A climate-driven hybrid energy storage integrated energy optimization method, system, device and storage medium
By employing climate-driven scenario generation and Monte Carlo decomposition optimization methods, the supply-demand imbalance caused by the uncertainty of renewable energy in integrated energy systems was addressed, enabling efficient energy storage system configuration and optimization, and improving system reliability and economy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XIAMEN UNIV
- Filing Date
- 2026-01-16
- Publication Date
- 2026-05-12
AI Technical Summary
Existing technologies are insufficient to effectively address the supply-demand imbalance caused by the climate uncertainty and load fluctuations of renewable energy in integrated energy systems. Traditional optimization models have high computational complexity and cost, making them difficult to meet actual needs.
By employing climate-driven scene generation technology combined with Monte Carlo decomposition optimization, and using Wasserstein generative adversarial networks to generate realistic and diverse time-series data of wind and solar resources, a capacity-scheduling joint optimization model for a hybrid energy storage integrated energy system is constructed. The Monte Carlo decomposition strategy is then used to solve the model, and the synergistic effect of battery energy storage and hydrogen energy storage is optimized by combining the energy storage duration integral index.
It improves the reliability and economy of integrated energy systems at different time scales, reduces computational complexity, provides a configuration ratio for short-term regulation and long-term energy transfer, and enhances the flexibility and economy of the system.
Smart Images
Figure CN121543836B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of integrated energy system planning and optimization technology, and more specifically to a climate-driven hybrid energy storage integrated energy optimization method, system, equipment and storage medium. Background Technology
[0002] In recent years, the proportion of renewable energy in the energy system has risen rapidly. Clean energy sources such as solar and wind power have significant intermittency and volatility, posing challenges to the safe and economical operation of integrated energy systems.
[0003] Integrated energy systems, as a novel energy utilization model, enable the interconnection and coordinated optimization of multiple energy forms such as electricity, heat, and hydrogen. By integrating heterogeneous energy units such as wind power, photovoltaic power generation, combined heat and power, and energy storage, integrated energy systems can improve energy utilization efficiency and reduce operating costs. At the same time, the coupled operation of multiple energy units increases the complexity of the model and places higher demands on the optimal scheduling of the system.
[0004] In integrated energy systems, not only does the load exhibit significant volatility and randomness, but the renewable energy generation side is also subject to strong uncertainties due to climate conditions. This dual uncertainty exposes the system to the risk of supply-demand imbalances at different time scales. For example, intraday load fluctuations and changes in solar irradiance may lead to short-term power shortages, while seasonal variations in wind power output can result in long-term energy imbalances.
[0005] To address this uncertainty, academic and engineering practices typically employ typical day methods, deterministic optimization, or scenario generation methods. However, typical day methods struggle to cover the diverse climate characteristics throughout the year, while deterministic methods may overlook extreme situations caused by fluctuations, leading to system planning results deviating from actual needs. Scenario generation methods improve the model's adaptability to uncertainty to some extent, but still suffer from problems such as unstable generated samples and excessive computational burden due to a large number of scenarios.
[0006] Energy storage technology is considered an important means to mitigate the fluctuations in renewable energy and improve system flexibility. Battery energy storage systems are characterized by fast response and high efficiency, making them suitable for hourly-level energy balancing and frequency regulation / peak shaving; hydrogen energy storage systems, on the other hand, have the potential for large-scale and long-term energy storage, enabling energy transfer on weekly, monthly, and even seasonal scales. However, how to rationally allocate and schedule short-term and long-term energy storage to complement each other's advantages is a key issue currently facing integrated energy system planning.
[0007] Furthermore, with the expansion of integrated energy systems and the dramatic increase in climate and operational data, the dimensionality and complexity of optimization models are constantly rising. Traditional mixed-integer programming methods suffer from significantly increased computational costs when dealing with high-dimensional annual time series and large-scale scenarios, and their solution speed and stability are insufficient to meet the needs of practical engineering applications.
[0008] Therefore, how to provide a climate-driven hybrid energy storage integrated energy optimization method, system, equipment and storage medium to improve model calculation efficiency and the rationality of energy storage system configuration while ensuring the realism and diversity of climate scenarios is an urgent problem to be solved by those skilled in the art. Summary of the Invention
[0009] In view of this, the present invention provides a climate-driven integrated energy optimization method, system, device, and storage medium for hybrid energy storage. By introducing climate-driven scene generation technology, it is possible to generate time-series data of wind and solar resources that are both realistic and diverse. On this basis, the Monte Carlo decomposition optimization method is used to solve the hourly capacity configuration and scheduling problem throughout the year, solving the problem of excessive computational burden of traditional optimization models in high-dimensional scenarios. At the same time, an energy storage duration integral index is proposed to quantify the synergistic effect of short-term battery energy storage and long-term hydrogen energy storage, thereby improving the reliability and economy of the integrated energy system at different time scales.
