A dynamic scheduling method and system of energy storage resources

Abstract

The application discloses a kind of dynamic scheduling method and system of energy storage resource, applied to electric power system dispatching field, including determining risk state estimation value based on meteorological prediction data and power grid operation data;Synchronization phasor data is analyzed to obtain disturbance entropy and disturbance complexity;Risk perception result representing risk accumulation state is obtained based on risk state estimation value, disturbance entropy and disturbance complexity;According to risk perception result, the economic characteristic parameters and operating characteristic parameters of energy storage resource are combined to adjust the objective function of energy storage optimization scheduling model;With disturbance entropy as constraint condition, the adjusted model is solved, and scheduling instruction is obtained and executed.The energy storage resource dynamic scheduling method provided by the application fuses meteorological prediction, power grid operation state and real-time monitoring information, constructs scheduling architecture of risk perception guided optimization target and real-time state constraint solution boundary, and realizes dynamic matching of scheduling strategy and system risk level.

Description

Technical Field

[0001] This invention relates to the field of power system dispatching technology, and in particular to a dynamic dispatching method and system for energy storage resources. Background Technology

[0002] With the increasing penetration rate of renewable energy, energy storage systems are playing an increasingly prominent role in coping with fluctuations in new energy output and ensuring the safe and stable operation of the power grid. In power grid dispatching, the dispatching strategy of energy storage resources directly affects the system's operational economy and power supply reliability. Existing energy storage dispatching methods are usually based on optimization models, aiming to minimize operating costs, and allocate the charging and discharging power of energy storage resources according to the system's operating status and forecast data.

[0003] Existing energy storage scheduling methods typically aim to minimize operating costs, using fixed weighting coefficients to weight operating efficiency and system safety in the optimization objective. This makes it difficult to feed real-time system operating status back into scheduling decisions. When operating conditions change, the scheduling strategy cannot dynamically adjust according to the actual risk level. This can easily lead to redundant reserve capacity under low-risk conditions, while under high-risk conditions, it may result in insufficient safety margins, making it difficult to adapt to complex operating scenarios. Summary of the Invention

[0004] This invention provides a dynamic scheduling method and system for energy storage resources to solve the technical problem that existing scheduling strategies are difficult to dynamically adjust and optimize according to the level of system operation risk, resulting in insufficient coordination between operation efficiency and safety margin, so as to achieve the effect of dynamic matching between scheduling strategy and system risk level.

[0005] To address the aforementioned technical problems, embodiments of the present invention provide a dynamic scheduling method for energy storage resources, comprising: Based on meteorological forecast data and power grid operation data, the risk status estimate of energy storage resources is determined; The disturbance entropy and disturbance complexity are obtained by analyzing the synchronization phasor data of the power grid. Based at least on the risk state estimate, the perturbation entropy, and the perturbation complexity, a risk perception result is obtained to characterize the risk accumulation state; The economic and operational characteristic parameters of the energy storage resources are obtained, and the objective function of the pre-constructed energy storage optimization scheduling model is adjusted based on the risk perception results, the economic and operational characteristic parameters. Using the disturbance entropy as a constraint, the adjusted energy storage optimization scheduling model is solved to obtain the scheduling instructions for the energy storage resources, and the energy storage resources are scheduled according to the scheduling instructions.

[0006] As one preferred embodiment, determining the risk status estimate of energy storage resources based on meteorological forecast data and power grid operation data includes: Acquire the meteorological forecast data and the power grid operation data; The meteorological forecast data and the power grid operation data are input into a pre-constructed state-space model; The state-space model is solved using the unscented Kalman filter algorithm to obtain the risk state estimate of the energy storage resource.

[0007] As one preferred embodiment, the analysis of the power grid's synchronization phasor data to obtain the disturbance entropy and disturbance complexity includes: The data sequence at the current moment is obtained by extracting the synchronization phasor data of the power grid through a sliding time window; The data sequence is reconstructed in phase space, decomposed into multiple subsequences, and the subsequences are mapped to permutation patterns; The frequency of occurrence of each of the aforementioned permutation patterns is statistically analyzed, and the information entropy is calculated based on the frequency as the perturbation entropy; The frequency distribution is constructed based on the occurrence frequency of each of the aforementioned permutation patterns, and the Jensen-Shannon divergence between the frequency distribution and the uniform distribution is calculated to obtain the perturbation complexity.

[0008] As one preferred embodiment, obtaining the risk perception result characterizing the risk accumulation state based at least on the risk state estimate, the perturbation entropy, and the perturbation complexity includes: Based on the pre-built fusion model, the risk state estimate, the perturbation entropy, and the perturbation complexity are fuzzified to obtain the corresponding membership degrees. Fuzzy inference is performed on the membership degrees corresponding to the risk state estimate, the perturbation entropy, and the perturbation complexity based on the fuzzy rule base to obtain fuzzy output; The fuzzy output is defuzzified to obtain the risk perception result.

[0009] As one preferred embodiment, the step of obtaining the economic and operational characteristic parameters of the energy storage resource, and adjusting the objective function of the pre-constructed energy storage optimization scheduling model based on the risk perception results, the economic and operational characteristic parameters, includes: Based on the risk perception results and the operational characteristic parameters, determine the coupling risk amplification coefficient; Based on the economic characteristic parameters and the coupling risk amplification coefficient, the standby opportunity cost of the energy storage resources is determined; Based on the risk perception results, determine the reserve capacity value multiplier; Obtain the basic objective function of the energy storage optimization scheduling model, which includes an operating cost item and a standby service cost item; An opportunity cost penalty term is constructed based on the said standby cost opportunity, and the coefficient of the standby service cost term is adjusted according to the said standby capacity value multiplier; The adjusted objective function is obtained based on the operating cost item, the opportunity cost penalty item, and the adjusted backup service cost item.

