Method and device for determining scheduling strategy of new energy system, and electronic equipment

Abstract

This invention discloses a method, apparatus, and electronic device for determining a scheduling strategy for a new energy system. The method includes: acquiring system operation data corresponding to the new energy system; determining multiple scenario data based on the system operation data; retrieving an objective function and objective constraints corresponding to the new energy system, wherein the objective function includes an economic cost sub-function minimizing economic costs and a carbon emission cost sub-function minimizing carbon emissions; the objective constraints include a total cost constraint determined based on a scheduling model, which includes one of the following: a robust model or a chance model; and solving the objective function under the objective constraints based on the system operation data and the multiple scenario data to obtain a scheduling strategy corresponding to the new energy system. This invention solves the technical problem in related technologies where it is difficult to consider variable scenarios when determining a scheduling strategy, leading to inaccurate strategies.

Description

Technical Field

[0001] This invention relates to the field of power grid dispatching, and more specifically, to a method, apparatus, and electronic equipment for determining dispatching strategies for a new energy system. Background Technology

[0002] As the penetration rate of new energy sources such as wind power and solar power in energy systems continues to rise, the operational complexity of these systems has increased significantly. The output of wind and solar power is highly random and volatile due to weather conditions, and coupled with the dynamic changes in electricity and heat load demands, this places higher demands on the adaptability of new energy system dispatch strategies.

[0003] In the current process of determining scheduling strategies for new energy systems, there is a common problem of incomplete scenario coverage. Many rely on scenario data from a single time period, resulting in scheduling strategies that cannot adapt to the ever-changing actual operating scenarios, lacking accuracy and reliability, and failing to meet the needs of efficient operation of new energy systems.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This invention provides a method, apparatus, and electronic device for determining scheduling strategies in a new energy system, which at least solves the technical problem in related technologies where it is difficult to consider changing scenarios when determining scheduling strategies, resulting in inaccurate strategies.

[0006] According to one aspect of the present invention, a method for determining a scheduling strategy for a new energy system is provided, comprising: acquiring system operation data corresponding to the new energy system, wherein the new energy system includes uncertain devices and deterministic devices, the uncertain devices include wind power energy devices, photovoltaic energy devices and load devices, and the deterministic devices include energy storage devices; determining multiple types of scenario data based on the system operation data, wherein the multiple types of scenario data include first-cycle scenario data, second-cycle scenario data and wind-solar related scenario data, the first cycle corresponding to the first-cycle scenario data is longer than the second cycle corresponding to the second-cycle scenario data, and the multiple types of scenario data represent predicted operation data under corresponding scenarios; and retrieving an objective function and objective constraints corresponding to the new energy system. The objective function includes an economic cost subfunction that minimizes economic costs and a carbon emission cost subfunction that minimizes carbon emissions. The objective constraint includes a total cost constraint determined based on a scheduling model. The scheduling model includes one of the following: a robust model and an opportunistic model. The robust model ensures that the total cost does not exceed a cost threshold in the first extreme scenario by maximizing uncertainty tolerance. The opportunistic model ensures that the total cost is lower than a cost threshold in the second extreme scenario by minimizing uncertainty tolerance. The uncertainty tolerance represents a quantitative parameter of the ability to withstand uncertainty factors, including the uncertain equipment. Based on the system operation data and the multi-scenario data, the objective function under the objective constraint is solved to obtain a scheduling strategy corresponding to the new energy system.

[0007] Optionally, based on the system operation data and the multi-scenario data, the objective function under the target constraint is solved to obtain the scheduling strategy corresponding to the new energy system, including: determining a mixed confidence interval, wherein the mixed confidence interval is determined based on a global constraint threshold and an extreme scenario constraint threshold, the global constraint threshold is used to constrain the overall probability distribution deviation of the multi-scenario data, and the extreme scenario constraint threshold is used to constrain the predicted operation data deviation of a single extreme scenario, so as to correct the deviation between the multi-scenario data and the actual operation data; based on the system operation data, the multi-scenario data, and the mixed confidence interval, the objective function under the target constraint is solved to obtain the scheduling strategy corresponding to the new energy system.

[0008] Optionally, based on the system operation data and the multi-scenario data, the objective function under the target constraint is solved to obtain the scheduling strategy corresponding to the new energy system. This includes: when the target constraint also includes a power balance constraint, determining the imbalance confidence interval of the new energy system, wherein the imbalance confidence interval represents the value range of the imbalance, used to quantify the dynamic data range of the system's internal power caused by fluctuation data changes, so as to correct the power balance constraint; the fluctuation data includes new energy output fluctuation data and load demand fluctuation data; the imbalance is determined based on the power supply and load; and based on the system operation data, the multi-scenario data, and the imbalance confidence interval, solving the objective function under the target constraint to obtain the scheduling strategy corresponding to the new energy system.

[0009] Optionally, determining the confidence interval of the imbalance in the new energy system includes: constructing an initial particle swarm based on the system operating data, wherein the initial particle swarm includes candidate values ​​of the imbalance corresponding to each particle; predicting the moment-to-moment particle state of each particle in the initial particle swarm based on the system state equation to obtain a particle state set; updating the initial weights corresponding to each particle in the initial particle swarm based on the particle state set to obtain the target weights corresponding to each particle at the target time; resampling particles whose corresponding target weights are lower than a threshold to obtain an equal-weighted particle swarm; and determining the confidence interval of the imbalance based on the equal-weighted particle swarm.

[0010] Optionally, before retrieving the objective function and objective constraints corresponding to the new energy system, the method further includes: determining the equipment operating parameters corresponding to each of the uncertain devices; determining the standardized parameter matrix corresponding to each of the uncertain devices based on the multiple equipment operating parameters; determining the entropy value corresponding to each of the uncertain devices based on the multiple standardized parameter matrices; determining the weight corresponding to each of the uncertain devices based on the multiple entropy values; determining the equivalent uncertainty based on the multiple weights; and determining the scheduling model based on the equivalent uncertainty.

[0011] Optionally, before determining the multiple types of scenario data based on the system operation data, the method further includes: determining the first edge cumulative distribution function corresponding to the daily wind power output and the second edge cumulative distribution function corresponding to the daily photovoltaic power output; obtaining the wind-solar joint distribution function based on the first edge cumulative distribution function and the second edge cumulative distribution function; and obtaining the wind-solar related scenario data based on the wind-solar joint distribution function.

[0012] Optionally, before determining multiple types of scenario data based on the system operation data, the method further includes: determining the output deviation data between the actual output data and the predicted output data of the new energy source; constructing a normal distribution corresponding to the output deviation fluctuation characteristics of the second cycle based on the output deviation data; determining multiple candidate cycle scenario data based on the normal distribution; and determining multiple second cycle scenario data from the multiple candidate cycle scenario data with the goal of maximizing the silhouette coefficient, wherein the distance between the multiple second cycle scenario data meets the clustering condition of maximizing the inter-class distance and minimizing the intra-class distance.

[0013] According to one aspect of the present invention, a scheduling strategy determination device for a new energy system is provided, comprising: an acquisition module, configured to acquire system operation data corresponding to the new energy system, wherein the new energy system includes uncertain devices and deterministic devices, the uncertain devices including wind power energy devices, photovoltaic energy devices and load devices, and the deterministic devices including energy storage devices; a first determination module, configured to determine multiple types of scenario data based on the system operation data, wherein the multiple types of scenario data include first-cycle scenario data, second-cycle scenario data and wind-solar related scenario data, the first cycle corresponding to the first-cycle scenario data is longer than the second cycle corresponding to the second-cycle scenario data, and the multiple types of scenario data represent predicted operation data under corresponding scenarios; and a retrieval module, configured to retrieve the target function corresponding to the new energy system. The system comprises a number of parameters and objective constraints, wherein the objective function includes an economic cost subfunction that minimizes economic costs and a carbon emission cost subfunction that minimizes carbon emissions. The objective constraints include a total cost constraint determined based on a scheduling model, which includes one of the following: a robust model and an opportunistic model. The robust model ensures that the total cost does not exceed a cost threshold in the first extreme scenario by maximizing uncertainty tolerance, and the opportunistic model ensures that the total cost is lower than a cost threshold in the second extreme scenario by minimizing uncertainty tolerance. The uncertainty tolerance represents a quantitative parameter of the ability to withstand uncertainty factors, including the uncertain equipment. A second determining module is used to solve the objective function under the objective constraints based on the system operation data and the multi-scenario data to obtain a scheduling strategy corresponding to the new energy system.

[0014] According to one aspect of the present invention, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the scheduling strategy determination method for a new energy system as described in any of the preceding embodiments.

[0015] According to one aspect of the present invention, a computer-readable storage medium is provided, wherein when the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is able to perform the scheduling strategy determination method for a new energy system as described above.