[0010] To achieve the above objectives, the present invention adopts the following technical solution:
[0011] A climate-driven integrated energy optimization approach for hybrid energy storage includes:
[0012] Step 1: Generate hourly scenes of scenery throughout the year, covering seasonal features, based on a Wasserstein generative adversarial network that integrates gradient penalty and mini-batch discriminant terms;
[0013] Step 2: Using the annual wind and solar hourly scenarios as random input, construct a joint optimization model of capacity-schedule for a hybrid energy storage integrated energy system with annualized total cost as the objective, including the coupling of electricity, hydrogen energy, and thermal energy, as well as equipment operation constraints. Then, use the Monte Carlo decomposition strategy to solve the joint optimization model of capacity-schedule for the hybrid energy storage integrated energy system, and obtain the optimal capacity configuration and the hourly scheduling strategy for each scenario under the optimal capacity configuration.
[0014] Step 3: Perform an integral evaluation of the energy storage duration on the hourly scheduling strategy to obtain the utilization distribution of battery energy storage and hydrogen energy storage in the hybrid energy storage integrated energy system at different time scales.
[0015] Optionally, in step 1, the Wasserstein generative adversarial network that integrates gradient penalty and mini-batch discriminant terms is as follows:
[0016] Using generators With critic The adversarial structure;
[0017] Mini-batch discrimination is introduced into the WGAN-GP architecture; among which, the generator... Using wind and light as two-channel inputs respectively, and upsampling through 1D convolution, downsampling, GLU residuals, and deconvolution, the output is a two-channel annual sequence; Critics It is a three-layer fully connected network, with the middle layer splicing mini-batch distance features;
[0018] The loss function is as follows:
[0019] The basic goal of WGAN-GP:
[0020]
[0021]
[0022] in, , These are the base loss values for the critic and the generator, respectively. The distribution of the actual climate data sample is denoted as . ; The distribution of the fake samples generated by the generator is denoted as . ; Represents the mathematical expectation; This represents the critic's rating output for the input sample; This is the gradient penalty coefficient, used to adjust the strength of the gradient penalty term; The random interpolation sampling points between the real sample and the generated sample are defined as follows: ,in for Random numbers between; This indicates that the critiquer output is relative to the input. The gradient; express Norm;
[0023] Small batch discrimination criteria: Set up a critic The middle layer The features are Define the intra-batch similarity measure :
[0024]
[0025] in, and These represent the first and second digits in the current batch. The and the first The feature vector of each sample in the intermediate layer of the critique; Indicates the Euclidean distance between features;
[0026] By incorporating generator and critique losses, the following is formed:
[0027]
[0028]
[0029] in, and These represent the total losses of the generator and the critic after adding mini-batch discriminant terms, respectively. This refers to the batch size during model training. and These are the weight coefficients for the mini-batch items in the generator and the critique, respectively, used to adjust the strength of the diversity constraint; This is an exponential function used to map distance to similarity weights.
[0030] Optionally, in step 2, the capacity-dispatch joint optimization model for the hybrid energy storage integrated energy system is as follows:
[0031] Decision variables and parameters:
[0032] Key capacity decision: Photovoltaic installed capacity installed capacity of wind turbines Electrolytic cell installed capacity Fuel cell installed capacity Battery installed capacity Hydrogen storage equivalent energy ;
[0033] Hourly variables in a climate scenario: , , , , , ;in, This refers to the hourly power generation of photovoltaic systems. This refers to the hourly power generation of wind power. This refers to the power consumption of the electrolytic cell; Power generation for fuel cells; The charging power for battery energy storage; This refers to the battery's energy storage discharge power. For time period numbering, ; This refers to the hourly state of charge of the battery. This represents the hourly equivalent state of charge of hydrogen energy storage. , These refer to the power purchased from the power grid and the power sold to the power grid, respectively.
