Industrial and commercial energy storage-based demand control method and system

Abstract

The application relates to the technical field of energy storage control, and discloses a demand control method and system based on industrial and commercial energy storage, which comprises the following steps: acquiring and standardizing real-time load demand, energy storage state and power grid scheduling data; establishing a load prediction model to generate demand baseline parameters and construct a dynamic adjustment rule library; checking the energy storage output in real time, analyzing the dynamic response trend of the power difference value, and generating demand control indexes to determine the charging and discharging strategy. The system comprises a load monitoring module, a strategy generation module, a dynamic execution module and a control determination module, which respectively realize data acquisition and processing, model construction, output checking and analysis and strategy determination functions. Through multi-dimensional data integration, dynamic rule library optimization and multi-constraint coordination technology, the precision and economy of industrial and commercial energy storage demand control are improved, the demand charge can be effectively reduced, the power utilization rate is improved, and the application is suitable for energy storage management and power grid collaborative optimization of industrial and commercial users.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy storage control, in particular to a demand control method and system based on industrial and commercial energy storage. BACKGROUND

[0002] Under the background of global energy transformation, the scientificity and economy of electricity management in industrial and commercial fields, as an important subject of energy consumption, are increasingly concerned. The traditional industrial and commercial electricity mode mainly relies on direct power supply from the power grid. With the continuous growth of electricity demand and the complication of electricity price policy, this mode gradually exposes many problems. On the one hand, industrial and commercial users are under pressure of high demand charge, which is calculated according to the maximum electricity demand of users. If the user's electricity load fluctuates greatly, it is easy to cause demand to exceed the standard, thereby increasing the cost of electricity. On the other hand, the load regulation capacity of the power grid for industrial and commercial users is limited. During the peak period of electricity consumption, the load pressure of the power grid increases, which may cause power supply shortage and unstable voltage, affecting the normal production and operation of users. In addition, during the valley period, there is a problem of waste of electric energy. Moreover, the existing energy storage system often lacks effective coordination with user load demand and power grid dispatching information in industrial and commercial applications, and cannot dynamically adjust the charging and discharging strategy according to real-time data, resulting in that the role of the energy storage system cannot be fully played.

[0003] At the same time, with the continuous development of smart grid and energy storage technology, higher requirements are put forward for the intelligent control of industrial and commercial energy storage systems. Traditional demand control methods are mostly based on fixed thresholds and simple rules, which cannot adapt to the dynamic changes of load demand and real-time adjustments of power grid dispatching. For example, in the face of time-of-use electricity price policy, the traditional method is difficult to adjust the charging and discharging behavior of the energy storage system in a timely manner according to the electricity price signal, and cannot fully utilize the low-price electric energy in the valley period to reduce the cost of electricity. Moreover, the existing system lacks the ability in data processing and analysis, and cannot effectively integrate and mine a large amount of real-time load data, energy storage state data and power grid dispatching data, making it difficult to establish an accurate load prediction model and a dynamic adjustment rule library, resulting in low precision and effectiveness of demand control.

[0004] In addition, industrial and commercial users have a variety of electricity equipment, and the load types are complex and diverse, with great differences in power curves and electricity characteristics of different equipment, which also increases the difficulty of demand control. Traditional control methods cannot finely manage different types of loads, and it is difficult to achieve optimal allocation of loads and reasonable scheduling of energy storage systems. At the same time, the running state of the energy storage system itself, such as SOC value, charging and discharging rate and equipment health, will also affect the effect of demand control. The traditional method often ignores the dynamic changes of these factors, and cannot adjust the control strategy in a timely manner to ensure the safe and stable operation of the energy storage system. SUMMARY

[0005] The present application aims to provide a demand control method and system based on industrial and commercial energy storage to solve the problems raised in the background.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions: a demand control method based on industrial and commercial energy storage, the method comprising:

[0007] Obtaining real-time load demand data, energy storage system operation state data and power grid dispatching information data of industrial and commercial users, and performing data standardization processing on the load demand data, energy storage state data and power grid dispatching data;

[0008] Based on the standardized data, a load prediction model is established to generate demand baseline parameters, and a control strategy set is formed according to the feature optimization of the demand baseline parameters, and a dynamic adjustment rule library is constructed in combination with the power grid dispatching information;

[0009] According to the dynamic adjustment rule library, the output of the energy storage system is matched and verified in real time, the power difference between the load demand and the energy storage output is identified by using multi-constraint coordination technology, and the trend of the dynamic response state of the power difference is analyzed;

[0010] Based on the verification result of the dynamic adjustment rule library, a demand control index is generated, and the charging and discharging adjustment strategy of the energy storage system is determined according to the demand control index.

[0011] Preferably, the real-time load demand data includes time-periodic equipment power curve and load type identification; the energy storage system operation state data includes energy storage SOC value, charging and discharging rate and equipment health degree parameter; the power grid dispatching information data includes time-of-use price signal and power grid carrying margin parameter; the load demand data, energy storage state data and power grid dispatching data are standardized, the standardization processing includes data cleaning, time stamp alignment, abnormal data interpolation correction and dimension unification; the load demand data is processed to generate a load feature matrix, and the energy storage state data is processed to generate an energy storage state vector; the load feature matrix and the power grid dispatching information are associated by time period to form a multi-dimensional constraint condition set.

[0012] Preferably, establishing a load prediction model to generate demand baseline parameters comprises the following steps:

[0013] Statistical feature parameters are extracted from the standardized load feature matrix, the statistical feature parameters include time period average value, power range value and change rate parameter, and each statistical feature parameter is combined to generate demand baseline parameters according to a preset proportion;

[0014] A dynamic threshold is set for the demand baseline parameters in the multi-dimensional constraint condition set, the feature parameters meeting the constraint boundary are selected to form an initial strategy set, and the feature parameters exceeding the constraint boundary are excluded;

[0015] a time series comparison analysis is performed on each feature parameter in the initial strategy set, a difference amount of the load demand and the energy storage output in adjacent periods is extracted, an absolute value of the difference amount is calculated and marked as a period deviation parameter;

[0016] The period deviation parameter is compared with a preset fluctuation threshold, feature parameters exceeding the fluctuation threshold are screened and added to the dynamic adjustment rule library, and the price sensitivity feature is supplemented to the dynamic adjustment rule library according to the power grid dispatching information.

[0017] Preferably, the trend analysis of the dynamic response state of the power difference value includes the following steps:

[0018] The instantaneous value of the load demand power and the instantaneous value of the energy storage output power are obtained in real time, the absolute value of the power difference value is calculated, and if the power difference value exceeds a preset allowable range, it is marked as a demand fluctuation event;

[0019] The number of demand fluctuation events in a preset period is counted, and the SOC change rate of the energy storage system and the real-time price fluctuation amplitude of the power grid are obtained synchronously;

[0020] The demand fluctuation index is calculated according to the number of demand fluctuation events, the SOC change rate and the price fluctuation amplitude, and if the demand fluctuation index exceeds a preset critical value, an energy storage output adjustment instruction is triggered.