[0016] In this embodiment of the invention, system operation data corresponding to a new energy system is obtained. The new energy system includes uncertain devices and deterministic devices. Uncertain devices include wind power equipment, photovoltaic equipment, and load equipment. Deterministic devices include energy storage equipment. Based on the system operation data, multiple types of scenario data are determined. These multiple types of scenario data include first-cycle scenario data, second-cycle scenario data, and wind-solar related scenario data. The first-cycle period corresponding to the first-cycle scenario data is longer than the second-cycle period corresponding to the second-cycle scenario data. The multiple types of scenario data represent predicted operation data under corresponding scenarios. An objective function and objective constraints corresponding to the new energy system are retrieved. The objective function includes... The objective function is calculated by minimizing economic costs and carbon emissions, with the target constraints including the total cost constraint determined by the scheduling model. The scheduling model includes one of the following: a robust model or an opportunistic model. The robust model ensures that the total cost does not exceed the cost threshold in the first extreme scenario by maximizing the uncertainty tolerance, while the opportunistic model ensures that the total cost is below the cost threshold in the second extreme scenario by minimizing the uncertainty tolerance. The uncertainty tolerance is a quantitative parameter representing the ability to withstand uncertainties, including uncertain equipment. Based on system operation data and data from multiple scenarios, the objective function under the target constraints is solved to obtain the scheduling strategy corresponding to the new energy system. By employing a multi-dimensional scenario-layered construction and dual-model collaborative optimization approach, this method generates multi-scenarios data covering the first cycle, the second cycle, and wind-solar related scenarios. It also combines robust and opportunistic models to quantify uncertainty tolerance and construct total cost constraints that adapt to different extreme scenarios. This achieves the goal of accurately adapting to diverse operational scenarios, thereby improving the robustness and economy of scheduling strategies in different scenarios and ensuring the effectiveness of the strategies across all scenarios. This solves the technical problem in related technologies where it is difficult to consider diverse scenarios when determining scheduling strategies, leading to inaccurate strategies. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0018] Figure 1 This is a flowchart of a method for determining the scheduling strategy of a new energy system according to an embodiment of the present invention;

[0019] Figure 2This is an overall technical architecture diagram provided by an optional embodiment of the present invention;

[0020] Figure 3 This is a structural block diagram of a scheduling strategy determination device for a new energy system according to an embodiment of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] Example 1

[0024] According to an embodiment of the present invention, an embodiment of a method for determining a scheduling strategy for a new energy system is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0025] Figure 1 This is a flowchart of a method for determining the scheduling strategy of a new energy system according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0026] Step S102: Obtain system operation data corresponding to the new energy system. The new energy system includes uncertain devices and deterministic devices. Uncertain devices include wind power energy devices, photovoltaic energy devices and load devices. Deterministic devices include energy storage devices.

[0027] Among them, the new energy system refers to an integrated energy system that integrates new energy power generation equipment, energy storage equipment and load equipment to realize the integrated energy production, storage and consumption. It can integrate multiple energy forms to meet load demand, such as park energy systems that include wind power, photovoltaic, energy storage and industrial loads.

[0028] Among them, uncertain equipment refers to equipment whose power or demand has random fluctuation characteristics and cannot be predicted completely accurately. The source of fluctuation in system operation can be clearly identified, such as wind power equipment, photovoltaic equipment, and load equipment.

[0029] Among them, wind power energy equipment refers to power generation equipment that uses wind energy to convert into electrical energy, and can provide clean electricity, such as wind turbine units.

[0030] Photovoltaic energy equipment refers to power generation equipment that converts light energy into electrical energy, enabling efficient utilization of solar energy resources, such as photovoltaic module arrays.

[0031] Load equipment refers to terminal equipment that consumes energy such as electricity and heat, and can reflect the energy demand of the system, such as industrial electrical equipment and residential heating equipment.

[0032] Among them, deterministic equipment refers to equipment whose operating status can be precisely controlled and whose output or charging / discharging power fluctuations can be controlled, thus mitigating the impact of fluctuations in uncertain equipment, such as energy storage equipment.

[0033] Among them, energy storage equipment refers to equipment with the function of storing and releasing electrical energy, which can realize peak shaving and valley filling and power regulation, such as lithium battery energy storage systems.

[0034] Among them, system operation data refers to the set of parameters that characterize the operating status of each device in the new energy system. It can provide data support for subsequent scenario construction and model solving, such as wind power output data, photovoltaic power data, energy storage charging and discharging data, and load demand data.

[0035] In this step, all operating parameters of both uncertain and deterministic devices in the new energy system are collected to form a complete system operation data set.

[0036] This step provides the foundational data that supports the entire scheduling strategy determination process. Subsequent scenario construction and objective function solving rely on the real and comprehensive data from this step, avoiding scenario distortion or model solution deviation caused by missing data, and laying a data foundation for the accuracy of the scheduling strategy.

[0037] Step S104: Based on the system operation data, determine multiple types of scenario data, including first-cycle scenario data, second-cycle scenario data, and wind-and-sun related scenario data. The first cycle corresponding to the first-cycle scenario data is longer than the second cycle corresponding to the second-cycle scenario data. Multiple types of scenario data represent the predicted operation data under the corresponding scenario.

[0038] Among them, multi-scenario data refers to a set of predictive operation data generated based on system operation data, covering different time scales and operating conditions, which can comprehensively characterize the system's variable operating state, such as first-cycle scenario data, second-cycle scenario data, and wind-solar related scenario data.

[0039] The first-cycle scenario data refers to the forecast operation data with a longer time scale, which can be adapted to medium and long-term scheduling needs, such as monthly and quarterly forecast data of new energy output and load demand.

[0040] The second-cycle scenario data refers to the predictive operation data with a shorter time scale, which can be adapted to short-term scheduling needs, such as the 24-hour forecast data of new energy output and load demand within a day.

[0041] Among them, wind-solar related scenario data refers to the predicted operation data that takes into account the correlation between wind power and solar power output, and can reflect the coupled fluctuation characteristics of wind and solar power output, such as scenario data of low wind power output and low solar power output superimposed.

[0042] Among them, predictive operating data refers to the parameters of future equipment operating status derived from historical operating data, which can provide input conditions for the scheduling model, such as the predicted wind power output and load demand values ​​for a certain period of time.

[0043] In this step, based on the system operation data, according to different time periods and the correlation analysis of wind and solar power output, multiple types of scenario data are generated, including medium and long term, intraday and coupled operating conditions. Each type of scenario data corresponds to a set of predicted operating parameters.

[0044] This step addresses the limitation of traditional scheduling strategies relying on single-scenario data. Multi-scenario data covers both normal and extreme operating conditions of the system, enabling subsequent scheduling models to solve problems under diverse operating conditions. This avoids insufficient strategy adaptability due to a single scenario and improves the scheduling strategy's ability to cover varied operating scenarios.

[0045] Step S106: Retrieve the objective function and objective constraints corresponding to the new energy system. The objective function includes an economic cost subfunction that minimizes economic costs and a carbon emission cost subfunction that minimizes carbon emissions. The objective constraints include the total cost constraint determined based on the scheduling model. The scheduling model includes one of the following: a robust model and an opportunistic model. The robust model ensures that the total cost does not exceed the cost threshold in the first extreme scenario by maximizing the uncertainty tolerance. The opportunistic model ensures that the total cost is lower than the cost threshold in the second extreme scenario by minimizing the uncertainty tolerance. The uncertainty tolerance represents a quantitative parameter of the ability to withstand uncertainty factors, including uncertain equipment.

[0046] The objective function refers to the mathematical function used to characterize the optimization objective of the new energy system scheduling. It can clarify the optimization direction of the scheduling strategy, such as a composite objective function that includes economic cost sub-functions and carbon emission cost sub-functions.

[0047] Among them, the economic cost sub-function refers to a mathematical expression that aims to minimize the economic cost of operating a new energy system. It can quantify the economics of scheduling strategies, such as a cost calculation function that includes electricity purchase costs and operation and maintenance costs.

[0048] Among them, the carbon emission cost sub-function refers to a mathematical expression aimed at minimizing the total carbon emissions of the new energy system, which can achieve low-carbon scheduling, such as the carbon emission cost function calculated based on carbon quotas.

[0049] Among them, the objective constraint refers to the set of conditions that limit the scope of the scheduling strategy solution and can ensure the feasibility of the scheduling strategy, such as the total cost constraint.

[0050] Among them, total cost constraint refers to the constraint condition that limits the upper or lower limit of the total operating cost of a new energy system, which can balance economy and stability, such as the total cost threshold constraint in extreme scenarios.

[0051] Among them, the scheduling model refers to the mathematical model used to solve the scheduling strategy, which can adapt to the optimization needs of different operating scenarios, such as the robust model and the opportunistic model.

[0052] Robust models, in particular, refer to scheduling optimization models that focus on handling extremely unfavorable scenarios. They ensure the system's operational stability under extreme conditions, such as models that guarantee the total cost in the first extreme scenario does not exceed a threshold. The first extreme scenario can be an extremely unfavorable operating scenario, representing the worst-case scenario where uncertainties and equipment fluctuations overlap in the renewable energy system. Robust models are needed to maximize uncertainty tolerance and ensure the total cost does not exceed a threshold. Extremely unfavorable operating scenarios, such as when renewable energy output is significantly lower than expected and load demand is significantly higher than expected, can verify the risk resilience of scheduling strategies, for example, in the scenario of a winter cold wave coinciding with a peak evening rush hour.