[0034] Objective function:
[0035] Minimize annualized total cost:
[0036]
[0037]
[0038]
[0039]
[0040]
[0041] in, This represents the system's annualized investment cost; Annual equipment maintenance costs; This represents the annual net cost of electricity purchases. It is a collection of equipment, including photovoltaics, wind turbines, electrolyzers, fuel cells, batteries, and hydrogen storage containers; For equipment The installed capacity; For equipment The unit investment price per unit capacity; For equipment The capital recovery coefficient; The discount rate; For equipment Service life; For equipment exist Operating power during a given time period; For equipment The unit power operation and maintenance cost per unit; for Time-of-use electricity pricing for specific time periods; for The on-grid electricity price for a given time period;
[0042] Electricity and hydrogen balance and equipment constraints:
[0043] Power balance:
[0044]
[0045] in, for System base electrical load demand during a given period;
[0046] Hydrogen energy balance, in terms of equivalent energy:
[0047]
[0048]
[0049]
[0050]
[0051]
[0052] in, and They are respectively Hydrogen charging and discharging power of hydrogen storage during specific time periods; This represents the self-loss coefficient for hydrogen energy storage. and These represent the efficiencies of the hydrogen storage and release processes, respectively. and These are the maximum rated power limits for hydrogen storage and hydrogen release, respectively. and for The binary variables represent the hydrogen storage and hydrogen release states respectively, and are used to ensure that the two are mutually exclusive.
[0053] Constraints of Electrolyzers and Fuel Cells:
[0054]
[0055]
[0056]
[0057]
[0058] in, The efficiency of hydrogen production in the electrolyzer; For fuel cell power generation efficiency; This is the lower heating value of hydrogen.
[0059] Constraints of lithium battery energy storage systems:
[0060]
[0061]
[0062]
[0063]
[0064]
[0065] in, This is the battery self-discharge rate coefficient; and These refer to the battery charging and discharging efficiency, respectively. and These are the maximum charging and discharging power limits for the battery, respectively. and for A binary variable used to ensure that charging and discharging are mutually exclusive;
[0066] Fan output constraints:
[0067]
[0068] in, for Real-time wind speed during the specified time period; To cut in wind speed; To cut off the wind speed; Rated wind speed; This refers to the rated power of the fan. The overall efficiency of wind turbine power generation;
[0069] Photovoltaic output constraints:
[0070]
[0071] in, Light intensity under standard test conditions; for The actual solar radiation intensity during the time period; For the overall efficiency of photovoltaic systems;
[0072] Power grid interconnection mutual exclusion constraints:
[0073]
[0074]
[0075]
[0076] in: and These refer to the power purchased from the power grid and the power sold to the power grid, respectively. and These are the maximum permissible power sales and power purchase capacity at the grid connection point, respectively. and for Binary variables ensure that electricity sales and purchases do not occur simultaneously.
[0077] Optionally, in step 2, the Monte Carlo decomposition strategy is used to solve the capacity-scheduling joint optimization model of the hybrid energy storage integrated energy system to obtain the optimal capacity configuration and the hourly scheduling strategy for each scenario under the optimal capacity configuration, specifically:
[0078] The generated hourly scenery scenes throughout the year are sampled multiple times to form multiple sub-models of the year's duration and solved independently. The capacity configuration distribution is then aggregated, and the distribution center is selected as the optimal capacity configuration based on the solution set distribution.
[0079] Given a fixed optimal capacity configuration, solve for the hourly scheduling strategy for each scenario under the optimal capacity configuration.
[0080] Optionally, in step 3, the time-by-time scheduling strategy is evaluated by energy storage duration integration to obtain the utilization distribution of battery energy storage and hydrogen energy storage in the hybrid energy storage integrated energy system at different time scales, specifically:
[0081] Based on the time-by-time scheduling strategy, the time-by-time energy curve of energy storage is generated;
[0082] The hourly energy curve of energy storage is discretized along the vertical axis and scanned row by row. The product of energy storage and duration in each segment is recorded as the energy storage duration integral index. The energy storage duration distribution map is obtained by accumulating the data for the whole year and all scenarios.
[0083] This invention also provides a climate-driven hybrid energy storage integrated energy optimization system utilizing a climate-driven hybrid energy storage integrated energy optimization method, comprising:
[0084] The year-round landscape scene generation module is used to generate year-round landscape scenes covering seasonal features based on a Wasserstein generative adversarial network that integrates gradient penalty and mini-batch discriminant terms.
