Demand response optimization method for power marketing business

Abstract

The invention relates to the technical field of power system demand side management, in particular to a demand response optimization method for power marketing business. A multi-scene dynamic reference library is constructed and a feature matching priority is set by collecting historical power consumption data of a user and associating external features. And based on a weighted matching result of the current scene and the reference library, determining an optimal scene and calculating a single / multi-dimensional deviation index, and in combination with a threshold value, judging a graded abnormal user and triggering a response mark. And for abnormal users, establishing a standby load resource pool scheduling model, and optimizing and compensating load deviation by integrating energy storage, adjustable load and other resources. Meanwhile, a dynamic credit rating is generated in combination with the user historical fulfillment rate and the load prediction accuracy, motivation such as electricity price discount and priority scheduling is implemented for high-credit users, punishment measures such as electricity price floating and permission limitation are taken for low-credit users, and a closed-loop optimized demand response management system is formed. Accurate optimization of demand response and user behavior regulation and control in power marketing are realized.

Description

Technical Field

[0001] This invention relates to the field of demand-side management technology for power systems, and more specifically to a demand response optimization method for power marketing operations. Background Technology

[0002] Against the backdrop of power market reform and smart grid development, demand response, as a core component of power marketing, aims to achieve a balance between power grid supply and demand and optimal resource allocation by guiding users to adjust their electricity consumption behavior.

[0003] Existing technologies suffer from numerous bottlenecks that urgently need to be addressed. On the one hand, traditional demand response baseline load calculations lack dynamism. For example, they do not fully integrate multi-dimensional data such as weather, calendar, and electricity prices, resulting in insufficient matching between the baseline model and actual scenarios, making it difficult to accurately identify abnormal user electricity consumption. For instance, CN113256031B relies solely on historical average load or single-dimensional data, considering only historical user electricity consumption behavior and lacking a dynamic correlation analysis mechanism. This prevents the baseline model from adapting to the load fluctuation characteristics of different scenarios in real time.

[0004] On the other hand, existing incentive and compensation mechanisms have significant shortcomings. For example, CN113256031B has obvious weaknesses in anomaly identification and incentive mechanism design. Focusing only on a single dimension or a simple combination of indicators makes it difficult to capture complex anomalies in user electricity consumption behavior. In terms of incentive mechanisms, CN113256031B lacks a dynamic credit management system and does not link historical performance rates, load forecast accuracy, and incentive strategies, resulting in insignificant compensation differences between high-credit and low-credit users. Summary of the Invention

[0005] This invention aims to provide a demand response optimization method for power marketing operations. By constructing a multi-dimensional dynamic benchmark library, accurately identifying anomalies, intelligent compensation scheduling, and credit rating management, it solves the problems of non-dynamic demand response benchmarks, incomplete anomaly identification, unoptimized compensation, and inaccurate incentives in existing technologies, thereby improving the management efficiency of power demand response and the safety of power grid operation.

[0006] The specific technical solution of the present invention is as follows:

[0007] The technical solution of this invention is to provide a demand response optimization method for electricity marketing operations, comprising:

[0008] By collecting users' historical electricity consumption data and linking it with weather, calendar, and electricity price data, a dynamic benchmark library for multiple scenarios is constructed, and matching priorities are set.

[0009] Based on the user's historical electricity consumption data, weather, calendar, and electricity price data for the current scenario, the system matches the data with a multi-scenario dynamic benchmark library according to priority to determine the optimal scenario, and calculates single-dimensional deviation indicators and multi-dimensional comprehensive deviation indicators; it sets multi-dimensional comprehensive anomaly thresholds for judgment, outputs a tiered list of abnormal users, and triggers anomaly response flags;

[0010] For single-dimensional deviation indicators and multi-dimensional comprehensive deviation indicators of abnormal users, a scheduling model for the standby load resource pool is established, and the optimal scheduling instruction is output to compensate for the load deviation of abnormal users.

[0011] The system generates a dynamic credit rating for users by combining historical response fulfillment rates and load forecast accuracy. Users with high credit ratings receive premium incentives, while users with low credit ratings are subject to punitive adjustments.

[0012] As a further optimization of the method of the present invention, the construction of the multi-scenario dynamic benchmark library includes:

[0013] The user's historical electricity consumption data is aligned with meteorological features, calendar features, and electricity price features according to the time dimension to form a feature matrix. The meteorological features include temperature, humidity, wind speed, and precipitation; the calendar features include holidays, weekends, seasons, and special events; and the electricity price features include real-time electricity price, historical electricity price, time-of-use electricity price type, and electricity price volatility.

