A long-short time nested operation simulation method, device, medium and product considering extreme weather and frequency safety

Abstract

The application discloses a long-short time nested operation simulation method and device considering extreme weather and frequency safety, a medium and a product, relates to the technical field of power system planning and operation, and comprises the following steps: based on a WGAN-GP model and a meteorology-power mapping model, constructing all-year source and load time series considering extreme weather scenarios; considering system frequency response characteristics, establishing an all-year time sequence operation simulation model considering dynamic frequency safety constraints, and taking the all-year source and load time series as input data of the all-year time sequence operation simulation model; and based on long-short time nesting and a linear unit combination model, accelerating the solution of the all-year time sequence operation simulation model to obtain power system operation simulation results. The application expands the extreme scenario data set, solves the low-efficiency problem caused by the frequency safety constraint in long-period simulation, realizes the collaborative optimization of extreme weather scenario coverage and dynamic frequency safety guarantee, and provides support for power system dispatching and energy storage planning.

Description

Technical Field

[0001] This application relates to the field of power system planning and operation simulation technology, and in particular to a long-term and short-term nested operation simulation method, equipment, medium and product that takes into account extreme weather and frequency security. Background Technology

[0002] Driven by the climate and energy crises, countries worldwide are vigorously developing renewable energy sources such as wind and solar power to achieve a low-carbon energy transition. The grid connection of large-scale renewable energy sources, such as the Sakhalin wind and solar power base, distributed photovoltaics, and offshore wind power, has highlighted the mismatch between energy supply and demand on the source and load sides, making power balance increasingly difficult. Furthermore, wind and solar power generation, connected to the grid via converters, possesses only weak inertia, weakening the system's frequency stability support capability. In addition, extreme weather events such as significant fluctuations in wind and solar output and sudden load surges pose threats to the safe operation of the power system from both the source and load sides, further reducing the system's resilience. For example, the 2021 Texas blackout highlighted the risk of severe supply and demand imbalances caused by extreme weather. The UK's "8.9", Australia's "9.28", and Spain and Portugal's "4.28" blackouts verified the frequency security risks under the background of high-proportion renewable energy integration. Therefore, how to construct a power system operation strategy that balances power balance and dynamic frequency security in the context of high-proportion renewable energy and frequent extreme weather is of great significance for guiding the safe and reliable operation of the system.

[0003] Power system time-series operation simulation is an effective means to address the aforementioned challenges. Currently, power balance analysis is typically based on traditional 8760-hour annual time-series operation simulation models, which are embedded in planning to ensure time-series accuracy. However, with the introduction of diverse flexible resources and complex constraints, its massive computational scale faces challenges such as slow convergence speed and memory overflow, limiting its application in relevant scenarios. Meanwhile, traditional methods that only consider power balance can no longer meet the dynamic frequency security requirements of high-proportion RES weak inertia systems. To compensate for this deficiency, some researchers have introduced frequency security constraints into the model, quantifying the system's frequency security boundary through key indicators such as the rate of frequency change and the frequency minimum point. However, such research focuses on economic dispatch, unit combination, and power supply-side configurations, with less attention paid to long-term power balance analysis. Traditional operation simulations neglect the impact of extreme weather and dynamic frequency security risks, and considering these risks makes model solving difficult and inefficient.

[0004] Furthermore, the abnormal coupling of meteorological factors such as temperature, wind speed, and sunlight has become the fundamental driving force behind power source-load fluctuations. The continuous increase in the proportion of renewable energy sources will lead to a significant increase in the sensitivity of both the power source and load sides to weather conditions. Existing model-driven or data-driven generation methods have limited data, making it difficult to cover diverse and continuous extreme scenarios. Historical extreme samples are sparse, and most research methods only perform data correlation without deeply revealing the physical processes by which meteorology cannot directly affect power sources and loads, lacking a deep characterization of the meteorological driving mechanism. Summary of the Invention

[0005] The purpose of this application is to provide a long-term and short-term nested operation simulation method, device, medium and product that takes into account extreme weather and frequency security, so as to solve the problems that traditional operation simulation ignores the impact of extreme weather and dynamic frequency security risks, and the model is difficult and inefficient after considering the above risks, as well as limited by the sparsity of historical extreme samples and lacks a deep characterization of meteorological driving mechanisms.

[0006] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a long-term and short-term nested simulation method that takes into account extreme weather and frequency security, including: Step A: Based on the WGAN-GP model and the meteorological-power mapping model, construct an annual source-load time series considering extreme weather scenarios; the meteorological-power data mapping model includes a wind power model, a photovoltaic power generation model, and a linear time-varying temperature-sensitive load model; Step B: Considering the system frequency response characteristics, establish a year-round time-series operation simulation model that takes into account dynamic frequency safety constraints, and use the year-round source-load time series as the input data for the year-round time-series operation simulation model; Step C: Based on the long-short time nesting and linear unit combination model, accelerate the solution of the annual time-series operation simulation model to obtain the power system operation simulation results.

[0007] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described long-short time nested simulation method that takes into account extreme weather and frequency security.

