Weather-Adaptive Multi-Dimensional Linkage Control Method and System for Wind, Solar, EV, and Energy Storage

Abstract

This invention provides a weather-adaptive multi-dimensional linkage regulation method and system for wind-solar-EV-energy storage, relating to the field of new energy and electric vehicle linkage regulation technology. The method includes the following steps: acquiring wind and solar power output data and real-time weather information; performing two-stage clustering processing on the wind and solar power output data, first clustering photovoltaic data and then wind power data, to generate cluster labels; constructing an electric vehicle (EV) travel characteristic model, and correcting vehicle driving energy consumption by integrating travel time probability distribution, fitting the relationship between temperature and battery capacity decay, and introducing a weather correction coefficient to predict charging demand and quantify V2G regulation potential; calculating real-time fluctuation levels based on wind and solar power output data, and constructing a three-dimensional scene identifier based on weather information, fluctuation levels, and cluster labels; and introducing a scene strategy switching mechanism based on the three-dimensional scene identifier to match differentiated wind-solar-EV-energy storage regulation strategies aimed at renewable energy consumption, power fluctuation mitigation, and power supply assurance.

Description

Technical Field

[0001] This invention relates to the field of new energy and electric vehicle linkage control technology, and in particular to a weather-adaptive multi-dimensional linkage control method and system for wind-solar-EV-energy storage. Background Technology

[0002] Driven by the global energy transition and dual-carbon goals, the distributed grid connection and consumption of renewable energy sources such as photovoltaics and wind power have become a core issue. However, their inherent intermittency and randomness have led to prominent source-load time mismatch problems, resulting in energy waste.

[0003] However, while existing energy regulation solutions combine energy storage and EV regulation, they focus on a single energy scenario, fail to consider weather and clustering results, and ignore the impact of the environment on battery capacity. Other solutions employ fixed-parameter clustering, which cannot adapt to the characteristics of wind and solar data under different weather conditions and lack scenario-differentiated strategies. As the number of EVs increases, their regulation potential is not effectively coupled with wind and solar characteristics and user needs, making it difficult to meet the multi-objective demands under high-penetration wind and solar access. Summary of the Invention

[0004] The purpose of this invention is to provide a weather-adaptive multi-dimensional linkage regulation method and system for wind-solar-EV-energy storage, in order to solve the problem mentioned in the background art that the existing energy regulation schemes have not effectively coupled with the characteristics of wind and solar power and user needs, making it difficult to meet the multi-objective requirements under high-penetration wind and solar power access.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a weather-adaptive multi-dimensional linkage regulation method for wind-solar-EV-energy storage, comprising the following steps: acquiring wind and solar power output data and real-time weather information; performing two-stage clustering processing on the wind and solar power output data, first photovoltaic clustering and then wind power clustering, to generate cluster labels; constructing an electric vehicle (EV) travel characteristic model, by integrating the probability distribution of travel time, fitting the temperature-battery capacity decay relationship using piecewise multinomial fitting, and introducing a weather correction coefficient to correct vehicle driving energy consumption, in order to predict the spatiotemporal distribution of EV charging demand and quantify V2G regulation potential; calculating the real-time fluctuation level based on the wind and solar power output data, and constructing a weather-fluctuation-cluster three-dimensional scene identifier based on the real-time weather information, the real-time fluctuation level, and the cluster labels; and introducing a scene strategy switching mechanism based on the weather-fluctuation-cluster three-dimensional scene identifier to match differentiated wind-solar-EV-energy storage regulation strategies aimed at renewable energy consumption, power fluctuation mitigation, and power supply assurance.

[0006] Optionally, the step of performing a two-stage clustering process on the wind and solar power output data, first clustering photovoltaic data and then clustering wind power data, to generate cluster labels specifically includes: using the Self-Organizing Map (SOM) algorithm to perform a first-stage clustering of the photovoltaic power output data in the wind and solar power output data to generate periodic photovoltaic power output cluster labels; using the Density-Based Clustering (DBSCAN) algorithm to perform a second-stage clustering of the wind power output data in the wind and solar power output data after the first-stage clustering to generate wind power output cluster labels for the corresponding period; and combining the photovoltaic power output cluster labels and the wind power output cluster labels to generate photovoltaic-dominated cluster labels, solar-wind fluctuation cluster labels, wind-solar attenuation cluster labels, and wind power-dominated cluster labels.

[0007] Optionally, the step of introducing a weather correction coefficient to correct vehicle driving energy consumption specifically includes: constructing a benchmark model reflecting the nonlinear changes in energy consumption of electric vehicles in different speed ranges; and optimizing the benchmark model by introducing a weather correction coefficient, wherein the weather correction coefficient is obtained by jointly analyzing EV energy consumption deviation and wind and solar fluctuation data, and calibrating using statistical confidence intervals, and its calculation formula is as follows: In the formula: Energy consumption coefficient This is a weather correction factor. This refers to energy loss caused by low motor efficiency in the low-speed range. This represents the linear relationship between energy consumption and speed in the medium-speed range. The energy consumption is dominated by the quadratic increase in air resistance at high speeds. This is a constant term, representing the basic energy consumption independent of speed.

[0008] Optionally, the step of calculating the real-time fluctuation level based on the wind and solar power output data specifically includes: the wind and solar power output data includes real-time wind and solar power generation data; based on the real-time wind and solar power generation data, a sliding window with a preset time is used to calculate the real-time power fluctuation coefficient, the calculation formula of which is: In the formula: This is the real-time power fluctuation coefficient. This represents the current real-time power. This represents the real-time power of the previous time window. The rated power is used as the reference value. Based on the average value of λ over several consecutive windows, the following levels are defined: if λ ≤ 5%, it is classified as a low fluctuation level; if 5% < λ ≤ 15%, it is classified as a medium fluctuation level; if λ > 15%, it is classified as a high fluctuation level.

[0009] Optionally, the step of switching the differentiated wind-solar-EV-energy storage control strategy based on the weather-fluctuation-cluster three-dimensional scene identifier with the goals of renewable energy consumption, power fluctuation mitigation, and power supply guarantee specifically includes: under the sunny-low fluctuation-solar-dominated clustering label mode, prioritizing electric vehicle (EV) charging according to the value coefficient K with the goal of maximizing wind and solar energy consumption; under the cloudy-medium fluctuation-solar-wind fluctuation clustering label mode, calling battery energy storage systems (BESS) and electric vehicles (EVs) within the first preset condition range to participate in mobile energy storage resource scheduling (V2G) with the goal of mitigating power fluctuations; under the cloudy-medium / high fluctuation-wind-dominated clustering label mode, calling battery energy storage systems (BESS) and electric vehicles (EVs) within the second preset condition range to participate in mobile energy storage resource scheduling (V2G) with the goal of stabilizing power supply; under the snow / sandstorm-high fluctuation-wind-solar attenuation clustering label mode, initiating emergency scheduling and calling battery energy storage systems (BESS) and electric vehicles (EVs) within the third preset condition range to participate in mobile energy storage resource scheduling (V2G); wherein, the first preset condition range < the second preset condition range < the third preset condition range.