[0053] Among them, the opportunistic model refers to a scheduling optimization model that focuses on tapping into the benefits of favorable scenarios. It can improve the economic efficiency of the system under optimal operating conditions, such as a model that ensures the total cost of the second extreme scenario is below a threshold. The second extreme scenario can be an extremely favorable operating scenario, representing the optimal operating condition for the coordination of uncertain equipment fluctuations in the renewable energy system. The opportunistic model needs to minimize the uncertainty tolerance to ensure the total cost is below the threshold. For example, in an extremely favorable operating condition where renewable energy output is significantly higher than expected and load demand is significantly lower than expected, the economic potential of the scheduling strategy can be explored. Another example is the operating scenario of high wind and solar activity combined with low load during midday in summer.

[0054] Uncertainty tolerance refers to a quantitative parameter that characterizes the system's ability to withstand uncertain equipment fluctuations. It can quantify the system's ability to resist fluctuations, for example, a tolerance parameter with a value of 0.05 to 0.12.

[0055] Among them, the cost threshold refers to the critical value that limits the total operating cost of the system. It can provide a clear optimization boundary for the scheduling model, such as the upper limit threshold of the cost in the first extreme scenario.

[0056] Uncertainty factors refer to factors that cause fluctuations in the system's operating status. These factors can clearly identify the sources of fluctuations that the model needs to address, such as fluctuations in wind power and photovoltaic output, as well as fluctuations in load demand.

[0057] In this step, the preset composite objective function and total cost constraints are retrieved, and the optimization logic of the robust model and the opportunistic model is clarified. The robust model copes with the first extreme scenario by maximizing the uncertainty tolerance, while the opportunistic model explores the benefits of the second extreme scenario by minimizing the uncertainty tolerance, thus constructing a scheduling model that adapts to different extreme conditions.

[0058] This step solves the problem that traditional single scheduling models cannot balance stability and economy. The dual-model architecture of robust model and chance model can deal with extremely unfavorable and extremely favorable scenarios respectively. The quantitative design of uncertainty tolerance makes the model more accurate in dealing with uncertainty factors, providing model support for solving scheduling strategies that balance robustness and economy.

[0059] Step S108: Based on system operation data and multi-scenario data, solve the objective function under the objective constraint to obtain the scheduling strategy corresponding to the new energy system.

[0060] Among them, scheduling strategy refers to the scheme obtained based on the solution of the objective function, which guides the operation of each device in the new energy system and can achieve the optimized operation of the system, such as the charging and discharging sequence of energy storage devices and the scheme for absorbing new energy output.

[0061] In this step, system operation data and multi-scenario data are input into the scheduling model. Under the boundary conditions of total cost constraints, a function with the objective of minimizing economic cost and carbon emission cost is solved, and finally a new energy system scheduling strategy adapted to multiple scenarios is generated.

[0062] This step combines basic data with multiple operating scenarios to solve the scheduling strategy. Due to the comprehensiveness of the input data and the adaptability of the model architecture, the solved scheduling strategy can cover different time periods and operating conditions. It can not only cope with extreme scenarios to ensure system stability, but also explore favorable scenarios to improve economic benefits. It effectively solves the technical problems of traditional scheduling strategies being unable to consider multiple changing scenarios and lacking accuracy.

[0063] Through steps S102-S108 above, system operation data corresponding to the new energy system is obtained. The new energy system includes uncertain and deterministic devices. Uncertain devices include wind power equipment, photovoltaic equipment, and load equipment. Deterministic devices include energy storage equipment. Based on the system operation data, multiple scenario data are determined. These multiple scenario data include first-cycle scenario data, second-cycle scenario data, and wind-solar related scenario data. The first-cycle period corresponding to the first-cycle scenario data is longer than the second-cycle period corresponding to the second-cycle scenario data. The multiple scenario data represent the predicted operation data under the corresponding scenario. The objective function and objective constraints corresponding to the new energy system are retrieved. The function includes an economic cost subfunction that minimizes economic costs and a carbon emission cost subfunction that minimizes carbon emissions. The objective constraint includes a total cost constraint determined based on the scheduling model. The scheduling model includes one of the following: a robust model and an opportunistic model. The robust model ensures that the total cost does not exceed the cost threshold in the first extreme scenario by maximizing the uncertainty tolerance, while the opportunistic model ensures that the total cost is below the cost threshold in the second extreme scenario by minimizing the uncertainty tolerance. The uncertainty tolerance represents a quantitative parameter of the ability to withstand uncertainty factors, including uncertain equipment. Based on system operation data and data from multiple scenarios, the objective function under the objective constraint is solved to obtain the scheduling strategy corresponding to the new energy system. By employing a multi-dimensional scenario-layered construction and dual-model collaborative optimization approach, this method generates multi-scenarios data covering the first cycle, the second cycle, and wind-solar related scenarios. It also combines robust and opportunistic models to quantify uncertainty tolerance and construct total cost constraints that adapt to different extreme scenarios. This achieves the goal of accurately adapting to diverse operational scenarios, thereby improving the robustness and economy of scheduling strategies in different scenarios and ensuring the effectiveness of the strategies across all scenarios. This solves the technical problem in related technologies where it is difficult to consider diverse scenarios when determining scheduling strategies, leading to inaccurate strategies.

[0064] As an optional embodiment, based on system operation data and multi-scenario data, an objective function under target constraints is solved to obtain a scheduling strategy corresponding to the new energy system. This includes: determining a mixed confidence interval, wherein the mixed confidence interval is determined based on a global constraint threshold and an extreme scenario constraint threshold. The global constraint threshold is used to constrain the overall probability distribution deviation of multi-scenario data, and the extreme scenario constraint threshold is used to constrain the predicted operation data deviation of a single extreme scenario, so as to correct the deviation between multi-scenario data and actual operation data; and based on system operation data, multi-scenario data, and the mixed confidence interval, an objective function under target constraints is solved to obtain a scheduling strategy corresponding to the new energy system.

[0065] Among them, the mixed confidence interval refers to the composite deviation constraint system constructed based on the global constraint threshold and the extreme scenario constraint threshold. It can accurately quantify the fitting deviation between the scenario data and the actual running data, and provide a reliable constraint basis for solving the objective function. For example, the 95% confidence level interval constructed based on the 1-norm and the ∞-norm.

[0066] Among them, the global constraint threshold refers to the critical value used to control the overall fluctuation of data in multiple scenarios. It can ensure that the statistical characteristics of the scenario set are consistent with the actual operating rules and avoid overall distortion of the entire scenario. For example, it can constrain the upper limit of the probability distribution of all scenarios from the benchmark value.

[0067] Among them, the extreme scenario constraint threshold refers to the deviation control threshold set for a single extreme scenario, which can limit the prediction data deviation of the first and second extreme scenarios and prevent extreme scenarios from excessively dominating the optimization results. For example, it can constrain the load prediction deviation of low wind and solar power output scenarios to not exceed 5%.

[0068] Among them, the overall probability distribution deviation refers to the degree of difference between the statistical distribution of data from multiple scenarios and the statistical distribution of actual operating data. It can reflect the overall fit of the scenario set, such as the deviation between the average wind power output of all scenarios and the actual historical average.

[0069] Among them, a single extreme scenario refers to an extremely favorable or extremely unfavorable operating condition of a new energy system, which can specifically verify the adaptability of the scheduling strategy, such as the first extreme scenario of low wind and solar output combined with high load in winter.

[0070] Among them, the predicted operating data deviation refers to the difference between the predicted operating data and the actual operating data under extreme scenarios. It can accurately locate the local distortion problem of scenario data, such as the difference between the predicted output and the actual output of photovoltaic power under a certain extreme scenario.

[0071] Among them, actual operating data refers to the parameter data generated by the actual operation of each device in the new energy system, which can serve as a benchmark for correcting scenario data, such as measured wind power output and load demand data.

[0072] In this embodiment, a hybrid confidence interval is constructed by using global constraint thresholds and extreme scenario constraint thresholds. The deviations between multi-scenario data and actual operating data are corrected from both the overall and local levels. Then, by combining system operating data and the corrected multi-scenario data, the objective function is solved under the objective constraints, and finally, the new energy system scheduling strategy is obtained.

[0073] This approach solves the problem that a single constraint cannot simultaneously ensure the overall fit of the scenario and the accuracy of extreme scenarios. The mixed confidence interval not only guarantees the statistical reliability of the entire scenario, but also controls the local deviation of extreme scenarios, making the scenario data input to the objective function closer to the real operating conditions. This improves the accuracy of the scheduling strategy solution and ensures that the strategy can balance economy and robustness in all scenarios.

[0074] As an optional implementation, based on system operation data and various scenario data, an objective function under target constraints is solved to obtain a scheduling strategy corresponding to the new energy system. This includes: when the target constraints also include power balance constraints, determining the confidence interval of the imbalance quantity of the new energy system, where the confidence interval of the imbalance quantity represents the range of values ​​of the imbalance quantity, used to quantify the dynamic data range of internal power caused by fluctuation data changes, so as to correct the power balance constraints. The fluctuation data includes fluctuation data of new energy output and fluctuation data of load demand, and the imbalance quantity is determined based on the power supply and load. Based on system operation data, various scenario data, and the confidence interval of the imbalance quantity, an objective function under target constraints is solved to obtain a scheduling strategy corresponding to the new energy system.

[0075] Among them, power balance constraints refer to the constraints that ensure the real-time matching of power supply and load in the new energy system. This can maintain the stability of system operation and avoid equipment failure caused by power imbalance. For example, it requires that the difference between the real-time power supply and the total load be controlled within the allowable range.

