Extreme meteorological operation method of wind and light storage base group system based on Markov

Abstract

The invention belongs to the technical field of power system risk assessment and decision making, and discloses a Markov-based wind and light storage base group system extreme meteorological operation method. According to the method, a meteorological-power generation-load-flexibility four-dimensional MDP state space is constructed, a meteorological condition layered transition probability model is established, LOLP, EENS, CVaR and RSI multi-dimensional risk indexes are calculated, 15-minute / 1-hour / 24-hour comprehensive risk indexes are implemented, and risk grading is performed based on a dynamic threshold value. According to the method, the technical blank of dynamic evaluation of the insurance supply risk of the GW-level wind and light storage base group system under the extreme weather is filled, the risk prediction accuracy is improved, the early warning time is advanced, a quantitative decision basis is provided for insurance supply mode switching, and the method has important engineering application value.

Description

Technical Field

[0001] This invention belongs to the field of power system risk assessment and decision-making technology, specifically relating to a Markov-based method for operating wind-solar-storage base cluster systems under extreme weather conditions. Background Technology

[0002] Under extreme weather conditions, wind, solar and energy storage cluster systems face significant supply risks: 1. Rapid power fluctuations and increased forecasting errors; 2. Lagging response of flexible resources within the province; 3. Supply and demand imbalance risks exhibit dynamic evolution characteristics.

[0003] For extreme weather conditions, current conventional methods use deterministic prediction models that combine numerical weather prediction (NWP) and statistical correction (MOS). These models establish physical conversion functions between wind speed, radiance, and wind and solar power to solve the power prediction problem under a single extreme weather event. However, due to the output of a single risk indicator, they lack a comprehensive assessment across multiple dimensions and time scales. Summary of the Invention

[0004] Firstly, in view of the shortcomings of the existing technology, the purpose of this application is to provide a Markov-based method for the operation of wind-solar-storage base systems in extreme weather conditions, thereby improving the accuracy of extreme weather risk assessment for wind-solar-storage base systems.

[0005] The objective of this application can be achieved through the following technical solutions: Markov-based extreme weather operation methods for wind-solar-storage reservoir systems include: Obtain the current weather conditions, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources within the province in the target area. Discretize the weather conditions, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources within the province to construct a four-dimensional finite state set. The corresponding meteorological state transition probability, power generation output transition probability, load state transition probability, and flexibility resource availability transition probability are obtained from the four-dimensional finite state set. The meteorological state transition probability under meteorological conditions is obtained by multiplying the meteorological state transition probability, power generation output transition probability, load state transition probability, and flexibility resource availability transition probability. Steady-state probability is obtained by the meteorological state transition probability of meteorological conditions, and risk indicators are calculated based on the steady-state probability. The risk indicators include power shortage probability, expected power shortage, and conditional risk value. The risk severity index is obtained by normalizing and weighting the power shortage probability, expected power shortage, and conditional risk value. A weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. Based on the real-time comprehensive risk index and dynamically adjusted risk thresholds, risk levels are classified, and the operation strategy of the wind, solar and energy storage base cluster system is switched according to the risk level classification results.

[0006] Furthermore, obtain the current weather conditions, power output status of the wind-solar-storage base system, load demand status, and available capacity status of flexible resources within the province in the target area, including: Weather conditions at time t for Among them, Normal: normal weather conditions; Light dust storm; Severe sandstorm; Light smog; Severe smog; Extreme: extreme complex weather conditions; Power output status of the wind-solar-storage base system at time t for ,in, ; ,in, For the i-th output level, The lower limit of the probability of power transfer for wind-solar-storage base cluster systems. N represents the upper limit of the power output transfer probability of the wind-solar-storage base cluster system, where N is a constant. Load demand status at time t for ,in, The load demand status is discretized into M levels corresponding to the status; Intra-provincial flexible resource availability status at time t , ,in, The load demand status is discretized into K levels corresponding to the status.

[0007] Furthermore, the probability of weather state transition, power generation transition, load state transition, and flexibility resource availability transition corresponding to weather conditions, wind-solar-storage base system output conditions, load demand conditions, and available capacity of provincial flexibility resources are obtained, including: Weather state transition probability The specific calculation formula is: in, Indicates the current weather conditions. Indicates the weather conditions at the next moment; Power generation output transfer probability The specific calculation formula is: ,in, : Expected change in power generation output under weather condition w; Standard deviation of output variation; This represents the i-th output level; This represents the j-th output level; Load state transition probability The specific calculation formula is: ; This represents the current load output level. The load output level for the next moment; Active resource availability transfer probability The specific calculation formula is as follows: , and These represent the i-th and j-th flexibility resource states, respectively. Margin The specific calculation formula is as follows: , Let t represent the load demand state. Let t represent the available capacity of flexible resources within the province. The power output status of the wind-solar-storage base system at time t.

[0008] Furthermore, a risk severity index is obtained by normalizing and weighting the probability of insufficient power supply, the expected power shortage, and the conditional risk value. ,include: in, This represents the probability of insufficient power supply. In anticipation of power shortage, Conditional risk value, To assess the time window, =1, This represents the largest power shortage in history under extreme circumstances. Maximum conditional risk value in extreme historical scenarios; Let be the weighting coefficient, satisfying .

[0009] Furthermore, a weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. ,include: in, The risk severity index is calculated over a 15-minute period. This represents the risk severity index over one hour. This represents the risk severity index over a 24-hour period. Time scale weights; .

[0010] Furthermore, risk levels are classified based on real-time comprehensive risk indices and dynamically adjusted risk thresholds, including: Risk level The specific formula for determination is: in, i 1 represents the low-to-medium risk threshold; i 2 represents the medium-to-high risk threshold; risk severity index Less than i 1， Risk level Corresponding to low risk Risk Severity Index In i 1 and i Risk level between 2 and 3 Corresponding to medium risk Risk Severity Index Greater than i 2. Risk Level Corresponding to high risk .