[0008] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described long-short time nested simulation method that takes into account extreme weather and frequency security.

[0009] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described long-short time nested simulation method that takes into account extreme weather and frequency security.

[0010] According to the specific embodiments provided in this application, this application has the following technical effects: First, this application leverages the powerful data distribution learning and unsupervised generation capabilities of the Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) model to capture the complex nonlinear relationships in meteorological time series. Addressing the scarcity of historical extreme weather data, this method can generate diverse extreme weather scenario sets based on limited historical extreme weather samples, thus overcoming the limitation of the number of historical data samples, significantly expanding the extreme scenario dataset, and ensuring that simulations can cover more potential extreme risk situations.

[0011] Secondly, this application establishes a meteorological-power mapping model, which goes beyond simple data statistical correlation. Instead, based on wind power models, photovoltaic power generation models, and linear time-varying temperature-sensitive load models, it utilizes the physical conversion relationship between meteorological elements (such as wind speed, irradiance, and temperature) and power system operating data (output and load) to transform meteorological data into power system operating data. This process deeply characterizes the physical driving mechanism of meteorological conditions on fluctuations on both the source and load sides, ensuring that the generated source and load data under extreme scenarios conforms to physical laws, thus improving the authenticity and reliability of the data.

[0012] Finally, based on the annual source-load time series including extreme weather conditions generated above, the annual time-series operation simulation model that takes into account dynamic frequency security constraints is used as input. Based on the long-short time nesting and linear unit combination model, the annual time-series operation simulation model is solved more quickly to obtain the power system operation simulation results. This allows the power system operation simulation results to truly reflect the system supply and demand status and frequency response characteristics under extreme weather conditions. This not only improves the accuracy of risk assessment of the power system under extreme weather conditions, but also provides more scientific and reliable data support for formulating operation plans, power planning and energy storage configuration to cope with extreme weather and frequency security. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 Flowchart of the long- and short-time nested simulation method for extreme weather and frequency security provided in this application; Figure 2 A schematic diagram of the long-short time nested execution simulation method provided in this application; Figure 3 This is a schematic diagram of multi-timescale temporal coupling provided in this application; Figure 4 A comparison chart of probability density distributions of the dominant historical meteorological data and generated data provided in this application; Figure 5 The operational curves for different extreme weather scenarios provided in this application; Figure 6 A comparison chart of typical daily operating results for different unit combination methods provided in this application; Figure 7 The diagram shows the daily power balance operation simulation results provided in this application; Figure 8 A comparison chart of frequency indicators before and after considering frequency security constraints provided for this application. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] To make the objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] like Figure 1 As shown, this application provides a method for nested long- and short-time simulation that takes into account extreme weather and frequency security, including: Step A: Based on the WGAN-GP model and the meteorological-power mapping model, construct an annual source-load time series considering extreme weather scenarios; the meteorological-power data mapping model includes a wind power model, a photovoltaic power generation model, and a linear time-varying temperature-sensitive load model.

[0018] Step B: Considering the system frequency response characteristics, establish a year-round time-series operation simulation model that takes into account dynamic frequency safety constraints, and use the year-round source-load time series as the input data for the year-round time-series operation simulation model.

[0019] Step C: Based on the long-short time nesting and linear unit combination model, accelerate the solution of the annual time-series operation simulation model to obtain the power system operation simulation results.

[0020] In an exemplary embodiment, step A specifically includes: Step A1: Generate a set of extreme weather meteorological scenarios based on the WGAN-GP model that incorporates Wasserstein distance and gradient penalty mechanism.

[0021] Step A2: Establish a meteorological-power data mapping model to convert the extreme weather meteorological scenario set into power system operation data under extreme weather conditions; the power system operation data consists of extreme system operation scenarios composed of wind power output, photovoltaic power output and load demand data.

[0022] Step A3: Replace the regular operating data for the corresponding dates in the system's annual operating scenario with the power system operating data to form an annual source-load time series including extreme weather.

[0023] In an exemplary embodiment, in step A2, meteorological data is converted into renewable energy output data through the wind power model and the photovoltaic power generation model, and temperature data is converted into load demand data through the linear time-varying temperature-sensitive load model.

[0024] In practical applications, step A proposes a method for constructing annual source-load time series considering extreme weather scenarios, based on the WGAN-GP model and the meteorological-electricity mapping model.

[0025] Step A specifically includes: Step A1: Extreme Weather Scene Generation Based on WGAN-GP. Generative Adversarial Networks (GANs) can understand the diversity of meteorological data, capture the complex nonlinear relationships of time series, and achieve unsupervised generation of extreme scenes. Considering that GAN models are prone to problems such as mode collapse and gradient vanishing during training, the WGAN-GP model is introduced. Wasserstein distance and gradient penalty are used to make the loss function smoother and more convergent. The specific steps include the following steps.

[0026] Step 1: Define the criteria for extreme weather conditions. To obtain a representative set of extreme weather scenarios, this application first formulates the following criteria for extreme weather scenarios based on the 27 climate extreme indexes defined by the Climate Change Monitoring and Indexing Expert Group.