[0010] Optionally, the scenario strategy switching mechanism specifically includes: starting 24 hours in advance, dividing the prediction period into a preset time window, performing lightweight clustering on the wind and solar power subsequences of each period, calculating the dynamic time warping similarity η between each subsequence and the basic clustering pattern, generating specific time period labels when η < 0.85, and using the basic clustering labels when η ≥ 0.85, forming a 24-hour clustering sequence containing basic labels and specific labels; and introducing a judgment coefficient based on the clustering sequence. The judgment coefficient comprehensively characterizes the fluctuation of wind and solar power over the next 24 hours. The calculation is performed by weighting the standard deviation and mean of photovoltaic and wind power outputs. When it is determined to be a high-fluctuation scenario, When it is determined to be a medium fluctuation scenario, A value ≤0.2 is considered a low-fluctuation scenario; the degree of agreement between the actual power sequence and the pre-simulated clustering labels is quantified using dynamic time warping similarity η. When the value is ≥0.8, the actual power is considered to be highly matched with the pre-simulated clustering, and the deviation is fine-tuned. When 0.6≤ When the value is less than 0.8, a local shift in the clustering pattern is determined, triggering a sub-policy reconstruction. When the value is less than 0.6, a fundamental change in the clustering pattern is determined, and a clustering transition is initiated. During strategy switching, a dynamic inertia coefficient γ based on the clustering transition intensity θ is introduced. By weighted fusion of the output power of the original strategy and the target strategy, a smooth transition of power commands is achieved. The value of the inertia coefficient γ and the power adjustment range limit are adjusted according to the fluctuation scenario.

[0011] Optionally, the strategy switching formula is: In the formula: For the output power of the new strategy, The inertia coefficient, The output power of the original strategy, The output power of the target strategy.

[0012] On the other hand, the present invention also provides a weather-adaptive wind-solar-EV-energy storage multi-dimensional linkage control system, comprising: an acquisition module for acquiring wind and solar power output data and real-time weather information; a clustering module for performing a two-stage clustering process on the wind and solar power output data, first photovoltaic clustering and then wind power clustering, to generate clustering labels; a model building module for constructing an electric vehicle (EV) travel characteristic model, which integrates the probability distribution of travel time, fits the temperature-battery capacity decay relationship using piecewise multinomial fitting, and introduces a weather correction coefficient to correct vehicle driving energy consumption, so as to predict the spatiotemporal distribution of EV charging demand and quantify V2G regulation potential; a three-dimensional scene identification construction module for calculating the real-time fluctuation level based on the wind and solar power output data, and constructing a weather-fluctuation-cluster three-dimensional scene identification based on the real-time weather information, the real-time fluctuation level, and the clustering labels; and a control strategy switching module for introducing a scene strategy switching mechanism based on the weather-fluctuation-cluster three-dimensional scene identification, matching differentiated wind-solar-EV-energy storage control strategies aimed at renewable energy consumption, power fluctuation smoothing, and power supply guarantee.

[0013] On the other hand, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described weather-adaptive wind-solar-EV-energy storage multi-dimensional linkage control method.

[0014] On the other hand, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described weather-adaptive wind-solar-EV-energy storage multi-dimensional linkage control method.

[0015] Compared with the prior art, the beneficial effects of the present invention are:

[0016] This application obtains wind and solar power output data and adopts a sequential clustering approach, first solar power and then wind power, to fully explore the coupling and correlation characteristics of wind and solar power output in the time dimension. By adapting the optimal clustering algorithm to different energy characteristics, it overcomes the shortcomings of single clustering algorithms in adapting to mixed energy data, and provides a core basis for constructing accurate scene identification.

[0017] This application constructs a multi-dimensional EV characteristic model that integrates travel patterns, temperature decay, and weather effects, enabling precise quantitative assessment of electric vehicle charging demand and V2G regulation potential. This transforms electric vehicles from unpredictable simple loads into intelligent mobile energy storage units that can be accurately predicted and flexibly dispatched. By quantifying their charging demand and regulation potential, it provides valuable flexible load-side resources for system dispatch, greatly enhancing the grid's ability to absorb and balance renewable energy fluctuations.

[0018] This application integrates weather type, power fluctuation level and clustering label in multiple dimensions to form a comprehensive scene identifier that can fully reflect the system's operating status. This breaks through the limitations of single-dimensional criteria and enables the system to fully understand the current and short-term operating status, providing a precise matching basis for differentiated energy regulation strategies.

[0019] This application uses a differentiated strategy matching mechanism based on three-dimensional scene identification, which enables the system to automatically select the optimal control target according to the actual operating status. Combined with smooth switching technology, it ensures the stability of strategy transition and ultimately achieves multi-objective collaborative optimization of efficient renewable energy consumption, stable grid operation and user demand satisfaction. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the method steps of the present invention.

[0021] Figure 2 This is a flowchart illustrating the overall process of the method of the present invention.

[0022] Figure 3 This is a flowchart of the scheduling process for the sunny weather mode of the present invention.

[0023] Figure 4 This is a flowchart of the multi-cloud mode scheduling process of the present invention.

[0024] Figure 5 This is a flowchart of the scheduling process for the rainy weather mode of the present invention.

[0025] Figure 6 This is a schematic diagram of the system structure of the present invention.

[0026] In the diagram: 10 - Acquisition module, 20 - Clustering module, 30 - Model building module, 40 - 3D scene identification building module, 50 - Control strategy switching module. Detailed Implementation

[0027] The present invention will now be clearly and completely described in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be used interchangeably where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] Those skilled in the art will understand that, unless explicitly stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in the specification of this application means the presence of features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0030] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0031] It should be understood that the sequence number and size of each step in this embodiment do not imply the order of execution. The execution order of each process is determined by its function and internal logic, and should not constitute any limitation on the implementation process of this application embodiment.

[0032] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0033] Please refer to Figures 1-5This invention discloses a weather-adaptive multi-dimensional linkage regulation method for wind-solar-EV-energy storage, comprising the following steps: acquiring wind and solar power output data and real-time weather information; performing a two-stage clustering process on the wind and solar power output data, first clustering photovoltaic power and then wind power power, to generate cluster labels; constructing an electric vehicle (EV) travel characteristic model, which predicts the spatiotemporal distribution of EV charging demand and quantifies V2G regulation potential by integrating travel time probability distribution, fitting the temperature-battery capacity decay relationship using piecewise multinomial fitting, and introducing a weather correction coefficient to correct vehicle driving energy consumption; calculating real-time fluctuation levels based on the wind and solar power output data, and constructing a weather-fluctuation-cluster three-dimensional scene identifier based on the real-time weather information, the real-time fluctuation levels, and the cluster labels; and introducing a scene strategy switching mechanism based on the weather-fluctuation-cluster three-dimensional scene identifier to match differentiated wind-solar-EV-energy storage regulation strategies aimed at renewable energy consumption, power fluctuation mitigation, and power supply assurance.

