Method, system and equipment for measuring and calculating energy storage scale of industrial park without photovoltaic access

Abstract

The invention belongs to the technical field of power system energy storage planning, and provides an industrial park energy storage scale measuring and calculating method, system and equipment without photovoltaic access, and the method comprises the steps: firstly collecting the historical load data, time-of-use electricity price, demand contract value and transformer capacity limitation of an industrial park, and generating a typical daily load curve; secondly, predicting a future set period load curve based on a Prophet algorithm, outputting a predicted value confidence interval and calculating a monthly demand safety threshold; and finally, constructing a mixed integer programming model containing demand soft constraint, and solving to obtain a Pareto optimal configuration scheme of the energy storage rated capacity and the maximum charging and discharging power by taking the minimization of the total cost of the whole life cycle as a target. According to the method, the mixed integer programming model containing the demand soft constraint and the Pareto solution set are adopted, accurate measurement and calculation of the energy storage scale are achieved, economical efficiency, reliability and dynamic adaptability are achieved, and the method is suitable for economical efficiency configuration and demand risk avoidance of industrial user side energy storage.

Description

Technical Field

[0001] This invention belongs to the field of power system energy storage planning technology, specifically involving a method, system and equipment for calculating the scale of energy storage in industrial parks without photovoltaic access. It is a method and system for calculating the capacity and power of energy storage systems in industrial parks without photovoltaic access, taking into account the dynamic characteristics of electricity load, time-of-use pricing mechanism and monthly maximum demand constraints. It is applicable to the economic configuration of energy storage on the industrial user side and demand risk avoidance. Background Technology

[0002] With the acceleration of industrialization, industrial parks have become core electricity consumers in the power system, characterized by highly volatile and continuous power loads. In industrial parks without photovoltaic grid connection, energy storage systems must simultaneously address multiple constraints, including dynamic power load, time-of-use pricing mechanisms, and monthly maximum demand assessments. The core challenge lies in:

[0003] Demand risk and cost game: User-side demand costs account for as much as 30%-50%. Traditional strategies based on fixed thresholds to control energy storage charging and discharging are difficult to adapt to production load fluctuations in real time, which can easily lead to penalties for exceeding demand limits or redundant energy storage investment.

[0004] Multi-timescale optimization coupling: Demand assessment is on a monthly cycle, while electricity price arbitrage and load regulation need to be completed within a day. Existing algorithms cannot achieve coordinated optimization of long-term and short-term goals.

[0005] Insufficient dynamic response capability: When sudden changes in industrial load (such as production line start-up and shutdown) occur, energy storage systems need to respond within seconds to suppress demand peaks, but traditional EMS rely on preset rules and lack flexibility.

[0006] To address the aforementioned challenges, the industry currently offers two main technical solutions, but both have significant drawbacks and cannot meet the energy storage optimization needs of industrial parks without photovoltaic grid connection.

[0007] 1. Energy storage scheduling strategy based on static model

[0008] Technical features: The system uses historical load data to train a fixed charging and discharging template, generates a scheduling plan with the goal of maximizing the peak-valley price difference, and sets a hard upper limit on monthly demand (such as 95% of the contract value).

[0009] Limitations: No load forecasting error feedback mechanism has been established, and demand control fails when the actual load deviates from the forecast value; the energy storage charging and discharging power is statically bound to the capacity configuration, making it impossible to adapt to dynamic adjustments in production plans.

[0010] 2. AI-driven dynamic optimization solution

[0011] Technical features: Introduces algorithms such as LSTM to predict short-term load, combines reinforcement learning to generate real-time charging and discharging strategies, and embeds a demand exceedance probability assessment model (such as Monte Carlo simulation).

[0012] Limitations: The optimization objective focuses on economic efficiency, ignoring the impact of energy storage cycle life decay on the total cost; demand emergency suppression relies on ex-post penalty mechanisms and lacks feedforward control for real-time load surges.

