Energy storage trusted capacity calculation method and system based on entropy risk and medium

Abstract

The invention discloses an energy storage credible capacity calculation method and system based on entropy risk and a medium, and the method comprises the steps: carrying out the probability distribution modeling based on the power related data of an uncertain variable, and forming a power system load duration curve; constructing an entropy risk value, introducing an auxiliary variable into the entropy risk value, and constructing an entropy risk value cone optimization model; constructing constraint conditions for the entropy risk value cone optimization model according to a power system load duration curve; based on the constraint condition, solving the objective function of the entropy risk value cone optimization model to obtain the discharge power and the charge power of the energy storage system; and according to the discharging power and the charging power of the energy storage system, calculating the energy storage credible capacity. The model can simulate energy storage to efficiently model the load supporting capacity of the power grid under an extreme event, so as to accurately evaluate the possible credible capacity and the power supply capacity boundary thereof, thereby improving the operation safety and capacity configuration efficiency of a power system under the extreme condition.

Description

Technical Field

[0001] This invention relates to the field of energy storage reliable capacity assessment technology, specifically to a method, system, and medium for calculating energy storage reliable capacity based on entropy risk. Background Technology

[0002] With the integration of a high proportion of renewable energy into the power system, the uncertainty of system operation has increased significantly. This is especially true after the large-scale integration of intermittent power sources such as wind and solar power, which has led to highly volatile and unpredictable net load. Against this backdrop, capacity adequacy assessment has become a crucial step in ensuring grid operational safety and developing capacity compensation mechanisms. Existing technologies primarily employ traditional capacity assessment methods such as the equivalent load substitution method. These methods typically rely on large amounts of historical data or scenario simulation results to construct discrete net load scenarios and determine the required reliable capacity level of the system by evaluating the expected load loss for each scenario. However, with the increase in system dimensionality and the enhanced continuity of variables, these methods face technical bottlenecks when dealing with high-dimensional and complex uncertainties, including large sample size requirements, insufficient modeling accuracy, and high algorithmic complexity. Their results exhibit poor stability and robustness, particularly in scenarios that need to reflect extreme tail events.

[0003] On the other hand, with the promotion and development of energy storage technology, the role of energy storage systems in capacity support and emergency response is becoming increasingly prominent. Existing methods often use static constraints or scheduling boundaries for coarse modeling of the role of energy storage, failing to fully consider the time-coupling characteristics of its charging and discharging process and its dynamic adjustment capabilities under extreme conditions. Furthermore, traditional models mostly prioritize economic efficiency as the primary optimization objective, neglecting the ability of energy storage resources to mitigate load loss risks in extreme risk scenarios, leading to an underestimation of its role in capacity assessment models.

[0004] Furthermore, regarding the measurement of tail risk, existing models generally employ traditional risk metrics such as VaR (Value at Risk) or CVaR (Conditional Value at Risk). While these metrics perform well in handling moderately deviating scenarios, they suffer from distortion in assessing extreme events and cannot accurately measure the actual impact of high-loss, low-probability scenarios on system reliability. Moreover, risk measurement methods like CVaR typically require constructing large-scale linear models or performing scenario-based reweighting when dealing with high-dimensional inputs, leading to increased overall computational complexity and hindering large-scale system applications.

[0005] Therefore, existing capacity assessment technologies have significant shortcomings in the following aspects: First, they are highly dependent on sample size when dealing with high-dimensional continuous uncertain variables (such as wind and solar power output and branch failure probability), resulting in low algorithm efficiency; second, they lack risk measurement tools with tail sensitivity in system extreme risk modeling, which cannot amplify the impact of extreme events on capacity planning decisions; and third, the temporal and dynamic charging and discharging characteristics of energy storage cannot be effectively modeled in the existing assessment framework, weakening the accurate expression of its risk mitigation capabilities.