[0034] This application obtains wind and solar power output data and adopts a sequential clustering approach, first solar power and then wind power, to fully explore the coupling and correlation characteristics of wind and solar power output in the time dimension. By adapting the optimal clustering algorithm to different energy characteristics, it overcomes the shortcomings of single clustering algorithms in adapting to mixed energy data, and provides a core basis for constructing accurate scene identification.

[0035] This application constructs a multi-dimensional EV characteristic model that integrates travel patterns, temperature decay, and weather effects, enabling precise quantitative assessment of electric vehicle charging demand and V2G regulation potential. This transforms electric vehicles from unpredictable simple loads into intelligent mobile energy storage units that can be accurately predicted and flexibly dispatched. By quantifying their charging demand and regulation potential, it provides valuable flexible load-side resources for system dispatch, greatly enhancing the grid's ability to absorb and balance renewable energy fluctuations.

[0036] This application integrates weather type, power fluctuation level and clustering label in multiple dimensions to form a comprehensive scene identifier that can fully reflect the system's operating status. This breaks through the limitations of single-dimensional criteria and enables the system to fully understand the current and short-term operating status, providing a precise matching basis for differentiated energy regulation strategies.

[0037] This application uses a differentiated strategy matching mechanism based on three-dimensional scene identification, which enables the system to automatically select the optimal control target according to the actual operating status. Combined with smooth switching technology, it ensures the stability of strategy transition and ultimately achieves multi-objective collaborative optimization of efficient renewable energy consumption, stable grid operation and user demand satisfaction.

[0038] In some embodiments, the step of performing a two-stage clustering process on the wind and solar power output data, first clustering photovoltaic data and then clustering wind power data, to generate cluster labels specifically includes: using the self-organizing map (SOM) algorithm to perform a first-stage clustering of the photovoltaic power output data in the wind and solar power output data to generate periodic photovoltaic power output cluster labels; using the density-based clustering (DBSCAN) algorithm to perform a second-stage clustering of the wind power output data in the wind and solar power output data after the first-stage clustering to generate wind power output cluster labels for the corresponding period; and combining the photovoltaic power output cluster labels and the wind power output cluster labels to generate photovoltaic-dominated cluster labels, solar-wind fluctuation cluster labels, wind-solar attenuation cluster labels, and wind power-dominated cluster labels.

[0039] Specifically, in the first stage of clustering, SOM (Signal-Oriented Model) is used to cluster photovoltaic (PV) power. Considering the significant periodicity of PV output and its power being primarily affected by weather conditions, exhibiting significant differences across different scenarios, the first stage of the two-stage clustering process prioritizes clustering analysis of PV output data. First, wavelet transform is used to decompose the PV power signal to 5 layers, retaining low-frequency effective components and removing high-frequency noise. Then, Kalman filtering is used for iterative correction to ensure that the signal-to-noise ratio of the input clustering data is improved to above 30dB. Its state equation is: The observation equation is The Self-Organizing Map (SOM) algorithm is an unsupervised neural network algorithm that simulates the self-organizing properties of neurons in the human brain to map high-dimensional data into a low-dimensional grid while preserving its topological structure. SOM is particularly useful in the initial stages of data processing. SOM is a single-layer neural network containing only an input layer and a computation layer.

[0040] SOM clustering includes the following steps:

[0041] Step 1, Initialize learning parameters Neighborhood and network weight The network weights are specifically the input samples. With output neurons The weight vector between them.

[0042] Step 2, Input learning samples .

[0043] Step 3, for each output layer neuron Calculate the Euclidean distance between each learning sample and the output neuron:

[0044] ;

[0045] Step 4: Select the neuron with the smallest distance as the winning neuron, and update the weight vector of the winning neuron and its neighboring neurons:

[0046] ;

[0047] Step 5, update the learning rate and neighborhood radius parameters:

[0048] ;

[0049] ;

[0050] in, and These are the initial learning rate and the initial neighborhood radius, and int[·] is the floor function. It is the current iteration number. It is a constant. It represents the total number of iterations.

[0051] Step 6: If the termination condition is not met, repeat from step 2 until the termination condition is met.

[0052] To achieve better clustering results, this paper uses the elbow method to determine the optimal number of clusters. The evaluation metric used in elbow clustering is the square of the sum of the distances from all sample points in the dataset to their central clusters.

[0053] Based on the clustering results of the first stage, the wind power samples are classified according to their date correspondence. Then, DBSCAN is used to perform a second stage of clustering on the wind power output samples of each category. DBSCAN is a density-based clustering algorithm that can discover clusters of arbitrary shapes and effectively identify outliers. Before clustering, the neighborhood radius needs to be determined. and minimum sample size .

[0054] DBSCAN clustering includes the following steps:

[0055] Step 1, Input Data Input neighborhood radius and minimum sample size All samples were marked as "unvisited".

[0056] Step 2, for each sample ,if If not visited, mark it. Visited. Calculate. of The neighborhood sample set, that is, in the... Centered on, The set of all samples within a space of radius is given by the following formula:

[0057] ;

[0058] in, , representing the distance between samples. The dimension of the sample.

[0059] if ,but Points that are not core points are marked as noise points, including both genuine noise points and misclassified boundary points; if ,but It is the core point, and Add it to the queue of clusters to be expanded.

[0060] Step 3: Take a sample from the cluster queue to be expanded. ,calculate .if Then traverse Each sample in ,if If not visited, mark it as visited and add it to the queue of clusters to be expanded; if If it is marked as noise, it will be assigned to the current cluster.

[0061] Step 4: Repeat steps 1-3 until the queue is empty.

[0062] In the first stage, this application utilizes the SOM algorithm to cluster photovoltaic (PV) power output data, effectively capturing the daily-cycle regularity of the data and forming periodic PV power output cluster labels. In the second stage, based on the PV clustering results, DBSCAN density clustering is performed on wind power data to accurately depict the fluctuation details of wind power under the same weather conditions on the same date. This phased, algorithm-based approach overcomes the shortcomings of single clustering models in adapting to mixed wind and solar data. It also leverages the complementary characteristics of the SOM clustering algorithm preserving topological structure and the DBSCAN clustering algorithm recognizing clusters of arbitrary shapes, deeply exploring the temporal coupling relationship between wind and solar power output. Ultimately, comprehensive cluster labels are generated for PV-dominated, solar-wind fluctuation, wind-dominated, and wind-solar attenuation clusters, providing refined and physically meaningful input for the subsequent construction of 3D scene labels. This fundamentally supports the accurate matching of differentiated control strategies and the optimization of system stability.