[0085] The hourly scheduling strategy solution module is used to construct a capacity-scheduling joint optimization model of a hybrid energy storage integrated energy system with annualized total cost as the objective, including the coupling of electricity, hydrogen energy, and thermal energy, as well as equipment operation constraints, using hourly wind and solar scenarios throughout the year as random input. The Monte Carlo decomposition strategy is then used to solve the capacity-scheduling joint optimization model of the hybrid energy storage integrated energy system to obtain the optimal capacity configuration and the hourly scheduling strategy for each scenario under the optimal capacity configuration.
[0086] Energy storage duration integral evaluation module: used to evaluate the energy storage duration integral of the hourly scheduling strategy, and obtain the utilization distribution of battery energy storage and hydrogen energy storage in the hybrid energy storage integrated energy system at different time scales.
[0087] The present invention also provides an electronic device, comprising:
[0088] Memory, used to store computer programs;
[0089] A processor is used to implement a climate-driven hybrid energy storage integrated energy optimization method when executing computer programs.
[0090] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a climate-driven hybrid energy storage integrated energy optimization method.
[0091] As can be seen from the above technical solutions, compared with the prior art, this invention discloses a climate-driven hybrid energy storage integrated energy optimization method, system, device, and storage medium. By using an improved generative adversarial network to generate climate-driven wind and solar scenarios, it can maintain temporal relevance and seasonal characteristics, avoiding the problems of insufficient samples and model collapse in traditional scenario generation methods. Through Monte Carlo decomposition optimization, the computational complexity of hourly multi-scenario models throughout the year is effectively reduced, ensuring the robustness of the optimization results. Through the energy storage duration integral index, the division of labor and synergistic effects of battery energy storage and hydrogen energy storage at different time scales can be intuitively revealed, providing a quantitative basis for the configuration ratio of short-term regulation and long-term energy transfer, thereby improving the economy, flexibility, and reliability of the integrated energy system. Attached Figure Description
[0092] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0093] Figure 1 This is a schematic diagram of the method flow provided by the present invention.
[0094] Figure 2 This is a schematic diagram of the Wasserstein generative adversarial network structure that integrates gradient penalty and mini-batch discriminant terms provided by the present invention.
[0095] Figure 3 This is a schematic diagram of the hybrid energy storage integrated energy system provided by the present invention.
[0096] Figure 4 This is a schematic diagram of the ESDI evaluation process provided by the present invention. Detailed Implementation
[0097] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0098] Example 1:
[0099] Embodiment 1 of this invention discloses a climate-driven hybrid energy storage integrated energy optimization method, such as... Figure 1 As shown, it includes:
[0100] Step 1: Based on the Wasserstein generative adversarial network that integrates gradient penalty and mini-batch discriminant terms, train it for spring, summer, autumn and winter respectively to learn seasonal features and generate a year-round wind and solar hourly scene (annual renewable energy power generation scene) that covers seasonal features.
[0101] A Wasserstein generative adversarial network (WGAN-GP-minibatch) integrating gradient penalty and mini-batch discriminant terms is used to improve the diversity and stability of climate samples, such as... Figure 2 As shown, specifically:
[0102] Using generators With critic The adversarial structure;
[0103] Mini-batch discrimination is introduced into the WGAN-GP architecture to encourage intra-batch sample diversity; among which, the generator... As shown in Table 1, wind and light are used as 2-channel inputs respectively, and after 1D convolution, downsampling, GLU residual and deconvolution upsampling, a 2-channel annual sequence is output; Critics As shown in Table 2, it is a three-layer fully connected network with the middle layer splicing mini-batch distance features;
[0104] Table 1 Generator structure
[0105]
[0106] Table 2 Critics structure
[0107]
[0108] The loss function is as follows:
[0109] The basic goal of WGAN-GP:
[0110]
[0111]
[0112] in, , These are the base loss values for the critic and the generator, respectively; The distribution of the actual climate data sample is denoted as . ; The distribution of the fake samples generated by the generator is denoted as . ; Represents the mathematical expectation; This represents the critic's rating output for the input sample; This is the gradient penalty coefficient, used to adjust the strength of the gradient penalty term; The random interpolation sampling points between the real sample and the generated sample are defined as follows: ,in for Random numbers between; This indicates that the critiquer output is relative to the input. The gradient; express Norm (Euclidean norm);
[0113] Small batch discrimination criteria: Set up a critic The middle layer The features are Define the intra-batch similarity measure :
[0114]
[0115] in, and These represent the th elements in the current batch. The and the first The feature vector of each sample in the intermediate layer of the critique; Indicates the Euclidean distance between features;
[0116] By incorporating generator and critique losses, the following is formed:
[0117]
[0118]
[0119] in, and These represent the total losses of the generator and the critic after adding mini-batch discriminant terms, respectively. This refers to the batch size during model training. and These are the weight coefficients for the mini-batch items in the generator and the critique, respectively, used to adjust the strength of the diversity constraint; The exponential function is used to map distance to similarity weights. This design improves the critiquer's sensitivity to intra-batch diversity, prompting the generator... Generate more diverse sequences to mitigate pattern collapse.