[0027] In the formula, Indicates whether an extreme scenario has occurred; Extreme weather is the dominant meteorological element; and These are the relative high and low thresholds for the corresponding meteorological elements.

[0028] Subsequently, based on the multi-year meteorological distribution characteristics and power system demand of the study area, different types of extreme weather were defined, and the criteria for judging extreme events were further quantified in conjunction with the criteria in step 1, as shown in Table 1.

[0029] Table 1. Criteria for Determining Extreme Weather

[0030] Step 2: For each type of historical extreme event, meteorological data such as temperature, wind speed, irradiance, and air density are used as real sample inputs. The generator generates candidate samples under random noise input. The discriminator distinguishes between true and false signals and continuously optimizes the two network parameters through backpropagation until the network parameters stabilize and reach Nash equilibrium. The overall objective function of WGAN is shown below.

[0031] In the formula, , The distributions of the real sample x and the noise z are respectively. , These represent the discriminant probabilities for classifying real samples and generated samples, respectively. For z to obey Expectations; For x to obey Expectations; This is a gradient penalty term. The weights represent the gradient penalty. This is the variable used for interpolation between generated samples and real samples.

[0032] After training, WGAN-GP can learn the distribution patterns of different types of extreme events and the coupling characteristics between meteorological elements, and obtain a set of extreme weather scenarios composed of meteorological data.

[0033] Step A2: Mapping Meteorological Data to Power System Operation. After acquiring meteorological data under different extreme weather scenarios, it is necessary to further transform it into data required for power system operation. Therefore, this application establishes a mapping model to map extreme weather scenarios composed of meteorological data into extreme system operation scenarios composed of wind power output, photovoltaic power output, and load demand data. Specifically, this includes the following steps.

[0034] Step 1: Wind power model. The output of the wind turbine mainly depends on the wind speed at the turbine height H. and air density The turbine output for each wind speed range is as follows.

[0035] The wind speed at the fan height H is obtained by the following formula.

[0036] In the formula, This refers to the output power of the fan. This refers to the rated power of the fan. Where A is the wind energy utilization coefficient; A is the effective swept area of ​​the wind turbine blades; h is the reference height for wind speed measurement. is the wind shear coefficient.

[0037] Step 2: Photovoltaic Power Generation Model. The formula for calculating the output power of photovoltaic power generation is as follows, which mainly depends on the solar irradiance R, referred to as downward shortwave radiation at the Earth's surface in meteorology.

[0038] In the formula, This refers to the rated output power of the photovoltaic system. The battery temperature under standard test conditions is 25°C. For the temperature of photovoltaic module cells; The radiation intensity under standard test conditions is 1000 W / m2; This is the temperature coefficient of a photovoltaic module, reflecting its efficiency at different temperatures.

[0039] Step 3: Linear Time-Varying Temperature-Sensitive Load Model. Based on the differences in load sensitivity to temperature, the daily electricity load affected by temperature is decomposed into a baseline load unaffected by temperature fluctuations and a temperature-sensitive load that changes with temperature. Considering the cumulative effect of temperature, the temperature data is corrected using the following correction formula.

[0040] In the formula, This is the temperature correction value for target day i at time t after taking into account the cumulative effect of temperature. The temperature at time t on the day before target day i; is the cumulative effect coefficient; p is the maximum cumulative number of days. Since the intensity of this effect mainly depends on the temperature within three days, p is taken as 2 here. (Based on load...) With equivalent temperature correlation coefficient With the goal of maximizing, iteratively solve for the optimal solution. The solution model is as follows: After correcting the initial temperature data, the least squares method is applied to the load and the corrected temperature in sequence. The influence of holidays is introduced to distinguish between weekdays and weekends, and independent modeling is performed for different times of the day, as shown below.

[0041] In the formula, Let be the load value at time t within the day, where t ranges from 1 to 24; To account for the influence of weekend factors, the characteristic indicators are set to 0 on weekdays and 1 on weekends; , , These are the calculated regression coefficients, representing the baseline load size, load temperature sensitivity, and weekend impact, respectively.

[0042] Step A3: Construction of the annual source-load time series including extreme weather. Based on the extreme operating data generated in Step A2, the regular operating data for the corresponding dates in the system's annual operating scenario are replaced by the extreme operating data according to the month and probability, forming a complete 8760-hour source-load series, thus realizing the construction of the annual operating scenario including extreme weather.

[0043] In an exemplary embodiment, step B specifically includes: Step B1: Construct a system frequency response model based on the rotor motion equation and derive dynamic frequency safety constraints; the dynamic frequency safety constraints include frequency change rate constraints and frequency minimum point constraints.

[0044] Step B2: Establish an optimization model with the objective function of minimizing the total system operating cost. Incorporate the dynamic frequency safety constraints, system power balance constraints, unit operation constraints, and energy storage operation constraints into the optimization model. Determine the annual time-series operation simulation model and use the annual source-load time series as the input data for the annual time-series operation simulation model.

[0045] In practical applications, step B considers the system frequency response characteristics and proposes a time-series operation simulation model that takes into account dynamic frequency security constraints, balancing power balance and frequency security requirements.