[0014] As a further optimization of the method of the present invention, the setting of the matching priority includes:

[0015] Strongly correlated features are assigned a weight of 0.5-0.6, including temperature, real-time electricity price, holidays, and seasons;

[0016] The moderately relevant features are assigned a weight of 0.3-0.4, including humidity, time-of-use electricity pricing type, and weekends;

[0017] Weakly correlated features are assigned a weight of 0.1-0.2, including wind speed, precipitation, special events, electricity price volatility, and historical electricity prices.

[0018] As a further optimization of the method of the present invention, the optimal scene determination process includes: feature priority weighting, calculation of matching value, and determination of optimal matching scene;

[0019] Among them, the feature priority weighting is as follows: 0.5-0.6 for strongly correlated features, 0.3-0.4 for moderately correlated features, and 0.1-0.2 for weakly correlated features;

[0020] Match value calculation: Calculated using the following formula:

[0021]

[0022] Where: D i x is the matching value between the current scene and the i-th scene in the benchmark library; ij x is the feature j value of the i-th scene in the benchmark library; cj w represents the feature j value of the current scene. j The priority weights for feature j;

[0023] Optimal matching scenario determination: Select the baseline scenario with the smallest matching value as the matching result for the current scenario.

[0024] As a further optimization of the method of the present invention, the single-dimensional deviation index includes electricity consumption behavior deviation index, meteorological deviation index, calendar deviation index, and electricity price deviation index. The multi-dimensional comprehensive deviation index integrates the single-dimensional deviation index using a weighted average method, and the calculation formula is as follows:

[0025]

[0026] Among them, S d The value of w represents a multi-dimensional comprehensive deviation from the index. j Let Δ be the priority weight of feature j. j The deviation index of feature j, i.e., the single-dimensional deviation index, where n is the number of features, n = 13.

[0027] As a further optimization of the method of the present invention, the output of the hierarchical abnormal user list includes:

[0028] Level 1 anomaly: The overall deviation index exceeds the threshold by 15% or any single dimension deviation index exceeds 20%;

[0029] Level 2 anomaly: The overall deviation index exceeds the threshold by 10%-15% and the single-dimensional deviation index is ≤20%;

[0030] Level 3 anomalies: The overall deviation index exceeds the threshold by 5%-10% and the single-dimensional deviation index is ≤15%.

[0031] As a further optimization of the method of the present invention, the objective function of the standby load resource pool scheduling model is:

[0032]

[0033] Wherein, ΔP i ξ represents the adjustment amount of the i-th load resource. i C represents the corresponding weighting coefficient. j Let λ be the unit adjustment cost of the j-th resource. c Weighting factors for user credit rating;

[0034] The constraints of the standby load resource pool scheduling model include physical constraints, time constraints, and economic constraints.

[0035] Among them, physical constraints: P min,i ≤P i ≤P max,i , where P min,i and P max,i Let P represent the minimum and maximum regulation capabilities of the i-th load resource, respectively. i This represents the adjustable load of the i-th load resource;

[0036] Time constraint: t start,i ≤t response,i ≤t end,i , where t start,i and t end,i Let t represent the available time window for the i-th load resource, and t represent the available time window for the i-th load resource. response,i Indicates its response time;

[0037] Economic constraints: Among them, C i C represents the unit adjustment cost of the i-th load resource. max This is the upper limit of the scheduling budget.

[0038] As a further optimization of the method of the present invention, the user dynamic credit rating assessment includes:

[0039] Based on three dimensions—historical response fulfillment rate, load forecast accuracy, and response stability—a fuzzy logic evaluation method is used to calculate a comprehensive credit score, which is then divided into three levels: A-high credit, B-medium credit, and C-low credit.

[0040] The incentive mechanism for high-credit users includes:

[0041] Electricity price discount incentives: Class A users enjoy a 5%-10% discount on electricity prices during peak hours;

[0042] Priority dispatch incentive: Class A users have priority in obtaining energy storage discharge or adjustable load dispatch resources;

[0043] Points reward mechanism: Users accumulate points based on their performance rate and deviation from the target, which can be used for electricity fee reduction or priority resource allocation;

[0044] The punitive adjustment mechanism for low-credit users includes:

[0045] Dynamic electricity price increase: C-level users will have their electricity price increased by 10%-15% during peak hours;

[0046] Dispatch authority restrictions: Class C users have their priority reduced by 50% in energy storage discharge and load dispatch;

[0047] Credit downgrade rules: A credit rating downgrade is triggered when the performance rate is below 70% for three consecutive cycles or the load forecasting error exceeds 15%.

[0048] As a further optimization of the method of the present invention, the abnormal response marker triggering strategy includes:

[0049] A level one anomaly triggers a load adjustment command and notifies the dispatch center.

[0050] A level 2 anomaly alert is sent to the user terminal.