[0013] Based on the above-mentioned existing technologies, it can be seen that the main technical shortcomings of the existing solutions are as follows:

[0014] 1. Demand constraints and economic efficiency are disconnected: The monthly demand surplus is not dynamically allocated to the daily dispatch, resulting in low energy storage capacity utilization or increased risk exposure;

[0015] 2. Insufficient model adaptability: Static parameter configuration cannot respond to adjustments in electricity price policies (such as fluctuations in time-of-use electricity prices in the spot market) and structural changes in load (such as new production lines);

[0016] 3. Limited system scalability: The lack of an open interface to accommodate diverse load management terminals (such as interruptible loads and flexible air conditioning) restricts the potential for improving overall energy efficiency. Summary of the Invention

[0017] The purpose of this invention is to overcome the existing defects and provide a method, system and equipment for calculating the energy storage capacity of industrial parks without photovoltaic access.

[0018] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0019] The primary objective of this invention is to provide a method for calculating the energy storage capacity of industrial parks without photovoltaic grid connection, comprising the following steps:

[0020] (1) Collect historical load data, time-of-use electricity price, demand contract value and transformer capacity limit of industrial park, perform outlier removal processing on the historical load data, and generate typical daily load curves through K-means clustering algorithm;

[0021] (2) Based on the typical daily load curve, the Prophet algorithm is used to predict the load curve for a future set period, output the confidence interval of the load prediction value, and calculate the monthly demand safety threshold based on the standard deviation of historical load data.

[0022] (3) Using the confidence interval of the load forecast value and the monthly demand safety threshold as input parameters, a mixed integer programming model containing demand soft constraints is constructed. The Pareto optimal configuration scheme of the rated capacity and maximum charging and discharging power of the energy storage system is obtained by minimizing the total cost of the energy storage system throughout its entire life cycle as the objective function. The total cost throughout its entire life cycle includes investment cost, electricity expenditure, energy storage replacement cost and capacity decay cost.

[0023] Furthermore, in step (2), before performing load forecasting, the Prophet algorithm first performs data preprocessing on the input data, specifically including...

[0024] Missing value handling: Automatically imput missing values;

[0025] Outlier handling: Identify and correct extreme values ​​using box plots or Z-Score methods;

[0026] Resampling and aggregation: If the data granularity is smaller than the daily level, it needs to be aggregated by day to match Prophet's daily forecasting characteristics.

[0027] Furthermore, in step (2), the Prophet algorithm is used to predict the load curve for a future set period, and the confidence interval of the load prediction value is output, including:

[0028] The time series is decomposed into a trend component, a seasonal component, a holiday effect, and an error term. The trend component adopts a piecewise linear trend, the seasonal component is modeled using Fourier series, and the holiday effect is modeled separately.

[0029] The model was optimized by adjusting parameters such as trend change points, seasonality intensity, and the impact of holidays.

[0030] The model is fitted to generate date data frames for future time points, and the predicted values ​​and confidence intervals for future time points are generated based on the fitted model.

[0031] Furthermore, in step (3), the constraint parameters of the mixed integer programming model include:

[0032] 1) Soft demand constraint:

[0033]

[0034] in, , Let be the charging and discharging power at time t, respectively. Let be the load power at time t. This is the upper limit of the contract demand. These are slack variables;

[0035] 2) Dynamic equilibrium of SOC:

[0036]

[0037] in, Let t be the state of charge. , These are the charging and discharging efficiencies, respectively. This refers to the nominal capacity of the energy storage system.

[0038] 3) SOH decay model:

[0039]

[0040] in, To accumulate the number of loops, The attenuation coefficient is... Replace the threshold for energy storage;

[0041] 4) Accumulated loop count:

[0042]

[0043] in, For the increment of the loop count, The counter is triggered when =1; Indicates whether charging / discharging occurs at time t. or The loop count is triggered at any time;

[0044] 5) Power and capacity boundaries

[0045]

[0046]

[0047]

[0048] in, For mutual exclusion constraints;

[0049] 6) Change decision trigger

[0050]

[0051] This refers to the amount of SOH restored in a single replacement.

[0052] Furthermore, in step (3), the objective function is expressed as:

[0053]

[0054] in, For investment costs, For electricity expenses, For energy storage replacement costs, Cost of capacity decay;

[0055] The Pareto optimal allocation scheme is solved using the ε-constraint method or the weighted sum method;

[0056] The ε-constraint method transforms the dual objective into a single objective, expressed as:

[0057]

[0058] The weighted sum method introduces weights ω∈[0,1], expressed as:

[0059] .

