A Simulation and Evaluation Method for Dynamic Load Regulation in Distribution Networks

Abstract

This invention relates to the field of electrical digital data processing technology, specifically to a method for simulating and evaluating power distribution networks with dynamic load regulation. The method includes: acquiring adjusted user electricity consumption data and power grid safety and economic indicators; determining each user's electricity experience sensitivity and the degree of electricity consumption deviation caused by the dynamic load regulation strategy based on historical user electricity consumption data; determining user-side experience impact indicators based on the user-side experience impact indicators; correcting the power grid safety and economic indicators based on the user-side experience impact indicators; and obtaining a comprehensive evaluation result of the power distribution network's dynamic load regulation by weighting and fusing the corrected power grid safety and economic indicators based on each user's historical electricity consumption scale. This invention quantifies user electricity experience sensitivity and load regulation deviation, establishes a predictive model for the potential impact of user behavior rebound on power grid operation, and achieves a closed-loop evaluation that balances power grid benefits and user experience.

Description

Technical Field

[0001] This invention relates to the field of electrical digital data processing technology, specifically to a method for simulating and evaluating power distribution networks with dynamic load regulation. Background Technology

[0002] In the operation of distribution networks, dynamic load regulation refers to the real-time, flexible, and orderly adjustment of adjustable loads on the user side based on real-time supply and demand conditions of the power grid, electricity price signals, or fluctuations in renewable energy output. Evaluating the distribution network simulation results of dynamic load regulation is a crucial step in verifying the effectiveness of regulation strategies. It is used not only to determine whether the strategies can achieve objectives such as peak shaving and valley filling, congestion mitigation, and network loss reduction, but also to verify the safety of power grid operation, ensuring that indicators such as node voltage and line load do not exceed limits during regulation. Currently, a multi-dimensional weighted fusion approach combining power grid safety indicators and economic indicators is generally used to evaluate the simulation results. Key quantitative indicators are extracted through power flow calculations and time-series simulations, and after normalization, a weighted sum is obtained to obtain a comprehensive evaluation value.

[0003] However, existing methods for evaluating the results of dynamic load regulation often rely solely on grid-side safety and economic indicators, neglecting the user's electricity experience, leading to poor results. In reality, a high-quality dynamic load regulation scheme not only needs to pursue grid safety, stability, and economic efficiency, but also needs to consider the user's actual electricity experience, achieving a two-way balance between grid benefits and user experience. If the regulation strategy excessively interferes with users' normal electricity consumption habits, resulting in a significant decline in user experience, it can easily trigger user resistance, leading to irrational behaviors such as electricity consumption rebound and malicious electricity use. Such behaviors will negatively impact the distribution network operation, causing voltage fluctuations, increased network losses, and a widening of the peak-valley load difference, thus contradicting the original intention of dynamic load regulation. Summary of the Invention

[0004] This invention provides a method for simulating and evaluating power distribution networks with dynamic load regulation to solve existing problems.

[0005] The present invention provides a method for simulating and evaluating power distribution networks with dynamic load regulation, which adopts the following technical solution: One embodiment of the present invention provides a method for simulation and evaluation of power distribution network dynamic load adjustment. The method includes: acquiring a load dynamic adjustment strategy; performing power flow simulation on a power distribution network simulation model based on the load dynamic adjustment strategy to acquire adjusted user electricity consumption data and power grid safety and economic indicators; determining the electricity experience sensitivity of each user and the degree of electricity deviation caused by the load dynamic adjustment strategy based on the user's historical electricity consumption data; determining user-side experience impact indicators based on the electricity experience sensitivity and the degree of electricity deviation; wherein, the electricity experience sensitivity is used to characterize the rigidity of the user's electricity consumption pattern under external environmental disturbances, and the degree of electricity deviation is used to characterize the deviation of the adjusted user electricity consumption data from the natural electricity consumption baseline in the unadjusted state; correcting the power grid safety and economic indicators based on the user-side experience impact indicators, and obtaining a comprehensive evaluation result of power distribution network load dynamic adjustment based on the weighted fusion correction of the power grid safety and economic indicators based on the historical electricity consumption scale of each user.

[0006] Furthermore, the method for obtaining the electricity consumption experience sensitivity includes: for each historical observation day of the target user, traversing all user nodes in the distribution network simulation model other than the target user, and obtaining the external environmental disturbance weight for each historical observation day based on the fluctuation degree of the electricity consumption curve of the other user nodes on the corresponding historical observation day relative to their historical average electricity consumption curve; performing a weighted average of the electricity consumption data of each historical observation day based on the external environmental disturbance weight to obtain a weighted benchmark electricity consumption curve; and obtaining the electricity consumption experience sensitivity based on the deviation degree of the electricity consumption data of each historical observation day relative to the weighted benchmark electricity consumption curve and the external environmental disturbance weight for each historical observation day.

[0007] Furthermore, the method for obtaining the degree of electricity consumption deviation includes: determining the natural electricity consumption baseline under the condition that the load dynamic adjustment strategy is not implemented based on the user's historical electricity consumption data; and obtaining the degree of electricity consumption deviation based on the degree of deviation between the adjusted user electricity consumption data and the natural electricity consumption baseline.

[0008] Further, determining the natural electricity consumption baseline under the condition of not implementing the load dynamic adjustment strategy based on the user's historical electricity consumption data includes: constructing a time-series electricity consumption dataset based on the user's historical electricity consumption data; training the user's electricity consumption prediction model with the historical electricity consumption sequence as input and the electricity consumption curve of the target time period as output to obtain the trained user's electricity consumption prediction model; and using the user's electricity consumption prediction model to perform electricity consumption prediction to obtain the natural electricity consumption baseline.

[0009] Furthermore, determining the user-side experience impact index based on the power consumption experience sensitivity and the power consumption deviation degree includes: weighting the power consumption deviation degree based on the power consumption experience sensitivity to obtain the user-side experience impact index; wherein, the user-side experience impact index is positively correlated with the power consumption experience sensitivity and the power consumption deviation degree.

[0010] Furthermore, the step of correcting the power grid safety and economic indicators based on the user-side experience impact indicators includes: determining the quantitative relationship between the user-side experience impact indicators and the potential impact of user behavior rebound on the power grid safety and economic indicators based on historical load adjustment data, and obtaining the impact correction coefficient; and using the impact correction coefficient to perform weighted correction on the power grid safety and economic indicators.

[0011] Furthermore, the step of determining the quantitative relationship between the user-side experience impact index and the potential impact of user behavior rebound on power grid safety and economic indicators based on historical load adjustment data, and obtaining the impact correction coefficient, includes: extracting the user-side experience impact index and the corresponding user electricity rebound magnitude in each adjustment event based on historical load dynamic adjustment data; wherein, the user electricity rebound magnitude is determined based on the difference in electricity consumption during a set period before and after adjustment; performing regression fitting on the user-side experience impact index and the user electricity rebound magnitude to establish the quantitative relationship; and determining the impact correction coefficient based on the quantitative relationship and the current user-side experience impact index.

[0012] Furthermore, the power grid safety and economic indicators corrected by weighted fusion based on the historical electricity consumption scale of each user include: determining the historical average electricity consumption of each user based on their historical electricity consumption data to obtain their historical electricity consumption scale; determining the evaluation weight of each user based on the ratio of their historical electricity consumption scale to the sum of all users' historical electricity consumption scales; and weighting and fusing the corrected power grid safety and economic indicators based on their evaluation weights to obtain a comprehensive evaluation result of the dynamic adjustment of the distribution network load.

[0013] Furthermore, the method for constructing the power distribution network simulation model includes: classifying user types based on user electricity consumption behavior characteristics, and constructing corresponding flexible load adjustable models for each type of flexible load; wherein, the flexible load adjustable models include air conditioning load models, water heater load models, charging pile load models, and interruptible load models; determining the adjustable power range, adjustable time period boundaries, and adjustment rate constraints of each of the flexible load adjustable models; and building a power distribution network simulation model that includes the flexible load adjustable models based on the power distribution network topology, line impedance parameters, transformer capacity, and distributed power generation output characteristics.

[0014] Furthermore, the method also includes: determining the applicability and / or optimizing the load dynamic adjustment strategy based on the comprehensive evaluation results of the power distribution network load dynamic adjustment.

[0015] The beneficial effects of the technical solution of the present invention are: In this embodiment, a load dynamic adjustment strategy is obtained; based on the load dynamic adjustment strategy, power flow simulation is performed on the distribution network simulation model to obtain adjusted user electricity consumption data and power grid safety and economic indicators; based on users' historical electricity consumption data, the electricity experience sensitivity of each user and the degree of electricity deviation caused by the load dynamic adjustment strategy are determined; based on the electricity experience sensitivity and the degree of electricity deviation, user-side experience impact indicators are determined; based on the user-side experience impact indicators, the power grid safety and economic indicators are corrected; and based on the weighted fusion correction of the power grid safety and economic indicators based on the historical electricity consumption scale of each user, a comprehensive evaluation result of the distribution network load dynamic adjustment is obtained.

[0016] This invention establishes a coupled evaluation mechanism for user-side experience impact and grid-side safety and economic indicators. By quantifying user electricity consumption experience sensitivity and load regulation deviation, it establishes a predictive model for the potential impact of user behavior rebound on grid operation, achieving a closed-loop evaluation that balances grid benefits and user experience. This overcomes the one-sidedness of traditional evaluation methods that only focus on the grid's immediate operating status while ignoring user behavior feedback. Furthermore, based on the weighting of external environmental disturbances to identify the rigidity of user electricity consumption patterns, it determines the daily weight by analyzing the fluctuation of other users' electricity consumption curves on high-disturbance days, and calculates user electricity consumption experience sensitivity accordingly. This allows for accurate differentiation between experience-sensitive users and flexible-tolerant users, providing a basis for differentiated negative impact assessments. The system provides a quantitative basis for load regulation strategy formulation. On the other hand, it uses historical load regulation data to fit the quantitative relationship between user-side experience impact indicators and electricity consumption rebound magnitude, transforming the degree of user experience impairment into a potential impact coefficient on grid safety and economic indicators. This allows the assessment results to reflect the long-term risks to grid operation caused by malicious electricity consumption behavior that may be triggered by the regulation strategy, enhancing the proactive nature of the assessment and risk prevention capabilities. Furthermore, based on the grid safety and economic indicators corrected by weighted fusion of historical electricity consumption scales of each user, the system reflects the differentiated impact of different users on the overall operation of the distribution network through the proportion of electricity consumption. This avoids the problem of underestimating the impact of large users due to equal weighting fusion, making the comprehensive assessment results more consistent with the actual operating mechanism of the distribution network. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating the power distribution network simulation and evaluation method for dynamic load regulation provided in this application embodiment; Figure 2 This is a daily electricity consumption curve of a user in a certain application scenario provided in an embodiment of this application. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following detailed description, in conjunction with the accompanying drawings and preferred embodiments, provides a specific implementation method, structure, features, and effects of the invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0021] The following description, in conjunction with the accompanying drawings, details a specific solution provided by the present invention.