[0006] In view of the above, this application is hereby submitted. Summary of the Invention

[0007] The technical problem this invention aims to solve is that existing capacity assessment technologies suffer from limitations in sampling dimensions, scheduling mechanism mismatch, and insufficient expression of optimization models when dealing with complex uncertainties. These issues make it difficult to meet the assessment requirements of new power systems that emphasize both reliability and flexibility, and result in inaccurate calculations of reliable energy storage capacity. The purpose of this invention is to provide a method, system, and medium for calculating reliable energy storage capacity based on entropy risk. This invention simultaneously considers high-dimensional uncertainty modeling, tail risk enhancement response (i.e., tail risk characteristics), and the dynamic characteristics of energy storage systems to create a novel capacity assessment method. Specifically, it establishes a load shedding risk model based on entropy risk value (EVaR) (i.e., entropy risk value cone optimization model) and integrates an energy storage scheduling model into it. This model can efficiently model the grid load support capacity of energy storage under extreme events, thereby accurately assessing the possible reliable capacity and its power supply capacity boundary, thus improving the operational safety and capacity allocation efficiency of the power system under extreme conditions.

[0008] This invention is achieved through the following technical solution:

[0009] In a first aspect, the present invention provides a method for calculating the reliable capacity of energy storage based on entropy risk, the method comprising:

[0010] Probability distribution modeling is performed based on power-related data with uncertain variables, and a power system load duration curve is generated.

[0011] Construct an entropy risk value and introduce auxiliary variables into the entropy risk value to build an entropy risk value cone optimization model;

[0012] Based on the power system load duration curve, construct constraints for the entropy risk value cone optimization model;

[0013] Based on the constraints, the objective function of the entropy risk value cone optimization model is solved to obtain the discharge power and charging power of the energy storage system.

[0014] The reliable capacity of the energy storage system is calculated based on its discharge and charging power.

[0015] Furthermore, the power-related data for uncertain variables include system load, wind and solar power output, and photovoltaic power output.

[0016] Furthermore, probability distribution modeling is performed based on power-related data with uncertain variables, and a power system load duration curve is generated, including:

[0017] The system load is modeled using a normal distribution to obtain the first probability distribution.

[0018] We use the Weibull distribution to model the probability distribution of wind and solar power output to obtain a second probability distribution.

[0019] The third probability distribution is obtained by using the WBeta distribution or the triangular distribution to model the probability distribution of photovoltaic power generation output.

[0020] Based on the first probability distribution, the second probability distribution, and the third probability distribution, multiple scenario samples are generated by random sampling and then stitched together to form the power system load duration curve.

[0021] Furthermore, the expression for the entropy risk value is:

[0022]

[0023] in, It is called the moment generating function of the random variable g(X), which reflects information about all the moments of g(X); α is the confidence level; t is the relaxation factor.

[0024] Furthermore, the entropy risk value cone optimization model introduces a first auxiliary variable z into the entropy risk value. i Second auxiliary variable The problem of minimizing the expression for entropy risk value is equivalently rewritten as a cone optimization model using the mathematical form of the exponential cone.

[0025] Among them, the first auxiliary variable z i This represents the upper bound of the risk value corresponding to each random variable;

[0026] Second auxiliary variable This represents the auxiliary cone variable of the exponential term.

[0027] Furthermore, the entropy risk value cone optimization model is as follows:

[0028]

[0029] In the formula, min{} is the objective function of the entropy risk value cone optimization model; α is the confidence level;

[0030] v hAs an auxiliary scene variable, it is used to characterize the contribution of the loss of the h-th sample (or hour) after the exponential function transformation; it maps the information of the original loss distribution to the exponential cone structure through constraints, as an approximate representation of the tail distribution;

[0031] This represents the load duration curve of the power system.

[0032] z represents a threshold-type decision variable under the loss distribution, used to construct an exponential cone constraint. The loss distribution, mathematically defined as the distribution function of the optimization objective, is applied here as the load duration curve.

[0033] t is the relaxation factor, which controls the scale of the exponential weighting; it corresponds to the denominator of the logarithmic transformation of the moment generating function in the entropy risk value, and plays a role in smoothing the tail risk curve. It is the set of real numbers.

[0034] Furthermore, the constraints include entropy risk value constraints and energy storage constraints;

[0035] Entropy risk value constraints include exponential cone constraints and scenario average constraints.