[0051] Level 3 anomalies generate abnormal behavior records and are incorporated into the basis for credit rating adjustments.

[0052] The beneficial effects of the technical solutions provided in this application include at least the following:

[0053] Beneficial effects:

[0054] By integrating historical user electricity consumption data with multi-dimensional external features such as weather, calendar, and electricity prices, a dynamic benchmark library is constructed and feature matching priorities are set. This overcomes the shortcomings of traditional methods, such as static benchmarks and single-dimensional matching, significantly improving the matching accuracy between current and historical scenarios. This lays a precise data foundation for subsequent anomaly identification and resource scheduling.

[0055] By calculating single-dimensional and multi-dimensional comprehensive deviation indicators, and classifying abnormal users into Level 1, Level 2, and Level 3 based on thresholds, comprehensive and precise identification of electricity consumption anomalies was achieved. Different response strategies were triggered at different levels, avoiding a "one-size-fits-all" approach and improving the targeting and efficiency of demand response.

[0056] The standby load resource pool scheduling model aims to minimize adjustment costs and deviations. It comprehensively considers physical, time, and economic constraints, and incorporates a user credit rating weighting factor to achieve optimal load resource allocation. This model efficiently compensates for load deviations from abnormal users, ensuring grid stability, while also controlling scheduling costs and improving resource utilization efficiency.

[0057] A dynamic credit rating system is built based on historical fulfillment rates and load forecasting accuracy, coupled with differentiated incentive and penalty measures, forming an "incentive-constraint" closed loop. This not only enhances the responsiveness of high-credit users but also motivates low-credit users to improve their behavior through a constraint mechanism, effectively improving the overall fulfillment rate and stability of demand response.

[0058] From multi-scenario benchmark construction and precise anomaly identification to resource optimization scheduling and dynamic credit management, a closed-loop demand response system has been formed. This system achieves "precision, intelligence, and marketization" in demand response in electricity marketing, ensuring both the balance of power grid supply and demand and safe operation, while also reducing overall electricity costs through user behavior regulation, providing an efficient and feasible solution for demand-side management of the power system. Attached Figure Description

[0059] Figure 1 A schematic diagram illustrating the overall process of demand response optimization methods for electricity marketing operations;

[0060] Figure 2 Detailed flowchart of steps S100 for the demand response optimization method for electricity marketing business;

[0061] Figure 3 Detailed flowchart of steps S200 for the demand response optimization method for electricity marketing business;

[0062] Figure 4 Detailed flowchart of S300 steps for demand response optimization method for electricity marketing business;

[0063] Figure 5 Detailed flowchart of the S400 steps for optimizing demand response for electricity marketing operations. Detailed Implementation

[0064] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0065] Existing technologies for demand response management typically suffer from the following technical bottlenecks: static baseline load calculation, limited anomaly identification dimensions, insufficient flexibility in compensation mechanisms, and rudimentary user incentive mechanisms. To address these issues, please refer to [link to relevant documentation / reference]. Figure 1 This illustrates a demand response optimization method for electricity marketing operations provided by an embodiment of the present invention, the method comprising:

[0066] S100: By collecting users' historical electricity consumption data and linking it with weather, calendar and electricity price data, a multi-scenario dynamic benchmark library is built, and matching priorities are set.

[0067] S200: Based on the user's historical electricity consumption data, weather, calendar and electricity price data in the current scenario, match with the multi-scenario dynamic benchmark library according to priority, determine the optimal scenario, and calculate single-dimensional deviation index and multi-dimensional comprehensive deviation index; set multi-dimensional comprehensive anomaly threshold for judgment, output a hierarchical abnormal user list, and trigger anomaly response marking.

[0068] S300: For single-dimensional deviation indicators and multi-dimensional comprehensive deviation indicators of abnormal users, establish a scheduling model for the standby load resource pool, output the optimal scheduling instruction, and compensate for the load deviation of abnormal users.

[0069] S400: Generates a dynamic credit rating for users by combining historical response fulfillment rate and load forecast accuracy. Users with high credit ratings receive premium incentives, while users with low credit ratings are subject to punitive adjustments.

[0070] The specific plan is as follows:

[0071] In the demand response optimization method for electricity marketing business, the purpose of S100 is to build a multi-scenario dynamic benchmark library and set the matching priority for each scenario in the multi-scenario dynamic benchmark library.

[0072] Please refer to Figure 2 The diagram illustrates a flowchart of an exemplary demand response optimization method S100 for electricity marketing operations, which includes:

[0073] S110: Collect historical electricity consumption data of users and extract external data features related to historical electricity consumption data of users.

[0074] User historical electricity consumption data refers to the user's electricity consumption, and this data is sourced from smart meters.