[0010] Another embodiment of the present invention provides a dynamic scheduling system for energy storage resources, comprising: The risk prediction module is used to determine the risk status estimate of energy storage resources based on meteorological forecast data and power grid operation data. The status monitoring module is used to analyze the synchronization phasor data of the power grid to obtain the disturbance entropy and disturbance complexity. The fusion decision module is used to obtain a risk perception result characterizing the risk accumulation state based at least on the risk state estimate, the perturbation entropy, and the perturbation complexity. The target reconstruction module is used to obtain the economic characteristic parameters and operating characteristic parameters of the energy storage resources, and adjust the objective function of the pre-built energy storage optimization scheduling model according to the risk perception results, the economic characteristic parameters and the operating characteristic parameters. The optimization scheduling module is used to solve the adjusted energy storage optimization scheduling model with the disturbance entropy as a constraint, obtain the scheduling instructions for the energy storage resources, and schedule the energy storage resources according to the scheduling instructions.

[0011] As one preferred embodiment, the risk prediction module includes: The data acquisition unit is used to acquire the meteorological forecast data and the power grid operation data; The data input unit is used to input the meteorological forecast data and the power grid operation data into a pre-built state space model; The risk solution unit is used to solve the state-space model using the unscented Kalman filter algorithm to obtain the risk state estimate of the energy storage resource.

[0012] As one preferred embodiment, the status monitoring module includes: The data extraction unit is used to extract the synchronization phasor data through a sliding time window to obtain the data sequence at the current moment; A phase space reconstruction unit is used to reconstruct the phase space of the data sequence, decompose the data sequence into multiple sub-sequences, and map the sub-sequences into permutation patterns; The perturbation entropy calculation unit is used to count the frequency of occurrence of each of the above permutation patterns and calculate the information entropy based on the frequency, which is used as the perturbation entropy. The perturbation complexity calculation unit is used to construct a frequency distribution based on the occurrence frequency of each of the permutation patterns, and calculate the Jensen-Shannon divergence between the frequency distribution and the uniform distribution to obtain the perturbation complexity.

[0013] As one preferred embodiment, the fusion decision module includes: The fuzzification processing unit is used to perform fuzzification processing on the risk state estimate, the perturbation entropy, and the perturbation complexity based on a pre-built fusion model to obtain the corresponding membership degrees. The fuzzy inference unit is used to perform fuzzy inference on the membership degree corresponding to the risk state estimate, the perturbation entropy, and the perturbation complexity respectively based on the fuzzy rule base, and obtain fuzzy output; The defuzzing processing unit is used to defuzzify the fuzzy output to obtain the risk perception result.

[0014] As one preferred embodiment, the target reconstruction module includes: The coupling risk amplification coefficient determination unit is used to determine the coupling risk amplification coefficient based on the risk perception results and the operating characteristic parameters; The standby opportunity cost determination unit is used to determine the standby opportunity cost of the energy storage resource based on the economic characteristic parameters and the coupling risk amplification coefficient. A reserve capacity value multiplier determination unit is used to determine the reserve capacity value multiplier based on the risk perception result. The basic function acquisition unit is used to acquire the basic objective function of the energy storage optimization scheduling model, wherein the basic objective function includes an operating cost item and a standby service cost item. A coefficient adjustment unit is used to construct an opportunity cost penalty item based on the standby cost opportunity, and to adjust the coefficient of the standby service cost item based on the standby capacity value multiplier. The function refactoring unit is used to obtain the adjusted objective function based on the operating cost item, the opportunity cost penalty item, and the adjusted standby service cost item.

[0015] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: (1) This invention determines the risk state estimate by acquiring meteorological forecast data and power grid operation data, obtains the disturbance entropy and disturbance complexity by analyzing synchronous phasor data, and then obtains the risk perception result characterizing the risk accumulation state based on the above parameters. This risk perception result can comprehensively reflect the impact of external uncertainties on system operation and provide a quantitative basis for scheduling decisions. On this basis, the objective function of the optimization model is adjusted according to the risk perception result combined with the economic characteristic parameters and operation characteristic parameters of energy storage resources, and the scheduling instructions are solved using the disturbance entropy as a constraint condition, so that the scheduling strategy can dynamically adjust the optimization orientation according to the actual risk level of the system.

[0016] (2) This invention deeply integrates risk prediction and state monitoring to construct a closed-loop scheduling architecture of "risk perception - target adjustment - constraint solving - instruction execution", enabling energy storage scheduling decisions to respond in real time to changes in system operating status. Compared with the fixed-weight scheduling method in the prior art, this invention can adaptively adjust the optimization target according to the system risk accumulation state, taking into account both operating efficiency and system safety in complex operating scenarios, and improving the adaptability of energy storage scheduling schemes to scenarios with high penetration of new energy. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a dynamic scheduling method for energy storage resources in one embodiment of the present invention. Figure 2 This is a schematic diagram of a dynamic scheduling system for energy storage resources in one embodiment of the present invention; Figure label: Among them, 11. Risk prediction module, 12. Status monitoring module, 13. Fusion decision module, 14. Target reconstruction module, and 15. Optimization scheduling module. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] In the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0020] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0021] One embodiment of the present invention provides a dynamic scheduling method for energy storage resources. For details, please refer to [link to relevant documentation]. Figure 1 , Figure 1 The diagram shown is a flowchart illustrating a dynamic scheduling method for energy storage resources according to one embodiment of the present invention, which includes steps S1 to S5: S1: Determine the risk status estimate of energy storage resources based on meteorological forecast data and power grid operation data; In the scheduling of energy storage resources, the impact of external uncertainties such as weather changes and fluctuations in new energy sources on the power grid is not immediately apparent, but rather accumulates gradually over a period of time, potentially leading to system instability. However, this accumulation process cannot be directly observed by sensors and must be estimated from observable data using mathematical models. Step S1, by fusing weather forecast data and power grid operation data, employs state estimation techniques to quantify the impact of external uncertainties on the power grid into a calculable indicator—a risk state estimate.

[0022] Meteorological forecast data is used to characterize the potential impact of external climate uncertainties on the power grid. In this embodiment, the interface is required to provide meteorological forecast data for at least the next 24 hours, with a spatial resolution of no less than 5 kilometers and a temporal resolution of no less than 1 hour. Power grid operation data is used to characterize the current internal operating status of the power grid. Meteorological forecast data and power grid operation data differ fundamentally in physical meaning and time scale, but both together determine the real-time risk level of the power grid.

[0023] Preferably, in one embodiment of the present invention, determining the risk status estimate of energy storage resources based on meteorological forecast data and power grid operation data includes: Obtain meteorological forecast data and power grid operation data; Meteorological forecast data and power grid operation data are input into a pre-built state-space model; The state-space model is solved using the unscented Kalman filter algorithm to obtain the risk state estimate of energy storage resources.