[0063] In some embodiments, the step of introducing a weather correction coefficient to correct vehicle driving energy consumption specifically includes: constructing a benchmark model reflecting the nonlinear changes in energy consumption of electric vehicles in different speed ranges; and optimizing the benchmark model by introducing a weather correction coefficient, wherein the weather correction coefficient is obtained by jointly analyzing EV energy consumption deviation and wind and solar fluctuation data, and calibrating using statistical confidence intervals, and its calculation formula is as follows: In the formula: Energy consumption coefficient This is a weather correction factor. This refers to energy loss caused by low motor efficiency in the low-speed range. This represents the linear relationship between energy consumption and speed in the medium-speed range. The energy consumption is dominated by the quadratic increase in air resistance at high speeds. This is a constant term, representing the basic energy consumption independent of speed.

[0064] Specifically, electric vehicles (EVs), as a mobile power source, play an auxiliary role in microgrids (MG). Based on the travel characteristics of EVs in residential areas, the following mathematical model is established. The probability density function of the EV's trip end time, which is also the EV's grid connection time, is shown below, where... =17.6, =3.4:

[0065] ;

[0066] The probability density of the EV's start time of travel, i.e. the end time of EV charging, is shown below, where, =9.24, =3.16:

[0067] ;

[0068] The probability density of the average daily driving mileage of EVs is shown below, where, =3.2, =0.88:

[0069] ;

[0070] The charging and discharging formulas for EVs are as follows:

[0071] ;

[0072] ;

[0073] During the journey, the EV's State of Charge (SOC) is as follows:

[0074] ;

[0075] in, It is the total electricity generated from renewable energy sources. It is the load power. , These are the charging and discharging power of the EV. It refers to the travel distance of EVs. It refers to the battery capacity of the EV. This refers to the energy consumption of an EV (km / kWh).

[0076] Currently, mainstream batteries include lithium iron phosphate batteries and ordinary lithium-ion batteries. Using the rated discharge capacity at a standard environment of 25℃ as the performance baseline, the capacity of this type of battery shows a significant decay trend as temperature decreases. Specific test data shows that when the ambient temperature drops to 0℃, its actual discharge capacity is approximately 80% of the baseline capacity; when the temperature further drops to -5℃, the capacity decays to below 70% of the baseline capacity. In the high-temperature range, the battery capacity shows a trend of first increasing and then decreasing: below 50℃, the capacity increases slightly due to the increased reaction rate; when the temperature exceeds 50℃, the capacity changes from positive growth to rapid decline. For ordinary lithium-ion batteries, the impact of charge-discharge cycle count on capacity is highly correlated with ambient temperature. Within the comfortable temperature range of around 24℃, the internal reaction of the battery is in a stable state, and the impact of charge-discharge cycles on capacity decay is relatively small, maintaining the cycle life within the ideal range. However, when the ambient temperature deviates from the comfortable range, the battery's cycle stability decreases significantly, and the capacity decays at an accelerated rate with increasing charge-discharge cycles. The core reason is that the battery interface impedance increases under extreme temperatures, leading to a greater proportion of irreversible capacity loss during charge and discharge. To quantify the impact of temperature on battery capacity, a multinomial model is used to fit the relationship between temperature and relative battery capacity, as shown below:

[0077] ;

[0078] in, It is a relative capacity. It's temperature. , , and These are the fitting coefficients of the polynomial, and the constant term. A positive value represents the baseline intercept of the fitted curve, and its absolute value is the largest among the four coefficients, dominating the baseline level of relative capacity in the normal temperature range; the first-order coefficient... It is a positive value and its absolute value is less than The linear effect of temperature on relative capacity is a key factor influencing relative capacity changes within the normal temperature range; the quadratic coefficient... Negative value and absolute value less than This is used to characterize the inhibitory effect on relative capacity after the temperature rises to a certain level, and can weaken the linear boosting effect of the first-order term; the coefficient of the cubic term It is positive and has the smallest absolute value among the four coefficients. It is only used for nonlinear correction of relative capacity change under extreme temperatures. Its influence on the overall capacity change is much smaller than that of the first-order and second-order terms.

[0079] The usage of air conditioning in EVs varies depending on the temperature. For electric vehicles, the air conditioning system consumes the most energy besides driving. Ambient temperature and humidity directly affect the use and energy consumption of the air conditioning system, further significantly impacting the driving range and charging needs of electric vehicles. To quantify the impact of intraday temperature variations on air conditioning energy consumption in charging load simulations, a piecewise polynomial model is used to fit the relationship between temperature and air conditioning power, as shown below:

[0080] ;

[0081] in, It is a relative capacity. , , and These are the fitting coefficients of the polynomial, and the constant term. The value is positive, serving as the baseline intercept of the fitted curve. Its absolute value is the largest among the four coefficients, dominating the baseline level of air conditioning power in the normal temperature range; the linear coefficient... Negative value and absolute value less than The linear suppression effect of temperature on air conditioner power is a key factor influencing air conditioner power variation within the normal temperature range; the quadratic coefficient... It is a positive value and its absolute value is less than This is used to characterize the effect of temperature deviation from normal temperature on the increase in air conditioning power, and can weaken the linear suppression effect of the linear term; the coefficient of the cubic term It is positive and has the smallest absolute value among the four coefficients. It is only used for nonlinear correction of air conditioner power changes under extreme temperatures. Its impact on the overall power change is much smaller than that of the first and second terms.

[0082] The energy consumption level of an EV exhibits a significant non-linear correlation with vehicle speed. The core reason for this is that at low speeds, the motor's efficiency is below its rated range, resulting in higher energy conversion losses. At medium speeds, the motor efficiency is at its optimal range, while air resistance and rolling resistance are low, leading to minimal energy consumption. At high speeds, air resistance increases with the square of speed, causing a sharp rise in energy consumption. Based on the vehicle's speed, the energy consumption coefficient of an EV at different speeds can be fitted using a polynomial, as shown below:

[0083] ;

[0084] in, , , , These are the fitting coefficients. Coefficient 'a' is positive, corresponding to energy loss due to insufficient motor efficiency at low speeds, and its impact gradually weakens as speed increases. Coefficient 'b' is negative, representing the energy reduction effect brought about by improved motor efficiency and reduced resistance at medium speeds, and is the main driving factor for energy consumption reduction in the medium-speed range. Coefficient 'c' is positive, corresponding to the energy increase effect caused by the surge in air resistance at high speeds, and its impact significantly increases with the square of speed. Coefficient 'd' is positive, serving as the baseline constant for energy consumption coefficients; its absolute value is the smallest among the four coefficients, and it is only used to correct the basic energy consumption level across the entire speed range. This model can effectively cover the conventional driving speed range of 20-120 km / h, accurately reflecting the influence of speed on EV energy consumption.

[0085] Different weather conditions significantly impact EV energy consumption by altering road surface conditions, environmental parameters, and the operating load of vehicle assistance systems. On sunny days, with ambient temperatures maintained between 25-30℃, dry road surfaces, a coefficient of friction ≥0.85, and no additional environmental interference factors, EV energy consumption is primarily determined by driving speed, while the load on assistance systems (such as air conditioning and lights) is at its lowest.