[0021] Preferably, the check result of the dynamic adjustment rule library in the historical operation data of the same user is extracted, the execution effect of each adjustment rule is weighted and calculated to generate an adjustment coefficient, and if the adjustment coefficient continuously exceeds a preset trigger number, it is determined that the energy storage charging and discharging strategy switching is executed.

[0022] Preferably, the present application also includes a demand control system based on industrial and commercial energy storage, which is applied to the implementation of the above-mentioned demand control method based on industrial and commercial energy storage, and the system includes a load monitoring module, a strategy generation module, a dynamic execution module and a control determination module.

[0023] The load monitoring module is used to obtain real-time load demand data, energy storage system operation state data and power grid dispatching information data and perform standardized processing;

[0024] The strategy generation module is used to establish a load prediction model to generate demand baseline parameters and construct a dynamic adjustment rule library;

[0025] The dynamic execution module is used to match and verify the energy storage output according to the dynamic adjustment rule library, and analyze the power difference value of the load demand and the energy storage output;

[0026] The control determination module is used to generate demand control indicators and determine the energy storage regulation strategy.

[0027] Preferably, the load monitoring module comprises a user-side acquisition unit and a power grid information receiving unit;

[0028] The user-side acquisition unit is configured to acquire the equipment power curve of the industrial and commercial user and the operating parameter of the energy storage system;

[0029] The power grid information receiving unit is configured to acquire the time-of-use electricity price signal and the power grid carrying margin data.

[0030] Preferably, the strategy generation module comprises a baseline modeling unit and a rule optimization unit;

[0031] The baseline modeling unit is configured to extract the time period average value, the power range value and the change rate parameter from the standardized data;

[0032] The rule optimization unit is configured to screen the characteristic parameters meeting the constraint boundary and supplement the electricity price sensitivity characteristics to the dynamic adjustment rule library.

[0033] Preferably, the dynamic execution module comprises a difference calculation unit and a trend analysis unit;

[0034] The difference calculation unit is configured to detect the power difference between the load demand and the energy storage output and mark the demand fluctuation event;

[0035] The trend analysis unit is configured to calculate the demand and quantity deviation index in combination with the SOC change rate and the electricity price fluctuation amplitude.

[0036] Preferably, the control determination module comprises a coefficient generation unit and a strategy switching unit;

[0037] The coefficient generation unit is configured to perform weighted calculation on the execution effect of the adjustment rule;

[0038] The strategy switching unit is configured to determine the energy storage charging and discharging strategy switching according to the continuous triggering number of the adjustment coefficient.

[0039] Compared with the prior art, the present application has the following beneficial effects:

[0040] The application provides a demand control method and system based on industrial and commercial energy storage, which significantly improves the accuracy and economy of demand control of industrial and commercial energy storage systems through multi-dimensional data integration and intelligent analysis. Specifically, the scheme obtains real-time load demand data, energy storage system operation state data and power grid dispatching information data, and performs standardization processing, realizing efficient integration of multi-element data, laying a foundation for subsequent accurate control. By establishing a load prediction model to generate demand baseline parameters, and combining with multi-dimensional constraint conditions to build a dynamic adjustment rule library, the load history and real-time change trend can be dynamically optimized to optimize the charging and discharging strategy of the energy storage system, effectively reduce the risk of demand exceeding, and reduce demand electricity charges. For example, by extracting statistical characteristic parameters and setting dynamic threshold values, abnormal periods of load fluctuation can be accurately identified, and energy storage output can be adjusted in advance to avoid demand exceeding caused by load mutation.

[0041] In the dynamic response mechanism, by real-time matching and verifying the energy storage system output, the power difference between load demand and energy storage output is identified, and the demand deviation index is calculated by combining parameters such as SOC change rate and electricity price fluctuation amplitude, which can quickly respond to load fluctuation events and trigger energy storage output adjustment instructions. The application of this multi-constraint coordination technology enables the energy storage system to flexibly adjust the charging and discharging strategy under different power grid dispatching conditions (such as time-of-use price signals) and energy storage states, achieving the optimization goal of charging low-cost electricity during low-peak periods and discharging to reduce dependence on the grid during peak periods, significantly improving the efficiency and economic benefits of electricity utilization. At the same time, by analyzing the execution effect of historical adjustment rules to generate adjustment coefficients, it can automatically determine whether to switch the charging and discharging strategy, realizing self-optimization of the control strategy and further improving the adaptability and intelligent level of the system.

[0042] From the system architecture, the hierarchical design of the load monitoring module, the strategy generation module, the dynamic execution module and the control determination module ensures the efficient cooperation of data acquisition, model construction, real-time control and strategy optimization. The independent setting of the user-side acquisition unit and the power grid information receiving unit ensures the real-time and accuracy of load data and power grid information; the cooperation of the baseline modeling unit and the rule optimization unit realizes the organic combination of load prediction and dynamic updating of rules; the linkage of the difference calculation unit and the trend analysis unit ensures the real-time monitoring and trend prediction of the power difference; the cooperation of the coefficient generation unit and the strategy switching unit provides a reliable basis for the adaptive adjustment of the control strategy. This modular design not only improves the scalability and maintainability of the system, but also realizes the whole-process closed-loop management of demand control through the deep cooperation of each module.

[0043] In addition, the scheme can discover abnormal state of the energy storage system in time by real-time monitoring of the operating state of the energy storage system (such as the SOC value, the charging and discharging rate, and the equipment health degree parameter), early warning and adjustment of the control strategy, guarantee of the safe and stable operation of the energy storage system, and prolongation of the service life of the equipment. Meanwhile, in combination with the grid carrying margin parameter, the output of the energy storage can be actively adjusted when the grid load is tight, the pressure of the grid is reduced, the stability and reliability of the grid operation are improved, and the two-way interaction and collaborative optimization of the user side and the grid side are realized. In summary, the application improves the demand control efficiency of the industrial and commercial energy storage system through the intelligent control of data driving, the real-time update of the dynamic rule base, the collaborative operation of multiple modules, and the two-way response to the demands of the user and the grid, has significant technical advantages and practical value in reducing the electricity cost, improving the electricity utilization rate, and guaranteeing the stability of the grid. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 A working principle diagram of the demand control method and system based on industrial and commercial energy storage is shown.

[0045] Figure 2 A flowchart of data standardization processing is shown.

[0046] Figure 3 A flowchart of power difference trend analysis is shown.