[0036] Furthermore, the expression for the entropy risk value constraint is:

[0037] Exponential cone constraint:

[0038]

[0039] Scenario average constraint:

[0040]

[0041] The expression for the energy storage constraint is:

[0042]

[0043] 0≤E h ≤E max ,

[0044]

[0045] Among them, L h This represents the system load. and η represents the charging power and discharging power of the energy storage system during time period h, respectively; η is the charging and discharging efficiency, i.e., the round-trip efficiency of the energy storage system; E h E represents the state of energy storage (SOC) at the end of time period h; h-1 This indicates the State of Storage (SOC) at the end of time period h-1; s h E is a binary variable representing the state of charge and discharge;max This represents the maximum value of the SOC state, i.e., the maximum battery capacity; p dc,max p is the maximum discharge power. c,max Maximum charging power.

[0046] Secondly, this invention also provides a reliable energy storage capacity calculation system based on entropy risk, the system comprising:

[0047] The data modeling unit is used to perform probability distribution modeling based on power-related data with uncertain variables, and to generate power system load duration curves; the power-related data with uncertain variables include system load, wind and solar power output, and photovoltaic power output;

[0048] The model building unit is used to construct the entropy risk value and introduce auxiliary scenario variables into the entropy risk value to construct the entropy risk value cone optimization model;

[0049] The constraint construction unit is used to construct constraints for the entropy risk value cone optimization model based on the power system load duration curve.

[0050] The model solving unit is used to solve the objective function of the entropy risk value cone optimization model based on constraints, so as to obtain the discharge power and charging power of the energy storage system.

[0051] The reliable capacity calculation unit is used to calculate the reliable capacity of the energy storage system based on the discharge power and charging power of the energy storage system.

[0052] Furthermore, the data modeling unit includes:

[0053] The first modeling subunit is used to model the system load using a normal distribution to obtain the first probability distribution;

[0054] The second modeling subunit is used to model the probability distribution of wind and solar power output using the Weibull distribution to obtain the second probability distribution.

[0055] The third modeling subunit is used to model the probability distribution of photovoltaic power generation output using WBeta distribution or triangular distribution to obtain the third probability distribution.

[0056] The splicing sub-unit is used to generate multiple scenario samples by random sampling based on the first probability distribution, the second probability distribution, and the third probability distribution, and then splice them into a power system load continuity curve.

[0057] Thirdly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for calculating the reliable capacity of energy storage based on entropy risk.

[0058] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0059] 1. This invention relates to a reliable energy storage capacity calculation method, system, and medium based on entropy risk. This invention is the first to introduce the entropy risk value (EVAR) into the capacity risk calculation and assessment modeling framework, which can effectively capture the tail loss characteristics under extreme scenarios and amplify the weight of high loss scenarios in the risk index, making the assessment results more in line with the safety assurance requirements of the power system under high uncertainty, and the accuracy of reliable energy storage capacity calculation is high.

[0060] 2. This invention relates to a method, system, and medium for calculating the reliable capacity of energy storage based on entropy risk. By constructing an exponential cone constraint form, this invention transforms the EVAR expression containing a nested structure of logarithms and exponents into a standard convex optimization problem, which can be directly applied to solve in efficient cone optimizers (such as MOSEK, SCS, etc.). This significantly improves the solvability, computational stability, and numerical accuracy of the model, and has good engineering practicality and algorithm scalability.

[0061] 3. This invention relates to a method, system, and medium for calculating the reliable capacity of energy storage based on entropy risk. The entropy risk value cone optimization model proposed in this invention introduces an auxiliary variable system, which realizes the analytical reconstruction and controllable scaling of risk constraints without affecting the rigor of the original risk definition. This makes the model have strong structural versatility and can adapt to various types of resource allocation and scheduling optimization problems.