[0086] In rainy conditions, EV energy consumption increases significantly compared to sunny days, with the increase being positively correlated with the rainfall level. Rainy weather typically results in ambient temperatures 3-5°C lower than sunny days, leading to a slight decrease in battery activity and a 2%-3% reduction in charge / discharge efficiency. Snowy weather is the most significant adverse weather condition affecting EV energy consumption, increasing it by 30%-50% compared to sunny days. When ice forms on the road surface, the coefficient of friction drops below 0.1, requiring frequent starts and stops to control speed, further increasing energy consumption by 10%-15% compared to snow-covered roads. On cloudy days, there is no precipitation, but ambient light intensity is reduced and temperatures are typically 2-4°C lower than on sunny days.

[0087] Based on the above analysis, a weather correction factor k is introduced to optimize the formula. The optimized formula is as follows:

[0088] ;

[0089] The correction coefficient was obtained by statistical analysis of historical measured data and calibration based on technical target constraints. The ideal scenario is a sunny day with solar radiation ≥800W / ㎡, dry road surface, and temperature of 15~25℃. Under this scenario, EV energy consumption is the lowest and wind and solar power output is the most stable. K=1.00 is set as the baseline value under this scenario.

[0090] The deviation ratio R = the comprehensive impact under a certain weather condition / the comprehensive impact under an ideal scenario, where the comprehensive impact = 0.6 × EV energy consumption impact + 0.4 × wind and solar fluctuation impact.

[0091] EV energy consumption impact = Actual EV energy consumption under this weather condition / EV energy consumption in an ideal scenario;

[0092] Impact of wind and solar power fluctuations = Wind and solar power output fluctuation coefficient under this weather condition / Wind and solar power output fluctuation coefficient under ideal scenario, where the fluctuation coefficient = standard deviation / mean × 100%.

[0093] The baseline k-value is the statistical interval of the deviation ratio R, taking the upper and lower limits of the 95% confidence interval.

[0094] For example, on a cloudy day, the impact of EV energy consumption is 1.05, the impact of wind and solar fluctuations is 1.10, and the overall impact is 0.6×1.05+0.4×1.10=1.07, with a confidence interval of [1.04, 1.08]. We take k as 1.06. The weather correction coefficients for various weather conditions are shown in Table 1 below.

[0095] Table 1: Weather correction coefficients for various weather conditions.

[0096]

[0097] At the same time, the optimal driving speed needs to be determined according to the road conditions under different weather conditions: the optimal driving speed is 60-80 km / h in sunny and cloudy weather; 50-60 km / h in light and moderate rain; 40-50 km / h in heavy rain and light snow; and 30-40 km / h in moderate snow and heavy snow / ice.

[0098] In terms of resource utilization and equipment lifespan protection, this application uses EV full-dimensional feature modeling to construct a precise model from multiple dimensions, making EV charging regulation more in line with the actual needs of users and vehicle characteristics. This not only improves the charging experience of high-value users, but also ensures the enthusiasm and reliability of EVs participating in V2G regulation.

[0099] In some embodiments, the step of calculating the real-time fluctuation level based on the wind and solar power output data specifically includes: the wind and solar power output data includes real-time wind and solar power generation data; based on the real-time wind and solar power generation data, a sliding window with a preset time is used to calculate the real-time power fluctuation coefficient, the calculation formula of which is: In the formula: This is the real-time power fluctuation coefficient. This represents the current real-time power. This represents the real-time power of the previous time window. The rated power is used as the reference value. Based on the average value of λ over several consecutive windows, the following levels are defined: if λ ≤ 5%, it is classified as a low fluctuation level; if 5% < λ ≤ 15%, it is classified as a medium fluctuation level; if λ > 15%, it is classified as a high fluctuation level.

[0100] Specifically, a driving architecture is constructed for weather type determination, fluctuation level quantification, cluster label matching, and resource linkage regulation, adapting to scenarios such as park micro-energy networks, industrial and commercial distributed energy, and residential V2G charging stations, achieving precise adaptation between strategies and scenarios. Weather type is determined by parameters such as solar radiation intensity, cloud coverage, precipitation, and dust concentration collected by distributed weather monitoring units, and is divided into five categories: sunny, cloudy, overcast and rainy, snowy, and dusty. Fluctuation level is calculated using a 5-minute sliding window to obtain the real-time fluctuation coefficient.

[0101] ;

[0102] The lower grade is determined by combining the average of three consecutive windows: Medium level: High-level: Clustering label matching calls the time period labels output by the time series pre-simulation module to form a three-dimensional scene identifier of weather-fluctuation-clustering.

[0103] This application effectively overcomes the shortcomings of traditional methods in responding to power surges with lag or misjudgment by using a sliding window-based real-time power fluctuation coefficient calculation and a multi-window mean level classification mechanism. Continuous window averaging smooths transient disturbances, ensuring the stability and reliability of fluctuation level determination. The low, medium, and high level classification forms a precise mapping with subsequent three-dimensional scene identification and control strategies, enabling the system to quickly identify fluctuation intensity and dynamically adjust mitigation strategies. For example, prioritizing absorption during low fluctuations and urgently calling upon energy storage during high fluctuations, thereby significantly improving the grid's adaptability to renewable energy volatility and operational stability.

[0104] In some embodiments, the step of switching the differentiated wind-solar-EV-energy storage control strategy based on the weather-fluctuation-cluster three-dimensional scene identifier with the goals of renewable energy consumption, power fluctuation mitigation, and power supply assurance specifically includes: under the sunny-low fluctuation-solar-dominated clustering label mode, prioritizing electric vehicle (EV) charging according to the value coefficient K with the goal of maximizing wind and solar energy consumption; under the cloudy-medium fluctuation-solar-wind fluctuation clustering label mode, calling battery energy storage systems (BESS) and electric vehicles (EVs) within the first preset condition range to participate in mobile energy storage resource scheduling (V2G) with the goal of mitigating power fluctuations; under the cloudy-medium / high fluctuation-wind-dominated clustering label mode, calling battery energy storage systems (BESS) and electric vehicles (EVs) within the second preset condition range to participate in mobile energy storage resource scheduling (V2G) with the goal of stabilizing power supply; under the snow / sandstorm-high fluctuation-wind-solar attenuation clustering label mode, initiating emergency scheduling and calling battery energy storage systems (BESS) and electric vehicles (EVs) within the third preset condition range to participate in mobile energy storage resource scheduling (V2G); wherein, the first preset condition range < the second preset condition range < the third preset condition range.