[0076] The confidence interval for imbalance refers to the range of values ​​for imbalance. It can quantify the dynamic range of internal power caused by fluctuations in data, providing a precise basis for correcting power balance constraints. For example, the power imbalance interval of [-15kW, 20kW] at a 95% confidence level.

[0077] Among them, the imbalance refers to the difference between the power supply and the load of the system. It can intuitively reflect the power balance status and provide a core quantitative indicator for the construction of confidence intervals. For example, when the power supply is 800kW and the load is 820kW, the imbalance is -20kW.

[0078] Among them, fluctuation data refers to the set of parameters that cause system power fluctuations. It can identify the source of power imbalance and provide support for imbalance analysis, such as fluctuation data of new energy output and fluctuation data of load demand.

[0079] Among them, the data on fluctuations in new energy output refers to the deviation between the actual output and the predicted output of new energy equipment such as wind power and photovoltaics. It can reflect the uncertainty of new energy output, such as the fluctuation data of wind power output dropping sharply by 50% compared with the predicted value.

[0080] Among them, load demand fluctuation data refers to the deviation between the actual demand and the predicted demand of various load devices, which can reflect the uncertainty on the load side, such as the fluctuation data of residential heating load increasing by 30% compared with the predicted value.

[0081] Among them, power supply refers to the total electrical energy provided by the new energy system to the load, which can characterize the system's power supply capacity, such as the total electrical energy released by wind power, photovoltaic power and energy storage.

[0082] Among them, load capacity refers to the total electrical energy consumed by various types of load equipment, which can characterize the scale of system energy demand, such as the total electrical energy consumption of industrial load and residential load.

[0083] In this embodiment, when the target constraint includes a power balance constraint, the imbalance is first determined based on the power supply and load, and then an imbalance confidence interval is constructed. This interval is used to quantify the dynamic range of internal power caused by fluctuations in renewable energy output and load demand, and the power balance constraint is corrected. Then, the objective function is solved under the target constraint by combining system operation data, multi-scenario data and the confidence interval to obtain the renewable energy system scheduling strategy.

[0084] This approach overcomes the shortcomings of traditional power balance constraints in adapting to fluctuating data. The confidence interval of the imbalance quantity accurately quantifies the range of power imbalance caused by fluctuations, making the corrected power balance constraints more closely match actual operating conditions. This effectively reduces the constraint deviation caused by power fluctuations, thereby improving the reliability of the scheduling strategy and ensuring that the system can maintain power balance under fluctuating conditions, thus balancing operational stability and economy.

[0085] As an optional embodiment, determining the confidence interval of the imbalance in a new energy system includes: constructing an initial particle swarm based on system operating data, wherein the initial particle swarm includes candidate values ​​of the imbalance corresponding to each particle; predicting the moment-state of each particle in the initial particle swarm based on the system state equation to obtain a set of particle states; updating and initializing the initial weights corresponding to each particle in the initial particle swarm based on the set of particle states to obtain the target weights corresponding to each particle at the target time; resampling particles whose corresponding target weights are lower than a threshold to obtain an equal-weighted particle swarm; and determining the confidence interval of the imbalance based on the equal-weighted particle swarm.

[0086] The initial particle swarm refers to an initial set of multiple particles constructed based on system operation data, which can provide basic candidate samples for the calculation of the confidence interval of the imbalance quantity. For example, a particle swarm containing 200 particles with values ​​covering the imbalance quantity range of [-30kW, 30kW].

[0087] Among them, a particle refers to the basic unit that represents the candidate values ​​of the imbalance quantity, and can carry the possible values ​​and weight information of the imbalance quantity, such as a single particle corresponding to an imbalance quantity of -10kW or 5kW.

[0088] Among them, the candidate values ​​of the imbalance quantity refer to the possible values ​​of the imbalance quantity corresponding to the particle, which can cover the potential power imbalance range of the system, such as random values ​​within [-25kW, 25kW] set based on historical data.

[0089] Among them, the system state equation refers to the mathematical equation that describes the law of change of the unbalance quantity over time. It can accurately predict the state of the particle at the next moment, such as the dynamic change equation derived based on "unbalance quantity = power supply - load quantity".

[0090] Among them, the particle state at time point refers to the predicted value of the unbalance quantity corresponding to a single particle at a specific time point, which can reflect the dynamic change trend of the unbalance quantity. For example, the predicted value of the unbalance quantity of a certain particle at time t=10 is 8kW.

[0091] Among them, the particle state set refers to the summary of the states of all particles at the target time, which can fully cover the potential dynamic values ​​of the imbalance quantity, such as the set of predicted imbalance quantities of 200 particles at each time.

[0092] The initial weight refers to the weight value assigned to a particle during initialization, which can characterize the reliability of the corresponding value of the particle. For example, the initial weight of all particles is an equal weight allocation of 1 / 200.

[0093] The target time refers to a specific time node for particle state prediction and weight update, which can focus on the analysis of imbalances during a certain period, such as the 12th hour of a 24-hour day.

[0094] Among them, the target weight refers to the weight value of the particle after state prediction and observation update, which can accurately filter particles with high credibility. For example, after the update, the weight of a certain particle increases to 0.02 and the weight of another particle decreases to 0.001.

[0095] The threshold refers to the weight critical value used to filter effective particles, which can eliminate particles with low credibility and optimize the quality of the particle swarm. For example, the weight threshold is set to 0.003.

[0096] Resampling refers to the operation of removing particles with weights below a threshold and supplementing them with valid particles, which can improve the overall credibility of the particle swarm. For example, after removing low-weight particles, new particles can be generated based on high-weight particles.

[0097] Among them, the equal-weighted particle swarm refers to the set of particles in which all particles have equal weights after resampling. It can simplify the calculation process of confidence intervals. For example, a particle swarm in which 200 particles have a weight of 1 / 200 after resampling.

[0098] In this embodiment, an initial particle swarm containing candidate values ​​of imbalance is constructed based on system operation data. The time state of each particle is predicted by the system state equation to form a particle state set. The initial weight of each particle is updated according to the set to obtain the target weight. Particles with target weights lower than the threshold are removed and resampled to obtain an equal-weighted particle swarm. Finally, the confidence interval of imbalance is calculated based on the equal-weighted particle swarm.

[0099] This approach, relying on the sampling, updating, and resampling logic of particle filtering, solves the problem that traditional methods struggle to accurately quantify the range of dynamic imbalance. The resampling operation eliminates low-confidence particles, and the equal-weighted particle swarm ensures the objectivity of the confidence interval calculation, making the obtained imbalance confidence interval more closely match the actual fluctuation pattern of the system. This provides high-quality data support for the accurate correction of power balance constraints and further enhances the adaptability of scheduling strategies to power fluctuations.

[0100] As an optional embodiment, before retrieving the objective function and objective constraints corresponding to the new energy system, the method further includes: determining the equipment operating parameters corresponding to each uncertain device; determining the standardized parameter matrix corresponding to each uncertain device based on the multiple equipment operating parameters; determining the entropy value corresponding to each uncertain device based on the multiple standardized parameter matrices; determining the weight corresponding to each uncertain device based on the multiple entropy values; determining the equivalent uncertainty based on the multiple weights; and determining the scheduling model based on the equivalent uncertainty.

[0101] Among them, equipment operating parameters refer to the core quantitative indicators that characterize the uncertain operating status of equipment. They can reflect the output or demand fluctuation characteristics of the equipment and provide basic data for subsequent uncertainty analysis, such as wind power output, output value corresponding to photovoltaic irradiance, and real-time load demand.

[0102] The standardized parameter matrix refers to the set of matrices formed after normalizing the operating parameters of equipment with different dimensions and magnitudes. It can eliminate the influence of differences in dimensions on the calculation results and realize fair comparison and analysis of multiple parameters. For example, a 5×24-order standardized matrix composed of 24-hour operating parameters of 5 types of uncertain equipment.

[0103] Among them, entropy value refers to a quantitative index that characterizes the degree of dispersion of parameter fluctuations, calculated based on a standardized parameter matrix. It can objectively reflect the degree of operational fluctuation of a single type of uncertain equipment. For example, the entropy value of wind power output parameter is 0.75 and the entropy value of load demand parameter is 0.62.

[0104] Among them, the weight refers to the proportion coefficient calculated based on the entropy value, which represents the degree of influence of a single type of uncertain device on the overall uncertainty of the system. It can objectively allocate the uncertainty contribution of different devices and avoid subjective assignment bias. For example, the weight of wind power equipment is 0.32 and the weight of photovoltaic equipment is 0.25.

[0105] Among them, equivalent uncertainty refers to the normalized index calculated by combining the weights of each uncertain device and the fluctuation characteristics. It can integrate multi-dimensional and multi-type uncertainty factors into a single quantitative value, providing a unified uncertainty measurement standard for the construction of scheduling models. For example, the equivalent uncertainty is 0.12 (the value ranges from 0 to 1, and the larger the value, the higher the system uncertainty).

[0106] In this embodiment, the operating parameters of each type of uncertain device are first determined, and a standardized parameter matrix is ​​constructed based on these parameters. The entropy value of each type of device is calculated through the matrix. Then, the weight of the corresponding device is determined based on the entropy value. The equivalent uncertainty of the system is calculated by combining the weight and the fluctuation characteristics of the device. Finally, the appropriate scheduling model is determined based on the quantification result of the equivalent uncertainty.