[0060] Another objective of this invention is to provide a system for calculating the energy storage capacity of industrial parks without photovoltaic grid connection, comprising:

[0061] The data input and preprocessing module is used to collect historical load data, time-of-use electricity prices, demand contract values ​​and transformer capacity limits of the industrial park, perform outlier removal processing on the historical load data, and generate typical daily load curves through K-means clustering algorithm;

[0062] The load forecasting module is used to use the Prophet algorithm to predict the load curve for a set period in the future based on the typical daily load curve, output the confidence interval of the load forecast value, and calculate the monthly demand safety threshold based on the standard deviation of historical load data.

[0063] The model building and solution module is used to build a mixed integer programming model with soft demand constraints. With the goal of minimizing the total life cycle cost of the energy storage system, it solves for the Pareto optimal configuration scheme of the rated capacity and maximum charge and discharge power of the energy storage system. The total life cycle cost includes investment cost, electricity expenditure, energy storage replacement cost and capacity degradation cost.

[0064] Furthermore, the load forecasting module includes:

[0065] The data preparation unit is used to divide the input data into date-time series and load value columns, and to preprocess the input data, including missing value handling, outlier handling, and resampling and aggregation.

[0066] The model building unit is used to decompose the time series into a trend component, a seasonal component, a holiday effect, and an error term. The trend component adopts a piecewise linear trend, the seasonal component is modeled using Fourier series, and the holiday effect is modeled separately.

[0067] The parameter tuning unit optimizes the model by adjusting parameters such as trend change points, seasonal intensity, and the impact of holidays.

[0068] The model training and prediction unit is used to fit the constructed model, generate date data frames for future time points, and generate predicted values ​​and confidence intervals for future time points based on the fitted model.

[0069] Another object of the present invention is to provide an electronic device, including a processor and a memory storing a computer program, characterized in that, when the processor executes the computer program, it implements the method for calculating the energy storage scale of an industrial park without photovoltaic access provided by the first object of the present invention.

[0070] In combination with the above technical solutions, the beneficial effects of the present invention compared with the prior art are as follows:

[0071] Economic optimization: Compared with traditional methods, the optimal energy storage configuration scale calculated by this invention reduces energy storage configuration costs by 18% to 25% and shortens the investment payback period by 1.5 to 2 years;

[0072] Demand risk is controllable: the probability of exceeding the maximum monthly demand limit is ≤5% (with a tolerance of ±2% of the contract value);

[0073] Dynamic adaptability: The self-correction response time of the configuration scheme under the scenario of sudden load change is ≤24 hours.

[0074] The predictive model of this invention includes soft demand constraints, allowing for short-term demand overruns, and balances economic efficiency and reliability through slack variables, avoiding investment redundancy caused by traditional hard constraints. By dynamically linking the SOH-cycle number, the physical model of battery degradation is embedded into the economic objective, realizing the quantification of the entire life cycle cost. Mixed integer programming is used to accurately characterize charging and discharging decisions through binary variables, avoiding nonlinear problems. Pareto solution set is used to reveal the optimal trade-off between investment costs and electricity expenses through multi-objective optimization. Attached Figure Description

[0075] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0076] Figure 1 This is a flowchart of the method for calculating the energy storage capacity of an industrial park without photovoltaic access, provided in an embodiment of the present invention;

[0077] Figure 2 This is a two-stage optimization flowchart provided in the embodiments of the present invention;

[0078] Figure 3 This is a schematic diagram of the structure of the energy storage capacity calculation system for industrial parks without photovoltaic access provided in an embodiment of the present invention. Detailed Implementation

[0079] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0080] Example 1:

[0081] like Figure 1 The image shows an embodiment of the method for calculating the energy storage capacity of an industrial park without photovoltaic grid connection provided by the present invention, which specifically includes the following steps:

[0082] S1: Collect historical load data, time-of-use electricity price, demand contract value and transformer capacity limit of the industrial park, perform outlier removal processing on the historical load data, and generate typical daily load curves through K-means clustering algorithm;

[0083] S2: Based on the typical daily load curve, the Prophet algorithm is used to predict the load curve for a future set period, output the confidence interval of the load prediction value, and calculate the monthly demand safety threshold based on the standard deviation of historical load data.