[0047] Figure 4 A flowchart of charging and discharging strategy switching determination is shown. DETAILED DESCRIPTION

[0048] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0049] Please refer to Figures 1-4 The application relates to a demand control method based on industrial and commercial energy storage, which specifically comprises the following steps:

[0050] The system acquires the load demand data of industrial and commercial users, the operating state data of the energy storage system, and the dispatching information data of the power grid in real time through the user-side acquisition unit of the load monitoring module and the power grid information receiving unit. The load demand data includes time-division equipment power curve and load type identification, such as power fluctuation of production equipment in different shifts and equipment type (such as motor, heating equipment, etc.); the operating state data of the energy storage system includes energy storage SOC value, charging and discharging rate, and equipment health degree parameters, such as remaining battery percentage, current charging and discharging power value, and cell resistance change trend; the dispatching information data of the power grid includes time-of-use electricity price signal (such as peak-valley-flat period electricity price) and power grid carrying margin parameters (such as remaining capacity of the transformer area).

[0051] The above data is standardized, and the specific process includes: removing duplicate or invalid data through data cleaning, synchronizing different source data to a unified time coordinate system by timestamp alignment technology, correcting abnormal data points by cubic spline interpolation method, and eliminating data dimension differences by dimension unification processing (such as unifying power units to kW and electricity units to kWh). After processing, the load demand data generates a three-dimensional load feature matrix containing time period-power-type, the energy storage state data generates a one-dimensional energy storage state vector containing SOC, charging and discharging rate, and health degree, and the load feature matrix and the power grid dispatching information are associated according to time dimension to form a multi-dimensional constraint condition set containing time period, electricity price, and power grid capacity.

[0052] Based on the standardized data, the strategy generation module establishes a load prediction model through the baseline modeling unit. First, statistical feature parameters are extracted from the load feature matrix, including average power value of each period (reflecting typical load level), power range value (reflecting load fluctuation range), and change rate parameter (reflecting load increase and decrease rate), and demand baseline parameters are generated by combining them according to a preset proportion (such as average value accounting for 50%, range value accounting for 30%, and change rate accounting for 20%), for example, the baseline power of a certain period is the average value ± range value x adjustment coefficient. Set dynamic threshold for demand baseline parameters in the multi-dimensional constraint condition set, which is dynamically adjusted in combination with time-of-use electricity price and carrying margin in the power grid dispatching information (such as reducing the upper limit of baseline power in peak period to reduce the pressure on the power grid). Filter the feature parameters that meet the constraint boundary to form an initial strategy set, and remove the parameters that exceed the constraint (such as abnormal values that exceed the lower limit of baseline power in valley period). Perform time series comparison analysis on the feature parameters in the initial strategy set, calculate the absolute value of the difference between load demand and energy storage output in adjacent periods, and mark it as period deviation parameter. Compare the period deviation parameter with the preset fluctuation threshold (such as ±5% of the baseline power), add the parameters that exceed the threshold to the dynamic adjustment rule library, and supplement the electricity price sensitivity characteristics (such as higher sensitivity to power deviation in peak period) according to the time-of-use electricity price signal of the power grid to form a dynamic adjustment rule containing period, deviation threshold, and electricity price sensitivity coefficient.

[0053] The dynamic execution module obtains the instantaneous value of the load demand power and the instantaneous value of the energy storage output power in real time through the difference calculation unit, and calculates the absolute value of the difference between the two. If the difference exceeds the preset allowable range (such as ±10% of the baseline power), it is marked as a demand fluctuation event. The number of events in a preset period (such as 15 minutes) is counted, and the SOC change rate (such as the percentage of SOC decrease per minute) of the energy storage system and the real-time electricity price fluctuation amplitude (such as the change rate of the current period electricity price compared with the baseline value) of the power grid are obtained synchronously. Based on the above data, the trend analysis unit calculates the demand deviation index through multi-constraint coordination technology, the formula is: demand deviation index = event number × weight 1 + SOC change rate × weight 2 + electricity price fluctuation amplitude × weight 3 (the weight value is determined according to historical data training, such as weight 1 = 0.5, weight 2 = 0.3, weight 3 = 0.2). If the index exceeds the preset critical value (such as 0.8), the energy storage output adjustment instruction is triggered, and the next control link is entered.

[0054] Based on the verification results of the dynamic adjustment rule base, the system generates demand control indicators, including charge and discharge power thresholds, SOC adjustment intervals, etc. The coefficient generation unit of the control decision module extracts the verification results of the dynamic adjustment rules in the historical operation data of the same user, and generates adjustment coefficients by weighted calculation of the execution effect of each rule (such as assigning weights according to the proportion of power deviation reduction after adjustment). If the adjustment coefficient exceeds the preset trigger number (such as 3 times) continuously, the strategy switching unit determines to switch the energy storage charge and discharge strategy, for example, from "peak shaving as the main" to "valley filling as the main", or adjust the charge and discharge power slope to match the current load fluctuation characteristics.

[0055] Embodiment 1:

[0056] The system obtains the real-time load demand data, energy storage system operation state data and power grid dispatching information data of industrial and commercial users through the user side acquisition unit and the power grid information receiving unit of the load monitoring module. Among them, the real-time load demand data includes time-division equipment power curve and load type identification. The time-division equipment power curve is collected at a minute level, for example, for motors, heating equipment and other equipment in the production workshop, the power fluctuation in different time periods (such as morning shift, mid-shift and night shift) can be obtained, and the power value change in each 5-minute interval is obtained. The load type identification is obtained by associating the database with the equipment number, such as marking the motor as "M", the heating equipment as "H", and the lighting equipment as "L", etc., so as to classify and process different types of loads subsequently.

[0057] The energy storage system operating state data includes energy storage SOC value, charging and discharging rate, and equipment health degree parameter. The energy storage SOC value is monitored and uploaded by the battery management system (BMS) in real time, reflecting the remaining percentage of the battery pack, with an accuracy of ±1%; the charging and discharging rate is the power value of the current energy storage system charging and discharging, with a unit of kW, which can display the charging or discharging state of the energy storage system and the corresponding power size in real time; the equipment health degree parameter covers battery pack temperature, cell voltage consistency index, etc., for example, the temperature of each cell is obtained by deploying temperature sensors in the battery module, the temperature standard deviation is calculated to evaluate the temperature consistency, and the voltage consistency is judged by monitoring the cell voltage range, if the difference between the voltage of a cell and the average voltage exceeds the preset threshold, it is considered as abnormal health degree.

[0058] The grid dispatching information data includes time-of-use electricity price signal and grid carrying capacity parameter. The time-of-use electricity price signal is sent by the grid side according to the preset time period (such as peak period 8:00-12:00, flat period 12:00-18:00, valley period 22:00-6:00 the next day), containing the electricity price value of each period, such as peak period electricity price 0.8 yuan / kWh, valley period electricity price 0.3 yuan / kWh; the grid carrying capacity parameter is the current remaining capacity of the transformer area, with a unit of kW, reflecting the additional load capacity that the grid can bear in the current period, for example, the current remaining capacity of a transformer area is 100 kW, indicating that the transformer area can still access a total of 100 kW of load without exceeding the safe operation limit.