[0011] Furthermore, dynamically adjusted risk thresholds are used to classify risk levels, including: Update the low-to-medium risk threshold or the medium-to-high risk threshold according to the trend of risk changes; The update formulas for low-to-medium risk thresholds or medium-to-high risk thresholds are as follows: in, The static baseline threshold is k, which is an adjustment coefficient. For risk change trends; For low-to-medium risk thresholds, The updated low-to-medium risk thresholds are based on changing risk trends; for medium-to-high risk thresholds, The updated medium-to-high risk thresholds are based on changing risk trends. in, This is a static baseline threshold, targeting low- to medium-risk thresholds. Equal to the previous version i 1; For medium- to high-risk areas, Equal to the previous version i 2; k is an adjustment factor; This represents the trend of risk changes.

[0012] The operation strategy of the wind, solar and energy storage base system is switched according to the risk level classification, including: Operational strategy The formula for determining it is: .

[0013] Secondly, in view of the shortcomings of the existing technology, the purpose of this application is to provide a Markov-based extreme weather operation system for wind, solar and energy storage cluster systems, so as to improve the accuracy of extreme weather risk assessment for wind, solar and energy storage cluster systems.

[0014] The objective of this application can be achieved through the following technical solutions: The extreme weather operation system of the Markov-based wind-solar-storage base system includes: The input unit is used to obtain the current weather status, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources in the province in the target area. The weather status, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources in the province are discretized to construct a four-dimensional finite state set. The processing unit is used to obtain the corresponding meteorological state transition probability, power generation output transition probability, load state transition probability and flexibility resource availability transition probability based on the four-dimensional finite state set, and to obtain the meteorological state transition probability under meteorological conditions by multiplying the meteorological state transition probability, power generation output transition probability, load state transition probability and flexibility resource availability transition probability. Steady-state probability is obtained by the meteorological state transition probability of meteorological conditions, and risk indicators are calculated based on the steady-state probability. The risk indicators include power shortage probability, expected power shortage, and conditional risk value. The risk severity index is obtained by normalizing and weighting the power shortage probability, expected power shortage, and conditional risk value. A weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. The output unit is used to classify risk levels based on the real-time comprehensive risk index and dynamically adjusted risk thresholds, and to switch the operation strategy of the wind, solar and energy storage base cluster system according to the risk level classification results.

[0015] Furthermore, obtain the current weather conditions, power output status of the wind-solar-storage base system, load demand status, and available capacity status of flexible resources within the province in the target area, including: Weather conditions at time t for Among them, Normal: normal weather conditions; Light dust storm; Severe sandstorm; Light smog; Severe smog; Extreme: extreme complex weather conditions; Power output status of the wind-solar-storage base system at time t for ,in, ; ,in, For the i-th output level, The lower limit of the probability of power transfer for wind-solar-storage base cluster systems. N represents the upper limit of the power output transfer probability of the wind-solar-storage base cluster system, where N is a constant. Load demand status at time t for ,in, The load demand status is discretized into M levels corresponding to the status; Intra-provincial flexible resource availability status at time t , ,in, The load demand status is discretized into K levels corresponding to the status.

[0016] Furthermore, the probability of weather state transition, power generation transition, load state transition, and flexibility resource availability transition corresponding to weather conditions, wind-solar-storage base system output conditions, load demand conditions, and available capacity of provincial flexibility resources are obtained, including: Weather state transition probability The specific calculation formula is: in, Indicates the current weather conditions. Indicates the weather conditions at the next moment; Power generation output transfer probability The specific calculation formula is: ,in, : Expected change in power generation output under weather condition w; Standard deviation of output variation; This represents the i-th output level; This represents the j-th output level; Load state transition probability The specific calculation formula is: ; This represents the current load output level. The load output level for the next moment; Active resource availability transfer probability The specific calculation formula is as follows: , and These represent the i-th and j-th flexibility resource states, respectively. Margin The specific calculation formula is as follows: , Let t represent the load demand state. Let t represent the available capacity of flexible resources within the province. The power output status of the wind-solar-storage base system at time t.

[0017] Furthermore, a risk severity index is obtained by normalizing and weighting the probability of insufficient power supply, the expected power shortage, and the conditional risk value. ,include: in, This represents the probability of insufficient power supply. In anticipation of power shortage, Conditional risk value, To assess the time window, =1, This represents the largest power shortage in history under extreme circumstances. Maximum conditional risk value in extreme historical scenarios; Let be the weighting coefficient, satisfying .

[0018] Furthermore, a weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. ,include: in, The risk severity index is calculated over a 15-minute period. This represents the risk severity index over one hour. This represents the risk severity index over a 24-hour period. Time scale weights; .

[0019] Furthermore, risk levels are classified based on real-time comprehensive risk indices and dynamically adjusted risk thresholds, including: Risk level The specific formula for determination is: in, i 1 represents the low-to-medium risk threshold; i 2 represents the medium-to-high risk threshold; risk severity index Less than i 1， Risk level Corresponding to low risk Risk Severity Index In i 1 and i Risk level between 2 and 3 Corresponding to medium risk Risk Severity Index Greater than i 2. Risk Level Corresponding to high risk .

[0020] Furthermore, dynamically adjusted risk thresholds are used to classify risk levels, including: Update the low-to-medium risk threshold or the medium-to-high risk threshold according to the trend of risk changes; The update formulas for low-to-medium risk thresholds or medium-to-high risk thresholds are as follows: in, The static baseline threshold is k, which is an adjustment coefficient. For risk change trends; For low-to-medium risk thresholds, The updated low-to-medium risk thresholds are based on changing risk trends; for medium-to-high risk thresholds, The updated medium-to-high risk thresholds are based on changing risk trends. in, This is a static baseline threshold, targeting low- to medium-risk thresholds. Equal to the previous version i 1; For medium- to high-risk areas, Equal to the previous version i 2; k is an adjustment factor; This represents the trend of risk changes.