[0084] S3: Using the confidence interval of the load forecast and the monthly demand safety threshold as input parameters, a mixed integer programming model with soft demand constraints is constructed. The Pareto optimal configuration scheme of the rated capacity and maximum charging and discharging power of the energy storage system is obtained by minimizing the total cost of the energy storage system throughout its entire life cycle as the objective function. The total cost throughout the entire life cycle includes investment cost, electricity expenditure, energy storage replacement cost and capacity decay cost.

[0085] The implementation details of each step are explained in detail below.

[0086] Step 1: Data Input and Preprocessing

[0087] 1. Collect historical load data, time-of-use electricity price tables, demand contract values, and transformer capacity limits;

[0088] 2. Remove abnormal load data and generate typical daily load curves through K-means clustering.

[0089] Step 2: Use the Prophet algorithm to predict the load curve for the next 30 days. The Prophet algorithm is particularly suitable for processing time series data with strong seasonality, holiday effects and trend changes. Its core idea is to decompose the time series into three main parts: trend, seasonality and holiday effects, and make predictions through an additive model.

[0090] 1. Data Preparation

[0091] (1) Data format requirements

[0092] The input data consists of two columns:

[0093] ds: Date and time column (needs to be converted to pandas.datetime format)

[0094] y: Load value column (numerical data, which can be normalized to improve model stability)

[0095] (2) Data preprocessing

[0096] Missing value handling: Prophet can automatically handle missing values, eliminating the need for manual imputation.

[0097] Outlier handling: The model is sensitive to outliers. Extreme values ​​are identified and corrected using box plots or Z-score methods.

[0098] Resampling and Aggregation: If the data granularity is smaller than the daily level (e.g., hourly), it needs to be aggregated daily (e.g., summed or averaged) to match Prophet's daily forecasting characteristics.

[0099] 2. Model Building

[0100] The time series is decomposed into the following parts:

[0101] (1) Trend part (g(t)): describes the long-term trend of the time series. It adopts a piecewise linear trend and introduces change points in the model, allowing the trend slope to change at these points.

[0102] (2) Seasonal component (s(t)): represents periodic fluctuations, such as daily, weekly, and annual periodic changes. Fourier series is used to model seasonality, which can flexibly adapt to periodic changes of different frequencies.

[0103] (3) Holiday effect (h(t)): This deals with the impact of special events such as holidays on time series. These effects are usually non-periodic and cannot be accurately captured by seasonal models, so they need to be modeled separately.

[0104] (4) Error term (ε(t)): used to represent random fluctuations that the model fails to capture, usually assumed to be normally distributed white noise.

[0105] 3. Parameter tuning

[0106] (1) Adjustment of trend change points

[0107] Automatic detection: Prophet automatically identifies turning points based on the data by default.

[0108] Manual specification: Pass the known turning point date (such as policy change, equipment maintenance date) through the changepoints parameter.

[0109] (2) Seasonal intensity optimization

[0110] Adjust the seasonality_prior_scale: the larger the value, the stronger the model adapts to seasonal fluctuations; the smaller the value, the more it suppresses seasonal effects.

[0111] (3) Holidays affect calibration

[0112] The impact of holidays on forecasts can be adjusted by using holiday_prior_scale. You can manually create a holiday data frame containing information such as holiday name, date, and impact window, and then pass it to the model through the holiday parameter.

[0113] 4. Model Training and Prediction

[0114] (1) Use the fit method to fit the model. Prophet will automatically detect trends and seasonality in the data.

[0115] (2) Use the make_future_dataframe method to generate a date data frame for a future time point.

[0116] (3) Use the predict method to generate predicted values ​​and confidence intervals for future time points based on the fitted model.

[0117] 5. Output confidence interval ; Calculate the monthly demand safety threshold:

[0118]

[0119] Step 3: Construct a mixed integer programming model with soft constraints on demand, and solve for the Pareto optimal solution set that minimizes investment cost and electricity expenditure.

[0120] 1. Objective function:

[0121]

[0122] in,

[0123] (1) Investment cost

[0124]

[0125] Unit capacity cost (yuan / kWh), for example, the unit price of a lithium battery system is 800 yuan / kWh.