[0075] Meteorological conditions are one of the important factors affecting users' electricity consumption behavior. The features extracted from meteorological conditions include: temperature characteristics, humidity characteristics, wind speed characteristics, and precipitation characteristics.

[0076] Calendar data is one of the most important factors influencing users' electricity consumption behavior. Features extracted from calendar data include: holiday features, weekend features, seasonal features, and special event features.

[0077] Electricity price is the third most important factor influencing users' electricity consumption behavior. The characteristics of electricity price extraction include: real-time electricity price characteristics, historical electricity price characteristics, time-of-use electricity price characteristics, and electricity price fluctuation characteristics.

[0078] S120: Link users' historical electricity consumption data with external data features such as weather, calendar, and electricity prices to form a dynamic benchmark library for multiple scenarios.

[0079] To accurately characterize the correlation between external data features such as weather, calendar, and electricity prices and users' electricity consumption behavior, external data features such as weather, calendar, and electricity prices are aligned with users' historical electricity consumption data according to the time dimension to form a dynamic benchmark library for multiple scenarios.

[0080] In one possible implementation, the constructed feature matrix is:

[0081]

[0082] The symbol mapping table is as follows:

[0083]

[0084] S130: Set the matching priority of each feature in the multi-scenario dynamic benchmark library.

[0085] In one possible implementation, the matching priority of each feature in the multi-scenario dynamic benchmark library is as follows:

[0086] Strong correlation feature: temperature (T) t Real-time electricity price (Rp) t Holidays t Season (S) t ).

[0087] Moderately relevant characteristics: Humidity (H t Time-of-use pricing type (Tp) t ), weekend (Wk t ).

[0088] Weakly correlated feature: wind speed (W) t ), precipitation (R) t Special events (E) t Electricity price fluctuations (V) t ), historical electricity price (Rh t ).

[0089] In the demand response optimization method for electricity marketing business, S200 uses a multi-scenario dynamic benchmark library to match the current user's electricity consumption behavior in real time and identify abnormal users through multi-dimensional deviation indicators.

[0090] Please refer to Figure 3 The document illustrates a flowchart of an exemplary demand response optimization method S200 for electricity marketing operations, which includes:

[0091] S210: Match the current scene with scenes in the multi-scene dynamic benchmark library to determine the optimal scene.

[0092] Collect and extract user electricity consumption data and related external features in the current scenario to construct a complete feature matrix for the current scenario. Match this matrix with a multi-scenario dynamic benchmark library to determine the closest historical scenario.

[0093] In one possible implementation, the process of determining the optimal scenario includes: feature priority weighting, calculation of matching values, and determination of the optimal matching scenario, to ensure that the final matched scenario can accurately reflect the current user's electricity consumption behavior pattern.

[0094] Based on the above implementation methods, specifically:

[0095] Feature priority weighting: The feature matrix is ​​weighted according to the matching priority set in S130. In one possible implementation, strongly correlated features are weighted by 0.5-0.6, moderately correlated features by 0.3-0.4, and weakly correlated features by 0.1-0.2, such that the sum of the weights of strongly correlated features, moderately correlated features, and weakly correlated features equals 1.

[0096] Match value calculation: In one possible implementation, the formula for calculating the match value is:

[0097]

[0098] Where: D i x is the matching value between the current scene and the i-th scene in the benchmark library; ij x is the feature j value of the i-th scene in the benchmark library; cj w represents the feature j value of the current scene. j represents the priority weight of feature j.

[0099] Optimal matching scenario determination: Select the baseline scenario with the smallest matching value as the matching result for the current scenario.

[0100] S220: Calculate the single-dimensional deviation index and the multi-dimensional comprehensive deviation index between the current scenario and the optimal scenario.

[0101] After determining the optimal scenario, single-dimensional deviation indicators and multi-dimensional comprehensive deviation indicators are calculated between the current scenario and the optimal scenario to quantify the degree of abnormality in user electricity consumption behavior. The multi-dimensional comprehensive deviation indicator covers multiple single-dimensional deviation indicators such as electricity consumption behavior, weather, calendar, and electricity price, which can comprehensively reflect the changes in user electricity consumption under different environments.

[0102] Single-dimensional deviation indicators include: electricity consumption behavior deviation indicators, meteorological deviation indicators, calendar deviation indicators, and electricity price deviation indicators.

[0103] Specifically, in terms of electricity consumption behavior, the electricity consumption behavior deviation index measures the difference between a user's current electricity consumption pattern and the optimal scenario. In terms of weather, the weather deviation index reflects the difference between current weather conditions and the optimal scenario. In terms of calendar, the calendar deviation index measures the degree of matching between current calendar characteristics and the optimal scenario. In terms of electricity price, the electricity price deviation index measures the difference between the current electricity price environment and the optimal scenario. In one possible implementation, the deviation indices for electricity consumption behavior, weather, calendar, and electricity price are measured by calculating the absolute difference or standard deviation between the current scenario data and the optimal scenario.

