Polymerization main body flexible load prediction regulation and control method based on energy storage economy

Abstract

The invention discloses a polymer main body flexible load prediction regulation and control method based on energy storage economy, and relates to the related field of energy storage optimization, and the method comprises the steps: carrying out the load curve construction of a flexible load of an energy storage main body, including a total load curve and a unit load curve; introducing an option contract, taking the load curve as a reference, executing a game between the risk power and the stable power, executing stable power-based internal shrinkage processing on the load curve, and determining a flexible resource-based stable option and risk power curve; by comparing the electricity price curve of the short time zone with the energy storage charge and discharge cost, carrying out charge and discharge excitation-based fine tuning processing on the risk electric power curve, and determining a regulation and control electric power curve; and performing energy storage management of the polymerization main body by stabilizing the option and regulating and controlling the electric power curve. The technical problem that the energy storage economical efficiency is reduced due to the fact that existing flexible load predictive regulation and control are insufficient in accuracy is solved, and the technical effects of improving the accuracy of flexible load predictive regulation and control and improving the energy storage economical efficiency are achieved.

Description

Technical Field

[0001] This application relates to the field of energy storage optimization, and in particular to a method for predicting and controlling flexible loads of aggregated entities based on the economics of energy storage. Background Technology

[0002] With a high proportion of new energy sources being integrated into the power system, accurate forecasting and economic regulation of flexible loads by aggregated entities play a crucial role in maintaining stable grid operation and improving market transaction efficiency. Currently, traditional forecasting and regulation methods based on historical data statistical analysis and simple rule settings are mainly employed. These methods involve analyzing past load and electricity price data, combined with pre-set fixed rules, to predict load changes and implement control strategies. However, the actual operation of the power system is affected by various complex factors, such as the randomness of renewable energy output, dynamic changes in user electricity consumption behavior, and frequent fluctuations in market electricity prices. This makes it difficult for statistical analysis based on historical data to accurately capture real-time changes, resulting in significant deviations between forecast results and actual conditions. Simultaneously, simple rule settings lack flexibility and adaptability, failing to dynamically adjust according to real-time market information and system status, leading to a mismatch between control strategies and actual demand, affecting the economic viability of energy storage, and hindering the efficient forecasting and accurate regulation of flexible loads.

[0003] Currently, the accuracy of flexible load forecasting and control is insufficient, which leads to a decrease in the economic efficiency of energy storage. Summary of the Invention

[0004] This application provides a method for predicting and controlling flexible loads of aggregated entities based on energy storage economics. It employs techniques such as constructing total load curves and unit load curves for flexible loads of energy storage entities, introducing option contracts, conducting risk and stable power game based on the load curves, determining stable option and risk power curves through curve shrinking, comparing short-time time zone electricity price curves with energy storage charging and discharging costs, fine-tuning the risk power curve to determine the control power curve, and managing aggregated entity energy storage based on the stable option and control power curves. These techniques address the technical problem of insufficient accuracy in existing flexible load prediction and control methods, which leads to reduced energy storage economics. The method achieves the technical effect of improving the accuracy of flexible load prediction and control and enhancing the economics of energy storage.

[0005] This application provides a method for predicting and regulating flexible loads of an aggregated entity based on the economics of energy storage, comprising: constructing a load curve for the flexible load of the energy storage entity, wherein the load curve includes a total load curve and a unit load curve; introducing an option contract, using the load curve as a benchmark, performing a game between risky power and stable power, performing a shrinking process based on stable power on the load curve, and determining a stable option and a risky power curve based on flexible resources; by comparing the electricity price curve of a short time zone with the energy storage charging and discharging cost, performing a fine-tuning process based on charging and discharging incentives on the risky power curve, and determining a regulating power curve; and using the stable option and the regulating power curve to perform energy storage management of the aggregated entity.

[0006] In a possible implementation, a load curve is constructed for the flexible load of the energy storage entity, and the following processing is performed: load elements are set, wherein the load elements include at least electrical elasticity, time elasticity, power elasticity and cost elasticity, and the elasticity range is defined by upper and lower limits; based on the load elements, an interval curve is constructed for the aggregation entity to determine the total load curve.

[0007] In a possible implementation, after determining the total load curve, the following process is performed: the total load curve of the aggregate is deconstructed into multiple unit load curves, wherein the smallest indivisible electrical action unit is used as the basis for deconstruction, and the load elements of each unit load curve measure the electrical consumption behavior; the total load curve and the multiple unit load curves are added to the load curve.

[0008] In a possible implementation, the following processing is performed: deploying option contracts in the energy storage grid based on dynamic game coordination between the risky power portion and the stable power portion of the aggregated flexible load.