[0046] Step B specifically includes: Step B1: System Frequency Response Modeling. Based on the frequency response process, derive the frequency safety constraint formula that satisfies the system requirements. This includes the following steps.

[0047] Step 1: The inertia support and primary frequency regulation response required for system frequency regulation are mainly provided by the synchronous machine and energy storage. To simplify the calculation, the virtual inertia delay of the energy storage is ignored here. The virtual inertia of the energy storage participates in the frequency response the instant the disturbance occurs. The system frequency response is jointly generated by the synchronous machine inertia, the virtual inertia of the energy storage, and the PFR of both. The rotor motion equation and the calculation formula for the system inertia are shown below.

[0048] In the formula, The system inertia is determined by the synchronous machine inertia. and energy storage virtual inertia composition; This is a 0-1 variable representing the unit's start-up and shutdown status (1 represents the start-up status). This refers to the rated capacity of the unit. Let g be the inertial time constant of the unit; Virtual inertia provided for energy storage (ES); , These are the dynamic response power of the energy storage and the synchronous PFR, respectively; This is for frequency deviation; The rated frequency; This refers to unbalanced power.

[0049] Step 2: The frequency security risks faced by the system mainly stem from system power imbalances caused by emergencies such as sudden load increases and generator shutdowns. Frequency security indicators mainly include: 1) Rate of Change of Frequency (RoCoF); 2) Minimum frequency. Based on the rotor motion equations, considering the combined effect of the synchronous machine inertia and the virtual inertia of the energy storage, the following constraints are established.

[0050] (1) RoCoF constraint The rate of frequency change is greatest at the initial moment of the disturbance, which is related to the system's inertia support, and the constraints are shown below.

[0051] In the formula, , These are the actual RoCoF value after the system is disturbed and the maximum RoCoF limit specified by the system.

[0052] (2) Minimum frequency constraint When the system is subjected to a large disturbance, in order to prevent the frequency from dropping below the low-frequency load reduction limit and causing subsequent accidents, the minimum value of the system frequency needs to be limited, as shown in the following constraints.

[0053] In the formula, This is the moment of lowest frequency. This represents the maximum permissible frequency deviation.

[0054] Based on the PFR ramping model of energy storage and synchronizing machine, by integrating both sides of the rotor motion equation and substituting them into the above equation, the constraints are obtained after equivalent transformation as shown below.

[0055] Since the above constraint is nonlinear, it can be transformed into a second-order cone constraint over a time interval t as follows: In the formula, A, B, and C are equivalent variables, and the specific formulas are shown below.

[0056] In the formula, and This refers to the total reserve capacity of the energy storage and synchro unit (PFR). and The ramp-up time constant for the energy storage and synchronizing unit PFR; and This refers to the response delay time of the energy storage and synchronization machine PFR.

[0057] Step B2: Model a year-round time-series operation simulation model considering dynamic frequency security. Dynamic frequency security constraints are introduced into the conventional power balance and various operational constraints to ensure safe system operation.

[0058] Step 1: Objective Function. The model uses the total system operating cost as an example. The optimization approach focuses on minimizing costs, encompassing conventional unit operating costs, load shedding losses, and renewable energy curtailment penalties, to comprehensively analyze the synergistic operational benefits of multi-energy systems. The specific formula is shown below.

[0059] In the formula, t is the index of the corresponding time. This represents 8760 hours per year; G and RE represent the sets of thermal power units and renewable energy units, respectively, with g and re being the subscripts of the corresponding units. , , , These represent the fuel cost, start-up cost, shutdown cost, and startup cost of a thermal power unit, respectively. , These represent the costs of curtailment and load shedding of renewable energy units, respectively. , , These are the output of thermal power units, the power curtailment of renewable energy units, and the power shedding from loads; , , These represent the start-up and stop-down states (0-1 variables), the start-up action (0-1 variables, where 1 represents start-up at time t), and the stop-down action (0-1 variables, where 1 represents stop-down at time t), respectively.

[0060] Step 2: Constraints. To ensure the frequency safety of the system under power disturbances, in addition to satisfying the frequency change rate and minimum frequency constraints established in Step B1, the model also needs to include the following constraints: (1) System power balance constraints In the formula, , These represent the output and load demand of renewable energy units, respectively; ES represents a collection of energy storage units, with es being the subscript of the corresponding unit. , These represent the energy storage charging / discharging power, respectively.

[0061] (2) Unit operating constraints The ramp-up constraints and minimum start / stop time constraints for thermal power units are shown below.

[0062] In the formula, and Adjust the ramp rate (up / down) for thermal power unit g; and The continuous start-up and shutdown time of thermal power units; and This is the minimum continuous start / stop time for thermal power units.

[0063] Thermal power units, in addition to providing energy supply, must provide inertial support and primary frequency regulation. Their output must not exceed the maximum and minimum output limits, and the PFR reserves they provide generally must not exceed their limits. Renewable energy units cannot exceed their actual output upper limit, with the constraints shown in the following formulas.