[0021] The operation strategy of the wind, solar and energy storage base system is switched according to the risk level classification, including: Operational strategy The formula for determining it is: .

[0022] Thirdly, in view of the shortcomings of the existing technology, the purpose of this application is to provide a computer device to improve the accuracy of extreme weather risk assessment for wind, solar and energy storage cluster systems.

[0023] The objective of this application can be achieved through the following technical solutions: A computer device, comprising: one or more processors; The processor is used to store one or more programs; When the one or more programs are executed by the one or more processors, the first aspect of the method for extreme weather operation of a Markov-based wind-solar-storage base system is implemented.

[0024] Fourthly, in view of the shortcomings of the existing technology, the purpose of this application is to provide a computer-readable storage medium to improve the accuracy of extreme weather risk assessment for wind, solar and energy storage cluster systems.

[0025] The objective of this application can be achieved through the following technical solutions: A computer-readable storage medium having a computer program thereon, which, when executed, implements the first aspect of the Markov-based wind-solar-storage base cluster system extreme weather operation method.

[0026] The beneficial effects of this application are: This application obtains the comprehensive state vector of the wind-solar-storage base system. The comprehensive state vector is a four-dimensional MDP discrete state space composed of the current meteorological state, the power output state of the wind-solar-storage base system, the load demand state, and the available capacity of flexible resources in the province. The four-dimensional transition probabilities (meteorological, power output, load, and flexible resource availability transition probabilities) are multiplied to obtain the meteorological state transition probability under meteorological conditions. This accurately depicts the dynamic evolution of the state of the wind-solar-storage base system under extreme meteorological conditions, thus laying a dynamic foundation model for the real-time quantitative assessment of power supply insufficiency risk. This application obtains steady-state probability through meteorological state transition probability and calculates risk indicators such as probability of insufficient power supply (LOLP), expected power shortage (EENS), and conditional value of risk (CVaR). It constructs a four-in-one risk indicator system of LOLP, EENS, CVaR, and RSI, transforming the traditional deterministic prediction model that relies on physical transformation functions into a probabilistic risk extrapolation, realizing the quantification from "whether there is a power shortage" to "when there is a power shortage and what is the probability of a power shortage". This application obtains a risk severity index by normalizing and weighting the probability of insufficient power supply, expected power shortage, and conditional risk value, thereby achieving a comprehensive quantification of multi-dimensional heterogeneous risk indicators and providing a single, reliable assessment basis for subsequent decision-making. This application integrates risk assessment results from multiple time scales and calculates a real-time comprehensive risk index through weighted fusion, achieving full coverage from second-level emergency response to day-ahead strategic planning; through parallel computing and adaptive weight fusion, the risk warning time is extended, gaining critical time for the activation of slow backup resources and cross-provincial support coordination; This application classifies risk levels by using a real-time comprehensive risk index and dynamically adjusted risk thresholds, mapping these levels to a three-tiered supply guarantee model. This eliminates the response delays, inconsistent standards, and decision-making errors caused by traditional reliance on manual experience. By introducing dynamic thresholds, it achieves proactive early warning (automatically lowering the trigger threshold when risks worsen) and avoids frequent switching of supply guarantee models due to noise or random fluctuations, thus achieving an optimal balance between sensitivity and stability. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 Overall flowchart of the supply guarantee risk assessment system based on MDP; Figure 2 Comparison chart of system experimental verification results. Detailed Implementation

[0029] 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.

[0030] like Figure 1 As shown, the extreme weather operation method for a Markov-based wind-solar-storage reservoir system includes: S1. Obtain the current weather status, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources within the province in the target area. Discretize the weather status, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources within the province to construct a four-dimensional finite state set.

[0031] Specifically, the construction of the Markov state space based on meteorological conditions includes the definition of system states, where the system state definition is represented by a defined comprehensive state vector. For this application, the comprehensive state vector is defined to include the meteorological state at time t, the power output state of the wind-solar-storage base system, the load demand state, and the available capacity state of flexible resources within the province. That is, the comprehensive state vector of the wind-solar-storage base system in this application is a four-dimensional finite state set; where: S101, Integrated State Vector of Wind-Solar-Storage Base Cluster System : in, This represents the weather conditions at time t; This indicates the power output status of the wind-solar-storage base system at time t; This represents the load demand status at time t; This indicates the available capacity status of provincial flexibility resources at time t.

[0032] In this application, the comprehensive state vector of the wind-solar-storage base system at time t is discussed. The weather conditions at time t Power output status of the wind-solar-storage base system at time t Load demand status at time t and the status of available capacity of flexible resources within the province at time t. The integrated state vector of the wind-solar-storage base system at time t The discrete state space of the four-dimensional MDP consists of the meteorological state at time t, the power output state of the wind, solar and energy storage base system at time t, the load demand state at time t, and the available capacity of flexible resources in the province at time t.