[0062] 4. This invention relates to a reliable capacity calculation method, system, and medium for energy storage based on entropy risk. Based on the Monte Carlo method, this invention constructs a joint sample of system load and renewable output on a continuous time scale, forming a fine-grained net load duration curve. Compared with the traditional discrete scenario method, it has higher temporal integrity and statistical coverage, which is beneficial for characterizing the energy storage charging and discharging behavior and system risk evolution path on an 8760-hour scale.

[0063] 5. This invention relates to a method, system, and medium for calculating the reliable capacity of energy storage based on entropy risk. This invention constructs an energy storage scheduling model based on the EVAR form (i.e., an entropy risk value cone optimization model), which can solve for the maximum reliable capacity that energy storage can provide under a specific confidence level. This significantly improves the modeling accuracy of the role of energy storage systems in capacity planning and ancillary services, realizes a system-level energy storage configuration scheme with risk awareness, and has high promotion and application value. Attached Figure Description

[0064] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0065] Figure 1This is a flowchart of the energy storage reliable capacity calculation method based on entropy risk of the present invention;

[0066] Figure 2 This invention provides the reliable capacity of energy storage at different ratios and permeability.

[0067] Figure 3 This invention provides a reliable capacity for energy storage (1:4) under different new energy installation ratios and penetration rates.

[0068] Figure 4 This is a block diagram of the energy storage reliable capacity calculation system based on entropy risk according to the present invention. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0070] This invention presents a novel capacity assessment method that simultaneously considers high-dimensional uncertainty modeling, tail risk enhancement response (i.e., tail risk characteristics), and the dynamic characteristics of energy storage systems. Specifically, it establishes a load shedding risk model (i.e., an entropy risk cone optimization model) based on entropy value-at-risk (EVaR) and integrates an energy storage scheduling model. This model can efficiently model the grid load support capacity of energy storage under extreme events, thereby accurately assessing the possible credible capacity and its power supply capacity boundary, thus improving the operational safety and capacity allocation efficiency of the power system under extreme conditions. This invention also avoids the scenario sampling dilemmas existing in traditional methods such as the equivalent load substitution method, ensuring the operational safety and capacity constraint compliance of the system under predetermined confidence levels.

[0071] The core design features of this invention are:

[0072] (1) The entropic value-at-Risk (EVaR) is introduced into the capacity risk assessment, and a load shedding risk model (i.e., the entropic value cone optimization model) based on the entropic value-at-Risk (EVaR) is constructed. This method can amplify the proportion of extreme scenarios with large losses according to the size of the tail loss, so that the energy storage scheduling strategy can minimize the extreme loss value in the system with load shedding risk scenarios instead of scheduling according to its own economic strategy.

[0073] (2) By transforming the EVAR risk expression into a convex optimization model that can be solved by an efficient cone optimizer (such as MOSEK) through the exponential cone equivalent transformation, it is suitable for the optimization scheduling needs in the operation of the power system.

[0074] (3) The method proposed in this invention transforms the reliable capacity estimation of traditional power system continuous sampling into a continuous time optimization form, which can fully reflect the continuous change of energy storage state of charge.

[0075] Example 1

[0076] like Figure 1 As shown, the present invention provides a method for calculating the reliable capacity of energy storage based on entropy risk. This method includes:

[0077] Step 1: Perform probability distribution modeling based on power-related data with uncertain variables, and generate a power system load duration curve; where the power-related data with uncertain variables include system load, wind and solar power output, and photovoltaic power output.

[0078] In this embodiment, the system load L h It can be modeled as a normal distribution or historical distribution kernel estimation; wind power output W h Photovoltaic power output S h Modeling can be done using a beta distribution, a Weibull distribution, or by fitting an empirical distribution. Step 1 specifically includes:

[0079] Step 11: Model the system load using a normal distribution to obtain the first probability distribution;

[0080] Specifically, system load is typically fitted to a multi-period distribution model based on historical data, with different distribution characteristics observed in different time periods (e.g., daytime, nighttime, seasonality). A normal distribution is generally used. Modeling hourly load values, μ h σ represents the expected load in hour h. h The standard deviation is represented by the standard deviation. Multi-peak fitting or kernel density estimation can also be introduced to further improve the model accuracy.