[0120] It also includes comparing the generated wind and solar sequences with historical measurements using three types of indicators: autocorrelation (ACF), cumulative distribution function (CDF), and spectral density (SD), in order to assess the degree of matching between temporal correlation, distribution consistency, and frequency domain characteristics.
[0121] Step 2: Using the annual wind and solar hourly scenarios as random inputs, construct a joint optimization model of capacity-schedule for a hybrid energy storage integrated energy system (multi-scenario mixed integer linear programming model) with annualized total cost as the objective and including coupling of electricity, hydrogen energy, and thermal energy as well as equipment operation constraints. Then, use the Monte Carlo decomposition strategy to solve the joint optimization model of capacity-schedule for the hybrid energy storage integrated energy system to obtain the optimal capacity configuration and the hourly scheduling strategy for each scenario under the optimal capacity configuration.
[0122] Hybrid energy storage integrated energy system, such as Figure 3 As shown, it includes: Battery Energy Storage System (BESS) and Hydrogen Energy Storage System (P2H2P); wherein: BESS is mainly used for regulating intraday and daily-scale fluctuations; Hydrogen Energy Storage System is used for energy balance on weekly, monthly and seasonal scales.
[0123] Decision variables and parameters:
[0124] Key capacity decision: Photovoltaic installed capacity installed capacity of wind turbines Electrolytic cell installed capacity Fuel cell installed capacity Battery installed capacity Hydrogen storage equivalent energy ;
[0125] Hourly variables in a climate scenario: , , , , , ;in, This refers to the hourly power generation of photovoltaic systems. This refers to the hourly power generation of wind power. This refers to the power consumption of the electrolytic cell; Power generation for fuel cells; The charging power for battery energy storage; This refers to the battery's energy storage discharge power. For time period numbering, ; This refers to the hourly state of charge of the battery. This represents the hourly equivalent state of charge of hydrogen energy storage. , These refer to the power purchased from the power grid and the power sold to the power grid, respectively.
[0126] Objective function:
[0127] Minimizing annualized total cost (TAC) includes equipment investment costs, maintenance costs, and grid interaction costs:
[0128]
[0129]
[0130]
[0131]
[0132]
[0133] in, This represents the system's annualized investment cost; Annual equipment maintenance costs; This represents the annual net cost of electricity purchases. It is a collection of equipment, including photovoltaics, wind turbines, electrolyzers, fuel cells, batteries, and hydrogen storage containers; For equipment The installed capacity; For equipment The unit investment price per unit capacity; For equipment The capital recovery coefficient; The discount rate; For equipment Service life; For equipment exist Operating power during a given time period; For equipment The unit power operation and maintenance cost per unit; for Time-of-use electricity pricing for specific time periods; for The on-grid electricity price for a given time period;
[0134] Electricity and hydrogen balance and equipment constraints:
[0135] Power balance:
[0136]
[0137] in, for System base electrical load demand during a given period;
[0138] Hydrogen energy balance, in terms of equivalent energy:
[0139]
[0140]
[0141]
[0142]
[0143]
[0144] in, and They are respectively Hydrogen charging and discharging power (equivalent to energy flow) of hydrogen energy storage during a given time period; This is the self-loss coefficient (storage efficiency) for hydrogen energy storage. and These represent the efficiencies of the hydrogen storage and release processes, respectively. and These are the maximum rated power limits for hydrogen storage and hydrogen release, respectively. and for The binary variables represent the hydrogen storage and hydrogen release states respectively, and are used to ensure that the two are mutually exclusive.
[0145] Constraints of Electrolyzers and Fuel Cells:
[0146]
[0147]
[0148]
[0149]
[0150] in, The efficiency of hydrogen production in the electrolyzer; For fuel cell power generation efficiency; This is the lower heating value of hydrogen.