[0022] like Figure 1 As shown in the figure, this application provides a method for simulating and evaluating power distribution networks with dynamic load regulation, including: Step S110: Obtain the load dynamic adjustment strategy.

[0023] The aforementioned dynamic load regulation strategy refers to a real-time power control scheme formulated for flexible loads on the user side (flexible loads include, but are not limited to, air conditioners, water heaters, charging piles, and interruptible industrial loads) during the operation of the distribution network. It achieves the goal of optimizing the supply and demand balance of the power grid by setting the start-stop periods, power adjustment range, and adjustment rate limits for each adjustable load. The dynamic load regulation strategy typically aims to reduce peak loads, fill load troughs, alleviate line congestion, reduce network losses, or smooth out fluctuations in distributed power generation output. During implementation, it is constrained by user comfort constraints, upper limits on equipment adjustment rates, and minimum start-stop times, among other technical boundary conditions. Its physical feasibility is verified through power flow calculations in a distribution network simulation model.

[0024] The aforementioned dynamic load adjustment strategies are primarily generated from two sources: grid-side dispatch instructions and market-side price signals. The former is automatically generated by the distribution network dispatch control center or Distributed Energy Resource Management System (DERMS) based on real-time monitoring of grid operation status (such as node voltage exceeding limits, line overload, and frequency deviation). The latter originates from economic incentive signals such as Real-Time Pricing (RTP), Time-of-Use Pricing (TOU), or Critical Peak Pricing (CPP) issued by the demand response management platform, as well as optimized dispatch plans formulated by the Virtual Power Plant (VPP) control center to aggregate distributed resources. Furthermore, in scenarios with a high proportion of renewable energy integration, this strategy can also originate from load tracking instructions issued by the Automatic Generation Control (AGC) system to track fluctuations in wind and solar power output, or from preventative adjustment plans formulated by the Distribution System Operator (DSO) based on ultra-short-term load forecasts and renewable energy power forecasts.

[0025] Step S120: Based on the load dynamic adjustment strategy, perform power flow simulation on the distribution network simulation model to obtain the adjusted user electricity consumption data and power grid safety and economic indicators.

[0026] The aforementioned distribution network simulation model is a digital mirror system used to characterize the physical structure and operating characteristics of the distribution network. It describes the network topology connections through node admittance matrices, integrates electrical parameters such as line impedance parameters, transformer ratios and capacity constraints, distributed generation output characteristics, and load power demand, and specifically includes a flexible load adjustable model to reflect the user-side adjustment potential. Using the distribution network topology as its framework and the electrical parameters and operating constraints of each component as attributes, the distribution network simulation model constructs a virtual experimental platform capable of simulating the evolution of the power grid state under different operating conditions, providing a computational platform for evaluating the physical feasibility of dynamic load adjustment strategies.

[0027] Power flow simulation is a numerical calculation process based on Kirchhoff's voltage and current laws to solve for steady-state operating parameters of the power grid. It establishes a set of nodal power balance equations (active and reactive power balance equations) and uses iterative algorithms such as the Newton-Raphson method or the forward-backward substitution method to solve for the voltage amplitude and phase angle of each node, the active and reactive power distribution of each branch, and the total system loss. In the above scheme, power flow simulation uses the dynamic load adjustment strategy as a boundary condition input to the distribution network simulation model. Through time-series power flow calculation, it obtains power grid safety and economic indicators such as nodal voltage, line load rate, and active power loss after the strategy is implemented. Simultaneously, it outputs the actual power consumption curves of each user under the adjustment state, providing basic data for subsequent user-side experience impact assessment.

[0028] The aforementioned adjusted user electricity consumption data refers to the actual power consumption time-series data of each user node within the evaluation period, obtained through distribution network power flow simulation calculations after implementing a specific load dynamic adjustment strategy. When flexible loads (such as air conditioners, water heaters, charging piles, etc.) respond to adjustment commands by adjusting power or shifting time periods, the power injection curves of each user obtained by the simulation model based on the node power balance equations reflect the actual state of user electricity consumption behavior deviating from the natural electricity consumption pattern under strategy intervention. This data is usually represented in discrete time series form, containing active and reactive power values ​​at each sampling time. It is a direct data source for quantifying the degree of user electricity consumption deviation and is used for comparative analysis with the natural electricity consumption baseline under unregulated conditions.

[0029] Power grid safety and economic indicators are a set of multi-dimensional quantitative parameters used to evaluate the immediate impact of dynamic load regulation strategies on the operational safety and economy of distribution networks. These typically include node voltage compliance rate (or voltage deviation), line and transformer load rates, grid active power losses, system frequency stability indicators, and power supply reliability indicators. Among these, the node voltage compliance rate reflects the ability of each node to maintain voltage within the allowable deviation range during regulation; the line and transformer load rate characterizes the utilization efficiency and overload risk of equipment after regulation; grid active power losses quantify the optimization effect of regulation strategies on reducing line and transformer resistance losses; system frequency stability indicators assess the degree of impact of power imbalance caused by regulation on system frequency deviation; and power supply reliability indicators measure the level of guarantee provided by regulation strategies for the continuous power supply capacity to users. These indicators, extracted through power flow simulation and normalized, constitute the benchmark parameters for evaluating the grid-side benefits of regulation strategies. Subsequent adjustments will incorporate user-side experience impact indicators for risk correction.

[0030] Optionally, the above-mentioned method for constructing the distribution network simulation model includes: classifying user types based on user electricity consumption behavior characteristics, and constructing corresponding flexible load adjustable models for each type of flexible load; wherein, the flexible load adjustable models include air conditioning load models, water heater load models, charging pile load models, and interruptible load models; determining the adjustable power range, adjustable time period boundaries, and adjustment rate constraints of each flexible load adjustable model; and building a distribution network simulation model that includes flexible load adjustable models based on the distribution network topology, line impedance parameters, transformer capacity, and distributed power generation output characteristics.

[0031] When constructing a power distribution network simulation model, classifying users based on their electricity consumption behavior is a prerequisite for ensuring the accuracy of flexible load modeling. This classification primarily relies on the temporal characteristics of the load curve, power density, adjustability potential, and load importance level, clustering users into three main categories: residential users, commercial users, and industrial users. Residential users are mainly temperature-controlled loads such as air conditioners and water heaters, whose electricity consumption behavior is dominated by outdoor temperature and human comfort preferences, exhibiting a clear bi-peak characteristic and seasonal fluctuations. Commercial users mainly include lighting, air conditioning, and charging pile loads; their peak electricity consumption is highly coupled with business hours, and there are significant differences between weekdays and weekends. Interruptible loads among industrial users (such as electrolytic aluminum and electric arc furnace steelmaking) have the characteristics of a large power adjustment range, fast response speed, but high continuity requirements. This classification allows for the establishment of differentiated flexible load response models for different user groups, avoiding the bias in adjustment potential assessment caused by using a uniform load model.

[0032] The aforementioned flexible load adjustable models are mathematical abstractions of the electrical characteristics and behavioral constraints of adjustable electrical equipment on the user side. For air conditioning loads, equivalent thermal parameters (ETP) models or state queue models are typically used. By establishing the thermodynamic balance equation between indoor temperature and cooling / heating power, the impact of temperature adjustment on compressor start-stop status and power consumption is quantified. The water heater load model, based on the water temperature thermodynamic equation and the water tank heat loss coefficient, describes the constraint relationship between heating element power adjustment and water temperature change rate and user water comfort. The charging pile load model needs to consider the electric vehicle battery state of charge (SOC), charging power level (slow charging / fast charging), and user charging urgency to establish an optimized scheduling space for charging power and charging duration. The interruptible load model is mainly for large industrial users. Through the contractually agreed interruption capacity, interruption duration upper limit, and interruption compensation mechanism, a quantitative relationship between power reduction and production loss costs is established. All types of models need to clearly define the upper and lower limits of their adjustable power, the allowed adjustment time window, and the power adjustment rate limit to simulate the physical constraints and user experience boundaries in the actual power grid control command execution process.

[0033] Determining the adjustable power range, adjustable time period boundaries, and adjustment rate constraints of flexible loads is a crucial step in establishing the feasible region of the simulation model. The adjustable power range refers to the maximum acceptable power adjustment range for a single load or aggregated load group without affecting the safe operation and core functions of the equipment. It is usually expressed as a percentage or absolute power value of the rated power. The adjustable time period boundaries are rigidly determined by the user's electricity consumption behavior. For example, residential water heaters are only allowed to preheat before bathing, and charging piles must complete the charging task within the parking period from vehicle entry to exit. The adjustment rate constraints reflect the physical limits of equipment power changes, such as the time interval limit for air conditioner compressor start-stop and the temperature change rate limit for industrial electric furnaces. The setting of these constraints needs to comprehensively consider the electrical characteristics of the equipment, the user comfort threshold, and the grid regulation requirements. By constructing an inequality constraint set embedded in the optimization scheduling model, it is ensured that the load dynamic adjustment strategy generated by the simulation is physically executable and will not cause excessive discomfort to the user side.