[0064] In the formula, and These represent the minimum and maximum output of the thermal power unit. For limiting the speed of thermal power unit governors, a value of 8% is generally used; This represents the upper limit of the actual output of renewable energy units.

[0065] (3) Energy storage operation constraints To facilitate collaborative analysis of different types of energy storage operations, a generalized model applicable to multi-timescale energy storage systems was established, as shown below. These represent the energy transfer constraints, upper and lower limits of charge / discharge power constraints, state of charge / discharge constraints, upper and lower limits of capacity constraints, and initial and final energy balance constraints, respectively.

[0066] In the formula, The time granularity for energy storage operation; , , These are energy storage capacity, initial energy storage capacity, and final energy storage capacity, respectively. / Represents the energy storage charging / discharging power; , / For energy storage, the self-release rate and charge / discharge efficiency; / Represents the charge / discharge power state of energy storage; / , / , , These are the upper limit of energy storage charging / discharging power capacity, the lower limit of charging / discharging power capacity, the upper limit of energy storage capacity, and the minimum energy storage ratio.

[0067] Considering that energy storage needs to provide frequency regulation support, additional power and capacity constraints are required. These represent the limits on PFR reserves provided by energy storage, the maximum limit on virtual inertia of energy storage, and the maximum power limit for virtual inertia support, respectively.

[0068] The adjusted upper and lower limits of the energy storage discharge power and the energy transfer constraints are shown below.

[0069] In the formula, Limiting the PFR of energy storage; Frequency modulation duration; Virtual inertia provided for energy storage; The maximum virtual inertia that energy storage can provide; The virtual inertial time constant for energy storage; This is the maximum inertial support power for energy storage.

[0070] In one exemplary embodiment, step C includes: Step C1: Calculate the distance between the sunrise power curves using the dynamic time warping method, and then use the hierarchical clustering algorithm to select a set of typical sunrise scenes.

[0071] Step C2: Based on the typical daily scenario set, and based on the long-short time nesting and linear unit combination model, according to the long-short time operation decoupling constraints, the annual operation simulation is decoupled into a long-time scale power balance simulation and a short-time scale power balance and frequency security simulation, and long-short time operation coupling constraints are established.

[0072] Step C3: Linearize the thermal power unit combination constraints in the short-time scale model within the long-short-time nested and linear unit combination model, and convert the unit start-up state and start-up / shutdown action variables into continuous variables for solution to obtain the power system operation simulation results.

[0073] In an exemplary embodiment, in step C2, the long-short-term operational coupling constraint includes: The output power of thermal power and renewable energy sources is limited to be consistent with the corresponding natural day for each day under extreme weather scenarios.

[0074] The energy storage capacity at the beginning of an extreme scenario is equal to the capacity of the previous day in the extreme scenario, and the capacity at the end of each day in the extreme weather scenario is equal to the capacity of the corresponding day.

[0075] In an exemplary embodiment, in step C3, the linearization process specifically includes: The on / off capacity is used to replace the start-up state and start-stop action variables of the thermal power unit. The inertia provided by the thermal power unit is obtained through online capacity calculation, and the online capacity of the unit at the initial moment of the current day is closely coupled with the long-term operating state.

[0076] In practical applications, step C proposes a simulation-based accelerated solution method based on long-short time nesting and linear unit combination models to achieve efficient model solving.

[0077] Step C specifically includes: Step C1: Typical daily screening based on DTW. The dynamic time warping method can align key feature points of different sequences through nonlinear scaling of the time axis, effectively mitigating the impact of peak-valley misalignment.

[0078] Step 1: Calculate the distance between the daily load and renewable energy output curves using a dynamic time warping algorithm. Assume there are Nm daily curves in month m, and each curve can be represented as: In the formula, This represents the load or new energy output value at time t on day i; the value of T is 24, representing the intraday time resolution.

[0079] The DTW distance between any two diurnal curves Xi and Xj is defined as follows: In the formula, W is the winding path that satisfies the constraints of boundary, continuity and monotonicity, which is used to allow nonlinear alignment of the time axis, and K is the length of the path.

[0080] Define cumulative distance function The optimal path W is the minimum cumulative distance from the starting point (1,1) to the point (p,q). It can be obtained through the following dynamic programming recursive formula: in The final DTW distance is .

[0081] Step 2: Select typical days month-by-month using hierarchical clustering algorithm. Using the DTW distance matrix obtained month-by-month in Step 1, construct a cluster tree using hierarchical clustering algorithm. Evaluate different clustering results using silhouette coefficients to determine the optimal number of clusters. Select the real day curve with the smallest average distance to other curves within its cluster as the typical day and assign it a weight coefficient: In the formula, The weight representing a typical day of type c; This represents the number of days included in this category. The resulting typical days retain the temporal correlation between wind and solar power output and load demand, and correspond to the natural days in the annual operation scenario, ensuring the establishment of the long-term and short-term operation coupling relationship.

[0082] Step C2: Long-term and short-term decoupling. For example... Figure 2 As shown, based on the generated annual operating scenario, the 8760-hour annual operating simulation is decoupled into a long-term daily power balance simulation and a short-term hourly intraday power balance simulation for typical days and extreme weather scenarios. Simultaneously, by embedding frequency security constraints at the short-term time scale, the system's supply and demand balance across multiple time scales and its frequency support capability under power disturbances are ensured.