[0059] After obtaining the above data, the load demand data, energy storage state data and grid dispatching data need to be standardized. The standardization process specifically includes data cleaning, timestamp alignment, abnormal data interpolation correction and dimension unification. The data cleaning link adopts a sliding window filtering algorithm, with a window length of 15 minutes, to process the device power curve and eliminate jump data points caused by signal interference, device start-stop transient state, etc. For example, a sharp peak data point appears in the device power curve of a certain period, which is significantly higher than the normal operating range. Through the mean value calculation within the sliding window, it is judged that the point is an abnormal value and is removed.

[0060] The timestamp alignment technology adopts a two-way matching algorithm based on timestamp to synchronize the load demand data, energy storage state data and grid dispatching information data to a unified time coordinate system. Specifically, taking 5 minutes as a time interval, the data of each type is sampled, and the data of different sources and different collection frequencies are aligned to the time interval, ensuring the time consistency of the data in subsequent analysis and processing. For example, the device power data collected every second is aggregated into 5-minute average values, which are one-to-one corresponding in time with the energy storage SOC data, grid time-of-use electricity price data collected every 5 minutes.

[0061] Abnormal data interpolation correction is used to deal with abnormal values that may occur in energy storage health parameters. Taking the cell voltage data as an example, if the voltage value of a certain cell is detected to be out of the normal range (for example, the normal voltage range is 3.2V-3.8V, and the detected value at a certain time is 2.5V), the estimated value at that time is calculated by using the normal voltage values before and after that time through the cubic spline interpolation algorithm to replace the abnormal data point, so as to ensure the continuity and reliability of the data.

[0062] Dimensionless processing is to convert data of different physical dimensions into dimensionless or unified dimension data. For power data, it is converted into kW; for electric quantity data, it is converted into kWh; for energy storage SOC, it is expressed in percentage; for electricity price, it is in yuan / kWh; for temperature in the equipment health parameter, it is in ℃, and for voltage, it is in V. At the same time, in order to facilitate subsequent model processing, the power, electricity price, SOC and other parameters are converted into dimensionless data in the interval [0, 1] by using the normalization method, and the specific formula is: normalized value = (original value-minimum value) / (maximum value-minimum value), wherein the minimum value and the maximum value are statistical values in the historical data.

[0063] After standardization processing, the load demand data generates a load feature matrix. The number of rows of the matrix is the number of time periods, and the number of columns is the feature dimension such as power and load type identification code. For example, if 288 5-minute periods in a day are taken as the number of rows, the columns include the power normalized value of each period and the one-hot encoding of the load type identification (for example, the motor type equipment code is [1, 0, 0], and the heating equipment code is [0, 1, 0]), a 288x(1+N) load feature matrix is formed, wherein N is the number of load types.

[0064] The energy storage state data processing generates an energy storage state vector, which contains SOC normalized value, charge and discharge rate normalized value and equipment health normalized value, etc., forming a one-dimensional vector of 1xM, and M is the number of energy storage state features.

[0065] Finally, the load feature matrix is associated with the grid dispatch information in the time period to form a set of multi-dimensional constraint conditions. Specifically, each time period in the load feature matrix is associated with the corresponding time-of-use electricity price and grid carrying capacity parameter, for example, the power in the load feature matrix of a certain time period is 0.6 (normalized value), the load type is motor type, and the corresponding grid dispatch information is peak period electricity price 0.8 yuan / kWh and grid carrying capacity 80kW, thereby forming a five-dimensional constraint condition including time period, power, load type, electricity price and grid carrying capacity, which provides multi-dimensional data support for subsequent establishment of load prediction model and generation of control strategy. Through the above standardization processing procedure, the accuracy, consistency and availability of the input data are ensured, which lays a foundation for the effective implementation of the entire demand control method.

[0066] Example 2:

[0067] The baseline modeling unit of the strategy generation module extracts statistical feature parameters and generates demand baseline parameters based on the standardized load feature matrix. The load feature matrix contains power data and load type identification information for each period, with a 5-minute period unit covering 24 hours a day. For example, for the working day data of a certain industrial and commercial user, the load feature matrix of 288 periods from 00:00-00:05 to 23:55-24:00 can be obtained, each period corresponding to a row of data, including the average power, maximum power, minimum power, and load type code of the period.

[0068] The statistical feature parameters are extracted from the load feature matrix, including period average, power range, and change rate parameters. The period average is the arithmetic mean of the power data in each period, reflecting the typical load level of the period. For example, if the power data collected in the 09:00-09:05 period is 198kW, 202kW, 200kW, 199kW, and 201kW, the period average is (198+202+200+199+201) / 5=200kW. The power range is the difference between the maximum and minimum power in each period, used to measure the fluctuation range of the load in the period. For example, if the maximum power in the period is 205kW and the minimum power is 195kW, the range is 205-195=10kW. The change rate parameter is the ratio of the power change to the time interval between adjacent periods, reflecting the increase or decrease rate of the load demand. Assuming that the average power of the 08:55-09:00 period is 170kW and the average power of the 09:00-09:05 period is 200kW, the change rate is (200-170) / 5=6kW / minute.

[0069] Each statistical feature parameter is combined according to a preset proportion to generate a demand baseline parameter. The preset proportion is determined according to historical data and control targets, for example, setting the period average to 50%, the power range to 30%, and the change rate parameter to 20%. For example, the period average of the 09:00-09:05 period is 200kW, the range is 10kW, and the change rate is 6kW / minute, so the demand baseline parameter is: baseline power = period average ± (range x 30% + change rate x 20% x time interval). The time interval is in hours, so 5 minutes is 1 / 12 hours, therefore baseline power = 200 ± (10 x 0.3 + 6 x 0.2 x 1 / 12) = 200 ± (3 + 0.1) = 200 ± 3.1kW, i.e. the demand baseline range of the period is 196.9kW to 203.1kW.

[0070] A dynamic threshold is set for the demand baseline parameter in the multi-dimensional constraint condition set. The setting of the dynamic threshold combines the time-of-use electricity price and the grid carrying capacity parameter in the grid dispatch information. For example, in the peak period (such as 09:00-12:00), the grid carrying capacity is low, and the upper limit of the demand baseline power is reduced by 10% to reduce the pressure on the grid. Assuming that the original upper limit of the baseline power is 203.1 kW, the reduced upper limit is 203.1 x (1-10%) = 182.79 kW, and the lower limit of the baseline remains unchanged at 196.9 kW. In the valley period (such as 22:00-06:00 the next day), the grid carrying capacity is high, and the baseline threshold can be appropriately relaxed, such as increasing the upper limit of the baseline power by 15% to fully utilize the low-price electricity during the valley period for energy storage charging.