[0107] This approach solves the problems of subjective quantification of uncertainties and difficulty in integrating multi-dimensional uncertainties in the construction of traditional scheduling models. The application of the entropy method ensures the objectivity of weight allocation, and the equivalent uncertainty achieves the normalization and integration of uncertainties from multiple devices. This allows the uncertainty tolerance setting of the scheduling model to better match the actual operating state of the system, thereby improving the adaptability of the scheduling model to different scenarios and laying a reliable model foundation for the accurate solution of the objective function.

[0108] As an optional embodiment, before determining the multi-type scenario data based on system operation data, the method further includes: determining the first edge cumulative distribution function corresponding to the daily output of wind power and the second edge cumulative distribution function corresponding to the daily output of photovoltaic power; obtaining the wind-solar joint distribution function based on the first edge cumulative distribution function and the second edge cumulative distribution function; and obtaining the wind-solar related scenario data based on the wind-solar joint distribution function.

[0109] Among them, wind power daily output refers to the actual output power of wind power energy equipment at different times of a day (such as 24 hours). It can reflect the daily fluctuation pattern of wind power and provide basic data for the construction of distribution functions, such as the wind power output sequence from 6:00 to 18:00 during the day and the data of low output period in the early morning.

[0110] The first marginal cumulative distribution function refers to a mathematical function that describes the probability that the daily output of wind power is less than or equal to a certain value. It can quantify the cumulative probability characteristics of the daily output of wind power, such as a distribution function in which the probability of wind power output being ≤80kW at a certain moment is 75%.

[0111] Among them, photovoltaic daily output refers to the actual output power of photovoltaic energy equipment at different times of the day, which can reflect the intraday characteristics of photovoltaics as light intensity changes, such as the rated output data of photovoltaics during the midday high light intensity period and the output data during the morning and evening low light intensity periods.

[0112] The second marginal cumulative distribution function is a mathematical function that describes the probability that the photovoltaic output value during the day is less than or equal to a certain value. It can accurately depict the probability distribution law of photovoltaic output during the day. For example, the distribution function that the probability of photovoltaic output ≤100kW at a certain moment is 80%.

[0113] Among them, the wind-solar joint distribution function refers to the joint mathematical function constructed based on the edge cumulative distribution functions of wind power and photovoltaic power, which describes the probability of coordinated fluctuations in the output of the two. It can capture the coupling correlation of wind and solar power output, such as the joint probability function of low wind power output and low photovoltaic power output superimposed.

[0114] In this embodiment, the first and second edge cumulative distribution functions corresponding to the daily output of wind power and solar power are first determined respectively. The wind-solar joint distribution function is obtained through the coupling analysis of the two edge distribution functions. Then, based on the joint distribution function, wind-solar related scene data that can reflect the coordinated fluctuation characteristics of wind and solar power output is generated, providing support for the construction of subsequent multi-type scene data.

[0115] This approach addresses the shortcomings of traditional scenario construction, which neglects the correlation between wind and solar output and only separately characterizes the fluctuations of the two energy types. The joint wind-solar distribution function accurately captures the coupling patterns of their outputs (such as low simultaneous wind and solar output on cloudy days and high wind and solar output superposition on sunny afternoons), making the generated wind-solar correlated scenario data closer to actual operating conditions. This avoids scenario distortion caused by severing the correlation between wind and solar, thereby improving the comprehensiveness and reliability of multi-scenario data and providing accurate data support for scheduling strategies to adapt to wind-solar coordinated fluctuation scenarios.

[0116] As an optional embodiment, before determining multiple types of scenario data based on system operation data, the method further includes: determining the output deviation data between the actual output data and the predicted output data of new energy; constructing a normal distribution corresponding to the output deviation fluctuation characteristics of the second cycle based on the output deviation data; determining multiple candidate cycle scenario data based on the normal distribution; and determining multiple second cycle scenario data from the multiple candidate cycle scenario data with the goal of maximizing the silhouette coefficient, wherein the distance between the multiple second cycle scenario data meets the clustering condition of maximizing the inter-class distance and minimizing the intra-class distance.

[0117] Among them, the actual output data of new energy refers to the output power / electricity data obtained by new energy equipment during operation. It can reflect the real operating status of the equipment and provide a benchmark for deviation calculation. For example, the measured output value of wind turbine units per hour and the real-time power generation data of photovoltaic module arrays.

[0118] Among them, the predicted output data refers to the output power / electricity data of new energy equipment estimated based on meteorological conditions, historical data, etc., which can be used as a benchmark for deviation analysis. For example, the intraday predicted output sequence of photovoltaic and wind power is derived based on sunrise and sunset times and wind speed forecasts.

[0119] Among them, the output deviation data refers to the difference between the actual output data of new energy and the predicted output data. It can quantify the degree of error in output prediction and provide core fluctuation indicators for the construction of short-cycle scenarios. For example, if the actual output of wind power is 80kW and the predicted output is 100kW in a certain period, the corresponding output deviation data is -20kW.

[0120] Among them, the output deviation fluctuation characteristics of the second cycle refer to the variation pattern and distribution characteristics of the output deviation within the second cycle (such as 24 hours within a day). It can accurately adapt to the deviation analysis needs of short cycle scheduling, such as the fluctuation pattern of large deviation in the morning and evening and small deviation in the midday period.

[0121] Among them, the normal distribution refers to a continuous probability distribution model that is constructed based on the output deviation data and characterizes the probability distribution law of the deviation. It can quantify the random fluctuation characteristics of the deviation, such as a normal distribution (N(0,σ²)) with a mean of 0 and a variance of σ².

[0122] Among them, candidate cycle scenario data refers to a set of scenario samples randomly generated based on a normal distribution, covering various output deviation conditions in the second cycle. It can comprehensively cover potential short-cycle deviation conditions, such as containing 50 intraday short-cycle scenario samples with different deviation levels.

[0123] The silhouette coefficient is a quantitative indicator used to evaluate the quality of clustering. Its value ranges from -1 to 1. It can accurately determine the rationality of scene clustering. For example, a silhouette coefficient of 0.8 indicates that the clustering effect is excellent (the closer to 1, the better the clustering effect).

[0124] Inter-class distance refers to the distance between different clusters (second-period scene groups), which can characterize the degree of difference between different scene clusters, such as the distance between two short-period scene clusters calculated by Euclidean distance.

[0125] Among them, intra-cluster distance refers to the distance between candidate periodic scene data within the same cluster, which can characterize the consistency of scenes within the cluster, such as the average distance of all short-period scene deviation data within the same cluster.

[0126] Among them, the clustering condition refers to the scenario selection criterion of "maximizing the distance between clusters and minimizing the distance within clusters". It can ensure that the selected scenarios are both significantly different and internally consistent. For example, based on this condition, three types of short-cycle scenarios with high, medium and low deviations can be selected.

[0127] In this embodiment, the output deviation data between the actual output data and the predicted output data of the new energy source is first calculated. Then, a normal distribution is constructed based on the output deviation fluctuation characteristics of the second cycle. Based on this normal distribution, multiple candidate cycle scenario data are generated. Then, with the goal of maximizing the silhouette coefficient, multiple second cycle scenario data are selected from the candidate data according to the clustering conditions of maximizing the inter-class distance and minimizing the intra-class distance.

[0128] This approach addresses the issues of incomplete coverage of output deviations, scene redundancy, or insufficient representativeness in traditional second-cycle scenario construction. The normal distribution accurately quantifies the probabilistic characteristics of short-cycle output deviations, ensuring that candidate scenarios cover various deviation conditions. The combination of the silhouette coefficient and clustering conditions filters out scenarios with significant differences and strong internal consistency, eliminating redundant samples. This makes the obtained second-cycle scenario data more closely aligned with short-cycle fluctuation patterns, providing high-quality data support for subsequent adaptation to intraday and other short-cycle scheduling needs, and further improving the accuracy of scheduling strategies in adapting to short-cycle fluctuation scenarios.

[0129] Based on the above embodiments and optional embodiments, an optional implementation method is provided, which is described in detail below.

[0130] In related technologies, with the introduction of dual-carbon targets, the proportion of new energy sources such as wind power and photovoltaics in integrated energy systems (IES) has been increasing year by year. The randomness and volatility of new energy output, coupled with the uncertainty of load demand, leads to three major challenges for IES operation: unstable power supply, insufficient reliability, uneconomical costs, and mismatched market transactions. Currently, there are three main shortcomings in handling uncertainty:

[0131] 1) Unclear classification of uncertainties makes it difficult to adapt to multi-level electricity market transactions. Most studies simply treat renewable energy uncertainty as a problem of single power fluctuations, without distinguishing between medium- and long-term power uncertainty (fluctuations in total renewable energy generation in quarters and months, making it difficult to sign medium- and long-term contracts) and short-term power uncertainty (intraday renewable energy output forecast errors, leading to difficulties in intraday spot clearing). Traditional stochastic optimization models short-term intraday fluctuations, which are difficult to support annual (quarterly) electricity trading; medium- and long-term planning ignores short-term intraday power deviations, causing medium- and long-term contract electricity to decouple from short-term actual output, increasing the risk of market default (e.g., a renewable energy base's quarterly contract fulfillment rate was only 82% due to ignoring medium- and long-term power fluctuations).