[0034] After completing the modeling and constraint definition of flexible loads, they can be integrated with the distribution network physical system model to construct a complete simulation environment. This integration process uses the distribution network topology (radial or weak ring network structure) as the connecting framework, describes the line impedance parameters (resistance, reactance) and transformer capacity constraints through the node admittance matrix, injects the output characteristic curves of distributed power sources (photovoltaics, wind power, energy storage) as time-varying power sources into the corresponding nodes, and connects the aforementioned flexible load model as a controllable load resource to the user-side nodes. Under this integrated architecture, the distribution network simulation model can simulate the power flow distribution evolution under different load dynamic adjustment strategies. By solving the node voltage equation and branch power equation, it obtains the grid safety and economic indicators after the strategy execution (including node voltage deviation, line and transformer load rate, system active power loss) and the actual power consumption data of each user node, providing a unified calculation benchmark and data interface for subsequent user-side experience impact assessment and grid-side safety and economic assessment.

[0035] It is understood that power flow simulation technology is a relatively mature existing technology. For the specific implementation scheme and working principle of the above step S120, which performs power flow simulation on the distribution network simulation model based on the load dynamic adjustment strategy, please refer to the relevant technology. This application embodiment will not repeat it.

[0036] Step S130: Determine the user's electricity experience sensitivity and the degree of electricity deviation caused by the load dynamic adjustment strategy based on the user's historical electricity consumption data. Determine the user-side experience impact index based on the user's electricity experience sensitivity and the degree of electricity deviation. Among them, the user's electricity experience sensitivity is used to characterize the rigidity of the user's electricity consumption pattern under external environmental disturbances, and the degree of electricity deviation is used to characterize the deviation of the user's electricity consumption data after adjustment from the natural electricity consumption baseline in the unadjusted state.

[0037] The aforementioned historical electricity consumption data is primarily collected via load curve data from smart metering terminals (such as smart meters and electricity information collection terminals) deployed on the user side at a fixed sampling frequency (usually every 15 minutes or 1 hour). This data is then transmitted to the distribution network data center via the communication networks of Advanced Metering Systems (AMI) or distribution automation systems (such as power line carrier, public wireless networks, or private fiber optic networks). The historical electricity consumption data mainly includes user identification information, timestamps, and active and reactive power measurements. After collection, the data undergoes cleaning processing (including missing value imputation, outlier removal, and noise filtering) to eliminate measurement errors caused by communication packet loss or equipment failure, forming continuous and reliable daily electricity consumption curves or hourly electricity consumption sequences. This serves as the foundational dataset for subsequent electricity behavior analysis and model training.

[0038] Regarding user privacy and data compliance, the aforementioned sensitive data involving user identification and electricity consumption behavior strictly adheres to relevant regulations throughout the entire lifecycle of collection, storage, and processing. Before data collection, explicit consent from users is obtained through user service agreements or specific authorization documents. During the data preprocessing stage, user identification is desensitized or anonymized (e.g., using hash encoding to replace real user identity information and removing precise geographical location information) to ensure that datasets used for model training and simulation evaluation cannot be traced back to a specific natural person. This ensures that while optimizing the operation of the distribution network by utilizing user electricity consumption characteristics, it fully protects users' personal privacy rights and data security.

[0039] The aforementioned electricity experience sensitivity is a user-side inherent characteristic parameter characterizing the rigidity of user electricity consumption behavior and the comfort tolerance threshold. Its quantification reflects the stability of a user's existing electricity consumption patterns under strong external environmental disturbances (such as drastic fluctuations in electricity prices, extreme temperature changes, or grid excitation signals). Electricity experience sensitivity can be determined by analyzing the stability of behavioral patterns in historical user electricity consumption data. High electricity experience sensitivity indicates that the user is experience-sensitive, their electricity consumption arrangements are less affected by external factors, their daily energy consumption patterns are stable, they have high requirements for electricity comfort and convenience, and tend to refuse to change their electricity consumption habits to ensure a good experience. Low electricity experience sensitivity, on the other hand, indicates that the user is flexible and tolerant, their electricity consumption arrangements are more adaptable to changes in the external environment, they have a higher tolerance for load regulation disturbances, and their adjustment potential is greater. As a bridge connecting users' subjective preferences and objective regulation strategies, electricity experience sensitivity is used to identify the differences in subjective feelings of different users towards the same load regulation strategy, providing a user-side characteristic benchmark for subsequently quantifying the differentiated impact of regulation strategies on user experience.

[0040] The embodiments of this application can obtain the above-mentioned power consumption experience sensitivity in at least one of the following ways: The first approach: a time-series analysis method based on the weight of external environmental disturbances; Optionally, the method for obtaining the above-mentioned sensitivity to electricity consumption experience includes: for each historical observation day of the target user, traversing all user nodes in the distribution network simulation model other than the target user, and obtaining the external environmental disturbance weight for each historical observation day based on the fluctuation of the electricity consumption curve of the other user nodes on the corresponding historical observation day relative to their historical average electricity consumption curve; performing a weighted average of the electricity consumption data of each historical observation day based on the external environmental disturbance weight to obtain a weighted benchmark electricity consumption curve; and obtaining the sensitivity to electricity consumption experience based on the deviation of the electricity consumption data of each historical observation day from the weighted benchmark electricity consumption curve and the external environmental disturbance weight for each historical observation day.

[0041] The aforementioned external environmental disturbance weight is a quantitative parameter that identifies the level of external environmental pressure on a specific date based on the deviation of group electricity consumption behavior. Its technical principle is that when there are significant fluctuations in electricity prices, extreme weather conditions, or grid incentive signals in the distribution network, the electricity consumption patterns of most users will deviate from their historical average behavior in a synchronous manner. By traversing the remaining user nodes except for the target user and calculating the integral of the time-series deviation (or cumulative absolute error) of their electricity consumption curves on the corresponding historical observation days relative to their respective historical average electricity consumption curves, the system-level disturbance intensity of that day can be effectively captured. The larger the external environmental disturbance weight value, the stronger the external environmental pressure on that day. Such high-weight days, as naturally generated "stress test" scenarios, can objectively test the behavioral rigidity of target users when faced with economic incentives or environmental pressures that change their electricity consumption habits, avoiding the strategic misreporting or cognitive bias that may be caused by relying on users' subjective declarations of comfort preferences.

[0042] The aforementioned weighted baseline electricity consumption curve is a weighted average sequence characterizing the electricity consumption behavior patterns of target users under typical external environmental pressures. Its construction process differs from a simple arithmetic average; instead, it uses weighted averages of electricity consumption data from each historical observation day based on the aforementioned external environmental disturbance weights. This ensures that the electricity consumption characteristics of days with high disturbances have a greater weight in the baseline construction. The technical purpose of this weighting mechanism is to capture the "normal" electricity consumption patterns of users when facing strong external stimuli, rather than a mixed average including a large number of low-pressure, stable days. This establishes a reference baseline that reflects the user's true comfort threshold and the rigidity of their electricity consumption habits, providing a benchmark for subsequent quantification of user behavior stability under different pressure levels.

[0043] In the above implementation, the final quantification of electricity consumption experience sensitivity is achieved by coupling the deviation degree of each historical observation day with the corresponding external environmental disturbance weights. Specifically, for each historical observation day, the deviation measure (such as time-domain integral absolute difference, root mean square error, or shape similarity) between its actual electricity consumption curve and the weighted benchmark electricity consumption curve is calculated, and then the deviation degree is weighted and combined using the external environmental disturbance weights. In terms of technical implementation, if an inverse proportional mapping function (such as an exponential decay function) is used to convert the deviation degree into a stability score, then if a high-weight day (strong disturbance day) shows a small electricity consumption deviation, it will significantly increase the final sensitivity value, indicating that the user belongs to the experience-sensitive type who still adheres to the established electricity consumption pattern when the external environment changes drastically. Conversely, it indicates that the user has strong electricity consumption flexibility. This scheme achieves objective reverse inference of the user's subjective experience preferences by passively observing the user's real response to naturally occurring power grid disturbance events.

[0044] The above-mentioned technical solutions require the distribution network to have accumulated historical measurement data covering multiple users and time periods, and the external environment must have sufficient volatility to generate effective daily samples of strong disturbances. Compared with user preference acquisition methods based on questionnaires or laboratory simulations, the advantages of the above solutions are that they use price signals and meteorological disturbances naturally generated in actual power grid operation as incentive sources, which can avoid the social expectation bias when users actively disclose their preferences. Furthermore, through multi-user cross-validation (defining environmental disturbances by the behavior of other users), the individual differences of users and common system shocks are effectively separated. However, it should be noted that in scenarios where the distribution network load is highly homogeneous (such as a single industrial load) or the external environment is extremely stable (such as a constant temperature environment without seasonal temperature changes), this method may need to be combined with other auxiliary means due to the lack of effective disturbance samples.

[0045] Furthermore, the calculation of the aforementioned external environmental disturbance weights can be replaced by other distance metrics instead of the absolute difference of the time-series integral. For example, Dynamic Time Warping (DTW) distance can be used to capture the phase shift of the electricity consumption curve rather than the amplitude difference, or Pearson correlation coefficient can be used to measure the shape similarity between the daily curve and the average curve. The construction of the weighted reference curve can also introduce the entropy weight method or Principal Component Analysis (PCA) to process high-dimensional time-series data. The quantification of the degree of deviation can also use Friesian distance or cosine similarity, which emphasize the geometric characteristics or trend consistency of the curve. These can be adaptively adjusted according to the load composition characteristics and data quality of the specific distribution network. It is understood that the aforementioned Dynamic Time Warping distance, Pearson correlation coefficient, entropy weight method, principal component analysis, Friesian distance, cosine similarity, etc., are all mature existing technologies. For the specific implementation schemes and working principles of each technology, please refer to the relevant technologies. The embodiments of this application will not be repeated here.

[0046] The second approach is to use a statistical analysis method based on the user's own electricity consumption variation coefficient. In this implementation, the temporal stability of a user's electricity consumption behavior can be directly measured by calculating the root mean square error or standard deviation of the user's historical electricity consumption curve and its own historical average electricity consumption curve at each time period. The coefficient of variation (the ratio of the standard deviation to the mean) of the user's electricity consumption curve is used as a sensitivity measure. The smaller the coefficient of variation, the more stable the user's electricity consumption habits and the higher the resistance to changes in electricity consumption patterns. Conversely, the larger the coefficient of variation, the stronger the randomness of the user's electricity consumption behavior and the higher the tolerance to regulatory interference.