[0083] (1) System operating costs Step 1: Reconstruct system operating costs into long-term operating costs Operating costs for typical days and extreme weather , The weighted average of the three factors is shown in the formula below.

[0084] In the formula, w is the long-term operating cost weighting factor, which can be adjusted to adapt to the simulation requirements under different conditions.

[0085] In this model, a typical day can be equated to an extreme weather scenario with a duration of 1 day, and its operating costs and variables are similar in meaning to those of extreme weather. The long-term operating costs and extreme weather operating costs in the objective function are as follows.

[0086] In the formula, D and E represent the set of extreme weather scenarios and the set of natural days within the year, respectively, and the subscripts d and e represent the corresponding set elements. This refers to the installed capacity of thermal power units; , , These are the daily utilization hours of thermal power units, the amount of electricity abandoned by renewable energy units, and the load shedding amount, respectively. This represents an extreme weather scenario, where t is the index of the corresponding time.

[0087] (2) Long-term and short-term operational decoupling constraints The system power balance constraint in step B2 is decoupled into a horizontal long-term power balance constraint and a short-term power balance constraint.

[0088] In the formula, Total number of hours in a day; Electricity generated by renewable energy units; This represents the daily load.

[0089] Over long timescales, the system achieves power balance through a combination of thermal power generation and renewable energy generation, along with energy storage to store or release electricity, with constraints as shown below. The operating status of thermal power units is constrained by annual and daily utilization hours.

[0090] In the formula, This is the upper limit for the daily utilization hours of thermal power units; This is the upper limit for the power generation of renewable energy units.

[0091] Energy storage operation constraints are established based on an energy storage generalization model, with electricity transfer on a daily basis. On a short timescale, unit operation constraints and frequency safety constraints remain consistent with previous standards.

[0092] (3) Long-term and short-term operation coupling constraints To ensure the consistency of simulations across multiple time scales, the operational status of typical days, extreme weather scenarios, and natural days needs to be constrained through the establishment of coupling relationships using key variables. Taking long-term energy storage as an example, the sum of its net daily charge and discharge should equal the charge and discharge of the corresponding natural day during medium- to long-term operation, as illustrated in the diagram below. Figure 3 As shown. Specifically, it is necessary to limit the output power of thermal power and renewable energy sources to be consistent with the corresponding natural day during extreme weather scenarios.

[0093] In the formula, t represents the time within a natural day d; For extreme weather scenario e, there is a set of natural days. A typical day can be obtained by simplifying the time scale to 1 day.

[0094] The energy storage capacity at the initial moment of the extreme scenario is equal to the capacity of the previous natural day of the extreme scenario, and the capacity at the end of each day in the extreme weather scenario is equal to the capacity of the corresponding natural day, with the constraints as follows.

[0095] In the formula, t represents the time within a natural day d; , Let i be the i-th natural day and the number of natural days within extreme weather scenario e, respectively. The energy storage capacity of the energy storage unit es at the initial moment of extreme weather scenario e.

[0096] Step C3: Linearization Method for Unit Combination Model. The unit combination constraints in the short-term timescale model are linearized. The start-up status and start-up / shutdown actions of thermal power units in the typical daily and extreme weather operating costs are replaced by the on / off capacity, and uniformly modeled as continuous variables. A long-short coupled constraint is constructed to connect the long-short timescale models, significantly improving the solution speed while maintaining accuracy.

[0097] The inertia provided by the thermal power unit can be obtained through online capacity calculation; the upper and lower limits of unit output, ramping constraints, minimum start / stop time constraints, and primary frequency regulation reserve limits are as follows.

[0098] The capacity change caused by start-up and shutdown operations is represented by the unit's online capacity. The start-up / shutdown operation process of thermal power units is modeled as follows.

[0099] In the simplified unit configuration, the initial online capacity of each unit must be closely coupled with its long-term capacity configuration. The long-term and short-term operational coupling constraints must also include the following formula to ensure that the initial online capacity of the units on the current day is closely coupled with their long-term operational status.

[0100] In the formula, This represents the minimum power output rate of a thermal power unit. , , This represents the online capacity and start / stop capacity of the thermal power unit at time t under extreme scenarios. The start-up capacity is the capacity of the unit that starts up at time t, and the stop-down capacity is the capacity of the unit that stops at time t.

[0101] To address the dual risks of frequency instability and supply-demand imbalance faced by power systems with low inertia and frequent extreme weather events due to high renewable energy penetration, this application proposes a long-term and short-term nested operation simulation method that considers both extreme weather and frequency security. First, to address the sparsity of historical extreme weather data, a Generative Adversarial Network (WGAN-GP) with Wasserstein distance and gradient penalty mechanisms is used to expand the extreme weather scenario set. A year-round source-load time series including extreme weather is constructed through a meteorological-power data mapping model. Second, a year-round time-series operation simulation model with embedded dynamic frequency security constraints is established, balancing supply-demand balance and frequency stability. Finally, to reduce model size, an accelerated solution method combining long-term and short-term nesting and linear unit combinations is proposed, dividing the year-round simulation into long-term power balance and short-term power balance and frequency security problems for solution. Results show that this application can significantly improve solution efficiency while maintaining computational accuracy, providing scientific guidance for the safe dispatch and planning of power systems.