[0132] 2) Difficult to generate, inefficient, and lacking in key relevance and adaptability. Lack of correlation between wind and solar output: Traditional scenario generation (independent Monte Carlo sampling) assumes that wind and solar outputs are independent, ignoring the statistical correlation of wind and solar resources within the same area (low wind speed during strong daytime sunlight and vice versa at night), resulting in low alignment between scenarios and reality. For example, scenario prediction for an industrial park showed a 15% probability of simultaneous high wind and solar output, but the actual probability was only 5%, leading to over-configuration of dispatch reserve capacity and a 12% increase in cost. Numerous sampling and clustering sample points: Random sampling (simple random sampling) yields many scenario combinations, requiring the generation of tens of thousands of scenario combinations to ensure accuracy. Time-consuming and inefficient: k-means clustering relies on subjectively determining the number of clusters (referring to a fixed k=5). When source-load characteristics change (seasonal changes), over-clustering or scenario distortion can easily occur, making it difficult to achieve a balance between accuracy and efficiency.

[0133] 3) Subjective quantification of multi-factor uncertainty weights makes it difficult to balance robustness and complexity in optimization models. Subjective weight assignment: Multi-factor uncertainties (wind, solar, electricity, heat and gas loads) are often weighted using expert scoring methods, subjectively determining the uncertainty weight of wind power to be 0.4 and solar power to be 0.3, ignoring the dynamic changes in the proportion of new energy sources and load types in actual operation. In a certain park, the proportion of wind power increased from 30% to 50%, but the subjective weight remained unchanged, resulting in insufficient conservatism in the dispatch strategy and an 8% increase in the wind curtailment rate. Poor applicability: Information gap decision theory (IGDT) only supports single-norm decision-making, which has weak applicability and is difficult to handle multi-norm (multi-factor coupling). Distributed robust optimization (DRO) takes into account both robustness and stochasticity, but often uses single-norm to construct confidence intervals, which can easily lead to excessive robustness and high cost or excessive stochasticity and low reliability. In addition, the scenario reduction does not use distance metric optimization, resulting in increased complexity (in one example, the DRO model took 2.5 hours to calculate, which is difficult to meet the intraday dispatch timeliness requirements).

[0134] 4) The accuracy of plant power imbalance sensing is low, and there is a lack of understanding of multi-load uncertainties. Plant power imbalance analysis often uses a simple judgment of measured values ​​minus predicted values, without considering the coupling of uncertainties among multiple types of loads such as electricity, heat, and gas, and without taking into account the indirect impact of temperature fluctuations caused by heat load inertia on electric loads (electric boilers). Furthermore, the lack of dynamic sensing algorithms (particle filtering) results in an imbalance state confidence level error exceeding 20%, making it difficult to support accurate energy storage configuration and load regulation.

[0135] In view of this, the optional embodiments of the present invention provide a comprehensive energy system scheduling method for accurate characterization and improved optimization of multi-dimensional uncertainties, especially for regional and park-level energy systems (IES) with high proportions of renewable energy grid connection. This method is adaptable to medium- and long-term contracts, day-ahead spot markets, and ancillary services in the electricity market, and provides uncertainty handling methods for classification characterization, scenario optimization, and decision adaptation of multi-level transactions. Figure 2This is an overall technical architecture diagram provided by an optional embodiment of the present invention, such as... Figure 2 As shown, it will be introduced below.

[0136] 1) Multidimensional uncertainty classification and characterization: Distinguish between medium- and long-term power uncertainty and short-term power uncertainty, and use probability scenario sets and normal distribution + scenario reduction processing respectively;

[0137] 2) High-precision scene generation: It integrates Frank-Copula correlation modeling, Latin hypercube sampling, and improved K-means clustering to achieve full scene coverage, high precision, and low redundancy;

[0138] 3) Improved uncertainty optimization: including improved IGDT dual model (adapted to multiple risk preferences), multi-norm DRO (balancing robustness and complexity), particle filter state perception (quantifying plant power imbalance), and outputting a dispatch strategy that matches the short-term contracts signed in the power market.

[0139] The specific steps are as follows:

[0140] S1, Obtain system operation data corresponding to the new energy system, wherein the new energy system includes uncertain equipment and deterministic equipment. Uncertain equipment includes wind power energy equipment, photovoltaic energy equipment and load equipment, and deterministic equipment includes energy storage equipment.

[0141] S2. Based on the system operation data, determine multiple types of scenario data, including first-cycle scenario data, second-cycle scenario data, and wind-and-sun related scenario data. The first cycle corresponding to the first-cycle scenario data is longer than the second cycle corresponding to the second-cycle scenario data. Multiple types of scenario data represent the predicted operation data under the corresponding scenario.

[0142] For the first cycle scenario data, the uncertainty of medium- and long-term electricity volume is considered, which is also called the probabilistic scenario set representation (matching medium- and long-term contract transactions).

[0143] To address the fluctuations in total renewable energy generation quarterly / monthly (affecting the signing of medium- and long-term contracts), a representation model is established using historical data-driven approaches, probability distribution fitting, and scenario set compression. This model includes, but is not limited to, the following:

[0144] 1) Basic Data Collection and Distribution Fitting: Collect daily renewable energy power generation data for the target area for three consecutive years (or longer), and use Weibull distribution (wind power) and beta distribution (photovoltaic) to fit the medium- and long-term power probability distribution.

[0145] Formula for daily wind power generation distribution:

[0146]

[0147] Where k is the shape parameter (representing the concentration of power generation fluctuations, with a value of 1.8 to 3.2), c is the scale parameter (representing the average power generation, with a value of 50 to 200 MWh / d), and x represents the daily power generation (unit: MWh / d).

[0148] Formula for daily photovoltaic power generation distribution:

[0149]

[0150] Where α and β are shape parameters (estimated by maximum likelihood estimation from historical data, with α ranging from 2.5 to 4.0 and β ranging from 3.0 to 5.0), and y is the standardized daily power generation (0 to 1).

[0151] 2) Generation and compression of probability scenario sets: Based on the above distribution, 1000-2000 sets of medium- and long-term power scenarios are generated. The Kantorovich distance scenario reduction method (minimizing the energy distance between scenarios) is used to compress the scenarios into 10-20 typical scenarios (high power scenario, medium power scenario, and low power scenario), and the probability of occurrence of each scenario is output (0.2 for high power scenario, 0.6 for medium power scenario, and 0.2 for low power scenario), supporting the signing of medium- and long-term contracts.

[0152] For the second-cycle scenario data, considering short-term power uncertainty, a normal distribution and scenario reduction representation (adapted to day-ahead spot trading) are used to construct a representation model for intraday (24h) renewable energy output prediction errors and load fluctuations, employing a normal distribution to describe the bias, Latin hypercube sampling, and improved K-means clustering.

[0153] 1) Modeling the distribution of prediction error: Define the deviation between the actual output of new energy and the predicted value as ζ=Pactual-Ppre. Assume that ζ follows a normal distribution ζ~N(μ, σ2), where μ is the mean deviation (based on historical data statistics, the value is -0.05~0.05pu), and σ is the standard deviation of the deviation (0.15~0.20pu for wind power and 0.12~0.18pu for photovoltaic).

[0154] 2) Latin hypercube sampling: The cumulative probability interval (0~1) of the bias distribution is uniformly divided into N layers (N=500). A sample point is randomly selected in each layer to obtain 500 sets of intraday source-load scenarios, so that the samples completely cover the entire interval of high bias, medium bias and low bias. The sampling efficiency is 3 times higher than that of traditional Monte Carlo.

[0155] 3) Improved K-means clustering (adaptive cluster number): The number of clusters K is adaptively determined with the goal of maximizing the silhouette coefficient (avoiding manual setting).

[0156] Calculate the silhouette coefficient corresponding to the number of candidate clusters (K=3~10). ,in Let i be the average distance between scene i and scenes in the same cluster. Let i be the average distance between scene i and its nearest neighbor clusters;

[0157] Select The largest K (for example, when K=5, the contour coefficient is 0.82, which is the maximum value) compresses 500 sets of scenes into 5 typical scenes, balancing accuracy (scene deviation ≤8%) and efficiency (solution time is shortened to less than 30 minutes).

[0158] For wind and solar power related scenario data, the correlation modeling of wind and solar power output is considered. The Frank-Copula joint distribution is taken into account. This study uses the Frank-Copula function to establish the joint distribution of wind power and solar power output in the same region to correct the scenario bias of independent sampling.

[0159] 1) Marginal distribution fitting: Kernel density estimation is performed on the intraday power output of wind power and photovoltaic power respectively to obtain the marginal cumulative distribution function. , ...;

[0160] 2) Construction of the Copula joint distribution: The correlation between the two is described by the Frank-Copula function:

[0161]

[0162] Where C(u,υ;χ) represents the binary Frank Copula function (connection function), which is the core function that describes the dependence of two random variables, with independent variables u and υ and parameter χ; , The marginal distribution values ​​are represented by χ, which is the correlation parameter (solved from historical data through maximum likelihood estimation; daytime χ = 3.5~4.2, indicating a strong negative correlation; nighttime χ = 1.2~1.8, indicating a weak correlation).

[0163] 3) Relevance scenario generation: Based on the joint distribution, the scenario is generated and the bias of independent sampling is corrected. In the example, independent sampling shows that the probability of wind power and photovoltaic power being at high output (>80% of rated power) at the same time is 12%, while the probability after correction by Copula joint distribution is only 4%, with a deviation of ≤3% from the actual operating data.