[0024] Specifically, meteorological forecast data is obtained from the meteorological data interface, including future wind speed, sunshine, temperature, and other elements, as well as the temporal gradient (change per unit time) and spatial gradient (change per unit distance) of these elements. Power grid operation data is obtained from the power grid SCADA / EMS interface, including real-time power flow data such as node voltage and line power, as well as the remaining transmission capacity and its rate of change of key transmission sections.

[0025] A spatial state model is pre-constructed, which consists of state equations and observation equations.

[0026] Let the state vector be ,in The equivalent load growth rate for key areas. This represents the rate of decline of the generator set's equivalent available reserve capacity. The state equations describe the evolution of unobservable state variables: in Process noise is composed of the temporal and spatial gradients of meteorological forecast data, the variance of real-time load fluctuations, and the rate of change of transmission section margin. Process noise reflects the driving effect of external uncertainties on the evolution of the system state.

[0027] Let the observation vector be... This corresponds to the active power of the line and the voltage amplitude at each key node. The observation equations describe the mapping relationship between observable variables and state variables: in, The noise is due to observations. Both the state equation and the observation equation are nonlinear, and their specific functional relationships are determined through offline identification or mechanistic modeling.

[0028] In this embodiment, the unscented Kalman filter algorithm is used to solve the state-space model. This algorithm approximates the probability distribution of the state using a carefully selected set of sampling points without requiring linearization of the nonlinear function. Unscented Kalman filtering is a state estimation method for nonlinear systems. Its core idea is to approximate the probability distribution of the state using a defined set of sampling points (called Sigma points), propagate these sampling points through a nonlinear function, and then reconstruct a new mean and covariance from the propagated sampling points.

[0029] During calculation, a Sigma point set is generated based on the state estimate and covariance matrix from the previous time step. These Sigma points are propagated through the state equation to obtain the Sigma point set for the prediction time. The process noise inputs in the state equation include the temporal and spatial gradients of the meteorological forecast data, the variance of real-time load fluctuations, and the rate of change of transmission section margin. The mean and covariance matrix of the predicted state are reconstructed based on the propagated Sigma points. The Sigma points are then mapped to the observation space through the observation equation to reconstruct the mean and covariance matrix of the predicted observations. The Kalman gain is calculated based on the predicted state covariance, predicted observation covariance, and cross-covariance. The state estimate is updated using actual observations (line power and node voltage acquired by SCADA / EMS) and the Kalman gain to obtain the optimal state estimate for the current time step. The weights are set according to the standard method of unscented Kalman filtering.

[0030] After obtaining the state estimate, it is necessary to extract the primary component (i.e., a comprehensive representation of the equivalent load growth rate and the equivalent reserve capacity decline rate) from the state vector, and normalize it using the logistic function to map the value to the [0,1] interval, thus obtaining the risk state estimate. The closer this value is to 1, the greater the accumulated internal stress within the power grid, and the higher the risk of future instability.

[0031] S2: Analyze the synchronization phasor data of the power grid to obtain the disturbance entropy and disturbance complexity; Under normal operating conditions, the voltage phase angle and frequency fluctuations of the power grid exhibit a certain degree of randomness. However, when the system approaches instability, these fluctuations show specific regularities—either the disorder increases sharply, or the fluctuation pattern tends to become singular. Step S2 uses two analytical methods, permutation entropy and Jensen-Shannon divergence, to characterize the real-time dynamic characteristics of the system from the two dimensions of "degree of disorder" and "pattern structure," respectively, providing refined state feedback for subsequent scheduling decisions.

[0032] Synchronization phasor data consists of dynamic power grid data collected by the synchronization phasor measurement unit, including voltage phase angle and frequency. Disturbance entropy is an index calculated based on the permutation entropy method, used to quantify the degree of dynamic disorder in the power grid at the current moment. Higher disturbance entropy indicates more random and chaotic dynamic fluctuations in the power grid; lower disturbance entropy indicates more regular and orderly dynamic fluctuations. Disturbance complexity is an index calculated based on the Jensen-Shannon divergence between the frequency distribution of permutation patterns and a uniform distribution, used to quantify the degree of structure of disturbance patterns. Disturbance complexity reflects whether the system's disturbance patterns are singular. When the probabilities of occurrence of each permutation pattern are similar, the disturbance complexity is low, indicating that the system disturbances are rich and diverse, possessing a certain degree of resilience and adaptability; when a few permutation patterns dominate, the disturbance complexity is high, indicating that the system disturbance patterns tend to be singular, structurally fragile, and prone to instability under specific disturbances.

[0033] Preferably, in one embodiment of the present invention, the analysis of the synchronization phasor data of the power grid to obtain the disturbance entropy and disturbance complexity includes: By extracting the synchronous phasor data of the power grid through a sliding time window, the data sequence at the current moment can be obtained; The data sequence is reconstructed in phase space, decomposing it into multiple subsequences and mapping the subsequences to permutation patterns; The frequency of occurrence of each permutation pattern is counted, and the information entropy is calculated based on the frequency as the perturbation entropy; Construct a frequency distribution based on the occurrence frequency of each permutation pattern, and calculate the Jensen-Shannon divergence between the frequency distribution and the uniform distribution to obtain the perturbation complexity.

[0034] Specifically, synchronization phasor data is obtained from the WAMS interface, including voltage phase angle time-series data or frequency time-series data of key nodes in the power grid (such as regional power grid interconnections and grid connection points of important power plants). The synchronization phasor data is collected by the synchronization phasor measurement unit, with a sampling frequency of no less than 50 times per second, and has a unified GPS / BeiDou time scale, which can reflect the dynamic response process of the power grid at the millisecond level.

[0035] Synchronous phasor data is extracted using a sliding time window to obtain the data sequence at the current moment. The length of the sliding window is set to a value that can cover the cycle of the main electromechanical oscillation modes of the power system. In this embodiment, 200 sampling points are selected, and the window slides forward over time, outputting a data sequence for each control cycle.

[0036] The data sequence is reconstructed in phase space. The embedding dimension m and time delay τ are set, and the data sequence is decomposed into multiple subsequences of length m. In this embodiment, m is 3 and τ is 1.