[0009] In a possible implementation, before executing the game between risky power and stable power, the following processing is performed: obtaining the real-time load status of the aggregation subject, performing fuzzy prediction of load trends, and determining the trend prediction result; based on the trend prediction result, performing a first narrowing process on the upper and lower limits of the load curve to determine the real-time load curve.

[0010] In a possible implementation, a game between risky power and stable power is performed, and the following processing is carried out: the option contract is activated, and a classification game processing based on stable power and risky power is performed on the real-time load curve to determine a first stable power component and a second risky power component; based on the first stable power component, the real-time load curve is subjected to a second shrinkage processing to determine the risky power curve with load uncertainty.

[0011] In a possible implementation, the following process is performed: the flexible resources based on the first stable power segment are packaged into dynamic stable options with a start time window, power curve shape, and duration.

[0012] In a possible implementation, the risk power curve is fine-tuned based on charging and discharging incentives by comparing the electricity price curve of a short time zone with the energy storage charging and discharging cost. The following processing is performed: a short time zone is set, wherein the short time zone is the time span for real-time fine-tuning; the risk power curve is fine-tuned according to the first short time zone to determine the control power curve, wherein the first short time zone is the next adjacent short time period of the real-time prediction time node.

[0013] In a possible implementation, the risk power curve is fine-tuned according to the first short-term time zone by performing the following process: retrieving the electricity price curve and energy storage charging and discharging cost of the first short-term time zone; comparing the electricity price curve and energy storage charging and discharging cost, and fine-tuning the risk power curve by measuring the difference.

[0014] In a possible implementation, the following process is performed: if the electricity price at the first node of the electricity price curve is higher than the energy storage charging and discharging cost, it is marked as a discharge incentive; according to the discharge incentive, the boundary of the first node of the risk power curve is adjusted upward, wherein the first node is the time node corresponding to the first node electricity price in the first short-time zone of the risk power curve.

[0015] In a possible implementation, the following process is performed: if the electricity price at the first node of the electricity price curve is lower than the energy storage charging and discharging cost, it is marked as a charging incentive; based on the charging incentive, the first node of the risk electricity curve is adjusted downward.

[0016] In a possible implementation, after fine-tuning the risk power curve, the following processing is performed: updating the charging and discharging incentives in real time according to the changes in the electricity price curve and the energy storage charging and discharging cost; and updating the risk power curve according to the real-time updated charging and discharging incentives.

[0017] In a possible implementation, the following processing is performed: integrating the stability option with the regulation power curve as the load forecast result; and submitting the power application for the distribution network based on the load forecast result.

[0018] In a possible implementation, the energy storage management of the aggregation entity is performed using the stability option and the regulation power curve, and the following processes are executed: establishing a flexible evaluation system, wherein the flexible evaluation system is defined by coupling the flexible evaluation dimensions based on power, frequency regulation assistance, and ramping; and performing a power assessment of the aggregation entity based on the load forecast results according to the flexible evaluation system to determine the load evaluation results.

[0019] In a possible implementation, the following process is performed: based on the load evaluation results and the load forecast results, power guidance for the aggregation entity is provided based on flexible load.

[0020] The proposed method for flexible load forecasting and control based on energy storage economics involves several steps. First, a load curve is constructed for the flexible load of the energy storage entity, comprising a total load curve and unit load curves. Next, an option contract is introduced, and a game between risky and stable power is performed based on the load curve. The load curve undergoes a contraction process based on stable power to determine the stable option and risky power curves based on flexible resources. Then, by comparing the electricity price curves of short-term time zones with the energy storage charging and discharging costs, the risky power curve is fine-tuned based on charging and discharging incentives to determine the control power curve. Finally, the energy storage of the aggregated entity is managed using the stable option and the control power curve. This method achieves the technical effect of improving the accuracy of flexible load forecasting and control and enhancing the economics of energy storage. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0022] Figure 1 A schematic flowchart of the flexible load prediction and control method based on energy storage economics provided in this application embodiment.

[0023] Figure 2 This is a schematic diagram of the process for constructing a load curve in the flexible load forecasting and control method based on energy storage economics provided in the embodiments of this application. Detailed Implementation

[0024] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0025] This application provides a method for predicting and controlling flexible loads based on energy storage economics, such as... Figure 1 As shown, the method includes: Step S100: Construct load curves for the flexible load of the energy storage entity, wherein the load curves include the total load curve and the unit load curve.

[0026] Specifically, key parameters describing load variability, such as electrical elasticity and time elasticity, are defined, and their upper and lower limits are used to delineate the variable range. Based on these parameters, a total load range curve for the aggregated entity is generated, and the total curve is then broken down into independent curves for the smallest electricity consumption units, forming a multi-level load data system. For example, in an energy storage system in an industrial park, the total load curve reflects the overall trend of flexible load changes for the entire park's energy storage entity, while the unit load curves correspond to the flexible load changes of different factories or production lines within the park.

