information collection system according to claim 8, characterized in that: The electricity price incentive scheme is based on the user load elasticity coefficient and the supply and demand status of the power grid. By dynamically adjusting the electricity price or providing demand response compensation, it encourages users to reduce their load during peak hours and increase their load during off-peak hours, thereby smoothing the peak-valley difference of the power grid; The design steps of the electricity price incentive plan include: The load elasticity coefficient calculated based on S2, the larger the absolute value of the elasticity coefficient, the more significant the impact of electricity price adjustment on the load; combined with the real-time supply and demand ratio : ,in is the net load, is the available power supply capacity, The larger it is, the tighter the supply and demand is, and the more electricity prices need to be raised to curb the load; Dynamic electricity price calculation formula: Peak electricity price , ,in is the base electricity price, is the electricity price adjustment coefficient, is the absolute value of the elastic modulus; Off-peak electricity prices , , encouraging users to increase electricity consumption by reducing prices; Demand Response Compensation: For users who actively participate in load shifting / reduction, compensation will be given according to the amount of reduction: ,in To actually reduce the load, is the compensation standard; the compensation amount increases with the increase of the supply-demand gap level; Once the plan is issued, real-time electricity prices and compensation policies will be pushed through smart meters or user apps, with peak and off-peak period adjustments announced one hour in advance.

10. The source-load interaction method based on the power consumption information collection system according to claim 8, characterized in that: The adjustable load instruction scheme sends precise control instructions based on its regulation potential and grid demand to achieve load time shifting or power reduction. The adjustable load instruction generation method includes: Instruction object filtering: Based on the S2 scalability potential assessment results, select users / devices with high scalability weights and a historical response rate ≥ 80%. Prioritize devices that have minimal impact on user experience. Specific instruction content: Transferable loads: Instructing equipment with time flexibility to transfer loads during specific periods, and providing supporting subsidies and incentives; It can reduce the load by sending power adjustment range or operation duration instructions to the equipment that can temporarily reduce the power, and at the same time limit the power of a single air conditioner to no more than 70% of the rated value; Calculation of instruction execution volume: Total target for a single adjustment : ; is the safety factor, ; The gap between supply and demand; Allocate adjustment amount according to user adjustable capacity ratio : ； For the Adjustable capacity for each user; Instruction issuance channel: Instructions are issued in the form of encrypted messages through the remote control interface of the electricity consumption information collection system to ensure that the instruction transmission delay is ≤10 seconds.

11. The source-load interaction method based on the power consumption information collection system according to claim 8, characterized in that: The distributed energy scheduling plan is used to coordinate the output of distributed energy, give priority to consuming renewable energy, reduce wind and solar power curtailment, and mitigate the impact of its output fluctuations on the power grid. The steps for formulating the distributed energy scheduling plan include: Output plan formulation: Formulate daily output plans based on S2 distributed energy output curve prediction , the planned photovoltaic output needs to match the changes in light intensity, and the planned wind power output needs to refer to the wind speed forecast; The planned output must meet the grid constraints: , The capacity accepted by the grid, and the ramp rate within 15 minutes is ≤ 10% of the rated capacity; Real-time scheduling strategy: Absorption priority, when When the amount is fully absorbed, no adjustment is required; Actual output of photovoltaic / wind power; Power curtailment control, when the actual output of photovoltaic / wind power When the local energy storage system is started to charge: If the energy storage capacity is full, the output will be limited according to the principle of "wind power prioritizing photovoltaic power"; Charging power for the local energy storage system; coordinated with load regulation: When the output of distributed energy resources drops sharply, an emergency load reduction command is immediately triggered to fill the output gap; when the output of distributed energy resources increases sharply, off-peak electricity price incentives are activated to attract users to increase their electricity consumption and absorb surplus electricity; Dispatching instruction issuance: Output limit instructions are sent to photovoltaic inverters and wind turbine controllers via the RS485 interface, with accuracy controlled within ±5% of rated capacity.

12. The source-load interaction method based on the power consumption information collection system according to claim 1, characterized in that: The strategy execution and feedback method: S41. Real-time monitoring of strategy execution results. By integrating data from the power consumption information collection system and the SCADA system, the monitoring screen is updated every 5 seconds, focusing on tracking user responses and grid status. In terms of user response, calculate the electricity price incentive response rate: ； During the calculation process, it is required that the transferable load is ≥70% and the reducible load is ≥80%, and the user APP satisfaction score is collected; In terms of grid status, monitor frequency, voltage and net load curves, and calculate the peak-to-valley difference changes after strategy execution : ； is the maximum load of the power grid during the peak period before the strategy is implemented, representing the peak pressure of the power grid before implementation; It is the maximum load of the power grid during peak hours after the strategy is executed, reflecting the effect of reducing peak load after demand response; When the command execution rate is less than 60% or the frequency deviation exceeds ±0.3Hz, an audible and visual alarm is triggered to notify the dispatcher to intervene; S42. Feedback data collection and analysis. Feedback data collection covers the user side, energy side, and grid side: the user side records the actual load curve through smart meters , calculate the adjustment amount : , The energy side records the deviation between the actual output and the planned output of distributed energy. : , To formulate a daily output plan, Actual output of photovoltaic / wind power; frequency fluctuation and voltage fluctuation collected on the grid side; during analysis, calculate the strategic benefit index : , is the duration, Effective; Evaluate elastic modulus deviation , is the actual elasticity coefficient calculated after the strategy is executed, To predict the elastic coefficient; The model needs to be modified; statistical response rate differences are calculated by user type to identify high-response and low-response user groups; S43. Dynamic adjustment and optimization of strategies, dynamic optimization of strategy parameters based on feedback analysis results: electricity price adjustment coefficient The adjustment formula is: ； is the updated electricity price adjustment coefficient, which affects the incentive intensity of the user-side electricity price; is the electricity price adjustment coefficient before the update, which is the basic value for adjustment; If the actual elastic coefficient is lower than the model prediction, it will increase , instruction allocation is tilted towards high-response users, reducing the quota for low-response users; distributed energy plan output is based on Correction; in the long-term iteration, update the user elasticity coefficient and adjustable potential model weekly, review the strategy effect monthly, and retain the strategy benefit index The system automatically matches the optimal strategy with the grid status by training the strategy selection model through the Q-learning algorithm; grants "green electricity certification" and additional discounts to high-response users, and increases the compensation standards for low-response industrial users to continuously improve the adaptability of the strategy.

Images