[0037] The values ​​in each subsequence are sorted according to their size, and the subsequences are mapped to corresponding permutation patterns based on the sorting results. For example, the subsequence [5,2,8] is sorted such that the second element is the smallest, the first element is the second smallest, and the third element is the largest, and the corresponding permutation pattern is denoted as [1,0,2]. Since the embedding dimension is 3, there are a total of m!=6 possible permutation patterns.

[0038] Then, the frequency of each permutation pattern corresponding to all subsequences within the sliding time window is calculated, and the probability of the i-th permutation pattern is denoted as . The entropy of the original permutation is calculated using the Shannon entropy formula: The theoretical maximum value of the entropy of the original permutation The maximum value is reached when all permutation patterns occur with equal probability.

[0039] Then, the entropy of the original permutation is normalized, and the formula is as follows: The disturbance entropy is obtained, with a value range of [0,1]. The larger the disturbance entropy, the higher the degree of dynamic disorder in the power grid; the smaller the disturbance entropy, the more regular the dynamic fluctuations.

[0040] Construct a frequency distribution based on the frequency of occurrence of each permutation pattern. In a uniform distribution U, each element has a probability of 1 / 6. Calculate the Jensen-Shannon divergence between the frequency distribution and the uniform distribution: Where H(·) is the Shannon entropy, and JSD has a numerical range of [0,1]. Normalization of JSD yields the perturbation complexity. Perturbation complexity reflects the degree of structure in the perturbation pattern. When the probabilities of different permutations are similar, JSD is smaller, indicating lower perturbation complexity and a rich variety of perturbations with resilience. Conversely, when a few permutations dominate, JSD is larger, indicating higher perturbation complexity and a single, fragile structure, suggesting the system is approaching a critical state.

[0041] Disturbance entropy reflects the current degree of disorder in the system, while disturbance complexity reflects the structural vulnerability of the system's disturbance patterns. Combining these two aspects provides a more comprehensive characterization of the real-time dynamic state of the power grid, offering refined state feedback for dispatching decisions. This is in contrast to traditional dispatching methods that rely solely on whether voltage and frequency limits are exceeded.

[0042] S3: Based at least on the risk state estimate, perturbation entropy, and perturbation complexity, obtain the risk perception results used to characterize the risk accumulation state; Step S3 integrates at least the risk state estimate reflecting future risk accumulation, the risk amplification coefficient reflecting the multi-factor coupling amplification effect, the perturbation entropy reflecting the current degree of disorder, and the perturbation complexity reflecting the current structural vulnerability into a unified decision-making guidance signal. These four indicators come from different data sources, reflect different time scales, and have different physical meanings. Simple superposition or weighted averaging cannot reflect the nonlinear coupling relationship between them. In this embodiment, a fuzzy logic reasoning method is used to simulate expert decision-making thinking and integrate multi-dimensional heterogeneous information into a single risk perception result.

[0043] The risk amplification factor is determined during the unscented Kalman filter estimation process in step S1. Specifically, a prediction error sequence is formed based on the difference between the observed values ​​predicted by the state-space model (such as predicted line power) and the observed values ​​actually collected by WAMS or SCADA. In this embodiment, error components generated by climate-related inputs and error components generated by load-related inputs are extracted from this error sequence to obtain two novel sequences. This separation can be achieved by setting the climate-related and load-related parts of the process noise to zero and then running the filter independently, which is a conventional technique in this field.

[0044] Wavelet coherence analysis was performed on these two novel sequences. Wavelet coherence is a mature time-frequency analysis method that can reveal the correlation between two non-stationary time series in the time and frequency domains. First, continuous wavelet transforms were performed on both sequences to obtain their respective wavelet spectra. Then, the cross-wavelet spectra were calculated, and subsequently, the wavelet coherence spectrum was calculated. The wavelet coherence spectrum is a matrix on the time-frequency plane, where each point has a value between 0 and 1, representing the degree of correlation between the two sequences at that time and frequency.

[0045] Within a predefined key frequency band (in this embodiment, the corresponding frequency band has a period of 30 minutes to 4 hours, calibrated by offline analysis of historical fault data), the values ​​of all points in the wavelet coherence spectrum whose values ​​exceed the significance threshold (0.8 in this embodiment) are integrated to obtain the original resonance intensity value. This original value is then normalized to the [0,1] interval using the logistic function to obtain the risk amplification coefficient. This coefficient reflects the intensity of the synchronous resonance between climate factors and load factors in the key frequency band; a higher value indicates a more significant coupling effect between the two types of uncertainties and a stronger amplification effect on system risk.

[0046] Preferably, in one embodiment of the present invention, a risk perception result for characterizing the risk accumulation state is obtained based at least on the risk state estimate, perturbation entropy, and perturbation complexity, including: Based on the pre-built fusion model, the risk state estimate, perturbation entropy and perturbation complexity are fuzzified to obtain the corresponding membership degree. Fuzzy inference is performed on the membership degrees corresponding to the risk state estimate, perturbation entropy, and perturbation complexity based on the fuzzy rule base to obtain fuzzy output; The fuzzy output is defuzzified to obtain the risk perception result.

[0047] Specifically, the first step is to pre-build a fusion model based on fuzzy logic reasoning. Fuzzy logic is a mature intelligent decision-making method widely used in industrial control. Its core is to convert precise numerical inputs into fuzzy concepts, perform reasoning based on expert rules, and then convert the reasoning results into precise numerical outputs.

[0048] Then, a fuzzification process is carried out. Specifically, the risk state estimate value R, the risk amplification factor A, the perturbation entropy H, and the perturbation complexity C are respectively input into the fusion model, and each input variable is fuzzified.

[0049] Three fuzzy sets are defined for each input variable: low, medium, and high. Each fuzzy set is described by a membership function. In this embodiment, a trapezoidal membership function is adopted. Taking the risk state estimate value as an example, when R ≤ 0.3, its membership degree belonging to "low" is 1; when 0.3 < R < 0.5, its membership degree belonging to "low" linearly decreases from 1 to 0, and at the same time, its membership degree belonging to "medium" linearly increases from 0 to 1; when 0.5 ≤ R ≤ 0.7, its membership degree belonging to "medium" is 1; when 0.7 < R < 0.9, its membership degree belonging to "medium" linearly decreases from 1 to 0, and at the same time, its membership degree belonging to "high" linearly increases from 0 to 1; when R ≥ 0.9, its membership degree belonging to "high" is 1. The fuzzification methods of other input variables are similar. The parameters of the membership function can be adjusted according to the actual system characteristics.