[0151] Constraints of lithium battery energy storage systems:
[0152]
[0153]
[0154]
[0155]
[0156]
[0157] in, This is the battery self-discharge rate coefficient; and These refer to the battery charging and discharging efficiency, respectively. and These are the maximum charging and discharging power limits for the battery, respectively. and for A binary variable used to ensure that charging and discharging are mutually exclusive;
[0158] Fan output constraints:
[0159]
[0160] in, for Real-time wind speed during the specified time period; To cut in wind speed; To cut off the wind speed; Rated wind speed; This refers to the rated power of the fan. The overall efficiency of wind turbine power generation;
[0161] Photovoltaic output constraints:
[0162]
[0163] in, The light intensity under standard test conditions (usually taken as...) ); for The actual solar radiation intensity during the time period; For the overall efficiency of photovoltaic systems;
[0164] Power grid interconnection mutual exclusion constraints:
[0165]
[0166]
[0167]
[0168] in: and These refer to the power purchased from the power grid and the power sold to the power grid, respectively. and These are the maximum permissible power sales and power purchase capacity at the grid connection point, respectively. and for Binary variables ensure that electricity sales and purchases do not occur simultaneously.
[0169] The Monte Carlo decomposition strategy is used to solve the capacity-scheduling joint optimization model of the hybrid energy storage integrated energy system, obtaining the optimal capacity configuration and the hourly scheduling strategy for each scenario under the optimal capacity configuration, specifically:
[0170] The generated hourly landscape scenes throughout the year (generator) Output) Perform multiple samplings (random sampling) The system is divided into several scenarios, forming multiple sub-models for the entire year (each scenario constitutes an 8760-hour MILP sub-model), which are solved independently and then aggregated to obtain the capacity configuration distribution. Based on the solution set distribution, the distribution center (geometric centroid) is selected as the optimal capacity configuration. ;
[0171] Fixed optimal capacity configuration Solve for the time-by-time scheduling strategy for each scenario under optimal capacity configuration. .
[0172] Step 3: Perform an Energy Storage Duration Integral (ESDI) evaluation on the hourly scheduling strategy to obtain the utilization distribution of battery energy storage and hydrogen energy storage in the hybrid energy storage integrated energy system at different time scales.
[0173] An integral evaluation of the energy storage duration is performed on the hourly dispatch strategy to obtain the utilization distribution of battery energy storage and hydrogen energy storage in the hybrid energy storage integrated energy system at different time scales, such as... Figure 4 As shown, specifically:
[0174] Based on the time-by-time scheduling strategy, the time-by-time energy curve of energy storage is generated;
[0175] The hourly energy curve of energy storage is discretized along the vertical axis and scanned row by row. The product of energy storage and duration in each segment is recorded as the energy storage duration integral index. The energy storage duration distribution map is obtained by accumulating the data for the whole year and all scenarios.
[0176] Specifically, set in the scene The hourly energy sequence of the lower energy storage is (Recommended) Discretize the vertical axis (SOC value) into several layers. Perform a lateral connectivity scan on each energy level and record the duration of each segment. With the energy amplitude of this layer The product is denoted as:
[0177]
[0178] The ESDI duration distribution was obtained by accumulating it throughout the year and across all scenarios:
[0179]
[0180] This method essentially calculates the distribution of the "energy-time area," which can separate and visualize the contributions of "short-term regulation (battery energy storage)" and "long-term balance (hydrogen energy storage)." The implementation steps are: "discrete the vertical axis → scan line by line → record ESDI → statistics / plotting."
[0181] Example 2:
[0182] Embodiment 2 of the present invention discloses a climate-driven hybrid energy storage integrated energy optimization system utilizing a climate-driven hybrid energy storage integrated energy optimization method, comprising:
[0183] The year-round landscape scene generation module is used to generate year-round landscape scenes covering seasonal features based on a Wasserstein generative adversarial network that integrates gradient penalty and mini-batch discriminant terms.
[0184] The hourly scheduling strategy solution module is used to construct a capacity-scheduling joint optimization model of a hybrid energy storage integrated energy system with annualized total cost as the objective, including the coupling of electricity, hydrogen energy, and thermal energy, as well as equipment operation constraints, using hourly wind and solar scenarios throughout the year as random input. The Monte Carlo decomposition strategy is then used to solve the capacity-scheduling joint optimization model of the hybrid energy storage integrated energy system to obtain the optimal capacity configuration and the hourly scheduling strategy for each scenario under the optimal capacity configuration.
[0185] Energy storage duration integral evaluation module: used to evaluate the energy storage duration integral of the hourly scheduling strategy, and obtain the utilization distribution of battery energy storage and hydrogen energy storage in the hybrid energy storage integrated energy system at different time scales.