[0104] Single-dimensional deviation indicators from electricity consumption behavior, weather, calendar, and electricity prices are integrated into a multi-dimensional comprehensive deviation indicator. One possible implementation method uses a weighted average method to integrate different single-dimensional deviation indicators into a multi-dimensional comprehensive deviation indicator, providing a more intuitive basis for anomaly assessment. The calculation formula for the multi-dimensional comprehensive deviation indicator is:

[0105]

[0106] Among them, S d The value of w represents a multi-dimensional comprehensive deviation from the index. j Let Δ be the priority weight of feature j. j The deviation index of feature j, where n is the number of features, n = 13.

[0107] S230: Output of hierarchical abnormal user list and abnormal response marking.

[0108] Based on historical electricity consumption data, meteorological data, calendar data, and electricity price data, a multi-dimensional comprehensive anomaly threshold and an arbitrary single-dimensional deviation index threshold are set. Based on the deviation between the multi-dimensional comprehensive anomaly threshold and the multi-dimensional comprehensive deviation index value, as well as the arbitrary single-dimensional deviation index, abnormal users are identified and classified into different levels.

[0109] In one possible implementation, the abnormal user classification method is as follows:

[0110] Level 1 anomaly: The overall deviation index exceeds the threshold by 15% or any single dimension deviation index exceeds 20%;

[0111] Level 2 anomaly: The overall deviation index exceeds the threshold by 10%-15% and the single-dimensional deviation index is ≤20%;

[0112] Level 3 anomaly: The overall deviation index exceeds the threshold by 5%-10%, and the deviation index of a single dimension is ≤15%.

[0113] Anomaly response flags trigger different response strategies based on the anomaly level:

[0114] Level 1 anomaly: Immediately trigger a load adjustment command and notify the dispatch center;

[0115] Level 2 Anomaly: Sends a warning message, requesting the user to adjust their electricity usage behavior;

[0116] Level 3 Anomaly: Records abnormal behavior for subsequent credit rating adjustments.

[0117] In the demand response optimization method for electricity marketing, S300 constructs a reserve load resource pool that includes energy storage, adjustable loads, and distributed power sources, establishes a scheduling optimization model that considers physical constraints, time constraints, and economic constraints, outputs the optimal scheduling command, and compensates for abnormal user load deviations.

[0118] Please refer to Figure 3 The document illustrates a flowchart of an exemplary demand response optimization method S200 for electricity marketing operations, which includes:

[0119] S310: Establish a standby load resource pool scheduling model.

[0120] The standby load resource pool comprises various types of controllable loads and energy storage resources to provide flexible load regulation capabilities in the event of abnormal electricity consumption. The main resource types include: energy storage systems, adjustable loads, distributed generation, and virtual power plants.

[0121] A scheduling model for the reserve load resource pool is constructed to ensure that, while meeting the stability requirements of the power system, load regulation capacity is maximized and scheduling costs are minimized. The core objective of the scheduling model for the reserve load resource pool is to minimize the load deviation of current abnormal users, while taking into account resource availability, response time, and regulation costs.

[0122] In one possible implementation, the objective function of the standby load resource pool scheduling model is:

[0123]

[0124] Wherein, ΔP i ξ represents the adjustment amount of the i-th load resource. i These are the corresponding weighting coefficients, used to balance the adjustment capabilities and response costs of different resources. j Let λ be the unit adjustment cost of the j-th resource. c As a weighting factor for user credit rating, λ is used for high-credit users. c The lower level reduces the impact of economic penalties.

[0125] The squared form of the objective function ensures the minimization of load deviation, while preventing excessive adjustment of individual resources and constraining economic costs, thereby maintaining system stability.

[0126] The constraints include:

[0127] Physical constraints: P min,i ≤P i ≤P max,i , where P min,i and P max,i Let P represent the minimum and maximum regulation capabilities of the i-th load resource, respectively. i This represents the adjustable load of the i-th load resource.

[0128] Time constraint: t start,i ≤t response,i ≤t end,i, where t start,i and t end,i Let t represent the available time window for the i-th load resource, and t represent the available time window for the i-th load resource. response,i This indicates its response time.

[0129] Economic constraints: Among them, C i C represents the unit adjustment cost of the i-th load resource. max This is the upper limit of the scheduling budget.

[0130] S320: Solve the standby load resource pool scheduling model to obtain the optimal scheduling instructions.

[0131] Based on the standby load resource pool scheduling model, in one possible implementation, mixed-integer linear programming or heuristic optimization algorithms are used to solve the problem. After the solution is obtained, the optimization results need to be converted into executable scheduling instructions.