[0033] S102, Indicates the weather conditions at time t: Among them, Normal: normal weather conditions; Light dust storm; Severe sandstorm; Light smog; Severe smog; Extreme: extreme complex weather conditions; In this application, the meteorological state at time t specifically includes normal weather, light dust storms, heavy dust storms, light haze, heavy haze, and other extreme combined weather conditions. Specifically, in practice, monitoring equipment from environmental monitoring stations, meteorological bureaus, or the base clusters themselves is used to monitor key indicators, such as real-time PM10 and PM2.5 concentrations for dust storms or haze, visibility reflecting the severity of the haze or dust storm, and the duration of the meteorological event. Normal weather can be understood as the absence of significant aerosols or extreme weather, and in some cases, it can be represented by PM10 < 150 μg / m³ and good visibility. Similarly, light dust storms, heavy dust storms, light haze, heavy haze, and other extreme combined weather conditions are defined as follows: Normal weather: no significant aerosols or extreme weather, PM10 < 150 μg / m³, visibility ≥ 10 km. Light dust storms: PM10 ≤ 150 μg / m³ < 300 μg / m³, visibility 5-10 km, duration < 6 hours. Heavy dust storms: PM10 ≥ 300 μg / m³ (up to 1500 μg / m³ during dust storms), visibility < 5 km (extremely < 500 m), duration 6-18 hours. Light haze: PM2.5 35-75 μg / m³, visibility 5-10 km, increased diffuse radiation. Heavy haze: PM2.5 ≥ 75 μg / m³ (up to 300+ μg / m³ in severe pollution), visibility < 5 km, significant attenuation of direct radiation. Extreme combined weather: Multiple extreme phenomena such as dust storms, haze, extreme cooling, and strong turbulence occur simultaneously.

[0034] S103, This represents the power output status of the wind-solar-storage base system at time t. Discretize into N levels, at this time ; ; Typically, N=10.

[0035] in, For the i-th output level, The lower limit of the probability of power transfer for wind-solar-storage base cluster systems. The upper limit of the power transfer probability of the wind-solar-storage base system is defined as N, where N is a constant. Specifically, in this application, the power output state of the wind-solar-storage base system is defined as follows: Discretized into N levels, such as In some cases, N=10 is set, meaning these 10 levels cover the rated capacity from 0% to 100%, with each level representing a 10% output range. In practice, however, N=9, 8, 7, 6, 5, etc., and the output level and range of the wind-solar-storage base system are determined by setting the N value as needed. In other words, this method is used to classify the output status of the wind-solar-storage base system. Discretize the representation.

[0036] S104, This represents the load demand status at time t. Discretize into M levels, at this time ; in, The load demand status is discretized into M levels. Specifically, in practice, implementers need to analyze historical load data and use statistical methods (e.g., cluster analysis or quantile division) to determine the upper and lower limits of the M levels to ensure that they can accurately represent the typical variation pattern of the load in a day or a year, and to divide the load demand status into M discretized levels. For example, based on the characteristics of the daily load curve, M=8 can be divided into characteristic periods such as the early morning extreme low, early morning low, morning peak, morning mid-load, midday adjustment, afternoon rebound, evening peak, and late night decline, as shown in Table 1.

[0037] Table 1. Schematic diagram of daily load curve characteristics For the M=8 selection, K-means clustering analysis can be performed using load data from over 3 years of historical data with a time resolution of 15 minutes. When K=8, the silhouette coefficient reaches its maximum value of 0.789, and the variance within the clusters exhibits an elbow inflection point, indicating that 8 clusters can optimally characterize the typical features of the daily load curve. The standard dispatching shifts in the power system consist of 3 shifts per day (night shift, morning shift, and evening shift), with 2-3 typical load conditions within each shift, totaling 8 different load levels, which are highly coupled with the daily dispatching cycle. Using 8 levels can capture more than 97% of the variation in the daily load curve, with a load prediction error of less than 5%.

[0038] (4) This represents the state of available capacity of flexible resources within the province at time t. Discretize into K levels, at this time ; in, The load demand status is discretized into K levels of status; specifically, in practice, as shown in the table below, K=5.

[0039] Table 2. State corresponding to the discretization level of load demand status At this point, considering the overall state as a multidimensional discrete state space, the total number of states is: For example, as mentioned above, the options are N=10, M=8, and K=5. =2400, achieving the optimal balance between accuracy and real-time performance; if k=4 is selected, the total number of states is 1920, resulting in a certain loss of accuracy; when k=6, the total number of states is 2880, which increases the extreme workload but provides limited improvement in accuracy.

[0040] The purpose of this application in constructing a Markov state space based on meteorological conditions is to obtain a comprehensive state vector for a wind-solar-storage base system. This comprehensive state vector is a four-dimensional MDP discrete state space composed of the meteorological state at time t, the power output state of the wind-solar-storage base system at time t, the load demand state at time t, and the available capacity of flexible resources within the province at time t. This allows for accurate characterization and dynamic evolution modeling of the wind-solar-storage base system state under extreme meteorological conditions. The effect is to lay a dynamic foundation model for real-time quantitative assessment of power shortage risks, calculation of risk indicators at multiple time scales, and quantitative decision support for power supply guarantee modes.

[0041] S2. Obtain the corresponding meteorological state transition probabilities, power generation output transition probabilities, load state transition probabilities, and flexibility resource availability transition probabilities based on the four-dimensional finite state set. Then, obtain the meteorological state transition probability under the given meteorological conditions by multiplying these probabilities by the product of their respective values. Specifically: S201. Construct a transition probability matrix based on meteorological conditions. Parameter description: The current and next system state.

[0042] Current weather conditions, i.e., normal weather, light dust storm. Severe dust storm Light haze Severe smog And other extreme and complex weather conditions.

[0043] : The probability of transitioning from state s to s' under meteorological conditions w.

[0044] In this application, the transition probability matrix of the wind-solar-storage base system under meteorological conditions is obtained through the state transition probability of the wind-solar-storage base system. This represents the state transition probability. This represents the probability that, under meteorological conditions w, the wind-solar-storage base system will transition from its current comprehensive state s to its next state s'.

[0045] S202. Meteorological condition-based hierarchical transition modeling: Constructing meteorological state transition probabilities based on meteorological conditions. Power generation output transfer Load state transition probability and the probability of flexible resource availability transfer The four-dimensional MDP discrete state space is formed.