[0164] S3, retrieve the objective function and objective constraints corresponding to the new energy system. The objective function includes an economic cost subfunction that minimizes economic cost and a carbon emission cost subfunction that minimizes carbon emissions. The objective constraints include total cost constraints and power balance constraints determined based on the scheduling model. The scheduling model includes one of the following: a robust model and an opportunistic model. The robust model ensures that the total cost does not exceed the cost threshold in the first extreme scenario by maximizing the uncertainty tolerance. The opportunistic model ensures that the total cost is lower than the cost threshold in the second extreme scenario by minimizing the uncertainty tolerance. The uncertainty tolerance represents a quantitative parameter of the ability to withstand uncertainty factors, including uncertain equipment.

[0165] For the scheduling model:

[0166] An improved IGDT is adopted: multi-factor weight quantization and robust and chance dual model.

[0167] Traditional IGDT can only handle single-factor problems. By using the entropy weight method to quantify the weights of multiple factors, a dual model can be constructed to meet different risk preferences.

[0168] 1) Multi-factor weight calculation (entropy weight method): Select wind power ( ), photovoltaic ( ), electrical load ( ), heat load ( ), gas load ( Five types of uncertainty parameters are used to calculate objective weights:

[0169] Standardized parameter matrix: Where i=1~5 is the parameter type, and j=1~24 is the time period;

[0170] Calculate the entropy value: ,in ;

[0171] Calculate the weights: ,satisfy (Example weight:) =0.32、 =0.25、 =0.23、 =0.12、 =0.08);

[0172] 2) Equivalent Uncertainty and Dual Model Construction: Defining Equivalent Uncertainty Construct two types of models:

[0173] For robust models:

[0174] Considering risk aversion, adapt to medium- to long-term contracts: maximize To ensure that the system cost does not exceed the threshold in the worst-case scenario;

[0175] F0 represents the deterministic cost. =0.03 is the robust deviation factor, which is suitable for conservative scenarios such as winter heating season and power supply to important loads;

[0176] For the opportunity model:

[0177] Considering risk utilization, adapt to day-ahead spot trading: minimize Ensure that system costs remain below a threshold under favorable scenarios;

[0178] , =0.03 is the opportunity bias factor, which is applicable to aggressive scenarios such as off-peak load periods and priority for renewable energy consumption.

[0179] The objective function can be a multi-objective optimization objective function.

[0180] The objectives are to minimize economic costs, minimize carbon emissions, and maximize market transaction compliance rates.

[0181] ,in

[0182] The total economic cost (including electricity / gas purchase cost, equipment operation and maintenance cost, and market transaction cost);

[0183] For carbon trading costs (tiered carbon pricing: When ≤L =λ For every L exceeding λυ, the amount increases by λ = 65 yuan / (υ=0.2)

[0184] This refers to the fulfillment rate of medium- and long-term contracts;

[0185] =0.5、 =0.3、 =0.2 is the target weight (can be adjusted as needed).

[0186] Objective constraints may also include:

[0187] 1) Power balance constraints:

[0188] (Electric power)

[0189] (Thermal power)

[0190] Among them, PWT wind power generation capacity and PPV photovoltaic power generation capacity... Power purchased by the power grid Energy storage system discharge power, PE, load electrical load power, PP2G electricity-to-gas power. Energy storage system charging power; heating power of QGB gas boiler, heating power of QAC heat pump, heat load power of QH. The heat release power of the thermal storage system.

[0191] 2) Market trading constraints: Medium- and long-term contract electricity volume satisfy Electricity purchased via spot trading recently Not exceeding the market declaration limit;

[0192] 3) Energy storage constraints: (Energy storage capacity) (Charging and discharging power).

[0193] S4. Determine the mixed confidence interval, which is determined based on the global constraint threshold and the extreme scenario constraint threshold. The global constraint threshold is used to constrain the overall probability distribution deviation of the multi-class scenario data, and the extreme scenario constraint threshold is used to constrain the deviation of the predicted running data of a single extreme scenario, so as to correct the deviation between the multi-class scenario data and the actual running data.

[0194] To address the imbalance between robustness and complexity caused by the traditional DRO single norm, a hybrid confidence interval is established using the 1-norm (controlling the overall scene bias) and the ∞-norm (controlling the maximum bias of a single scene):

[0195] ;

[0196] in, Let k be the probability of scenario k. The initial scene probability. (1-norm deviation) (∞-norm bias), calculated using historical data confidence levels (e.g., at a confidence level of 95%). =0.15、 =0.05). This interval avoids the optimization being dominated by a single extreme scenario (∞-norm constraint) while ensuring overall scenario coverage (1-norm constraint). In a certain example, the cost fluctuation range of the DRO model was reduced from ±15% to ±8%, and the robustness was significantly enhanced.

[0197] S5, determine the confidence interval of the unbalance of the new energy system. The confidence interval of the unbalance represents the range of values ​​of the unbalance. It is used to quantify the dynamic data range of the internal power of the system caused by the change of fluctuating data, so as to correct the power balance constraint. The fluctuating data includes the fluctuating data of the output of new energy and the fluctuating data of the load demand. The unbalance is determined based on the power supply and the load.

[0198] Optionally, to address the power imbalance caused by the uncertainty of multiple loads, particle filtering is used to achieve dynamic sensing and confidence estimation.

[0199] 1) Sensor network data acquisition: Install electrical, thermal, and gas load sensors to collect real-time power, temperature, air pressure, and other data, and generate an observation sequence Y={y1, y2, ..., yT};

[0200] 2) Particle sampling and weight update:

[0201] Initialize the particle swarm N=200 particles { , ,..., (Particle swarm optimization represents different values ​​of the plant's power imbalance), initialize weights ( =1 / N);

[0202] Predict particle states based on system state equations (e.g., plant power imbalance = total power supply - total load). Update weights based on observations: ,in To observe the likelihood function;

[0203] 3) Resampling and confidence estimation: Particles with weights below a threshold are resampled to obtain an equal-weighted particle swarm, and the confidence interval for the plant power imbalance is obtained. (At a 95% confidence level, the unbalance range is [-15kW, 20kW]), and the quantification accuracy is improved by 25% compared with the traditional method.

[0204] S6. Based on system operation data, multi-scenario data, mixed confidence intervals and imbalance confidence intervals, solve the objective function under the objective constraints to obtain the scheduling strategy corresponding to the new energy system.

[0205] The above optional implementation methods can achieve at least the following beneficial effects:

[0206] (1) Multi-classification adapts to the electricity market, reducing transaction risks. It separates medium- and long-term electricity uncertainty from short-term electricity uncertainty to support medium- and long-term contract signing, day-ahead spot clearing, and ancillary service dispatch. In a trial, after applying this invention to a certain new energy base, the performance rate of medium- and long-term contracts increased from 82% to 96%, and the market default cost decreased by RMB 450,000 per quarter; the power deviation of day-ahead spot clearing decreased from ±15% to ±8%, and the procurement cost of ancillary services decreased by 18%.

[0207] (2) Scene generation is both accurate and efficient. Through correlation correction by typical methods, the deviation between the wind and solar power output scene and the actual value is ≤3% through the Frank-Copla joint distribution, which is 70% lower than that of independent sampling; Adaptive clustering: The improved K-means uses the silhouette coefficient to adaptively select K, and the computation efficiency is increased by 5 times after scene compression (from 2.5 hours to 30 minutes), while maintaining scene accuracy (deviation ≤8%); Sampling coverage of independent sampling: Compared with traditional random sampling, Latin hypercube sampling improves scene coverage by 80%, avoiding scheduling failures caused by the omission of extreme scenes.

[0208] (3) The optimized model takes into account robustness, economy and risk preference. Objective weights are adopted, and the entropy weight method replaces expert scoring. The error of multi-factor weights is reduced from ±20% to ±5%. The scheduling cost of a certain park is reduced by 9% due to weight optimization. Dual model adaptation is adopted to improve the robust-chance dual model of IGDT. It can switch between different models. The robust model is used during the winter heating season, and the power supply reliability reaches 99.9%; the chance model is used in summer, and the renewable energy consumption rate is increased by 12%. Multi-norm DRO is adopted. The 1-norm and +∞-norm confidence intervals reduce the cost fluctuation range by 47%. Robustness and computational complexity are balanced (solution time ≤ 1 hour).

[0209] (4) High accuracy in sensing plant power imbalance. Particle filtering dynamically senses the uncertainty of multiple loads, and the confidence level error of the plant power imbalance is reduced from 20% to 5%. After being applied to a gas-fired power plant, the waste of natural gas caused by imbalance was reduced by 23%, and carbon emissions were reduced by 10%.

[0210] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0211] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0212] Example 2

[0213] According to embodiments of the present invention, an apparatus for implementing the above-described method for determining the scheduling strategy of a new energy system is also provided. Figure 3 This is a structural block diagram of a scheduling strategy determination device for a new energy system according to an embodiment of the present invention, such as... Figure 3 As shown, the device includes: an acquisition module 302, a first determination module 304, a retrieval module 306, and a second determination module 308. The device will be described in detail below.