[0047] The third approach: behavioral analysis based on historical demand response participation behavior; In this implementation, by retrospectively analyzing historical data of users' past participation in grid demand response events, the system statistically analyzes users' load reduction / transfer response rate, response delay time, and response volatility after receiving price or incentive signals. Users with lower response rates, longer response delays, or greater response volatility are considered to have higher sensitivity to electricity usage experiences, indicating stronger resistance to external regulatory interventions or greater difficulty in changing their electricity usage habits.

[0048] The fourth approach is to use a machine learning method based on multi-dimensional feature clustering to extract multi-dimensional feature vectors such as peak-valley difference rate, load rate, daily electricity consumption pattern entropy value, and seasonal fluctuation index from the user's historical electricity consumption curve. Using K-means clustering or hierarchical clustering algorithms, users are divided into experience-sensitive and flexible-tolerant groups. By calculating the Euclidean distance or Mahalanobis distance between the target user and the center of each type of cluster, the probability value of the user belonging to the experience-sensitive group is determined, which serves as a quantitative representation of electricity consumption experience sensitivity.

[0049] It is understood that the K-means clustering and hierarchical clustering algorithms mentioned above are mature clustering algorithms. For their clustering schemes and working principles, please refer to the relevant technologies. The embodiments in this application will not be repeated.

[0050] Optionally, the above method for obtaining the degree of electricity consumption deviation includes: determining the natural electricity consumption baseline under the condition of not implementing the load dynamic adjustment strategy based on the user's historical electricity consumption data; and obtaining the degree of electricity consumption deviation based on the degree of deviation between the adjusted user electricity consumption data and the natural electricity consumption baseline.

[0051] The aforementioned natural electricity consumption baseline refers to the inherent power consumption time-series curve formed by users based on their own living habits, production arrangements, and physical environment requirements under ideal conditions where no load dynamic adjustment strategies are implemented and there is no intervention from external price signals or grid incentives. The natural electricity consumption baseline represents the electricity consumption pattern that users naturally generate to meet comfort preferences, production process requirements, or basic living needs. It reflects the inherent regularity and stability characteristics of users' electricity consumption behavior and serves as a reference benchmark for assessing the intensity of load dynamic adjustment strategies' intervention in users' electricity consumption behavior. By quantifying the degree of deviation of actual electricity consumption data after adjustment from this baseline, the cumulative effect of adjustment strategies forcing users to deviate from their natural electricity consumption needs can be objectively measured.

[0052] The embodiments of this application can obtain the above-mentioned natural power consumption baseline in at least one of the following ways: The first method: using deep learning prediction methods based on Long Short-Term Memory (LSTM) networks; Optionally, the above-mentioned determination of the natural electricity consumption baseline under the condition of not implementing the load dynamic adjustment strategy based on the user's historical electricity consumption data includes: constructing a time-series electricity consumption dataset based on the user's historical electricity consumption data; training the user electricity consumption prediction model with the historical electricity consumption sequence as input and the electricity consumption curve of the target time period as output to obtain the trained user electricity consumption prediction model; and using the user electricity consumption prediction model to perform electricity consumption prediction to obtain the natural electricity consumption baseline.

[0053] The aforementioned time-series electricity consumption dataset is a sample set formed by structurally reorganizing and feature-engineering user historical electricity consumption data. Its construction process first involves quality control of the original measurement data (usually active power sequences sampled at fixed time intervals), including missing value imputation (such as using linear interpolation or trend filling after seasonal decomposition), outlier detection (identifying and correcting outliers based on the 3σ criterion or isolated forest algorithm), and normalization processing (such as Min-Max scaling or Z-score standardization) to eliminate dimensional differences and numerical range fluctuations. Subsequently, input-output sample pairs are constructed using the sliding window technique, setting the length of the historical observation sequence (such as electricity consumption data from the previous 7 days or the previous 24 hours) as the model input feature, and the 24-hour electricity consumption curve corresponding to the prediction day or the power sequence of a specific period as the target output, forming a time-series sample matrix suitable for supervised learning. The quality of this dataset directly determines the ability of the subsequent prediction model to capture the periodic and random characteristics of user electricity consumption.

[0054] The user electricity consumption prediction model employs a Long Short-Term Memory (LSTM) network architecture. By introducing a forget gate, input gate, and output gate, the LSTM network effectively solves the gradient vanishing and gradient exploding problems encountered by traditional recurrent neural networks when processing long sequences. It can learn long-term dependencies in user electricity consumption behavior (such as differences in weekday and weekend electricity consumption patterns, and the impact of seasonal temperature on air conditioning load) and short-term fluctuation characteristics. In terms of network topology design, a single hidden layer or multi-layer stacked LSTM structure can be used, with each layer containing tens to hundreds of memory units. End-to-end training is performed using the Backpropagation Through Time (BPTT) algorithm. Alternatively, a Bi-LSTM can be introduced to simultaneously utilize historical and future contextual information, or an attention mechanism can be combined to dynamically weight the contribution of different historical moments to the prediction output, improving the prediction accuracy for sudden events (such as holidays). It is understood that the aforementioned Backpropagation Through Time (BPTT) algorithm for Long Short-Term Memory Networks and Bidirectional LSTM (Bi-LSTM) are all mature existing technologies. For specific implementation schemes and working principles of each technology, please refer to the relevant technologies. The embodiments in this application will not be repeated.

[0055] The training process of the aforementioned user electricity consumption prediction model aims to minimize the loss function between the predicted output and the actual natural electricity consumption. Mean Squared Error (MSE) or Mean Absolute Error (MAE) is typically used as the loss metric. Adaptive learning rate optimization algorithms such as Adam or RMSprop are used for iterative parameter updates, and Dropout regularization or L2 weight decay is introduced to prevent overfitting. In terms of training strategy, time series cross-validation is used instead of random partitioning to ensure that the validation set is always located after the time series of the training set, thereby realistically evaluating the model's generalization ability to future unknown data. During training, the changing trend of the validation set loss is monitored, and an early stopping strategy is implemented to avoid overtraining. Finally, the model parameters with the best validation performance are saved as the trained user electricity consumption prediction model.

[0056] During the model deployment and baseline generation phase, the trained network is used to perform forward propagation calculations on the target prediction period. The input is the electricity consumption sequence of the most recent historical period and optional external features (such as weather forecasts and date codes). The output is the predicted value of the natural electricity consumption baseline under unregulated conditions. To improve the robustness of the baseline estimation, an ensemble learning strategy can be adopted, such as training multiple LSTM models with different initializations or architectures and taking the average of the predictions, or combining Bayesian deep learning to quantify the prediction uncertainty. When the confidence of the prediction interval is low, a data supplementation or model retraining mechanism is triggered to ensure that the natural electricity consumption baseline used for subsequent deviation calculations can accurately reflect the user's real electricity demand under uninterrupted conditions.

[0057] The second approach: The prediction method based on time series statistical models uses the Autoregressive Integrated Moving Average (ARIMA) model or its seasonal extension (Seasonal Auto Regressive Integrated Moving Average, SARIMA) to obtain the data. In this implementation, the model order can be determined by analyzing the autocorrelation and partial autocorrelation functions of historical electricity consumption data. After eliminating non-stationarity using difference operations, a linear prediction equation can be established. Alternatively, the Holt-Winters seasonal exponential smoothing method can be used to recursively decompose and smooth the trend, seasonal, and stochastic components of the electricity consumption series, thereby extrapolating to obtain the baseline load for the prediction period. The above implementation is suitable for scenarios where electricity consumption patterns exhibit significant periodicity and trends. Its parameters are highly interpretable and computationally complex, making it easy to deploy and implement on resource-constrained edge computing nodes.

[0058] It is understood that autocorrelation functions, Holt-Winters seasonality index smoothing methods, etc. are all mature existing technologies. For specific implementation schemes and working principles of each technology, please refer to the relevant technologies. The embodiments in this application will not be repeated.

[0059] The third method: obtaining the data based on the statistical average of historical data from the same period; In this embodiment, historical electricity consumption curves with similar attributes to the target prediction day (such as the same weekday type, similar weather conditions, or the same seasonal period) can be selected, and a typical daily electricity consumption pattern can be constructed as a natural electricity consumption baseline using arithmetic mean, weighted moving average, or median aggregation. This embodiment assumes that user electricity consumption behavior is repeatable under similar external conditions. It is simple to implement and does not require a complex model training process, making it suitable for scenarios with relatively stable electricity consumption patterns and sufficient historical data accumulation.

[0060] The fourth method: obtaining baseline estimation based on multiple regression analysis; In this implementation, exogenous variables affecting user electricity demand (such as ambient temperature, humidity, irradiance, date type coding, and economic indicators) can be used as explanatory variables to establish linear or nonlinear regression relationships with electricity consumption. Model coefficients are determined through least squares estimation or regularized regression (such as ridge regression or Lasso regression), and then the expected electricity consumption is calculated based on weather forecasts and calendar information for the predicted period as the natural electricity consumption baseline. This implementation can explicitly quantify the influence weight of external environmental factors on electricity consumption behavior and is suitable for baseline estimation of temperature-controlled loads (such as air conditioning and heating) significantly affected by weather conditions, but it relies on the accuracy and timeliness of exogenous variable data.

[0061] The aforementioned deviation in electricity consumption can be expressed in at least one of the following forms: The first method quantifies the cumulative power deviation based on time integration. This method integrates (or sums in the discrete time domain) the absolute difference in instantaneous power between the adjusted user power consumption data and the natural power consumption baseline at each sampling time to obtain the total power deviation within the evaluation period. Its physical meaning is the total amount of additional or reduced power consumption forced by the load dynamic adjustment strategy. It can intuitively reflect the cumulative intervention intensity of the adjustment strategy on user power consumption behavior and is suitable for evaluating the effectiveness of adjustment strategies aimed at peak shaving and valley filling or energy transfer.

[0062] The second approach employs a deviation measure based on statistical distance. This approach quantifies the overall deviation level of the two curves in terms of amplitude by calculating the root mean square error (RMSE) or mean absolute error (MAE) of the power difference between the regulated electricity consumption curve and the natural electricity consumption baseline at various times. This type of metric has different sensitivities to extreme deviation times (RMSE assigns higher weight to large deviations, while MAE remains linearly sensitive), making it suitable for assessment scenarios that focus on the risk of high power deviation during specific periods in the regulation process, and facilitating standardized comparisons across users and time periods.