[0027] like Figure 2 As shown, in one possible implementation, a load curve is constructed for the flexible load of the energy storage entity. Step S100 further includes step S110, setting load elements, wherein the load elements include at least electrical elasticity, time elasticity, power elasticity, and cost elasticity, with upper and lower limits defining the elasticity range. Specifically, in the software platform, four types of elastic constraints are set for each aggregation entity through a parameter configuration interface. Electrical elasticity refers to the allowable deviation range of load power from a benchmark value, such as ±10%; time elasticity represents the adjustability of the flexible load over time, such as the ability to advance or postpone certain production tasks within a certain time range; power elasticity refers to the ability of load power to change within a certain range, such as a maximum power change rate of 5% per unit time; cost elasticity reflects the load's sensitivity to electricity costs, such as the acceptable fluctuation ratio of electricity costs for users. These parameters are stored in the database as upper and lower limits, serving as boundary conditions for curve generation. For example, for a factory's production equipment, setting its electrical elasticity upper limit to 1.2 times the rated power and the lower limit to 0.8 times indicates that the equipment can flexibly adjust its electricity consumption within this power range.

[0028] Step S120: Based on the load factors, construct an interval curve for the aggregated entity to determine the total load curve. Specifically, read the preset elastic upper and lower limits, and calculate the upper and lower load values ​​for each moment according to the time series, such as 96 points / day, forming a strip-shaped interval curve. The upper load value = baseline load × upper limit of power elasticity, and the lower load value = baseline load × lower limit of power elasticity. In practice, call the load prediction baseline curve, and then combine the elasticity parameters to generate the total load interval using an interval multiplier. For example, if the baseline load is 100 kW at a certain moment, the upper limit of power elasticity is 1.2, and the lower limit is 0.8, then the load interval at that moment is [80, 120] kW. Connecting these consecutive time points yields the total load interval curve.

[0029] Step S130 involves deconstructing the total load curve of the aggregate entity into multiple unit load curves. The deconstruction is based on the smallest indivisible electrical action unit, and the load elements of each unit load curve measure the electricity consumption behavior. Specifically, the smallest indivisible electrical action unit refers to the basic unit of electricity consumption behavior that cannot be further subdivided. Based on the characteristics of the electrical equipment, such as air conditioners, charging piles, and industrial machine tools, the smallest indivisible unit is defined, and each unit independently records its load elements.

[0030] In practical implementation, firstly, the system has a built-in device signature library that stores the standard operating characteristics of various typical electrical devices. Each device signature contains key data fields such as power characteristics, time characteristics, start-stop characteristics, and elastic parameters. Power characteristics include steady-state power values ​​and power change patterns; time characteristics include typical working periods and working cycles; start-stop characteristics include start-up identification signals and shutdown association conditions; elastic parameters include the device's allowed power adjustment range and time shift window. During real-time deconstruction, a multi-level pattern recognition algorithm is used. The first layer is steady-state feature-based matching, scanning the steady-state plateau segment of the total load curve and matching it with the steady-state power values ​​in the signature library. The second layer is transient feature-based identification, analyzing the power change rate and waveform characteristics for load abrupt changes. The third layer is association rule-based verification, using the logical relationships between devices for cross-validation. The specific deconstruction algorithm uses iterative subtraction. Step 1: Initialize the remaining load as the total load curve. Step 2: Process the identified devices in descending order of confidence level. This includes extracting a typical power curve template for the device from the device signature library, aligning the power template to the actual time axis based on the identified operating period, subtracting the device's load curve from the remaining load, and recording the complete load curve and resilience parameters for that unit. Step 3: Repeat Step 2 for the remaining load until all identifiable devices have been extracted. Step 4: Classify the remaining load as either unknown load or base load.

[0031] Step S140: Add the total load curve and multiple unit load curves to the load curve. Specifically, store the total load interval curve and all unit curves in a time-series database and establish a hierarchical index. This is implemented by assigning a master data ID to the total load curve, and linking each unit load curve as a child record to the master ID via a foreign key, while also marking metadata such as curve type, time resolution, and elasticity parameter version. For example, a load curve table can be created in MySQL, containing fields such as curve ID, parent curve ID, timestamp, upper load limit, and lower load limit, supporting fast queries by time range and curve level.

[0032] Step S200: Introduce option contracts, and based on the load curve, perform a game between risky power and stable power. Perform inward processing on the load curve based on stable power to determine the stable option and risky power curve based on flexible resources. In this process, the option contracts are deployed in the energy storage grid by coordinating the dynamic game between the risky power part and the stable power part based on the flexible load of the aggregated entity.