[0105] Specifically, the control strategies are designed differently for different scenarios: the core objective of the solar-dominated clustering label for sunny days with low fluctuations is to maximize the absorption of solar power, while EV charging is based on a value coefficient. Sort, where, This represents the probability of participating in V2G, with values ​​ranging from [0,1]. It's the charging frequency. This is the emergency level, with a value range of [0,1]. Priority protection. The emergency response levels for the operating vehicles are shown in Table 2 below:

[0106] Table 2: Emergency Situation Level Table.

[0107]

[0108] BESS uses dynamic SOC constraints, and the formula for calculating the constraint interval is:

[0109] ;

[0110] ;

[0111] in, Accumulate the number of loops for BESS.

[0112] The "Cloudy-Moderate Fluctuation-Light and Wind Fluctuation" clustering label aims to mitigate fluctuations. The BESS response time is ≤150ms, and the charge / discharge power calculation formula is as follows:

[0113] ;

[0114] in, To maintain real-time fluctuations, BESS's SOC is kept between 30% and 80%; [call / request] Furthermore, EVs with no travel plans within 1 hour participating in V2G have sufficient discharge power. , The difference in power output between wind and solar power; the charging power of non-emergency EVs is reduced to 60% of the rated value.

[0115] The clustering label for "Rainy - Medium-High Fluctuation - Wind Power Dominant" focuses on power supply stability, switches to wind power dominant mode, and raises the BESSSOC lower limit to 30; the V2G start threshold for EV clusters is lowered to power supply gap > 15%, calling vehicles with K ≥ 0.7 and no travel plans within 2 hours, and the total discharge ≤ 40% of the fluctuation difference; grid interaction is only initiated when local resources are insufficient, and power purchase is prioritized during off-peak hours.

[0116] Emergency dispatch will be initiated when clustering tags for snow / dust storms, high elevation changes, and wind and light attenuation are used. Maintain 40%-80% capacity with a response time of ≤100ms; only EVs with higher emergency response levels are charged, while other EVs with SOC ≥ 80% are forced to participate in V2G; emergency power purchase is initiated when fluctuations exceed 30% of the rated value.

[0117] Historical weather data, cluster labels, and corresponding parameters of resource adjustment capabilities for different regions are pre-stored. During regulation, the matching parameters are directly called, and dynamic corrections are made based on real-time adjustment deviations. The formula for calculating the adjustment deviation is:

[0118] ;

[0119] when Increase the weight of high-response resources in a timely manner. It maintains the current weight, achieving dual protection through pre-storage and correction.

[0120] This application employs an energy regulation strategy based on 3D scene identifier matching with differentiated objectives. By dynamically adjusting the BESS and V2G calling strategies through progressive condition ranges, it ensures a high degree of alignment between the strategy and scene characteristics, while achieving a balance between system stability and economy through hierarchical management of resource scheduling. This effectively solves the problems of poor adaptability and low resource utilization efficiency of traditional regulation methods, significantly improving the robustness and overall operational efficiency of the power grid under high-proportion renewable energy integration.

[0121] In some embodiments, the scenario strategy switching mechanism specifically includes: starting 24 hours in advance, dividing the prediction period into predicted time periods using a preset time sliding window, performing lightweight clustering on the wind and solar power subsequences of each time period, calculating the dynamic time warping similarity η between each subsequence and the basic clustering pattern, generating specific time period labels when η < 0.85, and retaining the basic clustering labels when η ≥ 0.85, forming a 24-hour clustering sequence containing basic labels and specific labels; and introducing a judgment coefficient based on the clustering sequence. The judgment coefficient comprehensively characterizes the fluctuation of wind and solar power over the next 24 hours. The calculation is performed by weighting the standard deviation and mean of photovoltaic and wind power outputs. When it is determined to be a high-fluctuation scenario, When it is determined to be a medium fluctuation scenario, A value ≤0.2 is considered a low-fluctuation scenario; the degree of agreement between the actual power sequence and the pre-simulated clustering labels is quantified using dynamic time warping similarity η. When the value is ≥0.8, the actual power is considered to be highly matched with the pre-simulated clustering, and the deviation is fine-tuned. When 0.6≤ When the value is less than 0.8, a local shift in the clustering pattern is determined, triggering a sub-policy reconstruction. When the value is less than 0.6, a fundamental change in the clustering pattern is determined, and a clustering transition is initiated. During strategy switching, a dynamic inertia coefficient γ based on the clustering transition intensity θ is introduced. By weighted fusion of the output power of the original strategy and the target strategy, a smooth transition of power commands is achieved. The value of the inertia coefficient γ and the power adjustment range limit are adjusted according to the fluctuation scenario.

[0122] Specifically, when initiating two-stage clustering 24 hours in advance, the prediction period is divided into two-hour time windows. Lightweight clustering is performed separately for the wind and solar power subsequences of each period, and the dynamic time warping similarity η between the subsequence and the basic clustering pattern is calculated. When η < 0.85, the period is identified as an atypical period, triggering local clustering correction and generating a unique clustering label; when η ≥ 0.85, the basic clustering label is used. After the simulation, a 24-hour clustering sequence with "basic label + atypical period label" is output, providing precise input for the formulation of day-ahead control strategies.

[0123] ;

[0124] in, This is a preview of the typical power curve for clustering during this period. For dynamic time-normalized distance.

[0125] First, we introduce the judgment coefficient. The formula for comprehensively characterizing the fluctuation of renewable energy power over the next 24 hours is as follows:

[0126] ;

[0127] in, , The standard deviations of the theoretical power values ​​for photovoltaic and wind power are used to calculate the clustering weights for different time periods, assigning higher weights to periods with low similarity. , These are their means, The weighting coefficients are dynamically adjusted based on the time-period clustering results, with α increasing during daytime photovoltaic-dominated periods and increasing during nighttime wind-dominated periods. Lower; when The scenario was identified as a high-fluctuation scenario. A value ≤0.2 is considered a low-fluctuation scenario, while values ​​between 0.2 and 0.2 are considered medium-fluctuation scenarios. Simultaneously, the dominant pattern for the following day is determined by combining the two-stage clustering results, serving as the basic framework for the strategy. Intraday dynamic adjustments are made by comparing the actual power with the clustering expectations of the corresponding sub-strategy for each hour. Quantify the degree of agreement between the actual power sequence and the pre-defined clustering labels. When When the value is ≥0.8, the actual power is considered to be highly matched with the pre-simulated clustering, based solely on the cumulative deviation. Make fine adjustments:

[0128] ;

[0129] in, , The actual and theoretical power at time t are respectively. The time period is used as the weight, and T is the statistical period (1 hour); when 0.6 ≤ When the value is less than 0.8, a local shift in the clustering pattern is detected, triggering a sub-policy reconstruction. Lightweight clustering is performed based on real-time power data, updating only the features for the current time period, generating a temporary sub-policy, and embedding it into the basic framework. When the value is less than 0.6, it is determined that the clustering pattern has undergone a fundamental change, and the clustering transition process is initiated. The complete two-stage clustering algorithm is called again to update the dominant pattern, and a brand-new strategy sequence is generated based on the new pattern. At the same time, the transition time and triggering reason are recorded, such as sudden weather or equipment failure, as correction samples for subsequent clustering pre-play.