[0132] In one possible implementation, the conversion step includes:

[0133] The resource adjustment amount ΔP i This is mapped to specific operational instructions, such as the charging and discharging power of the energy storage system, the reduction ratio of adjustable load, and the output adjustment value of distributed power sources.

[0134] The execution order of instructions is sorted according to resource response time and user credit rating.

[0135] Commands are sent to the resource controller via IEC 60870-5-104 or MQTT protocol, ensuring that command execution delay is less than 1 second or less than 10 seconds.

[0136] In the demand response optimization method for electricity marketing, S400 constructs a dynamic credit assessment method based on users' historical performance rate and load forecast accuracy, implements electricity price discounts / priority dispatch incentives for high-credit users and electricity price increases / dispatch restrictions as penalties for low-credit users, thereby achieving precise incentives, dynamic constraints, and refined management of electricity marketing.

[0137] Please refer to Figure 3 The document illustrates a flowchart of an exemplary demand response optimization method S200 for electricity marketing operations, which includes:

[0138] S410: Construct a user credit rating assessment method.

[0139] User credit rating assessment is a crucial step in implementing precise incentive and penalty mechanisms. To construct a scientific and reasonable credit rating assessment method, two core indicators are comprehensively considered: the user's historical response and fulfillment rate and load forecasting accuracy. Data collection and feature extraction are conducted based on the database of the electricity marketing system.

[0140] Specifically, user credit rating assessment methods include:

[0141] Historical user response fulfillment rates are collected: Fulfillment rates reflect users' performance in past demand response events, specifically including whether users responded to load adjustment instructions on time and whether actual load reduction met expectations. These historical fulfillment rates are derived from the electricity marketing system and load management system, recording user response behavior under different times and electricity price conditions. In one possible implementation, a weighted average method is used to calculate the user's fulfillment rate.

[0142] Load forecast accuracy measures the deviation between actual user electricity consumption and predicted values, reflecting the predictability of user electricity consumption patterns. Forecast data is sourced from load management systems and smart meters. In one possible implementation, mean squared error or mean absolute percentage error is used to quantify the difference between actual user electricity consumption and predicted values.

[0143] The user's historical response fulfillment rate and load forecast accuracy are transformed into a quantifiable credit rating. In one possible implementation, a fuzzy logic evaluation method is used, which sets three evaluation dimensions—fulfillment rate, forecast accuracy, and response stability—and assigns weights to these three dimensions to ultimately calculate a comprehensive credit score.

[0144] Credit ratings are determined based on a comprehensive credit score. In one possible implementation, users are categorized into three levels: A, B, and C, corresponding to high, medium, and low credit, respectively. Level A users exhibit high fulfillment rates and stable load forecasting performance, Level B users show some fluctuations in fulfillment and forecasting, while Level C users demonstrate low fulfillment rates or large load forecasting errors.

[0145] S420: Design incentive mechanisms for high-credit users and punitive adjustment mechanisms for low-credit users.

[0146] Incentive mechanisms for high-credit users not only enhance users' willingness to respond but also improve the overall efficiency of demand response. Penalty adjustment mechanisms for low-credit users are crucial for ensuring fairness and fulfillment rates in demand response. Their goal is to constrain users with low fulfillment rates or large load forecasting errors through methods such as electricity price adjustments, dispatch authority restrictions, and credit downgrading, thereby prompting them to improve their electricity consumption behavior and enhance the overall efficiency of demand response.

[0147] In one possible implementation, the incentive mechanism for high-credit users includes: an electricity price discount incentive mechanism, a priority dispatch right incentive mechanism, and a points reward mechanism.

[0148] Specifically, the electricity price discount incentive mechanism allows high-credit users to enjoy lower electricity prices during specific periods, especially during peak load periods in the power system. Users can obtain additional electricity price benefits by actively reducing or shifting their load. For example, under the time-of-use pricing mechanism, high-credit users receive lower electricity prices than ordinary users during peak hours, encouraging them to reduce electricity demand and thus alleviating grid pressure. The priority dispatch incentive mechanism allows high-credit users to receive priority in the allocation of dispatch resources when demand response events occur. For example, they may receive priority in receiving discharge support from energy storage systems or priority in receiving dispatch instructions for adjustable loads. The points reward mechanism allows users to accumulate points by participating in demand response, fulfilling obligations on time, and maintaining accurate load forecasting, and redeem them for rewards in subsequent electricity consumption. For example, these rewards may include electricity bill reductions, priority access to dispatch resources, or even participation in electricity market transactions.