[0071] Based on the dynamic threshold, the characteristic parameters that meet the constraint boundary are selected to form an initial strategy set, and the characteristic parameters that exceed the constraint boundary are excluded. For example, in the peak period 09:00-09:05, the power value corresponding to a certain characteristic parameter is 205 kW, which exceeds the reduced upper limit of the baseline power 182.79 kW, so the parameter is excluded; if another characteristic parameter corresponds to a power value of 190 kW, which is between 196.9 kW and 182.79 kW (note that the logical relationship here, in fact, the new upper limit should be 182.79 kW, and the original lower limit 196.9 kW needs to be adjusted, in the example, assuming that the dynamic threshold only adjusts the upper limit, the lower limit remains unchanged, the parameter that meets the condition should be ≤182.79 kW and ≥196.9 kW, but in actual scenarios, the baseline range may be recalculated according to the threshold adjustment, here the adjusted baseline range is used as the reference), the parameter is retained in the initial strategy set.

[0072] Then, time series comparison analysis is performed on each characteristic parameter in the initial strategy set, and the difference between the load demand and the energy storage output in adjacent periods is extracted. Taking the adjacent 09:00-09:05 period and the 09:05-09:10 period as an example, assuming that the load demand power in the 09:00-09:05 period is 200 kW, the energy storage output power is 190 kW, and the difference is 200-190=10 kW; the load demand power in the 09:05-09:10 period is 210 kW, the energy storage output power is 195 kW, and the difference is 210-195=15 kW. The absolute values of the differences are 10 kW and 15 kW, respectively, which are marked as period deviation parameters.

[0073] The period deviation parameter is compared with a preset fluctuation threshold, and the preset fluctuation threshold is set according to the adjustment capacity of the energy storage system and the grid stability requirement, for example, it is set to ±5% of the baseline power. Taking the period of 09:00-09:05 as an example, the baseline power is 200kW, and the fluctuation threshold is 200*5%=10kW. The absolute value of the period deviation parameter of this period is 10kW, which is equal to the threshold, and can be considered as a critical state. The absolute value of the period deviation parameter of the period of 09:05-09:10 is 15kW, which exceeds the threshold of 10kW, so the feature parameter is added to the dynamic adjustment rule library. At the same time, according to the grid dispatching information, the price sensitivity feature is supplemented to the dynamic adjustment rule library. For example, if the current period is a peak period, the price sensitivity is high, and the deviation threshold adjustment coefficient of the dynamic adjustment rule in this period is set to 0.8 (that is, the actual allowed deviation threshold=baseline power*5%*0.8=8kW), so as to more strictly control the load fluctuation and reduce the impact on the grid in the peak period; if it is a valley period, the price sensitivity is low, and the deviation threshold adjustment coefficient can be set to 1.2, allowing a larger load fluctuation range to fully utilize low-cost electricity for energy storage charging.

[0074] Each rule in the dynamic adjustment rule library contains information such as period, deviation threshold, and price sensitivity feature. For example, a rule can be expressed as: in the peak period (09:00-12:00), the deviation threshold is 8kW, and the price sensitivity coefficient is 1.5 (used for weight adjustment in subsequent demand deviation index calculation); in the valley period (22:00-6:00 the next day), the deviation threshold is 12kW, and the price sensitivity coefficient is 0.7. Through continuous screening and supplementing of feature parameters, the dynamic adjustment rule library can be continuously optimized, thereby providing more accurate basis for real-time matching verification of energy storage system output, and ensuring that the demand control strategy can adapt to different periods of grid environment and user load characteristics.

[0075] Embodiment 3:

[0076] The dynamic execution module monitors and analyzes the power difference between the load demand and the energy storage output and its dynamic response state in real time through the difference calculation unit and the trend analysis unit. The difference calculation unit obtains the instantaneous values of the load demand power and the energy storage output power in real time at a millisecond level, both of which are instantaneous power data with a unit of kW, and are denoted as and , respectively, where represents the time point. The absolute value of the power difference is calculated, and the formula is:

[0077]

[0078] In the formula, is the power difference absolute value (unit: kW) at the time point . The preset allowed range is the fluctuation interval determined based on the demand baseline parameter, denoted as , wherein is the allowable deviation threshold (unit: kW) set by the dynamic adjustment rule base according to factors such as time period, electricity price, etc. If , it is marked as a demand fluctuation event, indicating that the difference between the current load demand and the energy storage output exceeds the normal adjustment range set by the system, and further analysis and control actions need to be triggered.

[0079] Taking the actual operation scenario of a certain industrial and commercial user as an example, it is assumed that in a certain peak period , the instantaneous value of the load demand power is , the instantaneous value of the energy storage output power is , and the allowable deviation threshold set by the dynamic adjustment rule base for this period is . The calculation can be obtained , since , it does not exceed the allowable range, and does not trigger the demand fluctuation event marker. Subsequently, at , the load demand suddenly increases to , and the energy storage output power is still , then , which exceeds the allowable deviation threshold, and it is marked as a demand fluctuation event at this time, and the time, power difference, etc. Information of the event is recorded.

[0080] The number of demand fluctuation events in a preset period is counted, and the preset period can be set according to the load characteristics of the user and the control accuracy requirement, for example, it is taken as . In this period, the system accumulates the number of events that meet , denoted as . The SOC change rate of the energy storage system and the real-time electricity price fluctuation amplitude of the power grid are obtained synchronously. The SOC change rate of the energy storage system is the change amount of SOC per unit time, denoted as , with a unit of , reflecting the charging and discharging intensity of the energy storage system in this period, for example represents a decrease of 0.5% per minute. The real-time electricity price fluctuation amplitude is the change rate of the current period electricity price compared with the reference electricity price, denoted as , with a unit of , and the reference electricity price is the standard electricity price of the power grid in this period, for example, the reference electricity price in the peak period is , if the real-time electricity price rises to , then .

[0081] According to the number of demand fluctuation events , the SOC change rate and the electricity price fluctuation amplitude , the demand deviation index is calculated by multi-constraint coordination technology, and the formula is:

[0082]

[0083] wherein, is the demand deviation index (dimensionless); , , are weight coefficients, respectively corresponding to the influence weight of the number of events, the SOC change rate, and the electricity price fluctuation amplitude, the values of which are determined by historical data training and satisfy . For example, set , , , it indicates that the influence of the number of demand fluctuation events on the demand deviation index is the largest, followed by the SOC change rate, and the electricity price fluctuation amplitude has the smallest influence.

[0084] Taking the statistical data in a preset period as an example, it is assumed that in , the number of demand fluctuation events , the SOC change rate of the energy storage (the absolute value is ), and the real-time electricity price fluctuation amplitude of the power grid (the absolute value is ). Substituting into the formula can obtain:

[0085]

[0086] The preset critical value is , for example, take , when the calculated demand deviation index , it is determined that the matching state of the current load demand and the energy storage output has exceeded the controllable range of the system, triggering the energy storage output adjustment instruction. The adjustment instruction is sent to the converter (PCS) of the energy storage system through the collaborative interaction of the dynamic execution module and the control determination module, for example, the instruction requires to increase the energy storage discharge power by , so as to reduce the power difference between the load demand and the energy storage output, and make gradually fall back to the range of the allowable deviation threshold .