[0063] The third approach: In scenarios where it is necessary to eliminate differences in electricity consumption among users, a relative deviation approach can be used. This approach obtains a percentage deviation index by dividing the absolute electricity deviation or statistical distance measure by the total electricity consumption or average power of the natural electricity consumption baseline. This normalization process makes the degree of deviation between users with different electricity consumption volumes comparable, avoiding the problem of large users being over-penalized due to large absolute deviations and small users being underestimated due to small absolute deviations. It is suitable for building a fair user experience impact assessment system.

[0064] The fourth form adopts a deviation form based on curve shape similarity. By calculating the cosine similarity, Pearson correlation coefficient, or Fraser distance between the adjusted electricity consumption curve and the natural electricity consumption baseline, it focuses on the shape characteristics, phase shift, or trend consistency of the electricity consumption curve rather than simple amplitude differences. This form is suitable for evaluating the degree of interference of the adjustment strategy on the rhythm of users' electricity consumption habits (such as the shift of peak electricity consumption periods) and can distinguish the differentiated impact of different adjustment modes such as overall shifting and peak shaving and valley filling on user experience.

[0065] Optionally, the above-mentioned determination of user-side experience impact indicators based on power consumption experience sensitivity and power consumption deviation degree includes: weighting the power consumption deviation degree based on power consumption experience sensitivity to obtain user-side experience impact indicators; wherein, the user-side experience impact indicators are positively correlated with power consumption experience sensitivity and power consumption deviation degree.

[0066] The weighted coupling mechanism between electricity experience sensitivity and the degree of electricity deviation is based on the technical principle of unified quantification of both subjective perception and objective physical quantities. Since the impact on user experience is essentially a comprehensive effect of objective electricity behavior intervention (degree of deviation) modulated by the user's subjective tolerance (experience sensitivity), the same amount of physical deviation produces significantly more subjective discomfort and resistance for users with high electricity experience sensitivity (i.e., those with rigid habits and high comfort requirements) than for users with low sensitivity and flexible tolerance. Therefore, by using electricity experience sensitivity as a weighting coefficient to weight the degree of electricity deviation, a personalized conversion is achieved, mapping the intensity of objective physical regulation to the degree of subjective experience impairment. This overcomes the limitations of traditional assessment methods that ignore individual user differences and apply a uniform deviation threshold to all users for a one-size-fits-all assessment.

[0067] The positive correlation between the user experience impact index and the sensitivity to electricity usage and the degree of electricity usage deviation characterizes the mathematical property that the index monotonically increases with the two input variables. That is, when the user's sensitivity to electricity usage increases or the degree of electricity usage deviation increases, the user experience impact index increases accordingly, indicating that the overall negative impact of the adjustment strategy on the user experience intensifies. This positive correlation can be achieved through a linear product relationship, or through a monotonically increasing nonlinear function or piecewise linear mapping, ensuring that highly sensitive users receive a higher experience impact score when experiencing the same physical deviation, thereby accurately identifying high-risk scenarios where the adjustment strategy may trigger strong user resistance.

[0068] At the level of specific technical implementation, in addition to simple arithmetic multiplication and weighting, various coupling function forms can be adopted to adapt to different user behavior assumptions and assessment accuracy requirements. For example, threshold-based logical weighting can be used, where the deviation from the electricity consumption level is determined to have no significant impact when it is below a specific user tolerance threshold, and the impact is amplified proportionally to the sensitivity level after exceeding the threshold, in order to simulate the nonlinear abrupt change characteristics of the user's psychological tolerance. Alternatively, power function weighting can be used, where the contribution elasticity of sensitivity and deviation to the final experience impact can be controlled by adjusting the exponential parameter. When the parameter is high, the impact of highly sensitive users is amplified exponentially, which is suitable for scenarios that strictly protect users with high comfort requirements. Saturation functions such as sigmoid or tanh can also be introduced to perform nonlinear transformation on the product result, so that when the deviation or sensitivity is extremely high, the experience impact index converges to the upper limit value, avoiding the excessive dominance of extreme outliers on subsequent risk assessment and ensuring the robustness of the assessment system.

[0069] The technical value of the aforementioned coupled quantification mechanism lies in establishing a quantitative bridge between the subjective experience of users and the objective adjustment strategies of the power grid. This allows the soft indicator of "user experience," which was originally difficult to observe directly, to be transformed into a calculable, comparable, and optimizable hard parameter through historical electricity consumption data mining and mathematical modeling. By embedding electricity consumption experience sensitivity as a priori user characteristic into the evaluation process, the evaluation results are made adaptive to user heterogeneity. This provides a quantitative decision-making basis for distribution network operators to formulate differentiated and refined dynamic load adjustment strategies (such as implementing flexible adjustment for highly sensitive users and deep adjustment for low-sensitive users). Thus, while ensuring the efficiency of power grid operation, it proactively prevents the risk of malicious electricity consumption rebound caused by the deterioration of user experience.

[0070] Step S140: Correct the power grid safety and economic indicators based on the user-side experience impact indicators, and obtain the comprehensive evaluation results of the power grid safety and economic indicators after weighted fusion based on the historical electricity consumption scale of each user.

[0071] Traditional power grid safety and economic indicators only characterize the direct and immediate optimization effects of load dynamic regulation strategies on distribution network operation, such as reducing network losses, improving voltage quality, and alleviating equipment overload. However, they fail to encompass the potential impact of user-side behavioral feedback triggered after the implementation of regulation strategies on power grid operation. By introducing user-side experience impact indicators as risk correction factors, the potential deterioration effect of malicious electricity consumption behavior (such as electricity consumption rebound and irrational high-power startup) caused by impaired user experience on power grid safety and economic indicators can be incorporated into the current evaluation cycle, achieving a comprehensive trade-off between the "explicit immediate benefits" and "implicit potential costs" of regulation strategies. This correction mechanism ensures that the evaluation results not only reflect the current optimization level of the power grid's operation but also possess the proactive ability to predict the risk of user behavior rebound. This guides distribution network operators to proactively avoid regulation schemes that may bring short-term reductions in network losses but could trigger strong user resistance and ultimately lead to long-term deterioration of power grid operation quality when formulating load regulation strategies, truly achieving multi-objective synergistic optimization of power grid operation safety, economy, and user-side acceptability.

[0072] The above step S140 can correct the power grid security and economic indicators in at least one of the following ways: The first method: Obtaining data through regression analysis based on historical load adjustment data; Optionally, the above-mentioned correction of power grid safety and economic indicators based on user-side experience impact indicators includes: determining the quantitative relationship between user-side experience impact indicators and the potential impact of user behavior rebound on power grid safety and economic indicators based on historical load regulation data, and obtaining the impact correction coefficient; and using the impact correction coefficient to perform weighted correction of power grid safety and economic indicators.

[0073] The aforementioned impact correction coefficient is a dimensionless risk transformation parameter constructed based on historical behavioral data mining. It is used to quantify the mapping strength of the potential deterioration effect of user experience impact indicators on power grid safety and economic indicators. The impact correction coefficient reflects the efficiency of the transformation of user experience impairment at a specific level (such as subjective discomfort caused by the forced change of electricity consumption patterns) into negative impacts on indicators such as power grid losses, voltage quality, and equipment load rate through malicious electricity consumption behaviors (such as retaliatory high-power startup and irrational continuous electricity consumption after the adjustment ends). Its value is strongly positively correlated with the rigidity of user behavior. When the user group has a low tolerance for adjustment strategies or historical data shows that a high experience impact is inevitably accompanied by a large-scale rebound in electricity consumption, the coefficient tends to be less than 1, indicating that the corresponding risk reduction value needs to be deducted in the original power grid benefit assessment.

[0074] One possible implementation of the weighted correction of power grid safety and economic indicators using the aforementioned impact correction coefficient is as follows: The original power grid safety and economic indicators are coupled with the impact correction coefficient through multiplication. That is, the corrected power grid safety and economic indicators are equal to the original indicator value multiplied by the coefficient. Under this mechanism, if the user-side experience impact indicator is low, causing the impact correction coefficient to approach 1, the original power grid safety and economic indicators remain essentially unchanged, indicating that the current adjustment strategy achieves power grid operation benefits without triggering significant subsequent risks. If the user-side experience impact indicator is high, causing the impact correction coefficient to be significantly less than 1 (or close to 0), the original power grid safety and economic indicators are significantly reduced, indicating that although the strategy may bring considerable reduction in network losses or congestion relief in the short term, the accompanying high user backlash risk will severely erode or even offset these immediate benefits. This weighted correction mechanism achieves a proactive unified quantification of the immediate optimization benefits of power grid operation and the potential backlash risk of user behavior.

[0075] Optionally, the above-mentioned determination of the quantitative relationship between user-side experience impact indicators and the potential impact of user behavior rebound on power grid safety and economic indicators based on historical load adjustment data, and the acquisition of impact correction coefficients, includes: extracting user-side experience impact indicators and corresponding user electricity rebound magnitudes in each adjustment event based on historical load dynamic adjustment data; wherein, the user electricity rebound magnitude is determined based on the difference in electricity consumption during a set period before and after adjustment; performing regression fitting on user-side experience impact indicators and user electricity rebound magnitudes to establish a quantitative relationship; and determining the impact correction coefficients based on the quantitative relationship and the current user-side experience impact indicators.

[0076] The aforementioned rebound in user electricity consumption is a key technical indicator for quantifying the intensity of irrational electricity consumption behavior after the termination of dynamic load adjustment strategies. It captures the compensatory or retaliatory electricity consumption effect caused by impaired user experience by comparing the difference in electricity consumption within a specific time window before and after the adjustment intervention. The rebound in user electricity consumption is usually defined as the absolute difference or relative rate of change between the total electricity consumption on the first day after the adjustment ends (or a specific recovery period such as 24 hours or 48 hours) and the electricity consumption in the baseline period of the same duration before the adjustment. It can also be derived from metrics such as peak power deviation and decrease in load curve similarity. Its technical essence lies in transforming the subjective resistance of users, which is difficult to observe directly, into an abnormal increase in electricity consumption that can be accurately measured, providing a dependent variable data basis for establishing a quantitative bridge between user experience and grid operation risk.