[0102] The following section will provide further explanation using specific data.

[0103] The time-series simulation method proposed in this application, based on historical meteorological data and system operation data of a certain province, was optimized and solved using MATLAB 2023b and the Gurobi-11.0.3 solver on a computer equipped with an Intel® Core i7-12800HX central processor and 32GB of memory.

[0104] The method proposed in step A is used to extend the generation of extreme scenarios. The probability density comparison results between the data generated by the WGAN-GP model and historical data under different extreme events are as follows: Figure 4 As shown in the figure, the generated data exhibits a high degree of consistency with historical samples in terms of overall distribution. After transformation using a mapping model, the system operation curve under extreme weather conditions is obtained as shown in the figure. Figure 5 As shown, during periods of extreme heat and no wind, load demand is high while wind power output is low; during periods of cloudy and rainy weather with no sunlight, photovoltaic power remains sluggish; wind power increases significantly during cold waves; and wind power exhibits fluctuations with sudden drops and increases during typhoons. Therefore, the scenario generation method proposed in this application can fully reflect the supply and demand characteristics and operating status under various extreme weather conditions, providing basic data support for subsequent operational simulations.

[0105] On non-extreme weather days, DTW combined with hierarchical clustering algorithm was used to select typical days. To verify the effectiveness of the clustering method proposed in step C, different clustering schemes were set up and the silhouette coefficient (SC) and Davidson-Bolding index (DBI) of different clustering schemes were compared, as shown in Table 2.

[0106] Table 2 Results of different clustering schemes and indicators

[0107] Analysis shows that the clustering method proposed in this application achieves an SC value of 0.78 and a DBI value of 0.82, indicating that this method performs best in terms of cluster density and inter-cluster separation. The selected typical days thus represent the output and load characteristics of non-extreme weather days throughout the year, while maintaining a complementary relationship with extreme scenarios.

[0108] The proposed long- and short-time nested operation simulation method, which considers extreme weather and frequency security, was tested on an improved IEEE-39 node system. This system comprises six thermal power plants totaling 1800MW, one 500MW hydropower station, one 800MW wind farm, one 600MW photovoltaic power station, and a 400MW energy storage system. The continuous discharge time was set to 10 hours. The peak system load for the planned horizontal year was 2500MW. All methods employed a linear unit combination approach. Table 3 shows the key parameter settings, with 10% of the load in each time period treated as a power disturbance.

[0109] Table 3 Relevant technical parameters for system operation

[0110] Five sets of methods were compared and analyzed. Details are shown in Table 4. Method II simulates multi-timescale operation of energy storage by connecting typical daily scenarios, while Scheme V is the method proposed in this application.

[0111] Table 4 Comparison Method Settings

[0112] Table 5 Optimization results of different methods

[0113] The optimization results of different methods are shown in Table 5. Method III, based on Method II, adopts the method of this application, maintaining accuracy while reducing the solution time from 176.62 seconds to 30.26 seconds, significantly improving computational efficiency. After incorporating frequency security constraints, Method V has an increased number of variables and constraints compared to Method III, and the solution time is slightly longer. However, compared to Method IV, the model size is significantly reduced, and the solution efficiency is significantly improved. Meanwhile, the system operating cost of Method V increases by 12.8% compared to Method III, indicating that frequency security constraints affect power generation plans.

[0114] Based on Method III, a comparison is made using a traditional unit combination method with binary 0-1 variables to verify the feasibility of the unit combination linearization method in step C. Table 6 shows the optimization results for both methods.

[0115] Table 6 Comparison of optimization results for different unit combination methods

[0116] As shown in the table, the objective functions of the two methods are basically the same, indicating high overall consistency. Although the total number of variables increases slightly, the number of binary variables is significantly reduced, shortening the solution time from 328.25 seconds to 30.26 seconds, an efficiency improvement of approximately 10 times. Simulation results of the two methods on some typical days are as follows: Figure 6 As shown in the figure, the output and SOC of thermal power units in a typical day are similar, indicating that the linearization method can significantly improve the solution efficiency while preserving the original operating mode of the system to the greatest extent.

[0117] Figure 7 The simulation results of the annual daily power balance operation of Method V are presented. It can be seen that the power generation of renewable energy is higher in summer, followed by spring, and lower in autumn and winter, showing obvious seasonal fluctuation characteristics. Figure 8The data illustrates the magnitudes of various frequency parameters for some typical intraday systems. After considering frequency constraints, the lowest frequency point remains approximately 49.5 Hz, and the RoCoF values ​​are all less than 0.5 Hz / s, both remaining within the set safety limits. Before considering frequency constraints, however, the frequency safety parameters all exceeded the set limits. Therefore, introducing frequency safety constraints effectively ensures the dynamic frequency safety of the system. Thus, the accelerated solution method proposed in this application can guarantee system frequency safety within controllable errors, achieving efficient solution to the optimization problem.