[0087] During the whole analysis process, the difference calculation unit and the trend analysis unit interact with real-time data and algorithm processing, realizing dynamic tracking of power difference and quantitative evaluation of demand deviation. The difference calculation unit ensures timely capture of load mutation events, and the trend analysis unit comprehensively judges the system running state through multi-dimensional parameter fusion, providing a scientific basis for real-time adjustment of energy storage output. This process does not rely on preset experimental effect data, but is based on actual collected running parameters and algorithm logic, realizing dynamic optimization of industrial and commercial energy storage system demand control, ensuring that the energy storage system can efficiently and stably participate in demand management under the conditions of grid dispatching constraints and user load fluctuations, improving the economy and reliability of power system operation.

[0088] Example 4:

[0089] The coefficient generation unit of the control decision module works with the strategy switching unit, evaluates the execution effect of the dynamic adjustment rule based on user historical operation data, and determines whether to trigger the switching of the energy storage charge and discharge strategy. The specific process is as follows:

[0090] The system extracts the verification results of each rule in the dynamic adjustment rule library from the historical operation database of the same user, including rule execution time, corresponding load period, power difference value before and after adjustment, energy storage SOC change, etc. Taking a dynamic adjustment rule suitable for peak period (9:00-12:00) as an example, the core parameters of this rule are: allowed deviation threshold 10kW, price sensitivity coefficient 1.2, and adjustment strategy is to increase energy storage discharge power by 5kW / time. In the past week's operation data, this rule was executed 10 times, and the power difference change and energy storage SOC response data were recorded after each execution.

[0091] For each rule execution, the coefficient generation unit calculates the execution effect of the rule. For example, the first execution, the power difference before adjustment is 15kW (load demand 215kW, energy storage output 200kW), and after executing the rule, the energy storage discharge power is increased by 5kW, and the power difference after adjustment is reduced to 10kW (energy storage output is changed to 205kW), the difference reduction ratio is (15-10) / 15≈33.3%. The second execution, the power difference before adjustment is 18kW (load demand 220kW, energy storage output 202kW), and after adjustment, the difference is reduced to 13kW, the reduction ratio is (18-13) / 18≈27.8%. Similarly, the difference reduction ratios of the third to tenth times are 40%, 25%, 30%, 35%, 28%, 32%, 38%, and 31% respectively.

[0092] To evaluate the overall execution effect of the rule, the coefficient generation unit performs weighted calculation on each execution effect. The weight setting is based on the time sequence of rule execution, for example, the first execution weight is 10%, and each subsequent execution weight increases by 5%, to reflect that the recent execution effect has a more significant impact on the current strategy adjustment. The specific weight allocation is as follows: the first time weight is 10%, the second time is 15%, the third time is 20%, the fourth time is 25%, the fifth time is 30%, the sixth time is 35%, the seventh time is 40%, the eighth time is 45%, the ninth time is 50%, and the tenth time is 55% (Note: the actual weight sum needs to be normalized, this is for simplicity of explanation).

[0093] Taking the first to the third execution as an example, the adjustment coefficient is calculated:

[0094] First: 33.3% x 10% = 3.33%

[0095] Second: 27.8% x 15% = 4.17%

[0096] Third: 40% x 20% = 8%

[0097] The cumulative contribution value of the first three times is 3.33% + 4.17% + 8% = 15.5%. Similarly, the contribution values of all 10 executions are calculated and summed up, and the adjustment coefficient of the rule is 32% (assuming the cumulative result). The adjustment coefficient reflects the average reduction ability of the power difference after the rule is executed, and the higher the value, the stronger the effectiveness of the rule.

[0098] The strategy switching unit determines whether to perform strategy switching according to the number of consecutive triggers of the adjustment coefficient. The preset trigger condition is that the adjustment coefficient exceeds the preset threshold (such as 30%) for 3 times in a row. Continuing with the above example, assuming that the adjustment coefficients in the last 4 execution periods are 35%, 33%, 31%, and 34%, all of which exceed the threshold of 30%, and the trigger is 4 times in a row (exceeding the preset 3 times), it is determined that the current charging and discharging strategy needs to be switched.

[0099] The specific logic of strategy switching is based on the matching degree of user load characteristics and grid scheduling information. For example, if the current strategy is "fixed power discharge strategy" (such as always discharging at 50kW power), in the case of continuous threshold exceeding of the adjustment coefficient, the system determines that this strategy has failed to effectively cope with load fluctuations, and switches to "dynamic slope tracking strategy". The dynamic slope tracking strategy automatically adjusts the slope of the energy storage discharge power (such as increasing or decreasing 2kW per minute) to more accurately track load fluctuations by monitoring the change rate of load demand in real time. The specific switching process is as follows:

[0100] The coefficient generation unit detects that the adjustment coefficient of a certain rule exceeds 30% for 4 times in a row, and sends a trigger signal to the strategy switching unit;

[0101] The policy switching unit calls the currently executed charging and discharging strategy parameters, such as fixed discharging power value, adjustment period, etc.

[0102] According to the rule characteristics with high adjustment coefficients in historical data (such as high price sensitivity, high load fluctuation period), the corresponding alternative strategy is matched from the strategy library, for example, the dynamic adjustment strategy based on PID control is selected;

[0103] The policy switching instruction is sent to the energy storage converter (PCS) through the control interface, including the parameter setting of the new strategy (such as the proportional coefficient, integral time, and differential time of the PID controller);

[0104] After the energy storage system switches to the new strategy, the coefficient generation unit reinitializes the execution effect statistics of the rule and starts to evaluate the effectiveness of the new strategy.

[0105] In another example, if a user frequently triggers the adjustment rule during the valley period (22:00-6:00) and the adjustment coefficient continuously exceeds the threshold, the system may determine that the current "night charging strategy" (such as fixed charging at 30kW power) cannot fully utilize the valley electricity resources, and then switch to an "adaptive charging strategy" that dynamically adjusts the charging power according to the real-time grid carrying capacity and the health of the energy storage (such as the maximum charging power is 80% of the remaining capacity of the grid). This switching can maximize the charging efficiency of the energy storage system while ensuring the safety of the grid, and reduce the user's electricity cost.

[0106] The whole process is driven by quantitative evaluation of historical data, avoiding control bias caused by subjective experience, and ensuring that the energy storage charging and discharging strategy can be dynamically optimized according to the actual operation effect. The weighting calculation method of the coefficient generation unit can flexibly adapt to the importance differences of different periods and different rules, and the policy switching unit realizes automatic decision-making through preset logic without manual intervention. This mechanism enables the demand control system to continuously learn the user's load pattern in the long-term operation, improves the utilization efficiency of energy storage resources and the accuracy of demand control, and realizes the two-way optimization of industrial and commercial users and the grid.