[0126] The rated capacity (kWh) of the energy storage system is obtained by solving the optimization model.

[0127] Unit power cost (yuan / kW), for example, the cost of inverter and installation is 500 yuan / kW.

[0128] Maximum charge / discharge power (kW) of the energy storage system is optimized in conjunction with its capacity.

[0129] (2) Electricity expenses

[0130]

[0131] Time-of-use electricity price (RMB / kWh), for example, the peak hour price is RMB 1.2 / kWh and the off-peak hour price is RMB 0.4 / kWh.

[0132] : Power exchange volume at time t (kWh) , The time step (e.g., 15 minutes).

[0133] : Demand over-limit penalty coefficient (yuan / kW), for example, the penalty for exceeding the contract demand is 50 yuan / kW.

[0134] Demand slack variable (kW): Allows short-term overruns but includes penalty costs.

[0135] (3) Energy storage replacement cost

[0136]

[0137] Cost per replacement (RMB / replacement), for example, the cost of replacing a battery pack is RMB 200,000 per replacement.

[0138] Replacement frequency (times), triggered by SOH decay (e.g., replace when SOH < 80%).

[0139] (4) Cost of capacity degradation over the entire life cycle

[0140]

[0141] Unit capacity degradation cost (yuan / kWh·year) reflects the revenue loss caused by capacity reduction (e.g., 100 yuan / kWh·year).

[0142] : Initial health state (e.g., 100%), SOH(t) is the health state at time t.

[0143] State of Health is defined as the ratio of remaining capacity to initial capacity (0~1).

[0144] 2. Constraint Parameters

[0145] (1) Soft constraint of demand

[0146]

[0147] , Charge / discharge power (kW) at time t , .

[0148] The load power (kW) at time t is generated by the Prophet algorithm.

[0149] Contractual demand limit (kW), for example, a user's maximum monthly demand contract with the power grid is 500kW.

[0150] Slack variables (allow short-run demand to exceed limits, but include penalty costs).

[0151] (2) Dynamic equilibrium of SOC

[0152]

[0153] : The state of charge at time t (0~1), SOC(0) is the initial value (e.g., 50%).

[0154] , Charge / discharge efficiency (e.g., 95%).

[0155] The nominal capacity (kWh) of the energy storage system is consistent with the physical meaning of Cmax.

[0156] (3) SOH decay model

[0157]

[0158] Total number of cycles (0.5 cycles for each charge / discharge)

[0159] : Attenuation coefficient (e.g., for lithium batteries, k≈0.001 / cycle), SOHmin=0.8 (replacement threshold)

[0160] (4) Accumulated number of cycles

[0161]

[0162] : Cycle count increment (0.5 times for each charge / discharge). The counter is triggered when the value is 1.

[0163] : Indicates whether charging / discharging occurs at time t. or The loop count is triggered at certain times.

[0164] (5) Power and capacity boundaries

[0165]

[0166]

[0167]

[0168] Mutual exclusion constraint to prevent simultaneous charging and discharging.

[0169] (6) Change decision trigger

[0170]

[0171] The amount of SOH restored in a single replacement (e.g., 0.2), meaning that the SOH increases by 20% after the replacement.

[0172] 3. Pareto optimal solution set solution method:

[0173] (1) ε-constraint method: Transform the dual objective into a single objective, such as fixing the upper limit of electricity expenditure, and optimizing investment + replacement cost:

[0174]

[0175] By adjusting Generate the Pareto front.

[0176] (2) Weighted sum method: Introduce weights ω∈[0,1]:

[0177]

[0178] like Figure 2 As shown, this embodiment of the invention realizes the feasibility and optimal calculation of energy storage scale in stages. The specific process logic is as follows:

[0179] 1. Data Input Stage: Based on core data such as historical electricity load data, monthly electricity price information, and maximum demand constraints, basic information such as electricity load, monthly electricity price, and maximum demand in scenarios without photovoltaics is entered to provide data support for subsequent energy storage scale calculations.

[0180] 2. First Phase: Preliminary Energy Storage Scale Assessment

[0181] First, based on historical data and rule-based algorithms, the initial energy storage scale, including energy storage capacity and charging / discharging strategies, is calculated to determine the initial energy storage scale.