[0081] Step 12: Use the Weibull distribution to model the probability distribution of wind and solar power output to obtain the second probability distribution;

[0082] Specifically, for wind power output, considering its significant influence from wind speed, a Weibull distribution is generally used to model the wind speed, i.e., v ~ Weibull(k,λ), and then the wind power output W is derived by combining it with the wind turbine power curve P(v). h The distribution pattern. Given the known wind speed distribution, wind power output samples over the time series can be obtained through Monte Carlo sampling, forming... Sample sequence.

[0083] Step 13: Use the WBeta distribution or triangular distribution to model the probability distribution of photovoltaic power generation output to obtain the third probability distribution;

[0084] Specifically, the output of photovoltaic power generation is mainly affected by solar radiation intensity and temperature, and its power generation can be expressed as follows:

[0085] S h =η1·G h ·A, where G h The horizontal solar irradiance at hour h typically follows a Beta or triangular distribution, where η1 is the photovoltaic module efficiency and A is the installed area. For annual sunshine conditions, an empirical distribution can also be constructed by kernel density estimation using measured statistics.

[0086] Step 14: Based on the first probability distribution, the second probability distribution, and the third probability distribution, perform random sampling to generate multiple scenario samples and stitch them together to form a power system load duration curve.

[0087] Specifically, assuming the total number of samples is N, and each sampling generates 8760 hours of time-series data (or higher resolution), the net payload corresponding to the i-th sampling is:

[0088]

[0089] In the formula, This represents the net load corresponding to the h-th sample (or hour) of the i-th sampling; The system load value for the h-th sample (or hour) of the i-th sampling; The wind power output of the h-th sample (or hour) in the i-th sampling; This represents the photovoltaic power generation of the h-th sample (or hour) in the i-th sampling.

[0090] Based on the above sampling, the power system load duration curve L is formed. h .

[0091] Step 2: Construct the entropy risk value and introduce auxiliary variables into the entropy risk value to build an entropy risk value cone optimization model;

[0092] In this embodiment, at confidence level 1-α, the expression for the entropy risk value is:

[0093]

[0094] Where α is the confidence level; t is the relaxation factor;

[0095] The moment-generating function (MGF) of a random variable g(X) reflects information about all the moments of g(X) and is an important tool for studying the tail characteristics of its distribution. The core role of the moment-generating function in entropy risk (E VaR) is to transform the originally difficult-to-handle tail probability constraint into a convex, differentiable optimization problem. According to Chernoff's bound inequality, the following relationship holds:

[0096]

[0097] in, Let g(X) represent the probability function of variable X, and in this invention, it represents the load duration curve L. h γ represents the confidence level, which is equivalent to the confidence level α in the text. It expresses expectation.

[0098] In other words, the moment generating function can provide an exponential upper bound for the tail probability. EVaR is obtained by taking the logarithm of this upper bound, dividing by t, and then minimizing the value for t > 0, thus providing a most compact convex upper bound estimate of VaR.

[0099] Therefore, the moment generating function It acts as a bridge in EVAR: it encodes tail risk information into an analytical expression, making the risk measure not only more numerically stable, but also more suitable for solving in an optimization framework.

[0100] In this embodiment, to enhance the numerical solvability of the model, a first auxiliary variable z will be introduced into the entropy risk value. i Second auxiliary variable The problem of minimizing the expression for entropy risk value is equivalently rewritten into a cone optimization model using the mathematical form of an exponential cone, namely the entropy risk value cone optimization model. Here, the first auxiliary variable z... i The second auxiliary variable represents the upper bound of the risk value corresponding to each random variable. This represents the auxiliary cone variable of the exponential term.