[0033] Specifically, by establishing a power finance contract mechanism, the physical load characteristics are transformed into tradable financial products, constructing a two-tiered decision-making framework to achieve load segmentation at the technical level and value allocation at the commercial level. The implementation employs a risk quantification model based on confidence intervals. It reads the load curve data generated in step S100, calculates the probability distribution of load values ​​for each time period based on statistical analysis of historical data, sets a risk tolerance threshold (e.g., 95% confidence level), and divides the load into a stable portion that is guaranteed to be available and a risky portion that requires market balancing. Regarding contract deployment, standardized option contract terms are generated, including parameters such as the strike price based on long-term power supply costs, the contract capacity of the stable power portion, and the exercise time window. These are deployed and recorded through a blockchain smart contract platform. For example, after analyzing the load data of an industrial user, the system determines that the portion with a 95% probability of not exceeding 500kW per day is considered stable power. A call option is issued for this portion, stipulating that when the spot price is higher than the strike price by 0.5 yuan / kWh, the option buyer can purchase this 500kW of power at a fixed price.

[0034] In one possible implementation, a game between risky and stable power is executed. Step S200 further includes step S210, which involves acquiring the real-time load status of the aggregation entity, performing fuzzy load trend prediction, and determining the trend prediction result. Specifically, load data is acquired in real time through smart meters and a data acquisition and monitoring system, with a sampling frequency of once per minute. A prediction engine based on the TS fuzzy model is used to fuzzify the input variables, including the current load value, load change rate, ambient temperature, and workday type. In the specific calculation, the activation intensity of each rule is calculated through a membership function, and a weighted average method is used for defuzzification to output a quantitative value of the load trend for the next 15 minutes.

[0035] Step S220: Based on the trend prediction results, the load curve undergoes a first narrowing process of its upper and lower limits to determine the real-time load curve. Specifically, based on the trend prediction results, an asymmetric interval narrowing strategy is adopted. For an upward trend, the lower limit remains unchanged, while the upper limit is narrowed, with the narrowing magnitude proportional to the trend strength. For a downward trend, the upper limit remains unchanged, while the lower limit is narrowed. A linear adjustment model is used: New upper limit = Original upper limit - Trend strength × (Original upper limit - Baseline value) × Narrowing coefficient. Simultaneously, coordinated adjustments are made for multiple consecutive time periods to ensure the temporal continuity of the load curve. Through dynamic interval adjustment, the load curve better reflects the real-time operating status.

[0036] Step S230: Activate the option contract and, using the real-time load curve, perform a classification game based on stable power and risky power to determine the first stable power portion and the second risky power portion. Specifically, at the beginning of each decision cycle, automatically activate expiring option contracts and execute the following classification process: obtain the spot market clearing price for that period from the blockchain platform; compare the spot price with the option exercise price; when the spot price is higher than the exercise price, the entire load for that period is classified as risky power, i.e., the buyer abandons the exercise; when the spot price is lower than or equal to the exercise price, the contract capacity portion is classified as stable power, and the excess portion is classified as risky power. Technically, a real-time price listener and an automatic classifier are used. The classification rule is expressed as follows: if the spot price > the exercise price, then risky power = total load, stable power = 0; otherwise, stable power = min(total load, contract capacity), risky power = total load - stable power, where min represents the minimum value. For example, if the total load is 550kW, the contracted capacity is 500kW, the execution price is 0.5 yuan, and the spot price is 0.45 yuan, then the stable power supply is 500kW, and the risky power supply is 50kW. Power risk is allocated in real time through the contract execution engine and price comparison mechanism.

[0037] Step S240: Based on the first stable power portion, the real-time load curve undergoes a second shrinkage process to determine the risk power curve for load uncertainty. Specifically, after completing the stable power allocation, the real-time load curve undergoes secondary processing, directly subtracting the allocated stable power portion from the original real-time load curve to generate a clean risk power curve. This is implemented using time-series data subtraction, performing the following operation at each time point: Risk load value = Total load value - Stable load value. Simultaneously, the risk power curve undergoes a rationality check to ensure non-negativity and continuity. For example, in a 96-point daily curve, the risk value is calculated point by point. When the total load at a certain point is 480kW and the stable power is 500kW, the risk load is set to 0kW to avoid negative values. The generated risk power curve is stored separately and its uncertainty characteristics are labeled, including parameters such as fluctuation range, probability of occurrence, and duration.

[0038] Step S250 involves packaging the flexible resources based on the first stable power component into a dynamic stable option, using the start-up time window, power curve shape, and duration. Specifically, three key technical parameters of the first stable power component are extracted: the start-up time window refers to the earliest and latest allowed start-up time; the power curve shape refers to mathematical descriptions such as constant power, ramp-up, ramp-down, and pulse; and the duration refers to the minimum and maximum operating times. During the packaging process, a template-based encapsulation technique is used. Standard contract templates, such as interruptible load contracts and adjustable load contracts, are selected based on resource characteristics, and then specific parameters are filled in to generate a machine-readable contract file. Simultaneously, the key value indicators of the option are calculated, including: flexibility value = adjustable capacity × adjustment range, and reliability value = availability probability × duration. A unique contract identifier is generated by registering the contract on the trading platform through a smart contract.