[0214] Acquisition module 302 is used to acquire system operation data corresponding to the new energy system. The new energy system includes uncertain devices and deterministic devices. Uncertain devices include wind power equipment, photovoltaic equipment, and load equipment. Deterministic devices include energy storage equipment. First determination module 304, connected to acquisition module 302, is used to determine multiple scenario data based on the system operation data. These multiple scenario data include first-cycle scenario data, second-cycle scenario data, and wind-solar related scenario data. The first-cycle scenario data corresponds to a first-cycle scenario data with a longer first-cycle period than the second-cycle scenario data. The multiple scenario data represent predicted operation data for the corresponding scenario. Retrieval module 306, connected to first determination module 304, is used to retrieve the target function corresponding to the new energy system. The system includes a number of objectives and constraints. The objective function includes an economic cost subfunction that minimizes economic costs and a carbon emission cost subfunction that minimizes carbon emissions. The objective constraints include a total cost constraint determined by a scheduling model, which includes one of the following: a robust model or an opportunistic model. The robust model ensures that the total cost does not exceed a cost threshold in the first extreme scenario by maximizing the uncertainty tolerance, while the opportunistic model ensures that the total cost is below a cost threshold in the second extreme scenario by minimizing the uncertainty tolerance. The uncertainty tolerance is a quantitative parameter representing the ability to withstand uncertainty factors, including uncertain equipment. The second determination module 308, connected to the above, is used to solve the objective function under the objective constraints based on system operation data and multi-scenario data to obtain a scheduling strategy corresponding to the new energy system.

[0215] It should be noted here that the above-mentioned acquisition module 302, first determination module 304, retrieval module 306 and second determination module 308 correspond to steps S102 to S108 in the method for determining the scheduling strategy of the new energy system. The multiple modules and the corresponding steps implement the same instances and application scenarios, but are not limited to the content disclosed in the above embodiment 1.

[0216] Example 3

[0217] According to another aspect of the present invention, an electronic device is also provided, comprising: a processor; and a memory for storing processor-executable instructions, wherein the processor is configured to execute instructions to implement the scheduling strategy determination method for any of the above-described new energy systems.

[0218] Example 4

[0219] According to another aspect of the present invention, a computer-readable storage medium is also provided, which, when the instructions in the computer-readable storage medium are executed by the processor of an electronic device, enables the electronic device to perform the scheduling strategy determination method of any of the above-described new energy systems.

[0220] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0221] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0222] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0223] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0224] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0225] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0226] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for determining the scheduling strategy of a new energy system, characterized in that, include: Acquire system operation data corresponding to the new energy system, wherein the new energy system includes uncertain devices and deterministic devices, the uncertain devices include wind power energy devices, photovoltaic energy devices and load devices, and the deterministic devices include energy storage devices; Based on the system operation data, multiple types of scenario data are determined, including first-cycle scenario data, second-cycle scenario data, and wind-and-sun related scenario data. The first cycle corresponding to the first-cycle scenario data is longer than the second cycle corresponding to the second-cycle scenario data. The multiple types of scenario data represent the predicted operation data under the corresponding scenario. The objective function and objective constraints corresponding to the new energy system are retrieved. The objective function includes an economic cost subfunction that minimizes economic costs and a carbon emission cost subfunction that minimizes carbon emissions. The objective constraints include a total cost constraint determined based on a scheduling model. The scheduling model includes one of the following: a robust model and an opportunistic model. The robust model ensures that the total cost does not exceed a cost threshold in the first extreme scenario by maximizing uncertainty tolerance. The opportunistic model ensures that the total cost is lower than a cost threshold in the second extreme scenario by minimizing uncertainty tolerance. The uncertainty tolerance represents a quantitative parameter of the ability to withstand uncertainty factors, and the uncertainty factors include the uncertain equipment. Based on the system operation data and the multi-scenario data, the objective function under the objective constraint is solved to obtain the scheduling strategy corresponding to the new energy system.

2. The method according to claim 1, characterized in that, Based on the system operation data and the multi-scenario data, the objective function under the objective constraint is solved to obtain the scheduling strategy corresponding to the new energy system, including: A mixed confidence interval is determined, wherein the mixed confidence interval is determined based on a global constraint threshold and an extreme scenario constraint threshold. The global constraint threshold is used to constrain the overall probability distribution deviation of the multi-scenarios data, and the extreme scenario constraint threshold is used to constrain the prediction operation data deviation of a single extreme scenario, so as to correct the deviation between the multi-scenarios data and the actual operation data. Based on the system operation data, multi-scenario data, and the mixed confidence interval, the objective function under the objective constraint is solved to obtain the scheduling strategy corresponding to the new energy system.

3. The method according to claim 1, characterized in that, Based on the system operation data and the multi-scenario data, the objective function under the objective constraint is solved to obtain the scheduling strategy corresponding to the new energy system, including: When the target constraint also includes a power balance constraint, the confidence interval of the imbalance of the new energy system is determined. The confidence interval of the imbalance represents the range of values ​​of the imbalance and is used to quantify the dynamic data range of the internal power of the system caused by the change of fluctuating data, so as to correct the power balance constraint. The fluctuating data includes the fluctuation data of new energy output and the fluctuation data of load demand. The imbalance is determined based on the power supply and load. Based on the system operation data, multi-scenario data, and the confidence interval of the imbalance, the objective function under the objective constraint is solved to obtain the scheduling strategy corresponding to the new energy system.

4. The method according to claim 3, characterized in that, Determining the confidence interval for the imbalance of the new energy system includes: Based on the system operation data, an initial particle swarm is constructed, wherein the initial particle swarm includes candidate values ​​of the imbalance quantity corresponding to each particle; Based on the system state equation, predict the instantaneous particle state of each particle in the initialized particle swarm to obtain a set of particle states; Based on the set of particle states, update the initial weights corresponding to each particle in the initialized particle swarm to obtain the target weights corresponding to each particle at the target time. Particles with target weights below a threshold are resampled to obtain an equal-weighted particle swarm. The confidence interval for the imbalance quantity is determined based on the equal-weighted particle swarm.

5. The method according to claim 1, characterized in that, Before retrieving the objective function and objective constraints corresponding to the new energy system, the process also includes: Determine the equipment operating parameters corresponding to the uncertain devices respectively; Based on the operating parameters of multiple devices, determine the standardized parameter matrix corresponding to each of the uncertain devices; Based on multiple standardized parameter matrices, the entropy values ​​corresponding to the uncertain devices are determined respectively; Based on multiple entropy values, the weights corresponding to the uncertain devices are determined respectively; The equivalent uncertainty is determined based on multiple weights; The scheduling model is determined based on the equivalent uncertainty.

6. The method according to claim 1, characterized in that, Before determining the various scenario data based on the system operation data, the process also includes: Determine the first edge cumulative distribution function corresponding to the intraday power output of wind power, and the second edge cumulative distribution function corresponding to the intraday power output of photovoltaic power; Based on the first edge cumulative distribution function and the second edge cumulative distribution function, the wind-solar joint distribution function is obtained; Based on the aforementioned wind-solar joint distribution function, the wind-solar related scene data is obtained.

7. The method according to any one of claims 1 to 6, characterized in that, Before determining the various scenario data based on the system operation data, the process also includes: Determine the output deviation data between the actual output data and the predicted output data of new energy sources; Based on the output deviation data, a normal distribution corresponding to the output deviation fluctuation characteristics of the second cycle is constructed; Based on the normal distribution, multiple candidate periodic scene data are determined; With the goal of maximizing the silhouette coefficient, multiple second-period scene data are determined from the multiple candidate periodic scene data, wherein the distance between the multiple second-period scene data meets the clustering condition of maximizing the inter-class distance and minimizing the intra-class distance.

8. A device for determining the scheduling strategy of a new energy system, characterized in that, include: The acquisition module is used to acquire system operation data corresponding to the new energy system. The new energy system includes uncertain devices and deterministic devices. The uncertain devices include wind power energy devices, photovoltaic energy devices and load devices. The deterministic devices include energy storage devices. The first determining module is used to determine multiple types of scenario data based on the system operation data. The multiple types of scenario data include first-cycle scenario data, second-cycle scenario data, and wind-and-sun related scenario data. The first cycle corresponding to the first-cycle scenario data is longer than the second cycle corresponding to the second-cycle scenario data. The multiple types of scenario data represent the predicted operation data under the corresponding scenario. The retrieval module is used to retrieve the objective function and objective constraints corresponding to the new energy system. The objective function includes an economic cost sub-function that minimizes economic costs and a carbon emission cost sub-function that minimizes carbon emissions. The objective constraints include a total cost constraint determined based on a scheduling model. The scheduling model includes one of the following: a robust model and an opportunistic model. The robust model ensures that the total cost does not exceed a cost threshold in the first extreme scenario by maximizing uncertainty tolerance. The opportunistic model ensures that the total cost is lower than a cost threshold in the second extreme scenario by minimizing uncertainty tolerance. The uncertainty tolerance represents a quantitative parameter of the ability to withstand uncertainty factors, and the uncertainty factors include the uncertain equipment. The second determining module is used to solve the objective function under the target constraint based on the system operation data and the multi-type scenario data, so as to obtain the scheduling strategy corresponding to the new energy system.

9. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the scheduling strategy determination method for a new energy system as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device is able to perform the scheduling strategy determination method for the new energy system as described in any one of claims 1 to 7.

Images