[0046] Specifically, for the purposes of this application, the comprehensive state vector of the wind-solar-storage base system comprises a four-dimensional MDP discrete state space consisting of the meteorological state at time t, the power output state of the wind-solar-storage base system at time t, the load demand state at time t, and the available capacity of provincial flexible resources at time t. The state transition probabilities of the wind-solar-storage base system are then defined. The corresponding meteorological state transition probabilities under meteorological conditions are respectively determined by the meteorological state transition probabilities. Power generation output transfer Load state transition probability and the probability of flexible resource availability transfer composition.

[0047] (1) Probability of weather state transition The specific calculation formula is: in, Indicates the current weather conditions. Indicates the weather conditions at the next moment; Based on the historical data and meteorological conditions of the wind and solar power storage base, the following measures were adopted: strong meteorological condition persistence (85% unchanged), allowing adjacent level jumps (10%), and low probability of cross-level jumps (5%).

[0048] Specifically, regarding the meteorological state transition probability of this application, meteorological conditions are divided into multiple levels based on PM10 and PM2.5 data. For this application, the probability that the current meteorological state is consistent with the meteorological state at the next moment is set to 85%. At this time, regarding the meteorological state transition probability, the meteorological state transition probability of the current meteorological state transitioning to the meteorological state at the next moment is 85%. Similarly, regarding the meteorological state transition probability, the meteorological state transition probability that allows the meteorological state to jump to an adjacent level is 10.0%, the meteorological state transition probability that allows the meteorological state to jump to a different level is 10.0%, and the meteorological state transition probability of other more distant jumps to different levels is 0.

[0049] S203, Power Generation Output Transfer Probability The specific calculation formula is: ; Detailed parameter definition: Expected output change under meteorological condition w, key parameters Normal weather: (Stable output) Light dust : (Photovoltaic intensity decreased by 30%) Severe sandstorm : (Photovoltaic sales plummeted by 60%) Light haze : (The scattering of haze is enhanced but still weakened) Severe smog : (Multispectral attenuation leads to a sharp drop in output) Extreme combined weather: (Wind and solar power both decline, wind turbines are affected by turbulence, and solar power is affected by dust and haze). Standard deviation of power output variation under meteorological condition w, key parameters This represents the i-th output level; This represents the j-th output level; Normal weather: (Small fluctuation range) Light dust : (Increased volatility) Severe sandstorm : (The fluctuations are dramatic, and the changes in output are difficult to predict) Light haze : (Increased volatility) Severe smog : (The fluctuations are dramatic, and the changes in output are difficult to predict) Extreme combined weather: (The fluctuations are the most dramatic, with sudden changes in output possible within minutes). Regarding the power generation output transfer probability under meteorological conditions, this application is relevant. A Gaussian distribution model with stratified meteorological conditions is adopted to dynamically set the expected changes in power generation output for different meteorological conditions W. , and the standard deviation of fluctuation The continuous output changes are discretized and mapped to N levels, accurately depicting the sudden changes and strong randomness of wind and solar power under extreme weather conditions.

[0050] Specifically, in modeling the probability of power generation shift, differentiated expected change parameters and fluctuation standard deviations were set for six different weather conditions. Under normal meteorological conditions, the expected change in power generation output is zero, indicating that the system maintains a stable operating state, and the standard deviation of fluctuation is set at 0.05, reflecting a small fluctuation range in power output. Under light dust storms, the expected change in power generation output is adjusted to -0.3 times the rated power, representing a slight downward trend in photovoltaic output, and the standard deviation of fluctuation expands to 0.15. Under heavy dust storms, the expected change in power generation output further decreases to -0.6 times the rated power, reflecting a significant decrease in photovoltaic output, and the standard deviation of fluctuation increases to 0.3, while the fluctuation of wind turbines is aggravated by turbulence. Under light haze conditions, the expected change in power generation output is set at 0.2 times the rated power, considering the enhanced effect of local scattered radiation, and the standard deviation of fluctuation is 0.2. Under heavy haze conditions, the expected change in power generation output turns to -0.5 times the rated power, reflecting the multispectral attenuation caused by high concentrations of particulate matter, and the standard deviation of fluctuation increases to 0.3. Under extreme combined meteorological conditions, the expected change in power generation output reaches -0.7 times the rated power, representing a sharp drop in both wind and solar output, and the standard deviation of fluctuation expands to 0.4, characterizing extreme fluctuation characteristics.

[0051] S204, Load transition probability: Regarding the load state transition probability under meteorological conditions in this application In this application, the system load demand is discretized into M levels, such as M=8 in some embodiments. Each level corresponds to a typical operating condition interval on the daily load curve, such as characteristic periods such as early morning trough, morning peak, mid-load, noon peak, evening peak, and night. In this application, the construction of load state transition probabilities under meteorological conditions is based on frequency statistics of historical load data. For example, by analyzing load time series of more than three consecutive years with a time resolution of 15 minutes, the load state transition frequency of each pair of adjacent time points (i→j) is counted to generate an M×M-dimensional original frequency matrix. Subsequently, the matrix is ​​normalized so that the sum of the probabilities in each row is strictly equal to 1, thereby obtaining a transition probability matrix that satisfies the Markov property. , where the matrix elements represent the probability of shifting to level j in the next time period if the load is at level i in the current time period; Under normal meteorological conditions, the power generation output state transition probability matrix presents a typical tridiagonal sparse structure; when the output level j at the next moment is equal to the current level i, that is, the system state remains unchanged, the transition probability is as high as 0.70; When there is a positive transition between adjacent output levels, such as j = i + 1, indicating that the output increases by 10% of the rated capacity, or a negative transition between adjacent levels, such as j = i - 1, indicating that the output decreases by 10% of the rated capacity, each is assigned a probability weight of 0.15; For the case of cross-level transfer, that is, the absolute difference between the next level and the current level is greater than 1, such as |j - i| > 1, the transition probability is strictly set to 0; the start and stop of some large production equipment will cause cross-level transfer, but this phenomenon is not common and generally there is a production plan. For the sake of simplicity, the cross-level transfer probability is set to 0 in this patent.