[0130] Optimizing the smoothness of strategy switching focuses on the continuity of clustering time sequence, in terms of inertia coefficient. The cluster transition intensity θ is introduced into the calculation:

[0131] ;

[0132] in, , These are typical curves for the new and old clusters, respectively. The duration of clustering. When When the value is ≤0.1 (clustering evolves slowly), The value is approached 0.7, and the gradual transition of the strategy is achieved through the following formula:

[0133] ;

[0134] in, For the output power of the new strategy, The output power of the original strategy, The output power of the target strategy. In high-fluctuation scenarios, take... =0.3, taken in low-fluctuation scenarios =0.7, ensuring the power adjustment rate does not exceed the energy storage state constraint range of EV and BESS. When When >0.1, The power level is rapidly reduced to 0.3 to accelerate the strategy's response speed, but the magnitude of a single power adjustment is limited to no more than 30% of the typical clustering curve's fluctuation range to avoid system shocks caused by sudden pattern changes. Furthermore, a clustering review is initiated daily at 23:00 to analyze the deviation between the actual clustering evolution and the pre-rendering. For the reward function, when At that time, the sliding window size was maintained for 2 hours the following day; when At that time, the sliding window was adjusted to 1.5 hours using the Q-learning algorithm, and the similarity threshold was set. The value was reduced from 0.85 to 0.82, allowing the clustering model to gradually adapt to the time-series characteristics of local renewable energy, fundamentally improving the stability and accuracy of strategy switching.

[0135] By deeply embedding the clustering process into the time dimension, the above-mentioned mechanism design can accurately capture intraday time differences, intraday correction can calibrate clustering deviations in real time, and strategy switching can adapt to the evolution intensity of clustering patterns. Ultimately, it achieves seamless coupling between clustering results and scheduling strategies on the time axis, ensuring rapid response in high-fluctuation scenarios and stable operation in low-fluctuation scenarios.

[0136] In some embodiments, the strategy switching formula is: In the formula: For the output power of the new strategy, The inertia coefficient, The output power of the original strategy, The output power of the target strategy.

[0137] In terms of performance regulation and system stability, the inertial regulation mechanism introduced by the cluster evolution smooth switching module completely changes the problem of sudden power change caused by the cliff-like switching of traditional strategies. Through the gradual transition design that adapts to the intensity of cluster evolution, it effectively suppresses oscillations such as system frequency fluctuations and voltage instability, while reducing the impact of frequent device start-ups and shutdowns.

[0138] Please refer to Figure 6 On the other hand, the present invention also provides a weather-adaptive wind-solar-EV-energy storage multi-dimensional linkage control system, comprising: an acquisition module for acquiring wind and solar power output data and real-time weather information; a clustering module for performing a two-stage clustering process on the wind and solar power output data, first photovoltaic clustering and then wind power clustering, to generate clustering labels; a model building module for constructing an electric vehicle (EV) travel characteristic model, which integrates the probability distribution of travel time, fits the temperature-battery capacity decay relationship using piecewise multinomial fitting, and introduces a weather correction coefficient to correct vehicle driving energy consumption, so as to predict the spatiotemporal distribution of EV charging demand and quantify V2G regulation potential; a three-dimensional scene identification construction module for calculating the real-time fluctuation level based on the wind and solar power output data, and constructing a weather-fluctuation-cluster three-dimensional scene identification based on the real-time weather information, the real-time fluctuation level, and the clustering labels; and a control strategy switching module for introducing a scene strategy switching mechanism based on the weather-fluctuation-cluster three-dimensional scene identification, matching differentiated wind-solar-EV-energy storage control strategies aimed at renewable energy consumption, power fluctuation smoothing, and power supply guarantee.

[0139] On the other hand, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described weather-adaptive wind-solar-EV-energy storage multi-dimensional linkage control method.

[0140] On the other hand, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described weather-adaptive wind-solar-EV-energy storage multi-dimensional linkage control method.

[0141] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0142] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, database, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0143] The above are merely embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention's specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A weather-adaptive multi-dimensional linkage control method for wind-solar-EV-energy storage, characterized by the following steps: include: Acquire wind and solar power output data and real-time weather information; The wind and solar power output data are subjected to a two-stage clustering process, first photovoltaic clustering and then wind power clustering, to generate cluster labels; We construct an electric vehicle (EV) travel characteristic model, which integrates the probability distribution of travel time, fits the relationship between temperature and battery capacity decay using piecewise polynomial fitting, and introduces a weather correction coefficient to correct vehicle driving energy consumption. The step of introducing a weather correction factor to correct vehicle driving energy consumption specifically includes: Construct a benchmark model that reflects the nonlinear changes in energy consumption of electric vehicles in different speed ranges; A weather correction factor is introduced to optimize the baseline model. This weather correction factor is obtained by jointly analyzing EV energy consumption deviation and wind and solar fluctuation data, and calibrated using statistical confidence intervals. Its calculation formula is as follows: ； In the formula: Energy consumption coefficient This is a weather correction factor. This refers to energy loss caused by low motor efficiency in the low-speed range. This represents the linear relationship between energy consumption and speed in the medium-speed range. The energy consumption is dominated by the quadratic increase in air resistance at high speeds. The constant term represents the basic energy consumption independent of speed, used to predict the spatiotemporal distribution of charging demand for electric vehicles (EVs) and quantify the V2G regulation potential. Calculate the real-time fluctuation level based on the wind and solar power output data; the step of calculating the real-time fluctuation level based on the wind and solar power output data specifically includes: The wind and solar power output data includes real-time power data of wind and solar power generation. Based on real-time wind and solar power generation data, a sliding window with a preset time is used to calculate the real-time power fluctuation coefficient. The calculation formula is as follows: ； In the formula: This is the real-time power fluctuation coefficient. This represents the current real-time power. This represents the real-time power of the previous time window. Rated power; Based on the mean λ value of several consecutive windows, the levels are divided as follows: If λ≤5%, it is judged as a low fluctuation level: If 5% < λ ≤ 15%, it is judged as a medium fluctuation level; If λ > 15%, it is determined to be a high fluctuation level. Based on the real-time weather information, the real-time fluctuation level, and the clustering label, a three-dimensional scene identifier of weather-fluctuation-clustering is constructed. Based on the aforementioned weather-fluctuation-clustering three-dimensional scene identifier, a scene strategy switching mechanism is introduced to match differentiated wind-solar-EV-energy storage control strategies with the goals of renewable energy consumption, power fluctuation smoothing, and power supply guarantee.