[0050] Fuzzy inference is performed on the membership degrees according to the preset fuzzy rule base. The fuzzy rule base consists of several "IF-THEN" rules, and each rule represents a risk judgment logic. The establishment of the rule base is based on expert experience or historical data mining. Some of the rules in this embodiment are as follows: Rule 1 (high-risk resonance mode): IF the risk state estimate value is high AND the risk amplification factor is high AND the perturbation entropy is high AND the perturbation complexity is low, THEN the risk perception result is extremely high. The situation corresponding to this rule is that the future risk accumulates severely, multiple factors produce a synchronous resonance amplification effect, the current system is highly chaotic and the perturbation mode is single, and the system is on the verge of critical instability.

[0051] Rule 2 (latency mode): IF the risk state estimate value is medium AND the risk amplification factor is low AND the perturbation entropy is low AND the perturbation complexity is high, THEN the risk perception result is medium. The situation corresponding to this rule is that the future risk has begun to accumulate, but there is no resonance amplification effect, the current system is stable and the perturbation mode is rich and resilient, and it is still within the controllable range.

[0052] Rule 3 (resonance amplification mode): IF the risk state estimate value is medium AND the risk amplification factor is high, THEN the risk perception result is high. This rule reflects the gain effect of the risk amplification factor: even if the basic risk is only at a medium level, if multiple factors produce synchronous resonance, the overall risk will still increase significantly.

[0053] Rule 4 (Structurally Vulnerable Mode): IF High perturbation entropy AND Low perturbation complexity THEN High risk perception result. This rule corresponds to the situation where, even if the future risk is not high, the current system is highly chaotic and structurally vulnerable, and the system is already on the verge of instability.

[0054] In this process, each rule calculates its activation level based on the input membership degree. For rules of the form "IF A AND B AND C AND D", the activation level is the minimum value of the four premise membership degrees; for rules of the form "IF A AND B", the activation level is the minimum value of the two premise membership degrees. This reasoning method (Mamdani reasoning) is the standard implementation method for fuzzy logic systems. Finally, the activation levels of each rule are aggregated with the corresponding rule's output fuzzy set to obtain the output fuzzy set. Five fuzzy sets are defined for the output variables: very low, low, medium, high, and very high. Each rule corresponds to one output fuzzy set, and its contribution is determined by the activation level.

[0055] The aggregated output fuzzy set is then defuzzified to obtain an accurate risk perception result. This embodiment employs the centroid method, which calculates the abscissa of the centroid of the region enclosed by the membership function curve of the output fuzzy set and the horizontal axis, using this as the output value. The risk perception result is a dimensionless value between 0 and 1; a higher value indicates a more severe accumulation of system risk. The centroid method is one of the most commonly used methods in defuzzification.

[0056] S4: Obtain the economic and operational characteristic parameters of energy storage resources, and adjust the objective function of the pre-built energy storage optimization scheduling model based on the risk perception results, economic and operational characteristic parameters. Risk perception results reflect the current risk level of the system, but translating this value into scheduling instructions for energy storage resources requires a bridge—namely, adjusting the objective function of the optimization model. When the risk perception result is high, scheduling decisions should tend to reserve spare capacity to ensure system safety; when the risk perception result is low, scheduling decisions should tend to participate in market transactions to improve operational efficiency. This step achieves dynamic adjustment of the objective function by constructing an opportunity cost penalty term and a spare capacity value multiplier associated with the risk perception result, enabling the scheduling strategy to adaptively change its optimization orientation according to the risk level.

[0057] Preferably, in one embodiment of the present invention, obtaining the economic characteristic parameters and operational characteristic parameters of energy storage resources, and adjusting the objective function of a pre-built energy storage optimization scheduling model based on risk perception results, economic characteristic parameters, and operational characteristic parameters, includes: Based on the risk perception results and operational characteristic parameters, determine the coupling risk amplification factor; The opportunity cost of energy storage resources for standby is determined based on economic characteristic parameters and coupling risk amplification coefficient. Based on the risk perception results, determine the reserve capacity value multiplier; Obtain the basic objective function of the energy storage optimization scheduling model. The basic objective function includes the operating cost item and the standby service cost item. Construct an opportunity cost penalty item based on the standby cost opportunity, and adjust the coefficient of the standby service cost item according to the standby capacity value multiplier; Based on the operating cost item, opportunity cost penalty item, and adjusted standby service cost item, the adjusted objective function is obtained.

[0058] Specifically, the first step is to obtain the economic and operational characteristic parameters of the energy storage resources. The economic characteristic parameters include the basic opportunity cost, which is determined through offline analysis of the average arbitrage profit obtainable by energy storage participating in the energy market. This reflects the potential profit lost by forgoing market trading opportunities. The operational characteristic parameters include weather sensitivity, which is determined through offline analysis of the maximum mutual information coefficient between historical energy storage operational data and meteorological data. The value ranges from 0 to 1, reflecting the degree to which the energy storage's operational status is affected by weather conditions; a higher value indicates that the energy storage resource is more susceptible to weather fluctuations.

[0059] Based on risk perception results and meteorological sensitivity, a coupling risk amplification coefficient is determined. The risk perception result is multiplied by the meteorological sensitivity to obtain an intermediate value, which is then transformed using a sigmoid nonlinear function. The sigmoid function exhibits slow output growth when the input value is small, but exponential growth when the input value approaches 1. This nonlinear relationship reflects the synergistic amplification effect of risk and meteorological sensitivity: for energy storage resources with high meteorological sensitivity, an increase in risk perception results significantly amplifies their opportunity cost.

[0060] The standby opportunity cost of energy storage resources is determined based on the basic opportunity cost and the coupling risk amplification factor. Multiplying the basic opportunity cost by the coupling risk amplification factor yields the standby opportunity cost. The standby opportunity cost reflects the potential loss incurred by using the energy storage resource for standby rather than participating in market transactions at the current risk level. When the perceived risk is high, the standby opportunity cost increases, and the incentive optimization model retains the energy storage resource as a standby resource; conversely, it decreases, incentivizing the energy storage resource to participate in the market.