[0077] The regression fitting described above is a statistical modeling method for establishing a functional mapping relationship between user experience impact indicators and the magnitude of user electricity consumption rebound. Its technical principle lies in using a sample set composed of historical adjustment events and an optimization algorithm to find the best fitting function that minimizes the error between the model's predicted value and the actual observed rebound magnitude. At the implementation level, classic Ordinary Least Squares (OLS) regression can be used, assuming a proportional relationship between the two, suitable for scenarios with relatively mild user behavior responses. Alternatively, multinomial regression or exponential regression can be used to capture the saturation or explosive characteristics of a sharp, non-linear increase in the rebound magnitude under high experience impact. For cases where there are anomalous outliers in historical data (such as abnormal electricity consumption caused by extreme weather combined with regulation), robust regression techniques such as M-estimation or the Random Sample Consensus (RANSAC) algorithm can be introduced. By reducing the weight of outlier samples, this ensures that the model parameters reflect the typical correlation between experience and rebound rather than being distorted by extreme values. It is understood that the aforementioned least squares linear regression, multinomial regression, exponential regression, and robust regression are all mature existing technologies. For specific implementation schemes and working principles of each technology, please refer to relevant technologies; this application's embodiments will not elaborate further.

[0078] Extracting and constructing samples of historical load dynamic adjustment data is a prerequisite for model training. This involves retrieving completed load adjustment event records from the log database of the distribution network energy management system or demand response management platform. For each historical event, the user-side experience impact index of each user at the time of adjustment is calculated back as the independent variable, and the rebound magnitude is calculated by comparing the actual electricity consumption within a set period after the adjustment with the baseline period before adjustment as the dependent variable, forming data pairs for regression analysis. Strict data quality control must be implemented during this process to remove historical samples associated with equipment failures, communication interruptions, extreme weather conditions, or other external emergencies. This ensures that the constructed regression model purely reflects the psychological feedback mechanism of user behavior to the adjustment strategy, rather than a spurious correlation influenced by confounding factors. Simultaneously, the time window definition for each event (e.g., 24 hours before adjustment to 24 hours after adjustment) must be standardized to ensure consistency in calculation methods across samples.

[0079] The process of determining the impact correction coefficient based on the quantitative relationship obtained from the fitting is essentially a technical step of transforming the predicted rebound in user behavior into a risk discount for power grid safety and economic indicators. The user-side experience impact index, calculated using the current load dynamic adjustment strategy to be evaluated, is substituted into a regression model trained with historical data to obtain the predicted user electricity consumption rebound magnitude. Then, through normalization mapping (such as comparing the rebound magnitude with the maximum rebound threshold allowed for safe operation of the power grid) or quantile transformation based on the historical rebound magnitude distribution, the predicted value is mapped to an impact correction coefficient between 0 and 1. This coefficient is numerically inversely proportional to the predicted rebound magnitude; that is, when the model predicts a significant rebound in user electricity consumption, the impact correction coefficient approaches 0, indicating a significant reduction in the original power grid safety and economic indicators to reflect potential risks; conversely, it approaches 1, retaining the original indicator value. This achieves proactive quantification of user behavior risks based on historical experience.

[0080] The second method: obtaining the result using a piecewise linear correction method based on a preset threshold; In this implementation, a tiered response mechanism is established to achieve differentiated risk weight allocation. Specifically, by pre-setting three threshold ranges (low, medium, and high) for user experience impact indicators, when the indicator is in the low impact range (below the first threshold), the risk of user behavior rebound is considered negligible, and the impact correction coefficient is set to 1 (i.e., no correction is made to the original power grid indicators). When the indicator is in the medium impact range, a linear interpolation method is used to calculate the correction coefficient, which decreases from 1 to a preset lower limit as the degree of experience impact increases. When the indicator is in the high impact range (above the second threshold), an exponential decay or fixed strong penalty coefficient is used to significantly reduce the power grid safety and economic indicators to reflect the high rebound risk. This implementation has low computational complexity and clear physical meaning, making it easy to deploy and implement quickly in the real-time control system of the distribution network. It is suitable for online evaluation scenarios with high requirements for computational timeliness.

[0081] The third method: obtaining the result using a flexible correction method based on fuzzy logic reasoning; In this implementation, the uncertainty of user behavior rebound risk is addressed by introducing linguistic variables and expert rules. Specifically, user experience impact indicators and power grid safety and economic indicators are fuzzified into low, medium, and high linguistic values, respectively. A fuzzy rule base is established in the form of "if the user experience impact is high and the power grid safety and economic indicators are excellent, then the corrected indicators are downgraded to medium level." The activation intensity of each rule is calculated using a membership function, and the centroid method or maximum membership method is used to defuzzify and obtain accurate impact correction coefficients. This implementation can integrate the experience and knowledge of power grid dispatching experts, handle scenarios where historical data is scarce or sudden changes in user behavior patterns cause statistical regularities to fail, and enhance the robustness and interpretability of the evaluation system under uncertain environments.

[0082] Optionally, the power grid security and economic indicators, which are corrected by weighted fusion based on the historical electricity consumption of each user, include: determining the historical average electricity consumption of each user based on their historical electricity consumption data to obtain their historical electricity consumption scale; determining the evaluation weight of each user based on the ratio of their historical electricity consumption scale to the sum of the historical electricity consumption scales of all users; and weighting and fusing the corrected power grid security and economic indicators based on their evaluation weights to obtain the comprehensive evaluation result of the dynamic adjustment of the distribution network load.

[0083] The aforementioned historical electricity consumption scale is a parameter representing the grid influence based on the statistical characteristics of users' long-term electricity consumption behavior. It typically uses the daily average electricity consumption or peak load of several historical days prior to the assessment period (such as the last 30 days or quarterly data) as the measurement benchmark. The technical essence of historical electricity consumption scale lies in reflecting the proportion and power contribution of a specific user in the total load composition of the distribution network. The larger the user's electricity consumption scale, the more significant the global impact of fluctuations in their electricity consumption patterns, response to adjustment strategies, and changes in user satisfaction on the power flow distribution, node voltage levels, and network losses. Therefore, using it as a fusion weight can objectively characterize the differentiated importance of different users in the overall operating status of the distribution network, avoiding the assessment distortion problem caused by using simple arithmetic averages, which underestimates the impact of large users and amplifies the impact of small users.

[0084] The determination of the above assessment weights is based on the principle of normalized proportional allocation. By calculating the ratio of the historical electricity consumption of a single user to the total historical electricity consumption of all users within the assessment scope, the influence of each user is mapped to a normalized weight coefficient. This coefficient satisfies the non-negativity and normalization constraints (i.e., the sum of the weights of each user is 1), and mathematically constitutes a probability measure. This allows the weighted fusion result to be interpreted as the expected value with electricity consumption as the probability distribution. This weight conforms to the physical mechanism of distribution network operation—the voltage quality, network loss level, and congestion degree of the distribution network are mainly dominated by the load characteristics of high-power users and their response behavior to regulation strategies, while the local fluctuations of small-scale users have a limited impact on global indicators. Therefore, weighting based on electricity consumption can ensure that the comprehensive assessment result truly reflects the dominant contradictions and key risk points of distribution network operation.

[0085] The weighted fusion technology employs a weighted summation operator, which multiplies each user's power grid safety and economic indicators (such as the corrected network loss reduction rate and voltage qualification rate) after the impact correction coefficient is applied, and then sums them up to obtain a system-level comprehensive evaluation result of the dynamic adjustment of the distribution network load. This process is mathematically equivalent to calculating the weighted arithmetic mean of the corrected indicators for the user group, where the weight vector is determined by the historical electricity consumption scale. The fusion result quantifies the overall applicability of the strategy after considering the risk of user behavior rebound.

[0086] In the above schemes, in addition to the weighting scheme based on historical average electricity consumption, a hierarchical weighting method based on load importance level can also be adopted. Users are divided into different levels according to load criticality (such as primary load, secondary load, tertiary load) or power supply reliability requirements, and are assigned differentiated basic weights, which are then fine-tuned in combination with the scale of electricity consumption. Alternatively, a topology weighting based on node electrical distance or voltage sensitivity can be adopted, giving higher weights to sensitive node users located at the end of the grid and with high voltage regulation difficulty, so as to reflect the difference in their marginal contribution to the safe and stable operation of the grid. Dynamic weighting in the time dimension can also be introduced, increasing the weight of large users during peak load periods to highlight their key role in alleviating congestion, and appropriately reducing the weight during off-peak load periods, thereby achieving adaptive matching of the evaluation system to the spatiotemporal characteristics of the grid operation status.

[0087] Optionally, the above-mentioned distribution network simulation and evaluation method for dynamic load regulation also includes: based on the comprehensive evaluation results of dynamic load regulation in the distribution network, determining the applicability of the dynamic load regulation strategy and / or optimizing it.

[0088] In the above scheme, the applicability determination of the comprehensive evaluation results of the dynamic adjustment of the distribution network load based on the preset evaluation threshold is a compliance review mechanism that compares the comprehensive evaluation results with a single threshold or a multi-level threshold range to determine whether the strategy achieves an acceptable balance between the benefits of grid operation and the risks of user behavior. When the evaluation result is higher than or equal to the preset qualified threshold, the strategy is deemed to be applicable to the current distribution network operation scenario and is allowed to enter the execution phase or maintain the current operation. When the evaluation result is lower than the threshold, the strategy is deemed to have an excessively high potential rebound risk caused by the deterioration of user experience, or the economic benefits of grid security are insufficient to offset the risk, triggering a strategy rejection or suspension mechanism to prevent high-risk strategies from causing irreversible negative impacts on the operation of the distribution network.

[0089] When the applicability judgment fails or the evaluation result is in the critical area, the parameters of the original load dynamic adjustment strategy are optimized by targeted correction and iterative optimization based on the user-side experience impact indicators and power grid safety and economic indicators decomposed from the comprehensive evaluation results. This includes reducing the adjustable load set based on the list of users with high electricity experience sensitivity, adjusting the power reduction magnitude or shift time based on the period of large electricity consumption deviation, redistributing the priority of adjustment tasks based on the weight of historical electricity consumption scale, or adjusting the price incentive level to change the user behavior response characteristics. Through the closed-loop feedback of strategy generation, simulation evaluation, and parameter correction, the Pareto optimal frontier of maximizing power grid operation benefits and minimizing user experience loss is gradually approached.