[0186] Example 3:
[0187] Embodiment 3 of the present invention also discloses an electronic device, comprising:
[0188] Memory, used to store computer programs;
[0189] A processor is used to implement a climate-driven hybrid energy storage integrated energy optimization method when executing computer programs.
[0190] Example 4:
[0191] Embodiment 4 of the present invention discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a climate-driven hybrid energy storage integrated energy optimization method.
[0192] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.
[0193] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A climate-driven integrated energy optimization method for hybrid energy storage, characterized in that, include: Step 1: Generate hourly scenes of scenery throughout the year, covering seasonal features, based on a Wasserstein generative adversarial network that integrates gradient penalty and mini-batch discriminant terms; Step 2: Using the annual wind and solar hourly scenarios as random input, construct a capacity-schedule joint optimization model for a hybrid energy storage integrated energy system with annualized total cost as the objective, including the coupling of electricity, hydrogen energy, and thermal energy, as well as equipment operation constraints. Then, use the Monte Carlo decomposition strategy to solve the capacity-schedule joint optimization model for the hybrid energy storage integrated energy system to obtain the optimal capacity configuration and the hourly scheduling strategy for each scenario under the optimal capacity configuration. Step 3: Perform an energy storage duration integral evaluation on the time-by-time scheduling strategy to obtain the utilization distribution of battery energy storage and hydrogen energy storage in the hybrid energy storage integrated energy system at different time scales; In step 1, the Wasserstein generative adversarial network that integrates gradient penalty and mini-batch discriminant terms is specifically as follows: Using generators With critic The adversarial structure; Mini-batch discrimination is introduced on the WGAN-GP architecture; wherein, the generator Using wind and light as two-channel inputs respectively, and after 1D convolution, downsampling, GLU residual and deconvolution upsampling, a two-channel annual sequence is output; the critic It is a three-layer fully connected network, with the middle layer splicing mini-batch distance features; The loss function is as follows: The basic goal of WGAN-GP: in, , These are the base loss values for the critic and the generator, respectively. The distribution of the actual climate data sample is denoted as . ; The distribution of the fake samples generated by the generator is denoted as . ; Represents the mathematical expectation; This represents the critic's rating output for the input sample; This is the gradient penalty coefficient, used to adjust the strength of the gradient penalty term; The random interpolation sampling points between the real sample and the generated sample are defined as follows: ,in for Random numbers between; This indicates that the critiquer output is relative to the input. The gradient; express Norm; Small batch discrimination criteria: Set up a critic The middle layer The features are Define the intra-batch similarity measure : in, and These represent the first and second digits in the current batch. The and the first The feature vector of each sample in the intermediate layer of the critique; Indicates the Euclidean distance between features; By incorporating generator and critique losses, the following is formed: in, and These represent the total losses of the generator and the critic after adding mini-batch discriminant terms, respectively. This refers to the batch size during model training. and These are the weight coefficients for the mini-batch items in the generator and the critique, respectively, used to adjust the strength of the diversity constraint; This is an exponential function used to map distance to similarity weights.
2. The climate-driven hybrid energy storage integrated energy optimization method according to claim 1, characterized in that, In step 2, the capacity-dispatch joint optimization model for the hybrid energy storage integrated energy system is as follows: Decision variables and parameters: Key capacity decision: Photovoltaic installed capacity installed capacity of wind turbines Electrolytic cell installed capacity Fuel cell installed capacity Battery installed capacity Hydrogen storage equivalent energy ; Hourly variables in a climate scenario: , , , , , ;in, The hourly power generation of photovoltaic power; This refers to the hourly power generation of wind power. This refers to the power consumption of the electrolytic cell; Power generation for fuel cells; The charging power for battery energy storage; This refers to the battery's energy storage discharge power. For time period numbering, ; This refers to the hourly state of charge of the battery. This represents the hourly equivalent state of charge of hydrogen energy storage. , These refer to the power purchased from the power grid and the power sold to the power grid, respectively. Objective function: Minimize annualized total cost: in, This represents the system's annualized investment cost; Annual equipment maintenance costs; This represents the annual net cost of electricity purchases. It is a collection of equipment, including photovoltaics, wind turbines, electrolyzers, fuel cells, batteries, and hydrogen storage containers; For equipment The installed capacity; For equipment The unit investment price per unit capacity; For equipment The capital recovery coefficient; The discount rate; For equipment Service life; For equipment exist Operating power during a given time period; For equipment The unit power operation and maintenance cost per unit; for Time-of-use electricity pricing for specific time