[0182] Taking into account constraints such as grid access limitations and energy storage charging and discharging power limitations, and considering grid constraints and economic assessments, the initial scale is verified and adjusted, and a preliminary optimized energy storage scale plan is output.

[0183] 3. Second Phase: Optimization of Fine-Scale Energy Storage

[0184] Advanced optimization methods such as genetic algorithms and particle swarm optimization are adopted to consider multi-timescale characteristics and multiple constraints, while also taking into account dimensions such as energy storage cost, benefits and environmental impact.

[0185] The parameters are continuously optimized through iterative calculations to determine the optimal energy storage scale, and the results are subjected to sensitivity analysis.

[0186] The system outputs the final energy storage scale and configuration recommendations, and then generates a special report to support energy storage configuration decisions for industrial parks.

[0187] Example 2:

[0188] like Figure 3 As shown in the figure, this embodiment of the invention provides a system for calculating the energy storage capacity of an industrial park without photovoltaic access, comprising:

[0189] The data input and preprocessing module is used to collect historical load data, time-of-use electricity prices, demand contract values ​​and transformer capacity limits of the industrial park, perform outlier removal processing on the historical load data, and generate typical daily load curves through K-means clustering algorithm;

[0190] The load forecasting module is used to use the Prophet algorithm to predict the load curve for a set period in the future based on the typical daily load curve, output the confidence interval of the load forecast value, and calculate the monthly demand safety threshold based on the standard deviation of historical load data.

[0191] The model building and solution module is used to build a mixed integer programming model with soft demand constraints. With the goal of minimizing the total life cycle cost of the energy storage system, it solves for the Pareto optimal configuration scheme of the rated capacity and maximum charge and discharge power of the energy storage system. The total life cycle cost includes investment cost, electricity expenditure, energy storage replacement cost and capacity degradation cost.

[0192] Preferably, the load forecasting module in this embodiment of the invention includes:

[0193] The data preparation unit is used to divide the input data into date-time series and load value columns, and to preprocess the input data, including missing value handling, outlier handling, and resampling and aggregation.

[0194] The model building unit is used to decompose the time series into a trend component, a seasonal component, a holiday effect, and an error term. The trend component adopts a piecewise linear trend, the seasonal component is modeled using Fourier series, and the holiday effect is modeled separately.

[0195] The parameter tuning unit optimizes the model by adjusting parameters such as trend change points, seasonal intensity, and the impact of holidays.

[0196] The model training and prediction unit is used to fit the constructed model, generate date data frames for future time points, and generate predicted values ​​and confidence intervals for future time points based on the fitted model.

[0197] Example 3:

[0198] The present invention provides an electronic device, including a processor and a memory storing a computer program, characterized in that the processor executes the computer program to implement the method for calculating the energy storage scale of an industrial park without photovoltaic access, which is the first objective of the present invention.

[0199] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated in the present invention, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0200] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0201] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for calculating the energy storage capacity of an industrial park without photovoltaic grid connection, characterized in that, The method includes the following steps: (1) Collect historical load data, time-of-use electricity price, demand contract value and transformer capacity limit of industrial park, perform outlier removal processing on the historical load data, and generate typical daily load curves through K-means clustering algorithm; (2) Based on the typical daily load curve, the Prophet algorithm is used to predict the load curve for a future set period, output the confidence interval of the load prediction value, and calculate the monthly demand safety threshold based on the standard deviation of historical load data. (3) Using the confidence interval of the load forecast value and the monthly demand safety threshold as input parameters, a mixed integer programming model containing demand soft constraints is constructed. The Pareto optimal configuration scheme of the rated capacity and maximum charging and discharging power of the energy storage system is obtained by minimizing the total cost of the energy storage system throughout its entire life cycle as the objective function. The total cost throughout its entire life cycle includes investment cost, electricity expenditure, energy storage replacement cost and capacity decay cost.

2. The method for calculating the energy storage capacity of an industrial park without photovoltaic grid connection as described in claim 1, characterized in that, In step (2), before performing load forecasting, the Prophet algorithm first performs data preprocessing on the input data, specifically including... Missing value handling: Automatically imput missing values; Outlier handling: Identify and correct extreme values ​​using box plots or Z-Score methods; Resampling and aggregation: If the data granularity is smaller than the daily level, it needs to be aggregated by day to match Prophet's daily forecasting characteristics.