[0107] Embodiment 5:

[0108] The modules of the demand control system realize specific functions through hardware deployment and software cooperation. The following describes the module composition and interaction process of the system in combination with specific application scenarios:

[0109] I. Implementation of the load monitoring module

[0110] The load monitoring module, serving as the system's data entry point, consists of a user-side acquisition unit and a power grid information receiving unit. The user-side acquisition unit is deployed within the power distribution system of industrial and commercial users, employing industrial-grade IoT sensors for data acquisition. For example, a Modbus protocol power transmitter is installed in the user's distribution cabinet. This device collects power data from each branch circuit in real time via current and voltage transformers, including active power, reactive power, and apparent power, at a sampling frequency of 1 second. The data is encoded in Modbus RTU format and transmitted to the data concentrator via an RS485 bus. For energy storage system operating parameters, the user-side acquisition unit communicates with the battery management system (BMS) via a CAN bus to obtain the energy storage SOC value, charge / discharge rate (e.g., current discharge power of 45kW), and equipment health parameters (e.g., average battery pack temperature of 25℃, cell voltage difference of 50mV).

[0111] The power grid information receiving unit acquires power grid dispatch information through a dedicated power communication network. Taking a regional power grid as an example, it sends time-of-use pricing signals and grid capacity margin parameters to users via the 104 protocol: the time-of-use pricing is divided into three periods: peak (8:00-12:00, price 0.85 yuan / kWh), flat (12:00-18:00, price 0.6 yuan / kWh), and valley (22:00-6:00 the next day, price 0.3 yuan / kWh); the grid capacity margin parameter updates the remaining capacity of the distribution area in real time. For example, if it displays 95kW at a certain moment, it means that the current maximum new load that the distribution area can withstand is 95kW. Both types of data are transmitted to the system's edge computing gateway through an encrypted channel and are stored synchronously in the local database with the user-side data.

[0112] II. Implementation of the Strategy Generation Module

[0113] The strategy generation module comprises a baseline modeling unit and a rule optimization unit, which implement data processing and strategy generation based on software algorithms. The baseline modeling unit utilizes Python's Pandas library to perform statistical analysis on the standardized load characteristic matrix. For example, for a user's load data during the weekday period of 9:00-9:05, it extracts the average power (200kW), range (20kW), and rate of change (5kW / minute) for that period, and generates demand baseline parameters (e.g., baseline power of 200±15kW) according to a preset ratio. The rule optimization unit employs a decision tree algorithm to filter feature parameters based on multi-dimensional constraints (e.g., time period, electricity price, grid capacity). During peak hours, the decision tree model prioritizes parameters with power values ​​below the baseline upper limit and embeds electricity price sensitivity features (e.g., higher priority for peak-hour adjustments) into the dynamic adjustment rule base.

[0114] Take a dynamic adjustment rule as an example. The rule optimization unit discovers through historical data training that when the grid carrying margin is lower than 100 kW and in the peak period, the probability of load demand exceeding the baseline power by 10% is high, so the rule is generated: “In the peak period and when the grid remaining capacity <100 kW, the allowed deviation threshold will be tightened from ±10% of the baseline power to ±8%, and the energy storage discharge power adjustment step will be increased from 5 kW to 8 kW”. The rule is formed through feature engineering and algorithm training and stored in the SQL table of the dynamic adjustment rule library. The fields include time interval, grid capacity threshold, deviation threshold coefficient, adjustment step, etc.

[0115] III. Implementation of dynamic execution module

[0116] The dynamic execution module is composed of a difference calculation unit and a trend analysis unit, responsible for real-time verification and trend analysis. The difference calculation unit is based on FPGA hardware to realize high-speed data processing, and calculates the power difference between load demand and energy storage output with microsecond-level delay. For example, when the instantaneous value of load demand is 210 kW and the instantaneous value of energy storage output is 185 kW, the difference calculation unit immediately obtains a difference of 25 kW, and compares it with the allowed deviation threshold (such as 20 kW in peak period) in the dynamic adjustment rule library. If it exceeds, it is marked as a demand fluctuation event, and the event timestamp and difference value are recorded.

[0117] The trend analysis unit uses MATLAB mathematical engine to realize demand deviation evaluation. In a 15-minute statistical period, assume that 3 demand fluctuation events are detected, the energy storage SOC change rate is -0.6% / minute (discharging state), and the grid real-time electricity price fluctuation amplitude is +15% (peak period electricity price up). The trend analysis unit calculates the demand deviation degree according to the preset weight logic (such as event frequency weight 0.5, SOC change rate weight 0.3, electricity price fluctuation amplitude weight 0.2), and determines whether to trigger the adjustment instruction. If the evaluation result exceeds the critical value, send an adjustment request to the control decision module through OPCUA protocol, including current difference, fluctuation frequency, energy storage state and other parameters.

[0118] IV. Implementation of control decision module

[0119] The coefficient generation unit and the strategy switching unit of the control determination module implement strategy optimization based on data mining and automated logic. The coefficient generation unit retrieves the execution records of a certain dynamic adjustment rule from the historical database, for example, the rule of "allowing deviation threshold 15 kW in valley period" was executed 20 times in the past month, and the power difference value was reduced by 20-40% each time. Through weighted calculation (such as the execution record in the past week accounting for 70%), the adjustment coefficient of the rule is 32%, indicating that its average efficiency is 32%. If the coefficient exceeds the preset threshold (such as 30%) for 5 consecutive times, the strategy switching unit determines that the current charging and discharging strategy (such as fixed charging power 40 kW) has reached the optimization bottleneck, and needs to switch the strategy.

[0120] The specific scenario of strategy switching is as follows: a user has been using the "fixed power charging strategy" in the valley period for a long time, but as the production equipment increases, the load fluctuation in the valley period intensifies, causing the dynamic adjustment rule to be triggered frequently and the adjustment coefficient to be continuously high. After the system detects this trend, the strategy switching unit retrieves the "load tracking charging strategy" from the strategy library. This strategy sets the energy storage charging power to 70% of the remaining capacity of the power grid (such as 120 kW) in real time to avoid overloading during charging. The switching instruction is sent to the energy storage converter (PCS) through industrial Ethernet, and the PCS reconfigures the control parameters after receiving it, adjusts the charging power slope, and feeds back a switching success signal to the system.

[0121] Five, interaction between modules and system integration

[0122] Each module realizes real-time data interaction through a message queue (such as RabbitMQ). The data collected by the load monitoring module is standardized and pushed to the "raw data queue" in JSON format; the strategy generation module reads data from the queue, completes baseline modeling and rule optimization, and pushes dynamic adjustment rules to the "control rule queue"; the dynamic execution module subscribes to the queue, obtains rules in real time and performs power difference verification, and pushes the analysis results to the "adjustment request queue"; the control determination module reads the request from the queue, performs coefficient calculation and strategy switching, and pushes the final instruction to the "execution instruction queue" for execution by the energy storage system.