[0101] Specifically, the entropy risk value cone optimization model is as follows:

[0102]

[0103] In the formula, min{} is the objective function of the entropy risk value cone optimization model; α is the confidence level;

[0104] v h As an auxiliary scene variable, it is used to characterize the contribution of the loss of the h-th sample (or hour) after the exponential function transformation; it maps the information of the original loss distribution to the exponential cone structure through constraints, serving as an approximate representation of the tail distribution; in other words, vh It is a "proxy variable" for the exponential weight loss in each scenario, thereby avoiding direct processing of nonlinear expectation forms;

[0105] above This is an exponential cone constraint. This structure originates from the logarithmic treatment of the moment generating function in the essential definition of EVAR. Since the expectation term can be approximated as the average value under discrete sampling, and the logarithmic-exponential combination is difficult to handle directly in the optimizer, a constraint is introduced by v... h It is then anchored to an exponential function, thereby allowing the original nonlinear objective to be replaced by a set of standard-form exponential cone constraints.

[0106] z represents a threshold-type decision variable under the loss distribution; its economic meaning is the "benchmark loss" or "cutoff level" set by the system at a given risk level. During optimization, z changes with t and the auxiliary scenario variable v. h The changes in the tail risk balance the severity of the tail risk.

[0107] t is a relaxation factor that controls the scale of the exponential weighting; it corresponds to the denominator of the logarithmic transformation of the moment generating function in the entropy risk value, and plays a role in smoothing the tail risk curve; a smaller t value will amplify the weight of the tail loss, resulting in a more conservative risk assessment; a larger t value will weaken the impact of tail risk. The existence of t ensures the convexity and numerical solvability of the model;

[0108] above The scenario average constraint is used to control the overall average value of auxiliary variables to not exceed the risk adjustment factor, i.e., the relaxation factor t, so as to strictly maintain the correspondence with the tail cumulative probability in the EVAR confidence definition.

[0109] The above technical solution addresses a convex optimization problem. This model maintains the rigorous definition of the EVAR risk metric at confidence level α, while avoiding numerical instability in solving log-exponential nested structures through exponential cone transformation, significantly improving solution efficiency and model scalability. Since its objective function and all constraints are convex, existing high-performance cone optimizers (such as MOSEK, SCS, or ECOS) can be directly called for rapid calculation of the confidence capacity in high-dimensional sample scenarios, demonstrating good practicality and engineering application prospects.

[0110] Step 3: Based on the power system load duration curve, construct constraints for the entropy risk value cone optimization model;

[0111] Specifically, the constraints include entropy risk value constraints and energy storage constraints; the entropy risk value constraints include exponential cone constraints and scenario average constraints.

[0112] The entropy risk value constraint has been introduced in the entropy risk value cone optimization model in step 2, and will not be repeated here.

[0113] Furthermore, the expression for energy storage constraints, i.e., the net load after energy storage scheduling, is:

[0114]

[0115] 0≤E h ≤E max ,

[0116]

[0117] Among them, L h This represents the system load. and η represents the charging power and discharging power of the energy storage system during time period h, respectively; η is the charging and discharging efficiency, i.e., the round-trip efficiency of the energy storage system; E h E represents the state of energy storage (SOC) at the end of time period h; h-1 This indicates the State of Storage (SOC) at the end of time period h-1; s h A binary variable representing the charging and discharging state;

[0118] Energy storage constraints ensure accurate calculation of net load, energy balance of energy storage, SOC limits, and operational feasibility.

[0119] Step 4: Based on the constraints, solve the objective function of the entropy risk cone optimization model to obtain the discharge power and charging power of the energy storage system.

[0120] In this embodiment, it can be directly applied to a high-efficiency cone optimizer (such as MOSEK, SCS, etc.) to solve for the discharge power and charging power of the energy storage system.

[0121] Step 5: Calculate the reliable energy storage capacity based on the discharge power and charging power of the energy storage system.

[0122] In this embodiment, the net load curves are first arranged in descending order, and the top one is selected. Each time period represents the period of most intense system load. The dispatch output of energy storage is extracted during these periods and summed. The formula for calculating the reliable capacity (UCAP) of energy storage is:

[0123]

[0124] in, This represents the set of indexes representing these peak periods; This represents the net output of the energy storage system during time period h. This method ensures that capacity certification reflects the system's actual capacity to handle peak demand.

[0125] After completing steps 1 to 5 above, the calculation and evaluation of the reliable energy storage capacity can be completed.