2. The weather-adaptive multi-dimensional linkage control method for wind-solar-EV-energy storage according to claim 1, characterized in that, The step of performing a two-stage clustering process on the wind and solar power output data, first clustering photovoltaic power and then clustering wind power power, to generate cluster labels specifically includes: The self-organizing map (SOM) algorithm is used to perform the first-stage clustering of the photovoltaic output data in the wind and solar power output data, generating periodic photovoltaic output cluster labels; The density-based clustering DBSCAN algorithm is used to perform a second-stage clustering of wind power output data in the wind and solar power output data after the first-stage clustering, generating wind power output clustering labels for the corresponding period. By combining the photovoltaic output clustering tags and the wind power output clustering tags, photovoltaic-dominant clustering tags, solar-wind fluctuation clustering tags, wind-solar attenuation clustering tags, and wind power-dominant clustering tags are generated.

3. The weather-adaptive multi-dimensional linkage control method for wind-solar-EV-energy storage according to claim 2, characterized in that, The steps of switching the differentiated wind-solar-EV-energy storage control strategy based on the weather-fluctuation-cluster three-dimensional scene identifier, with the goals of renewable energy consumption, power fluctuation smoothing, and power supply guarantee, specifically include: Under the clear-day-low-fluctuation-photovoltaic-dominant clustering label model, with the goal of maximizing the absorption of wind and solar energy, the charging priority of electric vehicles (EVs) is ranked according to the value coefficient K. In the clustering labeling mode of cloudy-medium fluctuation-solar-wind fluctuation, with the goal of smoothing power fluctuation, the battery energy storage system (BESS) and electric vehicle (EV) within the first preset condition range are called to participate in mobile energy storage resource scheduling (V2G). Under the clustering labeling mode of cloudy / rainy-medium / high fluctuation-wind power-dominated, with the goal of stable power supply, the battery energy storage system (BESS) and electric vehicle (EV) within the second preset condition range are called to participate in the mobile energy storage resource scheduling (V2G). In the snow / dust-high fluctuation-wind and solar attenuation clustering label mode, emergency dispatch is initiated, and the battery energy storage system (BESS) and electric vehicle (EV) within the third preset condition range are called to participate in mobile energy storage resource dispatch (V2G). Among them, the first preset condition range < the second preset condition range < the third preset condition range.

4. The weather-adaptive multi-dimensional linkage control method for wind-solar-EV-energy storage according to claim 1, characterized in that, The scenario strategy switching mechanism specifically includes: The process starts 24 hours in advance, and the prediction period is divided into time periods by a preset time sliding window. Lightweight clustering is performed on the wind and solar power subsequences of each time period. The dynamic time warping similarity η between each subsequence and the basic clustering pattern is calculated. When η < 0.85, a specific time period label is generated. When η ≥ 0.85, the basic clustering label is used to form a 24-hour clustering sequence containing the basic label and the specific label. Based on the clustering sequence, a judgment coefficient is introduced. The judgment coefficient comprehensively characterizes the fluctuation of wind and solar power over the next 24 hours. The calculation is performed by weighting the standard deviation and mean of photovoltaic and wind power outputs. When it is determined to be a high-fluctuation scenario, When it is determined to be a medium fluctuation scenario, A value ≤0.2 is considered a low-fluctuation scenario; The degree of agreement between the actual power sequence and the pre-formed clustering labels is quantified by dynamic time warping similarity η. When the value is ≥0.8, the actual power is considered to be highly matched with the pre-simulated clustering, and the deviation is fine-tuned. When 0.6≤ When the value is less than 0.8, a local shift in the clustering pattern is determined, triggering a sub-policy reconstruction. When the value is less than 0.6, it is determined that a fundamental change in the clustering pattern has occurred, and a clustering transition is initiated. During strategy switching, a dynamic inertia coefficient γ based on the cluster transition intensity θ is introduced. By weighted fusion of the output power of the original strategy and the target strategy, a smooth transition of power commands is achieved. The value of the inertia coefficient γ and the power adjustment range limit are adjusted according to the fluctuation scenario.

5. The weather-adaptive multi-dimensional linkage control method for wind-solar-EV-energy storage according to claim 1, characterized in that, The strategy switching formula is as follows: ； In the formula: For the output power of the new strategy, The inertia coefficient, The output power of the original strategy, The output power of the target strategy.

6. A weather-adaptive multi-dimensional linkage control system for wind-solar-EV-energy storage, characterized in that, include: The acquisition module is used to acquire wind and solar power output data and real-time weather information; The clustering module is used to perform a two-stage clustering process on the wind and solar power output data, first clustering photovoltaic power and then clustering wind power power, to generate cluster labels; The model building module is used to build a model of electric vehicle (EV) travel characteristics. It integrates the probability distribution of travel time, fits the relationship between temperature and battery capacity decay using piecewise polynomial fitting, and corrects the vehicle's driving energy consumption by introducing a weather correction coefficient. The step of introducing a weather correction factor to correct vehicle driving energy consumption specifically includes: Construct a benchmark model that reflects the nonlinear changes in energy consumption of electric vehicles in different speed ranges; A weather correction factor is introduced to optimize the baseline model. This weather correction factor is obtained by jointly analyzing EV energy consumption deviation and wind and solar fluctuation data, and calibrated using statistical confidence intervals. Its calculation formula is as follows: ； In the formula: Energy consumption coefficient This is a weather correction factor. This refers to energy loss caused by low motor efficiency in the low-speed range. This represents the linear relationship between energy consumption and speed in the medium-speed range. The energy consumption is dominated by the quadratic increase in air resistance at high speeds. The constant term represents the basic energy consumption independent of speed, used to predict the spatiotemporal distribution of charging demand for electric vehicles (EVs) and quantify the V2G regulation potential. A 3D scene identification construction module is used to calculate the real-time fluctuation level based on the wind and solar power output data; the step of calculating the real-time fluctuation level based on the wind and solar power output data specifically includes: The wind and solar power output data includes real-time power data of wind and solar power generation. Based on real-time wind and solar power generation data, a sliding window with a preset time is used to calculate the real-time power fluctuation coefficient. The calculation formula is as follows: ； In the formula: This is the real-time power fluctuation coefficient. This represents the current real-time power. This represents the real-time power of the previous time window. Rated power; Based on the mean λ value of several consecutive windows, the levels are divided as follows: If λ≤5%, it is judged as a low fluctuation level: If 5% < λ ≤ 15%, it is judged as a medium fluctuation level; If λ > 15%, it is determined to be a high fluctuation level. Based on the real-time weather information, the real-time fluctuation level, and the clustering label, a three-dimensional scene identifier of weather-fluctuation-clustering is constructed. The control strategy switching module is used to introduce a scenario strategy switching mechanism based on the weather-fluctuation-clustering three-dimensional scenario identifier, and match differentiated wind-solar-EV-energy storage control strategies with the goals of renewable energy consumption, power fluctuation smoothing, and power supply guarantee.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the weather-adaptive wind-solar-EV-energy storage multi-dimensional linkage control method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the weather-adaptive wind-solar-EV-energy storage multi-dimensional linkage control method according to any one of claims 1 to 5.

Images