[0052] S205, Transfer Probability of Flexibility Resource Availability : Among them, the margin: Parameter description: Margin > 0.2: The system margin is sufficient and the flexibility resource availability is high; 0 < Margin < 0.2: The margin is tight and the availability is medium; Margin < 0: Supply falls short of demand and the availability decreases.

[0053] This patent selects 0.2 as the margin limit value according to the power reliability management measures (tentative) standard issued by the National Development and Reform Commission. The standard requires that the load reserve capacity is 2% - 5% of the maximum power generation load, and the accident reserve capacity is about 10% of the maximum power generation load. In areas with a high proportion of power from other regions, new energy power generation, and non-interruptible users, the load reserve capacity should be appropriately increased. Since this patent considers the margin in extreme cases, the capacity limit must be further increased on the basis of the standard, so 0.2 is taken as the margin limit value.

[0054] Regarding the load state transition probability of this application under meteorological conditions , where, when Margin > 0.2, it indicates that the total capacity of system power generation and flexibility resources exceeds the load demand by more than 20%, and the reserve is abundant; when 0 < Margin < 0.2, the system is in a tight balance state and the reserve capacity is close to the critical point; when Margin < 0, there is a hard power shortage in the system and the contradiction between supply and demand intensifies.

[0055] This application obtains the comprehensive state vector of a wind-solar-storage base system. This comprehensive state vector is a four-dimensional discrete state space composed of the current meteorological state, the system's output state, load demand state, and the available capacity of flexible resources within the province. The four-dimensional transition probabilities (meteorological, power output, load, and flexible resource availability transition probabilities) are multiplied to obtain the meteorological state transition probabilities under specific meteorological conditions. This accurately characterizes the dynamic evolution of the wind-solar-storage base system's state under extreme weather conditions, thus laying a dynamic foundation model for the real-time quantitative assessment of power supply insufficiency risks. This method fills the technical gap in the dynamic assessment of power supply risks for GW-level wind-solar-storage base systems under extreme weather conditions.

[0056] S3. Obtain the steady-state probability by the meteorological state transition probability of meteorological conditions, and calculate the risk index based on the steady-state probability. The risk index includes the power supply insufficiency probability, the expected power shortage, and the conditional risk value. Normalize and weight the power supply insufficiency probability, the expected power shortage, and the conditional risk value to obtain the risk severity index. Weightedly calculate the risk severity indices at multiple time scales to obtain the real-time comprehensive risk index.

[0057] For the calculation of multi-timescale risk indicators in this application, the assessment time window for multi-timescale risk indicators is decoupled into three feature scales. Each scale independently calculates four basic indicators: probability of insufficient power supply (LOLP), expected power shortage (EENS), conditional value at risk (CVaR), and risk severity index (RSI). Then, a comprehensive risk index is generated by adaptive weight fusion. Specifically, S301, Probability of Insufficient Power Supply (LOLP) in: Detailed parameter description: The steady-state probability of the system being in state s at time t; Solving using MDP: ; Initial distribution: Based on the current system status.

[0058] Calculation method: In this application, for the probability of insufficient power supply (LOLP), the real-time power shortage risk probability is dynamically evolved based on a Markov decision process (MDP), and the steady-state probability in a four-dimensional state space is used. The solution precisely quantifies the instantaneous probability that the total capacity of power generation and flexible resources of a wind, solar and energy storage cluster cannot meet load demand within a specified time window under extreme weather conditions. This transforms deterministic power flow calculations into probabilistic risk projections, enabling a shift from "whether there is a power shortage" to "when there will be a power shortage and what is the probability of a power shortage."

[0059] S302, Expected Energy Shortage (EENS) Detailed parameter description: Evaluation time window (15 minutes / 1 hour / 24 hours) Power shortage in state s Represents the center power value for each state. Time step (usually 15 minutes) In this application, the Expected Energy Shortage (EENS) is a dynamic energy shortage assessment index based on Markov decision process. By integrating system state probability distribution, time-varying power shortage, and multi-timescale assessment windows, it accurately quantifies the total expected energy loss of a wind, solar, and energy storage cluster due to insufficient power generation and flexibility resources during a specific period under extreme weather conditions. This transforms the static annual average energy shortage index in traditional reliability assessment into a decision variable that can be extrapolated in real time at the minute level.

[0060] S303, Conditional Value at Risk (CVaR) in: α: Confidence level (typically 0.95) Value-at-risk cost coefficient at confidence level β: Loss: The loss function Cost coefficient Unit cost of power shortage (RMB / MWh), typically 5000-10000.

[0061] Frequency deviation penalty coefficient (yuan / Hz), usually 50,000.

[0062] In this application, Conditional Value at Risk (CVaR) is a tail economic loss metric designed for the risk of power system supply disruptions due to extreme weather. By integrating power shortage costs and frequency deviation penalties, it quantifies the expected economic loss in the "worst 5% scenario" at a 95% confidence level. This transfers the CVaR theory from the traditional financial field to the power system operation risk scenario and achieves minute-level online calculation based on the MDP state probability distribution, supporting the optimal trade-off between economic costs in dispatch decisions.

[0063] S304. The risk severity index (RSI) is obtained by normalizing and weighting the probability of insufficient power supply, the expected power shortage, and the conditional risk value. Detailed parameter description: The weighting coefficients satisfy the following conditions: ,suggestion: Normalized upper bound: LOLPmax=1.0 EENSmax: Maximum power shortage in historical extreme scenarios CVaRmax: Maximum CVaR in extreme historical scenarios In this application, a risk severity index RSI is constructed by normalizing and weighting the probability of insufficient power supply (LOLP), expected power shortage (EENS), and conditional value of risk (CVaR) and setting corresponding dynamic weights.