[0126] It should be noted that: an energy storage system can be simply called energy storage, which is similar to a generator unit; the power system is the entire network of power generation, transmission and distribution, which includes generator units.

[0127] In specific implementation, the steps are as follows: (1) wind, solar and load data acquisition; (2) power system load continuity curve modeling based on Monte Carlo sampling; (3) energy storage scheduling model programming based on entropy risk value (EVaR) optimization; (4) energy storage reliable capacity solution based on EvaR.

[0128] This embodiment is performed using Mosek based on Python to calculate the reliable capacity coefficient when the energy storage penetration rate is 0-20%. The results are as follows: When the energy storage penetration rate increases from 0% to 20%, the reliable capacity coefficients for energy storage with power-to-capacity ratios of 1:2, 1:4, and 1:8 decrease from 0.5 to 0.25, 0.8 to 0.35, and 1.0 to 0.45, respectively; Figure 2 and Figure 3 As shown.

[0129] from Figure 2 and Figure 3 The results above fully demonstrate the feasibility of large-scale promotion on engineering sites.

[0130] Example 2

[0131] like Figure 4 As shown, the difference between this embodiment and Embodiment 1 is that this embodiment provides a reliable energy storage capacity calculation system based on entropy risk, which corresponds one-to-one with the reliable energy storage capacity calculation method based on entropy risk in Embodiment 1; the system includes:

[0132] The data modeling unit is used to perform probability distribution modeling based on power-related data with uncertain variables, and to generate power system load duration curves; the power-related data with uncertain variables include system load, wind and solar power output, and photovoltaic power output;

[0133] The model building unit is used to construct the entropy risk value and introduce auxiliary scenario variables into the entropy risk value to construct the entropy risk value cone optimization model;

[0134] The constraint construction unit is used to construct constraints for the entropy risk value cone optimization model based on the power system load duration curve.

[0135] The model solving unit is used to solve the objective function of the entropy risk value cone optimization model based on constraints, so as to obtain the discharge power and charging power of the energy storage system.

[0136] The reliable capacity calculation unit is used to calculate the reliable capacity of the energy storage system based on the discharge power and charging power of the energy storage system.

[0137] As a further implementation, the data modeling unit includes:

[0138] The first modeling subunit is used to model the system load using a normal distribution to obtain the first probability distribution;

[0139] The second modeling subunit is used to model the probability distribution of wind and solar power output using the Weibull distribution to obtain the second probability distribution.

[0140] The third modeling subunit is used to model the probability distribution of photovoltaic power generation output using WBeta distribution or triangular distribution to obtain the third probability distribution.

[0141] The splicing sub-unit is used to generate multiple scenario samples by random sampling based on the first probability distribution, the second probability distribution, and the third probability distribution, and then splice them into a power system load continuity curve.

[0142] The execution process of each unit can be carried out according to the steps of the entropy risk-based reliable capacity calculation method for energy storage in Example 1, and will not be described in detail in this example.

[0143] Meanwhile, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for calculating the reliable capacity of energy storage based on entropy risk.

[0144] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0145] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0146] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0147] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0148] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for calculating the reliable capacity of energy storage based on entropy risk, characterized in that, The method includes: Probability distribution modeling is performed based on power-related data with uncertain variables, and a power system load duration curve is generated. Construct an entropy risk value and introduce auxiliary variables into the entropy risk value to construct an entropy risk value cone optimization model; Based on the power system load duration curve, construct constraints for the entropy risk value cone optimization model; Based on the aforementioned constraints, the objective function of the entropy risk cone optimization model is solved to obtain the discharge power and charging power of the energy storage system. The reliable capacity of the energy storage system is calculated based on its discharge and charging power.

2. The method for calculating the reliable capacity of energy storage based on entropy risk according to claim 1, characterized in that, The power-related data of the uncertain variables include system load, wind and solar power output, and photovoltaic power output.