[0149] To improve the accuracy and effectiveness of the incentive mechanism, incentive strategies such as electricity price discounts, priority dispatch rights, and points rewards are combined with multi-dimensional comprehensive deviation indicators. Specifically, deviation indicator thresholds are set for users with different credit ratings, and the incentive intensity is dynamically adjusted based on the actual deviation indicators. For example, for users with credit rating A, if their deviation indicator is below 5%, an additional electricity price discount is applied; if the deviation indicator is below 3%, additional points rewards are applied. Similarly, for users with credit rating B, if their deviation indicator is below 7%, they receive a certain electricity price discount, and if the deviation indicator is below 5%, they receive additional points rewards.

[0150] In one possible implementation, the punitive adjustment mechanism for low-credit users includes: an electricity price adjustment mechanism, dispatch authority restrictions, and a credit downgrade mechanism.

[0151] Specifically, the electricity price adjustment mechanism adopts a dynamic pricing strategy based on credit rating, setting different price fluctuation percentages according to the user's credit rating. For example, users with credit rating C may see a 10%-15% price increase during peak hours, while users with credit rating B may see a 5%-10% increase. Dispatch authority restrictions employ a strategy linking credit rating with dispatch priority; users with lower credit ratings have lower priority in load resource dispatching. For example, they have fewer opportunities to receive resource support in areas such as energy storage system discharge and adjustable load dispatching. The credit downgrade mechanism sets dynamic adjustment rules for credit ratings. For example, if a user's fulfillment rate is below 70% for three consecutive demand response cycles, their credit rating will automatically be downgraded by one level. Similarly, if a user's load forecasting error exceeds 15% for three consecutive periods, their credit rating will also be lowered.

[0152] To ensure the effectiveness and fairness of the punitive adjustment mechanism, a dynamic credit management strategy is established. Specifically, a real-time assessment system based on credit rating and deviation indicators is adopted to continuously monitor users' performance and dynamically adjust punitive measures according to the assessment results. For example, if a user's performance rate is low within a demand response cycle, but their deviation indicators are within a controllable range, mild punitive measures are taken, such as a slight increase in electricity price, rather than an immediate downgrade. Similarly, if a user's deviation indicators continue to exceed the limits, the severity of the penalties is gradually increased, such as increasing the electricity price fluctuation ratio, reducing dispatch priority, or even triggering a credit downgrade.

[0153] The S100-S400 system constructs a high-precision dynamic benchmark library through deep fusion of multi-source heterogeneous data and a priority-weighted matching algorithm, significantly improving the accuracy of scene identification. It innovatively adopts single / multi-dimensional deviation indicators and a three-level anomaly judgment mechanism, combined with a standby load resource pool scheduling model to achieve efficient resource compensation, balancing grid stability and economy. Based on a fuzzy logic credit evaluation system with historical performance rate and prediction accuracy, it forms a positive incentive and negative constraint closed loop through differentiated reward and punishment strategies, effectively promoting sustainable development of user behavior regulation and demand response. Overall, it has strong adaptability, closed-loop optimization capabilities, and long-term management value.

[0154] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0155] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0156] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps are disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0157] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein apply to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features of the invention herein.

[0158] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications or equivalent substitutions made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A demand response optimization method for electricity marketing operations, characterized in that, include: By collecting users' historical electricity consumption data and linking it with weather, calendar, and electricity price data, a dynamic benchmark library for multiple scenarios is constructed, and matching priorities are set. Based on the user's historical electricity consumption data, weather, calendar and electricity price data in the current scenario, the optimal scenario is determined by matching with a multi-scenario dynamic benchmark library according to priority, and single-dimensional deviation index and multi-dimensional comprehensive deviation index are calculated. Set multi-dimensional comprehensive anomaly thresholds for judgment, output a tiered list of abnormal users, and trigger anomaly response flags; For single-dimensional deviation indicators and multi-dimensional comprehensive deviation indicators of abnormal users, a scheduling model for the standby load resource pool is established, and the optimal scheduling instruction is output to compensate for the load deviation of abnormal users. The system generates a dynamic credit rating for users by combining historical response fulfillment rates and load forecast accuracy. Users with high credit ratings receive premium incentives, while users with low credit ratings are subject to punitive adjustments.

2. The demand response optimization method for electricity marketing operations according to claim 1, characterized in that, The construction of the multi-scenario dynamic benchmark library includes: The user's historical electricity consumption data is aligned with meteorological features, calendar features, and electricity price features according to the time dimension to form a feature matrix. The meteorological features include temperature, humidity, wind speed, and precipitation; the calendar features include holidays, weekends, seasons, and special events; and the electricity price features include real-time electricity price, historical electricity price, time-of-use electricity price type, and electricity price volatility.