[0061] Based on the risk perception results, a reserve capacity value multiplier is determined. The risk perception result is squared and then incremented by 1 to obtain the reserve capacity value multiplier. When the risk perception result is low, the reserve capacity value multiplier is close to 1; when the risk perception result increases, the reserve capacity value multiplier increases accordingly. This multiplier is used to adjust the coefficient of the reserve service cost term in the optimization model, thereby increasing the value weight of reserve services under high risk.

[0062] Obtain the basic objective function of the pre-built energy storage optimization scheduling model. This basic objective function aims to minimize the total system operating cost and typically includes an operating cost item and a backup service cost item. The operating cost item includes the generation cost of each power generation unit, the energy storage charging and discharging loss cost, etc.; the backup service cost item includes the cost of purchasing backup capacity from the market or the compensation cost of providing backup services by energy storage.

[0063] An opportunity cost penalty term is constructed based on the opportunity cost of backup. For each energy storage resource, the opportunity cost penalty term is the product of its trading power in the energy market and the opportunity cost of backup. This penalty term is added to the basic objective function. Its function is to penalize the energy storage resource's participation in market transactions during optimization. The penalty strength increases with the risk perception result, thus making the energy storage resource more inclined to reserve backup under high risk. At the same time, the coefficient of the backup service cost term in the basic objective function is adjusted according to the backup capacity value multiplier. That is, all terms related to the cost of purchasing backup services are multiplied by the backup capacity value multiplier. The effect of this adjustment is to increase the value weight of backup services during optimization, making the optimizer more willing to pay higher costs for backup capacity under higher risk.

[0064] The adjusted objective function is obtained by adding the operating cost, the adjusted backup service cost, and the opportunity cost penalty. When solving this objective function, the weights of each component are dynamically adjusted based on the risk perception results: when the risk is low, the opportunity cost penalty is small, the backup service cost coefficient is close to 1, and the optimization tends to minimize the operating cost; when the risk is high, the opportunity cost penalty increases significantly, the backup service cost coefficient is greater than 1, and the optimization tends to reserve backup capacity.

[0065] S5: Solve the adjusted energy storage optimization scheduling model with the disturbance entropy as the constraint to obtain the scheduling instructions for energy storage resources, and schedule the energy storage resources according to the scheduling instructions.

[0066] Step S5 combines the adjusted optimization model, guided by the risk perception results, with the disturbance entropy obtained from real-time state monitoring to form a complete closed loop for scheduling decisions. The disturbance entropy reflects the current dynamic disorder of the power grid and is used as a constraint to ensure that the calculated scheduling instructions do not lead to system deterioration. A predictive constraint mechanism is also introduced, using an embedded surrogate model to predict the future impact of scheduling instructions and avoid short-sighted decisions.

[0067] Specifically, a two-dimensional state phase space is first constructed, consisting of perturbation entropy H and perturbation complexity C. The safety boundary in this phase space is dynamically determined by the risk state estimate: as the risk state estimate increases, the safety boundary shrinks inward, and the area of ​​the safe region decreases. This means that when the future risk is high, even if the current perturbation entropy is at a low level, the system will automatically tighten its tolerance for the operating state, thereby increasing safety redundancy.

[0068] Obtain the preset perturbation entropy threshold This threshold is set based on power grid safety operation procedures and historical experience. The constraint condition is that the disturbance entropy is less than or equal to this threshold. A constraint set is constructed for the energy storage optimization scheduling model. This constraint ensures that no feasible scheduling instruction will cause the dynamic disorder of the system to exceed the safety limit.

[0069] Then, a pre-trained surrogate model is embedded into the optimization model. This surrogate model uses regression methods such as neural networks or support vector machines, with historical dispatch instructions and their corresponding grid response data as training samples, to learn the nonlinear mapping relationship between dispatch instructions and future state points (i.e., disturbance entropy and disturbance complexity). The input of the surrogate model is the dispatch instruction to be solved, and the output is the predicted future state point.

[0070] Based on the prediction results of the surrogate model, the drift rate of the state point towards the safety boundary is defined. Let d be the Euclidean distance from the current state point to the safety boundary, and the rate of change of this distance must satisfy... ε is a preset positive drift tolerance. The physical meaning of this constraint is that the state point cannot move towards the danger boundary at a speed exceeding the tolerance, thereby avoiding accelerated system instability due to improper scheduling instructions.

[0071] The adjusted objective function and the aforementioned constraint set are input into the optimization solver to solve the energy storage optimization scheduling model. The solver can be implemented using the interior-point method, mixed-integer programming algorithm, or commercial solvers (such as CPLEX or Gurobi), and the solution yields the scheduling instructions for energy storage resources, including the operating priority and power instructions for each energy storage unit.

[0072] Considering the possibility of data interruption or quality degradation during actual operation, this embodiment also provides a multi-modal operation mechanism. At the beginning of each control cycle, the data quality confidence level is calculated based on the difference between the latest timestamp of the data and the maximum allowable delay time. The maximum allowable delay time for latent heat data (risk status estimate) is set to ten minutes, and the maximum allowable delay time for entropy data is set to five seconds. The operating mode is selected based on the data quality confidence level. Normal mode: When the confidence level of both types of data quality is not lower than 0.9, execute the above complete process.

[0073] Pure feedback mode: When the confidence level of latent heat data is less than 0.9 while the confidence level of entropy data is not less than 0.9, the risk perception result is forcibly set to 0.5, and subsequent steps are executed based on this default value.

[0074] Pure feedforward prevention mode: When the confidence level of entropy data is below 0.9 and the confidence level of latent heat data is not below 0.9, the perturbation entropy constraint is ignored, and when the risk perception result exceeds 0.95, a conservative resource scheduling instruction is directly issued.

[0075] Highest level conservative mode: When the confidence level of both types of data is below 0.9, no optimization is performed, the highest level conservative scheduling instruction is issued directly and a system alarm is issued.