3. The method for calculating the reliable capacity of energy storage based on entropy risk according to claim 2, characterized in that, Probability distribution modeling is performed based on power-related data with uncertain variables, and a power system load duration curve is generated, including: The system load is modeled using a normal distribution to obtain the first probability distribution. We use the Weibull distribution to model the probability distribution of wind and solar power output to obtain a second probability distribution. The third probability distribution is obtained by using the WBeta distribution or the triangular distribution to model the probability distribution of photovoltaic power generation output. Based on the first probability distribution, the second probability distribution, and the third probability distribution, multiple scenario samples are generated by random sampling and then stitched together to form a power system load duration curve.

4. The method for calculating the reliable capacity of energy storage based on entropy risk according to claim 1, characterized in that, The expression for the entropy risk value is: in, It is called the moment generating function of the random variable g(X), which reflects information about all the moments of g(X); α is the confidence level; t is the relaxation factor.

5. The method for calculating the reliable capacity of energy storage based on entropy risk according to claim 1, characterized in that, The entropy risk value cone optimization model introduces a first auxiliary variable z into the entropy risk value. i Second auxiliary variable The problem of minimizing the expression for the entropy risk value is equivalently rewritten as a cone optimization model using the mathematical form of the exponential cone. Wherein, the first auxiliary variable z i This represents the upper bound of the risk value corresponding to each random variable; Second auxiliary variable This represents the auxiliary cone variable of the exponential term.

6. The method for calculating the reliable capacity of energy storage based on entropy risk according to claim 1, characterized in that, The entropy risk value cone optimization model is as follows: In the formula, min{} is the objective function of the entropy risk value cone optimization model; α is the confidence level; v h As an auxiliary scene variable, it is used to characterize the contribution of the loss of the h-th sample after the exponential function transformation; It maps the information of the original loss distribution to the exponential cone structure through constraints, serving as an approximate representation of the tail distribution; This represents the load duration curve of the power system. z represents a threshold-type decision variable under the loss distribution, used to construct an exponential cone constraint. The loss distribution, mathematically defined as the distribution function of the optimization objective, is applied here as the load duration curve. t is the relaxation factor, which controls the scale of the exponential weighting; It corresponds to the denominator of the logarithmic transformation of the moment generating function in the entropy risk value, and plays a role in smoothing the tail risk curve; It is the set of real numbers.

7. The method for calculating the reliable capacity of energy storage based on entropy risk according to claim 6, characterized in that, The constraints include entropy risk value constraints and energy storage constraints; The entropy risk value constraint includes the exponential cone constraint and the scenario average constraint.

8. The method for calculating the reliable capacity of energy storage based on entropy risk according to claim 7, characterized in that, The expression for the entropy risk value constraint is: Exponential cone constraint: Scenario average constraint: The expression for the energy storage constraint is: 0≤E h ≤E max , Among them, L h This represents the system load. and η represents the charging power and discharging power of the energy storage system during time period h, respectively; η is the charging and discharging efficiency, i.e., the round-trip efficiency of the energy storage system; E h E represents the energy storage state at the end of time period h; h-1 This indicates the energy storage state at the end of time period h-1; s h E is a binary variable representing the state of charge and discharge; max This represents the maximum value of the SOC state, i.e., the maximum battery capacity; p dc,max p is the maximum discharge power. c,max Maximum charging power.

9. A reliable capacity calculation system for energy storage based on entropy risk, characterized in that, The system includes: The data modeling unit is used to perform probability distribution modeling based on power-related data with uncertain variables, and to generate a power system load duration curve; the power-related data with uncertain variables includes system load, wind and solar power output, and photovoltaic power output; The model building unit is used to construct the entropy risk value and introduce auxiliary scenario variables into the entropy risk value to construct the entropy risk value cone optimization model. The constraint construction unit is used to construct constraints for the entropy risk value cone optimization model based on the power system load duration curve. The model solving unit is used to solve the objective function of the entropy risk value cone optimization model based on the constraints, so as to obtain the discharge power and charging power of the energy storage system. The reliable capacity calculation unit is used to calculate the reliable capacity of the energy storage system based on the discharge power and charging power of the energy storage system.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the energy storage reliable capacity calculation method based on entropy risk as described in any one of claims 1 to 8.

Images