3. The demand response optimization method for electricity marketing operations according to claim 2, characterized in that, The setting of matching priority includes: Strongly correlated features are assigned a weight of 0.5-0.6, including temperature, real-time electricity price, holidays, and seasons; The moderately relevant features are assigned a weight of 0.3-0.4, including humidity, time-of-use electricity pricing type, and weekends; Weakly correlated features are assigned a weight of 0.1-0.2, including wind speed, precipitation, special events, electricity price volatility, and historical electricity prices.

4. The demand response optimization method for electricity marketing operations according to claim 1, characterized in that, The optimal scene determination process includes: feature priority weighting, calculation of matching value, and determination of optimal matching scene; Among them, the feature priority weighting is as follows: 0.5-0.6 for strongly correlated features, 0.3-0.4 for moderately correlated features, and 0.1-0.2 for weakly correlated features; Match value calculation: Calculated using the following formula: Where: D i x is the matching value between the current scene and the i-th scene in the benchmark library; ij x is the feature j value of the i-th scene in the benchmark library; cj w represents the feature j value of the current scene. j The priority weights for feature j; Optimal matching scenario determination: Select the baseline scenario with the smallest matching value as the matching result for the current scenario.

5. The demand response optimization method for electricity marketing operations according to claim 1, characterized in that, The single-dimensional deviation indicators include electricity consumption behavior deviation indicators, meteorological deviation indicators, calendar deviation indicators, and electricity price deviation indicators. The multi-dimensional comprehensive deviation indicator integrates the single-dimensional deviation indicators using a weighted average method, and the calculation formula is as follows: Among them, S d The value of w represents a multi-dimensional comprehensive deviation from the index. j Let Δ be the priority weight of feature j. j The deviation index of feature j, i.e., the single-dimensional deviation index, where n is the number of features, n = 13.

6. The demand response optimization method for electricity marketing operations according to claim 5, characterized in that, The output of the tiered list of abnormal users includes: Level 1 anomaly: The overall deviation index exceeds the threshold by 15% or any single dimension deviation index exceeds 20%; Level 2 anomaly: The overall deviation index exceeds the threshold by 10%-15% and the single-dimensional deviation index is ≤20%; Level 3 anomalies: The overall deviation index exceeds the threshold by 5%-10% and the single-dimensional deviation index is ≤15%.

7. The demand response optimization method for electricity marketing operations according to claim 1, characterized in that, The objective function of the standby load resource pool scheduling model is: Where, ΔP i ξ represents the adjustment amount of the i-th load resource. i C represents the corresponding weighting coefficient. j Let λ be the unit adjustment cost of the j-th resource. c Weighting factors for user credit rating; The constraints of the standby load resource pool scheduling model include physical constraints, time constraints, and economic constraints. Among them, physical constraints: P min,i ≤P i ≤P max,i , where P min,i and P max,i Let P represent the minimum and maximum regulation capabilities of the i-th load resource, respectively. i This represents the adjustable load of the i-th load resource; Time constraint: t start,i ≤t response,i ≤t end,i , where t start,i and t end,i Let t represent the available time window for the i-th load resource, and t represent the available time window for the i-th load resource. response,i Indicates its response time; Economic constraints: Among them, C i C represents the unit adjustment cost of the i-th load resource. max This is the upper limit of the scheduling budget.

8. The demand response optimization method for electricity marketing business according to claim 1, characterized in that, The dynamic credit rating assessment for users includes: Based on three dimensions—historical response fulfillment rate, load forecast accuracy, and response stability—a fuzzy logic evaluation method is used to calculate a comprehensive credit score, which is then divided into three levels: A-high credit, B-medium credit, and C-low credit. The incentive mechanism for high-credit users includes: Electricity price discount incentives: Class A users enjoy a 5%-10% discount on electricity prices during peak hours; Priority dispatching incentive: Class A users have priority in obtaining energy storage discharge or adjustable load dispatching resources; Points reward mechanism: Users accumulate points based on their performance rate and deviation from the target, which can be used for electricity fee reduction or priority resource allocation; The punitive adjustment mechanism for low-credit users includes: Dynamic electricity price increase: C-level users will have their electricity price increased by 10%-15% during peak hours; Dispatch authority restrictions: Class C users have their priority reduced by 50% in energy storage discharge and load dispatch; Credit downgrade rules: A credit rating downgrade is triggered when the performance rate is below 70% for three consecutive cycles or the load forecasting error exceeds 15%.

9. The demand response optimization method for electricity marketing operations according to claim 1, characterized in that, The abnormal response flag triggering strategy includes: A level one anomaly triggers a load adjustment command and notifies the dispatch center. A level 2 anomaly alert is sent to the user terminal. Level 3 anomalies generate abnormal behavior records and are incorporated into the basis for credit rating adjustments.

Images