[0039] Step S300: By comparing the electricity price curve of a short time zone with the energy storage charging and discharging cost, the risk power curve is fine-tuned based on charging and discharging incentives to determine the control power curve.

[0040] Specifically, a dual-input, single-output control architecture is constructed. The input side receives real-time access to the nodal price curves of the electricity spot market and the full-cost model of the energy storage system. The output side generates economically optimized load control instructions. Technically, the following process is executed every 5 minutes: Forecasted electricity price data for the next 4 hours is obtained from the trading platform, along with dynamic cost data based on the battery management system's health status; cost-benefit analysis is used to calculate the charging and discharging economic indicators for each time period; and boundary adjustments are made to the risk power curve according to a preset control strategy. For example, if the peak electricity price reaches 0.8 yuan / kWh during the 14:30-15:00 period, while the energy storage discharge cost is only 0.3 yuan / kWh, an upper limit expansion operation is immediately performed on the risk load curve for that period, allowing more discharge power to participate in market trading.

[0041] In one possible implementation, by comparing the electricity price curve of a short-time zone with the energy storage charging and discharging cost, the risk power curve is fine-tuned based on charging and discharging incentives. Step S300 further includes step S310, setting a short-time zone, where the short-time zone is the time span for real-time fine-tuning. Specifically, the time range for optimization decisions is determined through a time window management mechanism. A sliding time window model is adopted, discretizing continuous time into multiple short-time zones, each with a fixed length of 4 hours, containing 48 5-minute intervals. The system maintains a circular buffer to store time series data, performing a window update every 5 minutes, removing the earliest time interval and adding the latest forecast data, always maintaining a 4-hour optimization perspective. The start and end times of the window are aligned with the electricity market trading hours; for example, if the current time is 10:05, the short-time zone is 10:05-14:05. A priority mechanism is also set, with intervals closer to the current time having higher weights and greater adjustment authority in the optimization calculation. The window parameters are dynamically adjustable through a configuration file, allowing for flexible modification of the duration and granularity based on market fluctuations.

[0042] Step S320: Based on the first short-time zone, fine-tune the risk power curve to determine the control power curve. The first short-time zone is the next adjacent short-time period after the real-time prediction time node. Specifically, a time-segmented priority control strategy is implemented, dividing the short-time zone into three decision levels: an emergency layer (0-30 minutes), an optimization layer (30-120 minutes), and a prediction layer (120-240 minutes), prioritizing control decisions for the emergency layer. Technically, a time-slicing algorithm is used, extracting the next complete trading period as the first short-time zone based on the current time. For example, if the current time is 10:03, and the electricity market trading periods are 10:00, 10:05, 10:10..., then the first short-time zone is determined to be the 5-minute period from 10:05 to 10:10. Fine-tuning the risk power curve for this period makes it more consistent with the actual electricity market situation and electricity demand, thereby determining a more reasonable control power curve.

[0043] In one possible implementation, the risk power curve is fine-tuned based on the first short-term time zone. Step S320 further includes step S321, retrieving the electricity price curve and energy storage charging / discharging cost of the first short-term time zone. Specifically, two data interfaces are called in parallel: one is the node electricity price data of the first short-term time zone obtained from the electricity market API interface, including 48 predicted prices at 5-minute granularity and their confidence intervals; the other is the real-time output of the energy storage cost calculation engine, which includes variable costs and fixed costs. Variable costs include battery depreciation costs and efficiency loss costs. Battery depreciation cost = initial cost / cycle life × current depth of discharge; efficiency loss cost = charging / discharging efficiency × electricity price difference; fixed costs include operation and maintenance costs, which = power × unit rate.

[0044] Step S322: Compare the electricity price curve with the energy storage charging and discharging cost, and fine-tune the risk electricity curve by measuring the difference. Specifically, establish an automatic decision-making mechanism based on threshold triggering. First, calculate the price difference index for each time period, i.e., price difference = electricity price - charging and discharging cost. Then, trigger the corresponding adjustment operation according to the preset bidirectional threshold. In terms of technical implementation, maintain a dynamic threshold parameter table, including charging trigger threshold and discharging trigger threshold. For each time period, if the price difference ≤ charging trigger threshold, trigger charging incentive; if the price difference ≥ discharging trigger threshold, trigger discharging incentive; otherwise, maintain the original state.