[0090] The above scheme constructs a closed-loop decision-making system for dynamic adjustment of distribution network load, breaking through the limitations of one-time design and rigid execution in the traditional open-loop strategy formulation mode, enabling the strategy to have the intelligent characteristic of self-correction based on evaluation feedback; a risk firewall is established through applicability judgment, and the adaptive evolution of strategy parameters is achieved through optimization and adjustment, ultimately improving the allocation efficiency of demand response resources and users' long-term willingness to participate, and realizing the coordinated optimization of safe and economical operation of distribution network and high-quality user service.

[0091] To facilitate understanding of the working principle of the above-mentioned distribution network simulation and evaluation method for dynamic load regulation, this application embodiment also provides a specific application case of this method in a certain application scenario. In this application scenario, the above-mentioned distribution network simulation and evaluation method for dynamic load regulation mainly includes: Step 1: Obtain the load dynamic adjustment strategy. Based on the load dynamic adjustment strategy, perform power flow simulation on the distribution network simulation model to obtain the adjusted user electricity consumption data and power grid safety and economic indicators. Collect all historical electricity consumption data from the user side, construct a flexible load model, execute dynamic load adjustment strategies, conduct distribution network power flow simulation, and obtain grid-side operation indicators (i.e., grid safety and economic indicators). Specifically, daily electricity consumption data from all users within the assessment period can be collected from smart meters. This data is then cleaned, including handling missing values, outliers, and noise smoothing, to form a data structure like... Figure 2 The user's daily electricity consumption curve shown ( Figure 2 The horizontal axis represents time, and the vertical axis represents electricity consumption.

[0092] Then, based on electricity consumption characteristics, users are categorized into residential, commercial, and industrial types. Adjustable flexible load models are constructed for each type, including air conditioners, water heaters, charging piles, and interruptible loads, clearly defining the adjustable power range, adjustable time periods, adjustment rates, and comfort constraints for each user. Based on the distribution network topology, line parameters, transformer capacity, and distributed generation output curves, a distribution network simulation model incorporating flexible loads is built. A preset dynamic load adjustment strategy is input, and simulation calculations are performed using optimal power flow or time-series power flow algorithms to obtain adjusted user electricity consumption data and grid safety and economic indicators.

[0093] Step 2: Determine the user's electricity experience sensitivity and the degree of electricity deviation caused by the load dynamic adjustment strategy based on the user's historical electricity consumption data; determine the user-side experience impact indicators based on the user's electricity experience sensitivity and the degree of electricity deviation. S2.1: Determine the power experience sensitivity of each user. ; The rationale for assigning weights to individual days in historical user electricity consumption data stems from the need to identify fluctuations in the external environment. When other users' electricity consumption curves show significant changes relative to their average curve on a given day, it usually indicates substantial fluctuations in electricity prices, sudden temperature changes, or grid stimulus signals. These days of strong environmental disturbances are key samples for examining users' true electricity consumption preferences and behavioral rigidity. On stable days with less environmental pressure, differences in users' electricity consumption habits are difficult to fully reflect; however, on days with drastic changes in external factors, whether users maintain their original electricity consumption patterns better reflects their true experience preferences and behavioral bottom lines.

[0094] For each historical observation day of the target user, the remaining user nodes in the distribution network simulation model (excluding the target user) are traversed. Based on the fluctuation of the electricity consumption curve of the remaining user nodes on the corresponding historical observation day relative to their historical average electricity consumption curve, the external environmental disturbance weights for each historical observation day are obtained. External environmental disturbance weights Indicates the first The first user history The external environmental disturbance weights for daily electricity consumption data are calculated using the following formula: ; in, This indicates traversing all except the first one. Other users of that user; Indicates the first The first user history Daily electricity consumption data curve; Indicates the first Historical average electricity consumption data curve for each user; Indicates the first The first user history The fluctuation of the daily electricity consumption data curve compared to the average electricity consumption curve.

[0095] A weighted average of electricity consumption data for each historical observation day is obtained by weighting the data based on external environmental disturbances. : ;

[0096] On days with significant fluctuations in external conditions (high-weight days), users face clear incentives to change their electricity consumption habits to gain economic benefits (such as responding to electricity prices) or to adapt to the environment (such as temperature changes). Dates with higher weights correspond to scenarios with stronger environmental disturbances. In these cases, the degree of overlap in users' electricity consumption curves directly reflects their comfort sensitivity and habit rigidity. If users maintain highly overlapping electricity consumption curves on high-weight days, it indicates that their electricity consumption patterns are less affected by external factors, their lifestyle and energy consumption patterns are stable, they have higher requirements for electricity comfort and convenience, and tend to refuse to change their electricity consumption habits to ensure a good experience; they are experience-sensitive users. Conversely, if the curve fluctuates significantly, it indicates that users have flexible electricity consumption arrangements, a higher tolerance for regulatory disturbances and changes in comfort, and greater adjustability. Therefore, electricity experience sensitivity can be obtained based on the deviation of electricity consumption data from the weighted benchmark electricity consumption curve on each historical observation day and the weight of external environmental disturbances on each historical observation day. : ; in, Indicates the first The user's sensitivity to electricity consumption is denoted by exp, which represents an exponential function with base e.

[0097] S2.2: Determine the degree of electricity consumption deviation caused by the current load dynamic adjustment strategy for each user. ; Electricity consumption deviation is a core indicator for measuring the intensity of user intervention in dynamic load regulation. Its physical meaning is the degree to which the strategy forces users to deviate from their natural electricity demand. Without regulatory intervention, users will consume electricity according to their own habits and needs, and this natural electricity consumption can be predicted using historical time-series data (LSTM). If current dynamic load regulation is implemented, user electricity consumption can be obtained through distribution network simulation.

[0098] This step first collects historical hourly electricity consumption data from users to construct a time-series electricity consumption dataset; then, the data is cleaned, normalized, and subjected to sliding window construction to form LSTM input samples; next, an LSTM network is built, trained with historical multi-day electricity consumption sequences as input and daily electricity consumption curves as output, to obtain a user electricity consumption prediction model; the model outputs the predicted electricity consumption of users under unregulated conditions; finally, the predicted electricity consumption (…) ) and the power consumption of the power distribution network after dynamic load adjustment ( By taking the difference at each time step, integrating, or summing, the total electricity consumption deviation can be obtained, which is the overall deviation of the adjustment from the user's electricity consumption. Therefore, based on the adjusted user electricity consumption data (i.e., the first value obtained from the distribution network simulation after load dynamic adjustment), the total electricity consumption deviation can be calculated. Electricity consumption data curve of individual users The degree of deviation from the natural electricity consumption baseline is used to obtain the degree of electricity consumption deviation. ; ; in, Indicates the degree of deviation in electricity usage; This represents the maximum and minimum value normalization function; This indicates the number of predictions obtained using LSTM. Electricity consumption data curves for individual users (natural electricity consumption baseline); This indicates that the current load dynamic adjustment strategy is obtained through distribution network simulation. Electricity consumption data curves for individual users.

[0099] S2.3: Determine user-side experience impact indicators based on electricity consumption experience sensitivity and electricity consumption deviation degree. ; The impact of the same deviation in electricity consumption varies significantly among different users, and the level of importance placed on user experience represents their sensitivity to changes in electricity usage habits. Higher importance translates to a more pronounced decline in experience from the same adjustment, resulting in stronger actual interference for the user. Evaluating the impact solely based on electricity consumption deviation ignores user subjective preferences, leading to distorted assessment results. Therefore, a weighted approach can be applied to the degree of electricity consumption deviation based on user experience sensitivity to obtain a user-side experience impact index. Among them, the user experience impact indicators are positively correlated with electricity experience sensitivity and the degree of electricity deviation: ; in, This indicates the metrics that affect user experience. Indicates sensitivity to electricity usage experience; This indicates the degree of deviation in electricity usage.

[0100] Step 3: Correct the power grid safety and economic indicators based on user experience impact indicators, and obtain the comprehensive evaluation results of power grid load dynamic adjustment based on the weighted and fused power grid safety and economic indicators of each user's historical electricity consumption scale. S3.1: Based on historical load regulation data, determine the quantitative relationship between user-side experience impact indicators and the potential impact of user behavior rebound on power grid safety and economic indicators, and obtain the impact correction coefficient. And use the impact correction coefficient to make weighted corrections to the power grid safety and economic indicators; A decline in user experience is a major trigger for malicious electricity use and a rebound in electricity consumption. The greater the impact on user experience, the more likely users are to engage in irrational electricity consumption, thereby impacting the safety and economic indicators of the power grid. Historical adjustment data shows a clear correlation between the degree of impact on user experience and the magnitude of subsequent electricity consumption rebounds. By fitting a linear function of these two factors to historical samples, the driving effect of a decline in user experience on malicious electricity use can be quantified.

[0101] This step first extracts data pairs for each user from historical load adjustment events: the independent variable is the user experience impact index calculated in the previous step, and the dependent variable is the corresponding malicious behavior quantification value (the rebound in electricity consumption one day after the adjustment compared to the day before the adjustment). Then, linear regression is used to fit a function for each user: ; in, This indicates the two days before and after the historical Ath load dynamic adjustment. Difference in daily electricity consumption per user This indicates that the historical load dynamic adjustment strategy for the Ath time is relevant to the first time. User-side experience impact metrics for individual user experiences; and This represents the unknown parameters obtained from the fitting.

[0102] Next, the least squares method is used to linearly fit the historical samples to obtain the unknown parameters. Finally, the current load dynamic adjustment strategy obtained above is applied to the first... Substituting the user-side experience impact metrics of individual user experiences into the above linear function yields a quantitative relationship and the current user-side experience impact metrics. Determine the impact correction factor : ; Traditional safety and economic indicator scores (such as those calculated based on line loss and voltage qualification rate) reflect the direct and immediate impact of regulation schemes on power grid operation, without considering the impact of degraded user experience. By weighting traditional safety and economic indicator scores with an impact correction coefficient, the negative impact of potential risks can be reduced on top of existing operational benefits. This makes the assessment results more comprehensive and forward-looking, reflecting both the immediate optimization effect of regulation strategies on the power grid and considering the long-term risks arising from user behavior feedback, achieving a unified trade-off between benefits and risks. Therefore, the following formula can be constructed to express the... Individual user's evaluation score for current load dynamic adjustment: ; in, Indicates the first Corrected power grid security and economic indicators for individual users; This represents the impact correction factor. This indicates the first [indicator] obtained using traditional power grid safety economic indicators. The evaluation score of each user for the current load dynamic adjustment.