[0064] S305, weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. The details are as follows: Detailed parameter description: The risk severity index is calculated over a 15-minute period. This represents the risk severity index over one hour. This represents the risk severity index over a 24-hour period. Time scale weights; Short-term (15 minutes): (Addressing immediate risks) Mid-term (1 hour): (Balanced scheduling capability) Long-term (24 hours): (Strategic decision support) For this application, four basic indicators—LOLP, EENS, CVaR, and RSI—are calculated in parallel across three characteristic time windows: 15-minute short-term, 1-hour medium-term, and 24-hour long-term. An adaptive weighted fusion method is then used to form a comprehensive risk index. ; Specifically, to address immediate risks, the probability of system instability within the next 15 minutes is assessed. This scale uses the current state distribution as the initial vector of the Markov chain, iteratively solving for 10 time steps (15 minutes / step) to quickly capture emergency risks caused by minute-level drops in photovoltaic output, sudden load increases, or insufficient short-term ramp-up of flexible resources. Its LOLP reflects the probability of immediate power shortage, EENS represents short-term power shortage, CVaR quantifies the risk of economic loss within 15 minutes, and RSI is a comprehensive short-term risk severity index. The weighting coefficient is as high as 0.5 to ensure the highest priority response to immediate risks.

[0065] To balance dispatch capabilities and assess system operational risks within the next hour, a short-term final-state distribution is used as the initial vector, iterating for 40 steps. The focus is on the adjustment potential of slow-dynamic resources such as thermal power plant start-up and shutdown, pumped reservoir water level changes, and energy storage SOC management. At this scale, EENS emphasizes dispatchable power shortages, and CVaR incorporates frequency deviation accumulation penalties with a weighting coefficient of 0.3, for rolling dispatch decisions and secondary allocation of reserve capacity.

[0066] Strategic decision support is provided for contingency planning and cross-day risk assessment at the strategic level. It iterates 96 times from the current moment, covering the complete daily load cycle and wind and solar resource fluctuation cycle, assessing the cumulative risk exposure under persistent extreme weather conditions. This scale implicitly considers slow variables such as unit combination optimization, medium- and long-term fuel constraints, and cross-provincial support coordination, with a weighting coefficient of 0.2, providing a basis for revising day-ahead plans and rehearsing supply guarantee plans.

[0067] This application obtains steady-state probability through meteorological state transition probability and calculates risk indicators such as probability of insufficient power supply (LOLP), expected power shortage (EENS), and conditional value of risk (CVaR). It constructs a four-in-one risk indicator system of LOLP, EENS, CVaR, and RSI, transforming the traditional deterministic prediction model that relies on physical transformation functions into a probabilistic risk extrapolation, realizing the quantification from "whether there is a power shortage" to "when there is a power shortage and what is the probability of a power shortage".

[0068] Simultaneously, the probability of insufficient power supply, expected power shortage, and conditional risk value are normalized and weighted to obtain a risk severity index, realizing the comprehensive quantification of multi-dimensional heterogeneous risk indicators and providing a single, reliable assessment basis for subsequent decision-making.

[0069] Simultaneously, risk assessment results from multiple time scales (15 minutes, 1 hour, and 24 hours) are integrated and weighted to obtain a real-time comprehensive risk index, achieving full coverage from second-level emergency response to day-ahead strategic planning. Through parallel computing and adaptive weight fusion, the risk warning time is extended to 75 minutes (45 minutes earlier than traditional methods), gaining critical time for the activation of slow backup resources and cross-provincial support coordination.

[0070] S4. Based on the real-time comprehensive risk index and dynamically adjusted risk thresholds, risk levels are classified, and the operation strategy of the wind-solar-storage base cluster system is switched according to the risk level classification results. Specifically: S401. Based on the real-time comprehensive risk index and dynamically adjusted risk thresholds, risk levels are classified, including: Risk level The specific formula for determination is: in, i 1 represents the low-to-medium risk threshold; i 2 represents the medium-to-high risk threshold; risk severity index Less than i 1， Risk level Corresponding to low risk Risk Severity Index In i 1 and i Risk level between 2 and 3 Corresponding to medium risk Risk Severity Index Greater than i 2. Risk Level Corresponding to high risk .

[0071] Detailed parameter description: i 1 = 0.3: Low-to-medium risk threshold Corresponding to LOLP≈5%, EENS<100MWh i 2=0.7: Medium-to-high risk threshold Corresponding to LOLP≈15%, EENS>500MWh This application sets a dual threshold dividing point to determine the risk severity index. The linear mapping ∈[0,1] to three segments: low, medium, and high. Specifically, when the low-risk threshold is reached, the medium-risk threshold is... i 1 = 0.3, corresponding to a normalized LOLP ≈ 5%, EENS < 100 MWh; that is, when When the value is less than 0.3, the system determines that the supply and demand balance margin is sufficient, maintains the economic operation mode to ensure supply, does not activate additional reserves, maintains the basic output of thermal power in the province, and keeps the energy storage SOC above 50% on standby. When the risk level changes from medium to high: i 2 = 0.7, corresponding to a normalized LOLP ≈ 15% and EENS > 500 MWh; that is... ≥0.7 triggers the highest level response, initiates inter-provincial support and orderly power consumption mode, applies for an additional 2-3GW of power transmission through inter-regional interconnection lines, and issues interruptible load shedding instructions; In some embodiments, when i When the value falls within the 0.3-0.7 range between the two thresholds, this is considered a medium-risk buffer zone. The system then activates the provincial resource coordination mode, utilizing thermal power plants for spinning standby, pumped storage plants for full-capacity generation, and energy storage plants for discharging, while closely monitoring the situation. Based on changing trends, prepare and upgrade contingency plans.