[0076] Step S5 uses real-time status monitoring results as a hard constraint on the optimization model to ensure that scheduling instructions do not cause the system's dynamic disorder to exceed the limit. Through an embedded proxy model and drift rate constraints, scheduling decisions become predictive, avoiding accelerated system instability due to short-sightedness. A multimodal operation mechanism enables adaptive degradation operation under different data quality conditions, ensuring that the system can execute reasonable safety strategies under various operating conditions.

[0077] Another embodiment of the present invention provides a dynamic scheduling system for energy storage resources. For details, please refer to [link to relevant documentation]. Figure 2 , Figure 2 The diagram shown is a flowchart illustrating a dynamic scheduling method for energy storage resources according to one embodiment of the present invention, which includes: Risk prediction module 11 is used to determine the risk status estimate of energy storage resources based on meteorological forecast data and power grid operation data; The status monitoring module 12 is used to analyze the synchronization phasor data of the power grid to obtain the disturbance entropy and disturbance complexity. The fusion decision module 13 is used to obtain risk perception results that characterize the risk accumulation state, based at least on the risk state estimate, perturbation entropy, and perturbation complexity. The target reconstruction module 14 is used to obtain the economic characteristic parameters and operating characteristic parameters of energy storage resources, and adjust the objective function of the pre-built energy storage optimization scheduling model based on the risk perception results, economic characteristic parameters and operating characteristic parameters. The optimization scheduling module 15 is used to solve the adjusted energy storage optimization scheduling model with disturbance entropy as a constraint, obtain the scheduling instructions for energy storage resources, and schedule the energy storage resources according to the scheduling instructions.

[0078] Preferably, in one embodiment of the present invention, the risk prediction module includes: The data acquisition unit is used to acquire meteorological forecast data and power grid operation data; The data input unit is used to input meteorological forecast data and power grid operation data into a pre-built state-space model; The risk solution unit is used to solve the state-space model using the unscented Kalman filter algorithm to obtain the risk state estimate of the energy storage resources.

[0079] Preferably, in one embodiment of the present invention, the status monitoring module includes: The data extraction unit is used to extract synchronous phasor data through a sliding time window to obtain the data sequence at the current moment. The phase space reconstruction unit is used to reconstruct the phase space of the data sequence, decompose the data sequence into multiple subsequences, and map the subsequences into permutation patterns; The perturbation entropy calculation unit is used to count the frequency of occurrence of each permutation pattern and calculate the information entropy based on the frequency, which is used as the perturbation entropy. The perturbation complexity calculation unit is used to construct a frequency distribution based on the occurrence frequency of each permutation pattern, and calculate the Jensen-Shannon divergence between the frequency distribution and the uniform distribution to obtain the perturbation complexity.

[0080] Preferably, in one embodiment of the present invention, the fusion decision module includes: The fuzzification processing unit is used to fuzzify the risk state estimate, perturbation entropy, and perturbation complexity based on the pre-built fusion model to obtain the corresponding membership degree. The fuzzy inference unit is used to perform fuzzy inference on the membership degrees corresponding to the risk state estimate, perturbation entropy, and perturbation complexity based on the fuzzy rule base, and obtain fuzzy output. The defuzzing processing unit is used to defuzzify the fuzzy output to obtain the risk perception result.

[0081] Preferably, in one embodiment of the present invention, the target reconstruction module includes: The coupling risk amplification factor determination unit is used to determine the coupling risk amplification factor based on the risk perception results and operational characteristic parameters. The standby opportunity cost determination unit is used to determine the standby opportunity cost of energy storage resources based on economic characteristic parameters and coupling risk amplification coefficient. The standby capacity value multiplier determination unit is used to determine the standby capacity value multiplier based on the risk perception results. The basic function acquisition unit is used to acquire the basic objective function of the energy storage optimization scheduling model. The basic objective function includes the operating cost item and the standby service cost item. The coefficient adjustment unit is used to construct an opportunity cost penalty item based on the standby cost opportunity and adjust the coefficient of the standby service cost item based on the standby capacity value multiplier. The function refactoring unit is used to obtain the adjusted objective function based on the operating cost item, the opportunity cost penalty item, and the adjusted standby service cost item.

[0082] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: (1) This invention introduces synchronous phasor data and uses the permutation entropy analysis method to extract feature parameters of two dimensions, perturbation entropy and perturbation complexity, from high-frequency measurement data. Perturbation entropy can quantify the dynamic disorder of the system at the current moment, and perturbation complexity can reflect the structured characteristics of the system's perturbation mode. The combination of the two can more comprehensively depict the real-time operating state of the system and provide a refined state perception basis for scheduling decisions.

[0083] (2) This invention integrates risk state estimates, disturbance entropy, disturbance complexity and other multi-source information by constructing a fusion model, and quantifies external meteorological uncertainty, internal risk accumulation status and real-time dynamic characteristics into risk perception results, thereby realizing the effective fusion of multi-dimensional risk information and improving the accuracy and reliability of risk perception.

[0084] (3) The present invention uses perturbation entropy as the constraint condition of the optimization model, and directly incorporates the real-time dynamic disorder of the system into the constraint boundary of the scheduling decision, so that the scheduling instructions obtained can ensure that the system maintains sufficient stability margin during dynamic operation, and avoid the risk of system instability caused by improper scheduling instructions.

[0085] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for dynamic scheduling of energy storage resources, characterized in that, include: Based on meteorological forecast data and power grid operation data, the risk status estimate of energy storage resources is determined; The disturbance entropy and disturbance complexity are obtained by analyzing the synchronization phasor data of the power grid. Based at least on the risk state estimate, the perturbation entropy, and the perturbation complexity, a risk perception result is obtained to characterize the risk accumulation state; The economic and operational characteristic parameters of the energy storage resources are obtained, and the objective function of the pre-constructed energy storage optimization scheduling model is adjusted based on the risk perception results, the economic and operational characteristic parameters. Using the disturbance entropy as a constraint, the adjusted energy storage optimization scheduling model is solved to obtain the scheduling instructions for the energy storage resources, and the energy storage resources are scheduled according to the scheduling instructions.

2. The dynamic scheduling method for energy storage resources as described in claim 1, characterized in that, The process of determining the risk status estimate of energy storage resources based on meteorological forecast data and power grid operation data includes: Acquire the meteorological forecast data and the power grid operation data; The meteorological forecast data and the power grid operation data are input into a pre-constructed state-space model; The state-space model is solved using the unscented Kalman filter algorithm to obtain the risk state estimate of the energy storage resource.