periods; for The on-grid electricity price for a given time period; Electricity and hydrogen balance and equipment constraints: Power balance: in, for System base electrical load demand during a given period; Hydrogen energy balance, in terms of equivalent energy: in, and They are respectively Hydrogen charging and discharging power of hydrogen storage during specific time periods; This represents the self-loss coefficient for hydrogen energy storage. and These represent the efficiencies of the hydrogen storage and release processes, respectively. and These are the maximum rated power limits for hydrogen storage and hydrogen release, respectively. and for The binary variables represent the hydrogen storage and hydrogen release states respectively, and are used to ensure that the two are mutually exclusive. Constraints of Electrolyzers and Fuel Cells: in, The efficiency of hydrogen production in the electrolyzer; For fuel cell power generation efficiency; This is the lower heating value of hydrogen. Constraints of lithium battery energy storage systems: in, This is the battery self-discharge rate coefficient; and These refer to the battery charging and discharging efficiency, respectively. and These are the maximum charging and discharging power limits for the battery, respectively. and for A binary variable used to ensure that charging and discharging are mutually exclusive; Fan output constraints: in, for Real-time wind speed during the specified time period; To cut in wind speed; To cut off the wind speed; Rated wind speed; This refers to the rated power of the fan. The overall efficiency of wind turbine power generation; Photovoltaic output constraints: in, Light intensity under standard test conditions; for The actual solar radiation intensity during the time period; For the overall efficiency of photovoltaic systems; Power grid interconnection mutual exclusion constraints: in: and These refer to the power purchased from the power grid and the power sold to the power grid, respectively. and These are the maximum permissible power sales and power purchase capacity at the grid connection point, respectively. and for Binary variables ensure that electricity sales and purchases do not occur simultaneously.
3. The climate-driven hybrid energy storage integrated energy optimization method according to claim 1, characterized in that, In step 2, the Monte Carlo decomposition strategy is used to solve the capacity-scheduling joint optimization model of the hybrid energy storage integrated energy system, obtaining the optimal capacity configuration and the hourly scheduling strategy for each scenario under the optimal capacity configuration, specifically: The generated hourly scenery scene throughout the year is sampled multiple times to form multiple sub-models of the year's duration and solved independently. The aggregated results are obtained to obtain the capacity configuration distribution. The distribution center is selected as the optimal capacity configuration based on the solution set distribution. With the optimal capacity configuration fixed, solve for the hourly scheduling strategy for each scenario under the optimal capacity configuration.
4. The climate-driven hybrid energy storage integrated energy optimization method according to claim 1, characterized in that, In step 3, the time-by-time scheduling strategy is evaluated by energy storage duration integration to obtain the utilization distribution of battery energy storage and hydrogen energy storage in the hybrid energy storage integrated energy system at different time scales, specifically: Based on the time-by-time scheduling strategy, an hourly energy curve for energy storage is generated; The hourly energy curve of the energy storage is discretized along the vertical axis and scanned row by row. The product of energy storage and duration of each segment is recorded as the energy storage duration integral index. The energy storage duration distribution map is obtained by accumulating the data for the whole year and all scenarios.
5. A climate-driven hybrid energy storage integrated energy optimization system utilizing the climate-driven hybrid energy storage integrated energy optimization method according to any one of claims 1-4, characterized in that, include: The year-round landscape scene generation module is used to generate year-round landscape scenes covering seasonal features based on a Wasserstein generative adversarial network that integrates gradient penalty and mini-batch discriminant terms. The hourly scheduling strategy solution module is used to construct a capacity-scheduling joint optimization model of the hybrid energy storage integrated energy system with the annual wind and solar hourly scenarios as random input, with the annualized total cost as the objective and including the coupling of electricity, hydrogen energy, and thermal energy as well as equipment operation constraints. The module then uses a Monte Carlo decomposition strategy to solve the capacity-scheduling joint optimization model of the hybrid energy storage integrated energy system, and obtains the optimal capacity configuration and the hourly scheduling strategy for each scenario under the optimal capacity configuration. Energy storage duration integral evaluation module: used to perform energy storage duration integral evaluation on the hourly scheduling strategy to obtain the utilization distribution of battery energy storage and hydrogen energy storage in the hybrid energy storage integrated energy system at different time scales.
6. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of a climate-driven hybrid energy storage integrated energy optimization method as described in any one of claims 1-4 when executing the computer program.
7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of a climate-driven hybrid energy storage integrated energy optimization method as described in any one of claims 1-4.