[0103] S3.2: Obtain the comprehensive evaluation results of the dynamic adjustment of the distribution network load based on the power grid security and economic indicators after weighted fusion correction of the historical electricity consumption scale of each user; The power grid, as a whole, is the result of the cumulative electricity consumption behavior of all users, but the importance of different users varies. Users with larger average daily electricity consumption have a more significant impact on the overall power grid through changes in their electricity consumption, user satisfaction, and behavioral feedback. Smaller users' electricity fluctuations have a limited impact on global indicators. Applying equal weighting to all users would overestimate the role of smaller users and underestimate the impact of larger users, leading to distorted evaluation. This step uses the average daily electricity consumption of each user as the weight to perform a weighted fusion of the evaluation scores obtained above for each user's assessment of current load dynamic adjustment, making the overall evaluation result more consistent with the operating mechanism of the distribution network. Therefore, the historical average electricity consumption of each user can be determined based on their historical electricity consumption data. Historical electricity consumption figures are obtained; based on the ratio of each user's historical electricity consumption to the sum of all users' historical electricity consumption figures, the evaluation weight of each user is determined; and based on each user's evaluation weight, the revised power grid security and economic indicators are applied. Weighted fusion is performed to obtain the comprehensive evaluation results of dynamic load regulation of the distribution network. : ; in, This indicates the comprehensive evaluation results of the dynamic adjustment of the distribution network load; Indicates the first Historical average daily electricity consumption of each user (historical electricity consumption scale). Indicates the first Individual user-corrected power grid security and economic indicators.

[0104] This concludes the comprehensive evaluation of the dynamic load adjustment strategy.

[0105] This invention is now complete.

[0106] In summary, in this embodiment, a load dynamic adjustment strategy is obtained; based on the load dynamic adjustment strategy, power flow simulation is performed on the distribution network simulation model to obtain adjusted user electricity consumption data and power grid safety and economic indicators; based on users' historical electricity consumption data, the electricity experience sensitivity of each user and the degree of electricity deviation caused by the load dynamic adjustment strategy are determined; based on the electricity experience sensitivity and the degree of electricity deviation, user-side experience impact indicators are determined; based on the user-side experience impact indicators, the power grid safety and economic indicators are corrected; and based on the weighted and fused power grid safety and economic indicators corrected by the historical electricity consumption scale of each user, a comprehensive evaluation result of distribution network load dynamic adjustment is obtained. This invention constructs a coupled evaluation mechanism for user-side experience impact and grid-side safety and economic indicators. By quantifying user electricity consumption experience sensitivity and load regulation deviation, it establishes a predictive model for the potential impact of user behavior rebound on grid operation, achieving a closed-loop evaluation that balances grid benefits and user experience. This overcomes the one-sidedness of traditional evaluation methods that only focus on the grid's immediate operating status while ignoring user behavior feedback. Furthermore, based on the weighting of external environmental disturbances, it identifies the rigidity of user electricity consumption patterns. By analyzing the fluctuation of other users' electricity consumption curves on high-disturbance days, it determines the daily weight and calculates the user's electricity consumption experience sensitivity accordingly. This accurately distinguishes between experience-sensitive users and flexible-tolerant users, enabling differentiated load management. The system provides a quantitative basis for formulating regulation strategies. On the other hand, it uses historical load regulation data to fit the quantitative relationship between user-side experience impact indicators and electricity consumption rebound magnitude, transforming the degree of user experience impairment into a potential impact coefficient on grid safety and economic indicators. This allows the assessment results to reflect the long-term risks to grid operation from malicious electricity consumption behaviors that may be triggered by regulation strategies, enhancing the proactive nature of the assessment and risk prevention capabilities. Furthermore, based on the grid safety and economic indicators corrected by weighted fusion of historical electricity consumption scales of each user, the system reflects the differentiated impact of different users on the overall operation of the distribution network through the proportion of electricity consumption. This avoids the problem of underestimating the impact of large users due to equal weighting fusion, making the comprehensive assessment results more consistent with the actual operating mechanism of the distribution network.

[0107] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for simulating and evaluating power distribution networks with dynamic load regulation, characterized in that, The method includes: Obtain dynamic load adjustment strategies; Based on the aforementioned load dynamic adjustment strategy, power flow simulation is performed on the distribution network simulation model to obtain adjusted user electricity consumption data and power grid safety and economic indicators. Based on users' historical electricity consumption data, the electricity experience sensitivity of each user and the degree of electricity deviation caused by the load dynamic adjustment strategy are determined. Based on the electricity experience sensitivity and the degree of electricity deviation, user-side experience impact indicators are determined. The electricity experience sensitivity is used to characterize the rigidity of users maintaining their electricity consumption patterns under external environmental disturbances, and the degree of electricity deviation is used to characterize the deviation of the adjusted user electricity consumption data from the natural electricity consumption baseline in the unadjusted state. Based on the user-side experience impact index, the power grid safety and economic index is corrected, and based on the power grid safety and economic index corrected by weighted fusion of the historical electricity consumption scale of each user, the comprehensive evaluation result of the dynamic adjustment of the distribution network load is obtained. The method for obtaining the sensitivity of electricity consumption experience includes: for each historical observation day of the target user, traversing all user nodes in the distribution network simulation model other than the target user, and obtaining the external environmental disturbance weight for each historical observation day based on the fluctuation degree of the electricity consumption curve of the other user nodes on the corresponding historical observation day relative to their historical average electricity consumption curve; performing a weighted average of the electricity consumption data of each historical observation day based on the external environmental disturbance weight to obtain a weighted benchmark electricity consumption curve; and obtaining the sensitivity of electricity consumption experience based on the deviation degree of the electricity consumption data of each historical observation day relative to the weighted benchmark electricity consumption curve and the external environmental disturbance weight for each historical observation day.

2. The distribution network simulation and evaluation method based on load dynamic adjustment as described in claim 1, characterized in that, The method for obtaining the degree of electricity consumption deviation includes: Based on the user's historical electricity consumption data, determine the natural electricity consumption baseline under the condition that the load dynamic adjustment strategy is not implemented; The degree of deviation of the user's electricity consumption data after adjustment from the natural electricity consumption baseline is obtained.

3. The distribution network simulation and evaluation method based on dynamic load regulation as described in claim 2, characterized in that, The step of determining the natural electricity consumption baseline under conditions where the load dynamic adjustment strategy is not implemented, based on the user's historical electricity consumption data, includes: A time-series electricity consumption dataset is constructed based on the user's historical electricity consumption data; Using historical electricity consumption sequences as input and electricity consumption curves for the target time period as output, the user electricity consumption prediction model is trained to obtain the trained user electricity consumption prediction model. Electricity consumption is predicted using the user electricity consumption prediction model to obtain the natural electricity consumption baseline.

4. The distribution network simulation and evaluation method based on dynamic load regulation as described in claim 1, characterized in that, The determination of user-side experience impact indicators based on the power consumption experience sensitivity and the degree of power consumption deviation includes: The degree of power consumption deviation is weighted based on the power consumption experience sensitivity to obtain the user-side experience impact index; wherein, the user-side experience impact index is positively correlated with the power consumption experience sensitivity and the degree of power consumption deviation.

5. The distribution network simulation and evaluation method based on dynamic load regulation as described in claim 1, characterized in that, The step of correcting the power grid safety and economic indicators based on the user-side experience impact indicators includes: Based on historical load regulation data, the quantitative relationship between the user-side experience impact index and the potential impact of user behavior rebound on power grid safety and economic indicators is determined, and the impact correction coefficient is obtained. The power grid safety and economic indicators are weighted and corrected using the aforementioned impact correction coefficient.

6. The distribution network simulation and evaluation method based on dynamic load regulation as described in claim 5, characterized in that, Based on historical load regulation data, the quantitative relationship between the user-side experience impact indicators and the potential impact of user behavior rebound on power grid safety and economic indicators is determined, and the impact correction coefficient is obtained, including: Based on historical load dynamic adjustment data, the user-side experience impact indicators and corresponding user electricity consumption rebound magnitudes are extracted for each adjustment event; wherein, the user electricity consumption rebound magnitude is determined based on the difference in electricity consumption during a set period before and after adjustment; The quantitative relationship is established by performing regression fitting between the user experience impact index and the user electricity consumption rebound magnitude. Based on the quantification relationship and the current user experience impact indicators, the impact correction coefficient is determined.

7. The distribution network simulation and evaluation method based on dynamic load regulation as described in claim 1, characterized in that, The power grid security and economic indicators, which are weighted and adjusted based on the historical electricity consumption of each user, include: Determine the historical average electricity consumption of each user based on their historical electricity consumption data, and obtain the historical electricity consumption scale. The evaluation weight of each user is determined based on the ratio of each user's historical electricity consumption to the sum of all users' historical electricity consumption. The revised power grid security and economic indicators are weighted and integrated based on the evaluation weights of each user to obtain the comprehensive evaluation result of the dynamic adjustment of the distribution network load.

8. A method for simulating and evaluating power distribution networks based on dynamic load regulation according to any one of claims 1-7, characterized in that, The method for constructing the power distribution network simulation model includes: User types are categorized based on user electricity consumption behavior characteristics, and corresponding adjustable flexible load models are constructed for each type of flexible load; wherein, the adjustable flexible load models include air conditioning load models, water heater load models, charging pile load models, and interruptible load models; Determine the adjustable power range, adjustable time period boundary, and adjustment rate constraint for each of the aforementioned flexible load adjustable models; Based on the distribution network topology, line impedance parameters, transformer capacity, and distributed power generation characteristics, a distribution network simulation model incorporating the aforementioned flexible load adjustable model is constructed.

9. A method for simulating and evaluating power distribution networks based on dynamic load regulation according to any one of claims 1-7, characterized in that, The method further includes: Based on the comprehensive evaluation results of the dynamic adjustment of the power distribution network load, the applicability of the dynamic load adjustment strategy is determined and / or optimized.

Images