3. The method for calculating the energy storage capacity of an industrial park without photovoltaic grid connection as described in claim 1, characterized in that, In step (2), the Prophet algorithm is used to predict the load curve for a future set period, and the confidence interval of the load prediction value is output, including: The time series is decomposed into a trend component, a seasonal component, a holiday effect, and an error term. The trend component adopts a piecewise linear trend, the seasonal component is modeled using Fourier series, and the holiday effect is modeled separately. The model was optimized by adjusting parameters such as trend change points, seasonality intensity, and the impact of holidays. The model is fitted to generate date data frames for future time points, and the predicted values ​​and confidence intervals for future time points are generated based on the fitted model.

4. The method for calculating the energy storage capacity of an industrial park without photovoltaic grid connection as described in claim 1, characterized in that, In step (3), the constraint parameters of the mixed integer programming model include: 1) Soft demand constraint: ； in, , Let be the charging and discharging power at time t, respectively. Let be the load power at time t. This is the upper limit of the contract demand. These are slack variables; 2) Dynamic equilibrium of SOC: ； in, Let t be the state of charge. , These are the charging and discharging efficiencies, respectively. This refers to the nominal capacity of the energy storage system. 3) SOH decay model: ； in, To accumulate the number of loops, The attenuation coefficient is... Replace the threshold for energy storage; 4) Accumulated loop count: ； in, For the increment of the loop count, The counter is triggered when =1; Indicates whether charging / discharging occurs at time t. or The loop count is triggered at any time; 5) Power and capacity boundaries ； ； ； in, For mutual exclusion constraints; 6) Change decision trigger ； This refers to the amount of SOH restored in a single replacement.

5. The method for calculating the energy storage capacity of an industrial park without photovoltaic grid connection as described in claim 1, characterized in that, In step (3), the objective function is expressed as: ； in, For investment costs, For electricity expenses, For energy storage replacement costs, Cost of capacity decay; The Pareto optimal allocation scheme is solved using the ε-constraint method or the weighted sum method; The ε-constraint method transforms the dual objective into a single objective, expressed as: ； The weighted sum method introduces weights ω∈[0,1], expressed as: 。 6. A system for calculating the energy storage capacity of an industrial park without photovoltaic access, characterized in that, The system includes: The data input and preprocessing module is used to collect historical load data, time-of-use electricity prices, demand contract values ​​and transformer capacity limits of the industrial park, perform outlier removal processing on the historical load data, and generate typical daily load curves through K-means clustering algorithm; The load forecasting module is used to use the Prophet algorithm to predict the load curve for a set period in the future based on the typical daily load curve, output the confidence interval of the load forecast value, and calculate the monthly demand safety threshold based on the standard deviation of historical load data. The model building and solution module is used to build a mixed integer programming model with soft demand constraints. With the goal of minimizing the total life cycle cost of the energy storage system, it solves for the Pareto optimal configuration scheme of the rated capacity and maximum charge and discharge power of the energy storage system. The total life cycle cost includes investment cost, electricity expenditure, energy storage replacement cost and capacity degradation cost.

7. The energy storage capacity calculation system for industrial parks without photovoltaic grid connection according to claim 6, characterized in that, The load forecasting module includes: The data preparation unit is used to divide the input data into date-time series and load value columns, and to preprocess the input data, including missing value handling, outlier handling, and resampling and aggregation. The model building unit is used to decompose the time series into a trend component, a seasonal component, a holiday effect, and an error term. The trend component adopts a piecewise linear trend, the seasonal component is modeled using Fourier series, and the holiday effect is modeled separately. The parameter tuning unit optimizes the model by adjusting parameters such as trend change points, seasonal intensity, and the impact of holidays. The model training and prediction unit is used to fit the constructed model, generate date data frames for future time points, and generate predicted values ​​and confidence intervals for future time points based on the fitted model.

8. An electronic device comprising a processor and a memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the method for calculating the energy storage capacity of an industrial park without photovoltaic access as described in any one of claims 1 to 5.

Images