3. The dynamic scheduling method for energy storage resources as described in claim 1, characterized in that, The analysis of the synchronization phasor data of the power grid yields the disturbance entropy and disturbance complexity, including: The data sequence at the current moment is obtained by extracting the synchronization phasor data of the power grid through a sliding time window; The data sequence is reconstructed in phase space, decomposed into multiple subsequences, and the subsequences are mapped to permutation patterns; The frequency of occurrence of each of the aforementioned permutation patterns is statistically analyzed, and the information entropy is calculated based on the frequency as the perturbation entropy; The frequency distribution is constructed based on the occurrence frequency of each of the aforementioned permutation patterns, and the Jensen-Shannon divergence between the frequency distribution and the uniform distribution is calculated to obtain the perturbation complexity.

4. The dynamic scheduling method for energy storage resources as described in claim 1, characterized in that, The process of obtaining risk perception results characterizing the risk accumulation state, based at least on the risk state estimate, the perturbation entropy, and the perturbation complexity, includes: Based on the pre-built fusion model, the risk state estimate, the perturbation entropy, and the perturbation complexity are fuzzified to obtain the corresponding membership degrees. Fuzzy inference is performed on the membership degrees corresponding to the risk state estimate, the perturbation entropy, and the perturbation complexity based on the fuzzy rule base to obtain fuzzy output; The fuzzy output is defuzzified to obtain the risk perception result.

5. The dynamic scheduling method for energy storage resources as described in claim 1, characterized in that, The step of obtaining the economic and operational characteristic parameters of the energy storage resources, and adjusting the objective function of the pre-constructed energy storage optimization scheduling model based on the risk perception results, the economic and operational characteristic parameters, includes: Based on the risk perception results and the operational characteristic parameters, determine the coupling risk amplification coefficient; Based on the economic characteristic parameters and the coupling risk amplification coefficient, the standby opportunity cost of the energy storage resources is determined; Based on the risk perception results, determine the reserve capacity value multiplier; Obtain the basic objective function of the energy storage optimization scheduling model, which includes an operating cost item and a standby service cost item; An opportunity cost penalty term is constructed based on the said standby cost opportunity, and the coefficient of the standby service cost term is adjusted according to the said standby capacity value multiplier; The adjusted objective function is obtained based on the operating cost item, the opportunity cost penalty item, and the adjusted backup service cost item.

6. A dynamic scheduling system for energy storage resources, characterized in that, include: The risk prediction module is used to determine the risk status estimate of energy storage resources based on meteorological forecast data and power grid operation data. The status monitoring module is used to analyze the synchronization phasor data of the power grid to obtain the disturbance entropy and disturbance complexity. The fusion decision module is used to obtain a risk perception result characterizing the risk accumulation state based at least on the risk state estimate, the perturbation entropy, and the perturbation complexity. The target reconstruction module is used to obtain the economic characteristic parameters and operating characteristic parameters of the energy storage resources, and adjust the objective function of the pre-built energy storage optimization scheduling model according to the risk perception results, the economic characteristic parameters and the operating characteristic parameters. The optimization scheduling module is used to solve the adjusted energy storage optimization scheduling model with the disturbance entropy as a constraint, obtain the scheduling instructions for the energy storage resources, and schedule the energy storage resources according to the scheduling instructions.

7. A dynamic scheduling system for energy storage resources as described in claim 6, characterized in that, The risk prediction module includes: The data acquisition unit is used to acquire the meteorological forecast data and the power grid operation data; The data input unit is used to input the meteorological forecast data and the power grid operation data into a pre-built state space model; The risk solution unit is used to solve the state-space model using the unscented Kalman filter algorithm to obtain the risk state estimate of the energy storage resource.

8. A dynamic scheduling system for energy storage resources as described in claim 6, characterized in that, The status monitoring module includes: The data extraction unit is used to extract the synchronization phasor data through a sliding time window to obtain the data sequence at the current moment; A phase space reconstruction unit is used to reconstruct the phase space of the data sequence, decompose the data sequence into multiple sub-sequences, and map the sub-sequences into permutation patterns; The perturbation entropy calculation unit is used to count the frequency of occurrence of each of the above permutation patterns and calculate the information entropy based on the frequency, which is used as the perturbation entropy. The perturbation complexity calculation unit is used to construct a frequency distribution based on the occurrence frequency of each of the permutation patterns, and calculate the Jensen-Shannon divergence between the frequency distribution and the uniform distribution to obtain the perturbation complexity.

9. A dynamic scheduling system for energy storage resources as described in claim 6, characterized in that, The fusion decision module includes: The fuzzification processing unit is used to perform fuzzification processing on the risk state estimate, the perturbation entropy, and the perturbation complexity based on a pre-built fusion model to obtain the corresponding membership degrees. The fuzzy inference unit is used to perform fuzzy inference on the membership degree corresponding to the risk state estimate, the perturbation entropy, and the perturbation complexity respectively based on the fuzzy rule base, and obtain fuzzy output; The defuzzing processing unit is used to defuzzify the fuzzy output to obtain the risk perception result.

10. A dynamic scheduling system for energy storage resources as described in claim 6, characterized in that, The target reconstruction module includes: The coupling risk amplification coefficient determination unit is used to determine the coupling risk amplification coefficient based on the risk perception results and the operating characteristic parameters; The standby opportunity cost determination unit is used to determine the standby opportunity cost of the energy storage resource based on the economic characteristic parameters and the coupling risk amplification coefficient. A reserve capacity value multiplier determination unit is used to determine the reserve capacity value multiplier based on the risk perception result. The basic function acquisition unit is used to acquire the basic objective function of the energy storage optimization scheduling model, wherein the basic objective function includes an operating cost item and a standby service cost item. A coefficient adjustment unit is used to construct an opportunity cost penalty item based on the standby cost opportunity, and to adjust the coefficient of the standby service cost item based on the standby capacity value multiplier. The function refactoring unit is used to obtain the adjusted objective function based on the operating cost item, the opportunity cost penalty item, and the adjusted standby service cost item.

Images