[0045] In one possible implementation, step S322 further includes step S3221: if the electricity price at the first node of the electricity price curve is higher than the energy storage charging and discharging cost, it is marked as a discharge incentive. Based on the discharge incentive, the boundary of the first node of the risk power curve is adjusted upwards, where the first node is the time node corresponding to the first node electricity price in the first short-time zone of the risk power curve. Specifically, the price difference is calculated as: first node electricity price - energy storage discharge cost. When the price difference exceeds a preset discharge trigger threshold, i.e., covering all costs and generating considerable profit, it is marked as a discharge incentive. After being marked as a discharge incentive, the load limit of the risk power curve at that first node is adjusted upwards, i.e., during periods of high electricity prices, energy storage discharge is used as much as possible to replace expensive grid electricity, thereby reducing the net load, i.e., the electricity obtained from the grid. The magnitude of the increase is determined by the real-time available discharge capacity of the energy storage system. The energy storage status is queried to obtain the current maximum dischargeable power. Then, the original upper limit value of the risk power curve at this node is modified to a new upper limit value. The calculation formula can be expressed as: new upper limit value = original upper limit value + maximum dischargeable power.

[0046] Step S3222: If the electricity price at the first node of the electricity price curve is lower than the energy storage charging and discharging cost, it is marked as a charging incentive. Based on the charging incentive, the boundary of the first node of the risk power curve is lowered. Specifically, the price difference is calculated as: first node electricity price - energy storage charging cost. When the price difference is lower than a preset charging trigger threshold, i.e., charging is significantly economical, it is marked as a charging incentive. After being marked as a charging incentive, the load lower limit of the risk power curve at that first node is lowered, i.e., during periods of low electricity price, cheap grid electricity is used as much as possible to charge the energy storage. The magnitude of the reduction is determined by the real-time available charging capacity of the energy storage system. The energy storage status is queried to obtain the current maximum rechargeable power. Then, the original lower limit of the risk power curve at that node is modified to a new lower limit, the calculation formula of which can be expressed as: new lower limit = original lower limit - maximum rechargeable power.

[0047] In one possible implementation, after fine-tuning the risk power curve, step S300 further includes step S330, updating the charging and discharging incentives in real time according to the changes in the electricity price curve and the energy storage charging and discharging costs. Specifically, a dynamic monitoring and evaluation module is established, which monitors two data streams in parallel. One is the external market stream, which continuously receives the latest node electricity price forecast data through the electricity market API and compares the latest electricity price curve with the electricity price curve used in the previous cycle. If the change in the electricity price of any node exceeds a preset price refresh threshold, it is determined to be a valid change. The other is the internal state stream, which receives real-time updates on the energy storage system status from the battery management system, including but not limited to changes in state of charge, health status, internal temperature, and available power. For example, when the state of charge drops to a certain level, it may lead to a decrease in maximum discharge capacity, or when the battery temperature is abnormal, the charging and discharging costs can be increased to protect the equipment. Once a change in either of the above data streams triggers the update condition, the price difference calculation and threshold comparison in step S322 are immediately re-executed for all affected time nodes. Specifically, for each first node requiring an update, the current price difference is recalculated as the latest node electricity price minus the latest energy storage charging and discharging cost. Based on the charging and discharging trigger thresholds, the node is re-determined and marked as either charging-incentivized, discharging-incentivized, or unincentivized. By establishing the system's dynamic response capability to changes in the market environment and internal state, the incentive state throughout the entire short-term time zone is ensured to remain consistent with the latest information.

[0048] Step S340: Update the risk power curve according to the real-time updated charging and discharging excitation. Specifically, load the latest charging and discharging excitation mapping table generated in step S330. For each time node where the excitation state changes, based on the new excitation marker, call the same boundary adjustment algorithm as in steps S3221 and S3222 to rewrite the boundaries of the corresponding node of the risk power curve. Encapsulate the curve update process into a transaction to ensure that either all updates succeed or all are rolled back, preventing intermediate states of data inconsistency. After the update is completed, perform a smoothness check to check whether the updated risk power curve is smooth and continuous in the time series, avoiding jumps in power commands due to drastic adjustments at a single node, and ensuring safe control of the energy storage device.

[0049] Step S400: Using the stability option and the control power curve, perform energy storage management of the aggregation entity.

[0050] Specifically, the controllable load corresponding to the stability option is superimposed with the finely adjusted risk power curve to generate a global load plan, which is then converted into control commands for energy storage devices, including charging, discharging, and standby. In practice, a command distribution engine is used to calculate the power required by each energy storage unit based on the load plan and distribute it to the battery management system via the Modbus / TCP protocol.