[0123] In terms of hardware deployment, the sensors and communication devices of the load monitoring module are distributed in the user site and the power grid side, the strategy generation module and the control determination module run on the edge computing server (configured as Inteli7 processor, 16 GB memory, 512 GB SSD), and the FPGA board card of the dynamic execution module is integrated in the industrial control cabinet. Each component is connected through a redundant network to ensure the stability and reliability of the system in complex industrial environments.

[0124] Through the above module design and collaborative mechanism, the demand control system realizes the full-process automation from data acquisition, strategy generation, dynamic execution to strategy optimization, can accurately adapt to the load characteristics of industrial and commercial users and the requirements of power grid dispatching, and provides a complete technical solution for efficient operation of the energy storage system and demand control.

[0125] It should be noted that, in this text, relational terms such as first and second are used merely to distinguish one entity or action from another, without necessarily requiring or implying that there is any such actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof are intended to cover non-exclusive inclusions, so that a process, method, article, or apparatus that comprises a list of elements does not only include those elements, but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0126] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A demand control method based on industrial and commercial energy storage, characterized in that, The method includes the following steps: Acquire real-time load demand data, energy storage system operation status data, and power grid dispatch information data from industrial and commercial users, and perform data standardization processing on the load demand data, energy storage status data, and power grid dispatch data. Based on the standardized data, a load forecasting model is established to generate demand baseline parameters. Based on the demand baseline parameters, feature optimization is performed to form a set of control strategies. Combined with power grid dispatch information, a dynamic adjustment rule base is constructed. The power output of the energy storage system is matched and verified in real time based on the dynamic adjustment rule base. Multi-constraint coordination technology is used to identify the power difference between load demand and energy storage output, and the dynamic response status of the power difference is analyzed for trends. Demand control indicators are generated based on the verification results of the dynamically adjusted rule base, and the charging and discharging regulation strategy of the energy storage system is determined based on the demand control indicators. The steps to establish a load forecasting model and generate demand baseline parameters are as follows: Statistical feature parameters are extracted from the standardized load feature matrix. These statistical feature parameters include time-period average, power range, and rate of change parameters. Each statistical feature parameter is combined according to a preset ratio to generate demand baseline parameters. Set dynamic thresholds for the baseline parameters of the demand in the multi-dimensional constraint set, filter the feature parameters that meet the constraint boundaries to form an initial strategy set, and remove the feature parameters that exceed the constraint boundaries. A time series comparative analysis was performed on each feature parameter in the initial strategy set to extract the difference between load demand and energy storage output in adjacent time periods. The absolute value of the difference was calculated and marked as the time period deviation parameter. The time period deviation parameter is compared with the preset fluctuation threshold, and the characteristic parameters that exceed the fluctuation threshold are added to the dynamic adjustment rule base. The electricity price sensitivity feature is added to the dynamic adjustment rule base according to the power grid dispatch information. The real-time load demand data includes time-segmented equipment power curves and load type identifiers; the energy storage system operation status data includes energy storage SOC value, charge / discharge rate, and equipment health parameters; the grid dispatch information data includes time-of-use electricity price signals and grid carrying capacity margin parameters; the load demand data, energy storage status data, and grid dispatch data are standardized, including data cleaning, timestamp alignment, abnormal data interpolation correction, and unit unification; after processing the load demand data, a load feature matrix is ​​generated, and after processing the energy storage status data, an energy storage status vector is generated; the load feature matrix and grid dispatch information are correlated by time period to form a multi-dimensional constraint set.

2. The demand control method based on industrial and commercial energy storage according to claim 1, characterized in that: Trend analysis of the dynamic response state of the power difference includes the following steps: The instantaneous values ​​of load demand power and energy storage output power are acquired in real time, and the absolute value of the power difference between the two is calculated. If the power difference exceeds the preset allowable range, it is marked as a demand fluctuation event. Count the number of demand fluctuation events within a preset period, and simultaneously obtain the SOC change rate of the energy storage system and the real-time electricity price fluctuation range of the power grid. The demand deviation index is calculated based on the number of demand fluctuation events, the rate of change of SOC, and the magnitude of electricity price fluctuations. If the demand deviation index exceeds the preset threshold, an energy storage output adjustment command is triggered.

3. The demand control method based on industrial and commercial energy storage according to claim 2, characterized in that: Extract the verification results of the dynamic adjustment rule base from the historical operation data of the same user, perform weighted calculation on the execution effect of each adjustment rule to generate an adjustment coefficient, and if the adjustment coefficient exceeds the preset number of triggers consecutively, it is determined to switch the energy storage charging and discharging strategy.

4. A demand control system based on industrial and commercial energy storage, wherein the system is applied to the demand control method based on industrial and commercial energy storage as described in any one of claims 1-3, characterized in that, The system includes a load monitoring module, a strategy generation module, a dynamic execution module, and a control decision module; The load monitoring module is used to acquire real-time load demand data, energy storage system operation status data, and power grid dispatch information data, and to perform standardized processing. The strategy generation module is used to establish a load forecasting model to generate demand baseline parameters and construct a dynamic adjustment rule base. The dynamic execution module is used to match and verify the energy storage output according to the dynamic adjustment rule base, and analyze the power difference between the load demand and the energy storage output. The control and determination module is used to generate demand control indicators and determine energy storage regulation strategies.

5. A demand control system based on industrial and commercial energy storage according to claim 4, characterized in that: The load monitoring module includes a user-side data acquisition unit and a power grid information receiving unit; The user-side acquisition unit is used to collect the equipment power curves and energy storage system operating parameters of industrial and commercial users. The power grid information receiving unit is used to acquire time-of-use electricity price signals and power grid capacity margin data.

6. A demand control system based on industrial and commercial energy storage according to claim 4, characterized in that: The strategy generation module includes a baseline modeling unit and a rule optimization unit; The baseline modeling unit is used to extract time-period averages, power ranges, and rate of change parameters from standardized data. The rule optimization unit is used to filter feature parameters that satisfy the constraint boundaries and supplement electricity price sensitivity features to the dynamic adjustment rule base.

7. A demand control system based on industrial and commercial energy storage according to claim 4, characterized in that: The dynamic execution module includes a difference calculation unit and a trend analysis unit; The difference calculation unit is used to detect the power difference between load demand and energy storage output and to mark demand fluctuation events. The trend analysis unit is used to calculate the demand deviation index by combining the SOC change rate and the electricity price fluctuation range.

8. A demand control system based on industrial and commercial energy storage according to claim 4, characterized in that: The control determination module includes a coefficient generation unit and a strategy switching unit; The coefficient generation unit is used to perform weighted calculations on the execution effect of the adjustment rules; The strategy switching unit is used to determine the energy storage charging and discharging strategy switching based on the number of consecutive triggers of the adjustment coefficient.

Images