[0118] In summary, the operational simulation method proposed in this application can effectively solve the operational simulation model with embedded frequency security, and ensures the accuracy of the solution while significantly improving the solution efficiency, thus providing reliable support for power system planning and energy storage configuration optimization.

[0119] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments. The computer device can be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (I / O), and a communication interface. The processor, memory, and I / O interface are connected via a system bus, and the communication interface is connected to the system bus via the I / O interface. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device stores data to be processed. The I / O interface of the computer device is used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communicating with an external terminal via a network connection. When the computer program is executed by the processor, it implements the above-described methods.

[0120] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0121] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0122] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0123] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by hardware related to computer program instructions. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0124] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0125] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0126] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A nested long- and short-time simulation method that takes into account extreme weather and frequency security, characterized in that, include: Step A: Based on the WGAN-GP model and the meteorological-electricity mapping model, construct the annual source-load time series considering extreme weather scenarios; The meteorological-power data mapping model includes a wind power model, a photovoltaic power generation model, and a linear time-varying temperature-sensitive load model; Step B: Considering the system frequency response characteristics, establish a year-round time-series operation simulation model that takes into account dynamic frequency safety constraints, and use the year-round source-load time series as the input data for the year-round time-series operation simulation model; Step C: Based on the long-short time nesting and linear unit combination model, accelerate the solution of the annual time-series operation simulation model to obtain the power system operation simulation results.

2. The long-term and short-term nested simulation method considering extreme weather and frequency security according to claim 1, characterized in that, Step A specifically includes: Step A1: Generate a set of extreme weather meteorological scenarios based on the WGAN-GP model that incorporates Wasserstein distance and gradient penalty mechanism; Step A2: Establish a meteorological-power data mapping model to convert the extreme weather meteorological scenario set into power system operation data under extreme weather conditions; the power system operation data consists of extreme system operation scenarios composed of wind power output, photovoltaic power output, and load demand data; Step A3: Replace the regular operating data for the corresponding dates in the system's annual operating scenario with the power system operating data to form an annual source-load time series including extreme weather.

3. The long-term and short-term nested simulation method considering extreme weather and frequency security according to claim 2, characterized in that, In step A2, meteorological data is converted into renewable energy output data through the wind power model and photovoltaic power generation model, and temperature data is converted into load demand data through the linear time-varying temperature-sensitive load model.

4. The long-term and short-term nested simulation method considering extreme weather and frequency security according to claim 1, characterized in that, Step B specifically includes: Step B1: Construct a system frequency response model based on the rotor motion equations and derive dynamic frequency safety constraints; the dynamic frequency safety constraints include frequency change rate constraints and frequency minimum point constraints; Step B2: Establish an optimization model with the objective function of minimizing the total system operating cost. Incorporate the dynamic frequency safety constraints, system power balance constraints, unit operation constraints, and energy storage operation constraints into the optimization model. Determine the annual time-series operation simulation model and use the annual source-load time series as the input data for the annual time-series operation simulation model.

5. The long-term and short-term nested simulation method considering extreme weather and frequency security according to claim 1, characterized in that, Step C includes: Step C1: Calculate the distance between the sunrise power curves using the dynamic time warping method, and then use the hierarchical clustering algorithm to select a set of typical daily scenes; Step C2: Based on the typical daily scenario set, and based on the long-short time nesting and linear unit combination model, according to the long-short time operation decoupling constraints, the annual operation simulation is decoupled into a long-time scale power balance simulation and a short-time scale power balance and frequency security simulation, and long-short time operation coupling constraints are established. Step C3: Linearize the thermal power unit combination constraints in the short-time scale model within the long-short-time nested and linear unit combination model, and convert the unit start-up state and start-up / shutdown action variables into continuous variables for solution to obtain the power system operation simulation results.

6. The long-term and short-term nested simulation method considering extreme weather and frequency security according to claim 5, characterized in that, In step C2, the long-term and short-term operational coupling constraints include: The output power of thermal power and renewable energy sources during each day in extreme weather scenarios shall be limited to be consistent with the corresponding natural day. The energy storage capacity at the beginning of an extreme scenario is equal to the capacity of the previous day in the extreme scenario, and the capacity at the end of each day in the extreme weather scenario is equal to the capacity of the corresponding day.

7. The long-term and short-term nested simulation method considering extreme weather and frequency security according to claim 5, characterized in that, In step C3, the linearization process specifically includes: The on / off capacity is used to replace the start-up state and start-stop action variables of the thermal power unit. The inertia provided by the thermal power unit is obtained through online capacity calculation, and the online capacity of the unit at the initial moment of the current day is closely coupled with the long-term operating state.

8. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the long-short time nested simulation method taking into account extreme weather and frequency security as described in any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the long-short nested simulation method that takes into account extreme weather and frequency security as described in any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the long-short nested simulation method that takes into account extreme weather and frequency security as described in any one of claims 1-7.

Images