[0051] In one possible implementation, step S400 further includes step S410, integrating the stability option and the control power curve as the load forecast result. Specifically, a structured forecast result object is constructed, which contains three core parts: a basic guarantee part, a flexible control part, and metadata and constraints. The basic guarantee part represents the stable power capacity and its time window, represented by the stability option and sold through contracts. The flexible control part describes the risk power range after economic fine-tuning, as described by the control power curve. This part is a range band, with upper and lower limits defining the power space that the aggregation entity can flexibly adjust in the market through energy storage charging and discharging. The metadata and constraints simultaneously record the key parameters on which the forecast result is based, such as the option exercise price, the energy storage charging and discharging cost threshold, and the version of the elasticity parameter. At the software level, a data structure for the load forecast result is created, containing fields such as stable capacity, control upper limit curve, control lower limit curve, timestamp, and version number. This result is synchronized to the reporting module and the evaluation module via a data bus.

[0052] Step S420 involves submitting power declarations for the distribution network based on the load forecast results. Specifically, based on the integrated load forecast results, declaration data conforming to the rules is generated and submitted to the power trading platform. For the stable portion, since it has already been sold through option contracts, there is no need for repeated declarations in the spot market. For the flexible control portion, which is the focus of the declaration, the control power curve is used as the basis for the declared power range. The upper limit of the declaration is the upper limit curve of the control curve, and the lower limit is the lower limit curve of the control curve. The declaration service is invoked through the standardized API interface of the power trading platform, and the declaration quantity, time granularity, node information, etc., are filled in according to the data format required by the platform, along with an identity authentication certificate, for automated declaration. Simultaneously, a copy of each declaration is recorded for post-settlement and strategy backtracking analysis.

[0053] Step S430: Establish a flexibility evaluation system, defined by the coupling of three dimensions: electrical energy, frequency regulation assistance, and ramping. Specifically, a multi-dimensional, quantitative benchmark is constructed to assess the comprehensive value and performance of the aggregated entity's flexible resources. The electrical energy dimension evaluates the entity's ability to transfer energy in time and space, with indicators including transferable capacity, transfer efficiency, charge / discharge depth, and duration. This dimension is related to peak shaving and arbitrage profits. The frequency regulation assistance dimension evaluates the entity's ability to respond to grid frequency fluctuations and provide frequency regulation services, with indicators including frequency regulation capacity, response speed, regulation accuracy, and continuous availability. This dimension is related to the revenue of the ancillary services market. The ramping dimension evaluates the entity's ability to cope with drastic changes in grid net load, with indicators including maximum ramp rate, ramping response delay, and sustainable ramping power. The flexibility evaluation system couples these three dimensions to calculate a comprehensive flexibility index, which is a weighted sum of the scores for the electrical energy, frequency regulation, and ramping dimensions. The weights can be dynamically adjusted based on market value, grid demand, or the aggregated entity's own strategy.

[0054] Step S440: Based on the flexibility evaluation system, perform a power assessment on the aggregated entity based on the load forecast results to determine the load evaluation results. Specifically, map the regulation power curve and the real-time state parameters of the energy storage system from the load forecast results onto three dimensions of the flexibility evaluation system. In the energy dimension, calculate the area enclosed by the regulation power curve to assess the total available regulation energy. In the frequency regulation auxiliary dimension, analyze the fluctuation characteristics and response granularity of the regulation power curve to assess whether it can meet the requirements of the frequency regulation signal. In the ramp-up dimension, calculate the maximum difference in power values ​​between adjacent time periods of the regulation power curve to assess its ramp-up capability. Output a structured load evaluation result object, including individual scores for each dimension, a comprehensive flexibility index, and improvement suggestions. For example, if the current strategy has a low frequency regulation score due to overly gradual power changes, it is recommended to increase power fluctuation reserves during specific time periods.

[0055] Step S450: Based on the load evaluation results and the load forecast results, provide power guidance to the aggregation entity based on flexible loads. Specifically, if the load evaluation results show a consistently low score in a certain dimension, while the market value of that dimension increases, a strategy adjustment suggestion is generated. For example, it is suggested that in the next cycle, the discharge trigger threshold be reduced from 0.2 yuan / kWh to 0.15 yuan / kWh to increase the flexibility of power changes, thereby improving the frequency regulation service application capability. The final load forecast results are sent to the energy storage management system or load controller as the benchmark for setting power commands. The battery management system controls the charging and discharging behavior of the energy storage and the start-up and shutdown of the flexible load according to the upper and lower limits of the curve and the timing requirements.

[0056] This application's embodiments employ techniques such as constructing a total load curve and unit load curve for the flexible load of the energy storage entity, introducing option contracts, conducting risk and stable power game based on the load curve, determining the stable option and risk power curve through curve shrinking processing, comparing the short-term time zone electricity price curve with the energy storage charging and discharging cost, fine-tuning the risk power curve to determine the control power curve, and performing aggregated entity energy storage management based on the stable option and control power curve. These techniques solve the technical problem of insufficient accuracy in existing flexible load forecasting and control, which leads to reduced energy storage economics, and achieve the technical effect of improving the accuracy of flexible load forecasting and control and enhancing the economics of energy storage.

[0057] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A flexible load forecasting and control method based on energy storage economics, characterized in that, The method includes: A load curve is constructed for the flexible load of the energy storage entity, which includes the total load curve and the unit load curve; By introducing option contracts and using the load curve as a benchmark, a game is played between risky power and stable power. The load curve is then subjected to a shrinkage process based on stable power to determine a stable option and risky power curve based on flexible resources. By comparing the electricity price curves of short time zones with the charging and discharging costs of energy storage, the risk power curve is fine-tuned based on charging and discharging incentives to determine the control power curve. The energy storage management of the aggregation entity is carried out using the stability option and the control power curve.

2. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 1, characterized in that, The load curve for the flexible load of the energy storage system is constructed, including: Define load elements, wherein the load elements include at least electrical elasticity, time elasticity, power elasticity and cost elasticity, and define the elasticity range with upper and lower limits; Based on the load factors, interval curves are constructed for the aggregate body to determine the total load curve.

3. The flexible load forecasting and control method based on energy storage economics as described in claim 2, characterized in that, After determining the total load curve, the following is included: The total load curve of the aggregate is deconstructed into multiple unit load curves. The smallest indivisible electrical action unit is used as the basis for deconstruction, and the load elements of each unit load curve measure the electrical consumption behavior. The total load curve and multiple unit load curves are added to the load curve.

4. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 1, characterized in that, Option contracts are deployed in the energy storage grid based on the dynamic game coordination between the risky power component and the stable power component of the aggregated flexible load.

5. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 4, characterized in that, Before implementing the game between risky power and stable power, the following should be included: Obtain the real-time load status of the aggregated entity, perform fuzzy load trend prediction, and determine the trend prediction results; Based on the trend prediction results, the load curve is subjected to a first narrowing process of the upper and lower limits of the interval to determine the real-time load curve.

6. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 5, characterized in that, The game between risky power and stable power includes: Activate the option contract, and use the real-time load curve to perform a classification game based on stable power and risky power to determine the first stable power component and the second risky power component; Based on the first stable power section, the real-time load curve is subjected to a second shrinkage process to determine the risk power curve for load uncertainty.

7. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 6, characterized in that, The flexible resources based on the first stable power component are packaged into dynamic stable options based on the start-up time window, power curve shape, and duration.

8. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 1, characterized in that, By comparing the electricity price curves of short-term time zones with the charging and discharging costs of energy storage, the risk power curve is fine-tuned based on charging and discharging incentives, including: Set a short time zone, wherein the short time zone is the time span for real-time fine-tuning; Based on the first short time zone, the risk power curve is fine-tuned to determine the control power curve, wherein the first short time zone is the next adjacent short time period after the real-time prediction time node.

9. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 8, characterized in that, Based on the first shortest time zone, the risk power curve is fine-tuned, including: Retrieve the electricity price curve and energy storage charging and discharging costs for the first shortest time zone; By comparing the electricity price curve with the energy storage charging and discharging cost, the risk power curve is fine-tuned by measuring the difference.

10. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 9, characterized in that, If the electricity price at the first node of the electricity price curve is higher than the energy storage charging and discharging cost, it is marked as a discharge incentive. Based on the discharge excitation, the first node of the risk power curve is adjusted upwards at the boundary, wherein the first node is the time node corresponding to the first node electricity price in the first short time zone of the risk power curve.

11. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 10, characterized in that, If the electricity price at the first node of the electricity price curve is lower than the energy storage charging and discharging cost, it is marked as a charging incentive. Based on the charging incentive, the first node of the risk power curve is adjusted downwards.

12. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 11, characterized in that, After fine-tuning the risk power curve, the following steps are included: The charging and discharging incentives are updated in real time according to the changes in the electricity price curve and the energy storage charging and discharging costs. The risk power curve is updated based on the real-time updated charging and discharging excitation.

13. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 1, characterized in that, The stability option and the control power curve are integrated as the load forecast result; The power supply application for the distribution network is based on the load forecast results.

14. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 13, characterized in that, Using the stability option and the control power curve, the energy storage management of the aggregation entity includes: A flexible evaluation system is established, wherein the flexible evaluation system is defined by the coupling of flexible evaluation dimensions based on electrical energy, frequency modulation assistance, and hill climbing. Based on the aforementioned flexible evaluation system, a power assessment is conducted on the aggregated entity based on the load forecast results to determine the load evaluation results.

15. The aggregated flexible load forecasting and control method based on energy storage economics as described in claim 14, characterized in that, Based on the load evaluation results and the load forecast results, power guidance based on flexible load is provided to the aggregation entity.

Images