[0072] S402. Dynamically adjusted risk thresholds are used to classify risk levels, including: Static baseline threshold Adjustment coefficient Risk change trend Among them, for the low-to-medium risk threshold, The updated low-to-medium risk thresholds are based on changing risk trends; for medium-to-high risk thresholds, The updated medium-to-high risk thresholds are based on changing risk trends. in, This is a static baseline threshold, targeting low- to medium-risk thresholds. Equal to the previous version i 1, such as i 1 = 0.3; for medium- to high-risk areas, Equal to the previous version i 2, such as i 2 = 0.7; In practice, k is an adjustment factor, usually taken as 0.1-0.2. In this example, 0.1 is chosen.

[0073] This application's dynamic threshold adjustment introduces risk change trends as a regulating factor, enabling the risk level classification threshold to adjust accordingly. The acceleration fluctuates in real time. When the risk deteriorates sharply, the trigger threshold is automatically lowered to achieve forward-looking early warning. When the risk is stable or alleviated, the threshold is appropriately raised to avoid over-response and frequent switching, thereby achieving the optimal balance between sensitivity and stability.

[0074] Specifically, for static thresholds, only focus on The value itself, when When the value is 0.69, the system still classifies it as medium risk and requires waiting for it to exceed 0.70 before upgrading the response; however, under extreme weather conditions... It could surge from 0.69 to 0.85 within 10 minutes, at which point... =+0.16, the static threshold causes a response lag of 15-20 minutes, missing the golden window for backup startup; In this application, the dynamic threshold is monitored... ,exist When the value reaches 0.68 and accelerates upward, the threshold is lowered to 0.66, triggering the high-risk contingency plan 2-3 cycles (30-45 minutes) in advance, thus gaining crucial time for preparations such as providing support from other provinces and issuing notices for orderly electricity use.

[0075] Meanwhile, traditional static thresholds in When the value is around 0.70, due to measurement noise or random fluctuations, It may repeatedly cross between 0.68 and 0.72, causing frequent switching of the power supply mode (such as switching 3-4 times within 30 minutes), which has a great impact on the power grid.

[0076] In this application, the dynamic threshold smooths the decision-making process in the boundary region through trend judgment, such as... It decreased from 0.71 to 0.69, but It remains positive (upward trend). i 1. Maintaining a low level (e.g., 0.66) keeps the system in a high-risk mode; only when... Continuously negative and Stable below i It only exits the response at 1 o'clock.

[0077] S403. Switch the operation strategy of the wind-solar-storage base system according to the risk level classification results, including: Build and run strategies, run strategies The formula for determining it is: In this application, a three-level supply guarantee model mapping system based on dynamic risk level classification is constructed. The operation strategy is automatically switched according to the real-time assessed risk severity index (Level=Low / Medium / High), realizing a hierarchical and adaptive supply guarantee response from base cluster autonomy to provincial resource-led. For example, when the risk level is Low, the system activates the "base cluster-led - province-assisted" mode. In this mode, the wind, solar and energy storage base clusters primarily focus on self-balancing, fully leveraging their new energy forecasting and autonomous control capabilities, and utilizing energy storage and station-level frequency regulation resources within the bases to achieve local supply and demand balance. Flexible resources within the province, such as thermal power and pumped storage, serve only as auxiliary reserves, maintaining basic output or standby status, and do not actively participate in the internal regulation of the base cluster, thus reducing resource idle losses.

[0078] For example, when the risk level rises to Medium, the system automatically switches to "coordinated dispatch" mode. At this time, the power base cluster and the province's flexible resources enter a state of two-way cooperation and joint optimization: the power base cluster shares its own power generation capacity, energy storage SOC, and ramp-up margin with the provincial dispatch in real time; the province's thermal power plants increase voltage in advance based on the power shortage prediction of the power base cluster, pumped storage switches from charging to discharging, and the energy storage discharge depth increases from 50% to 80%. Both parties share the responsibility of ensuring supply through collaborative optimization algorithms, achieving efficient allocation of 1-2GW level regulation capacity.

[0079] For example, when the risk level exceeds High, the system decisively switches to a "province-led - base cluster coordinated" mode. Provincial flexible resources take full control of supply: thermal power plants operate at full capacity or even exceed capacity, pumped storage systems discharge at full power, and energy storage systems are deeply discharged to 90% SOC; base clusters play a supporting role, providing surplus output and strictly executing provincial dispatch instructions, accepting orderly power consumption arrangements when necessary. This mode supports 2-3GW of inter-provincial power transmission support and the shelving of interruptible loads exceeding 500MW, ensuring the safety baseline of the large power grid.

[0080] That is, for this application, a risk severity index is established ( Deterministic quantification function between the three-tier supply guarantee model and the three-tier supply guarantee model This method transforms multidimensional and heterogeneous risk calculation results into directly executable scheduling instructions, completely eliminating response delays, inconsistent standards, and decision-making errors caused by traditional reliance on human experience. It supports the system in achieving adaptive, delay-free, and economically optimal supply guarantee mode switching under extreme weather conditions, thereby establishing a quantitative mapping relationship between risk indicators and supply guarantee modes, and providing a basis for adaptive switching.

[0081] This application classifies risk levels based on a real-time comprehensive risk index and dynamically adjusted risk thresholds, mapping them to a three-tiered supply guarantee model. This eliminates the response delays, inconsistent standards, and decision-making errors caused by traditional reliance on manual experience. By introducing dynamic thresholds, it achieves proactive early warning (automatically lowering the trigger threshold when risks worsen) and avoids frequent switching of supply guarantee models due to noise or random fluctuations, thus achieving an optimal balance between sensitivity and stability.

[0082] Experimental verification: A case study of a GW-level base cluster in Gansu Province was set up for verification, as follows: Project Background: Location: A wind, solar and energy storage base complex with a capacity of tens of millions of kilowatts in Gansu Province; Installed capacity: 6GW wind power, 4GW solar power, and 1GW energy storage; Flexible resources within the province: 8GW of thermal power and 2GW of pumped storage.

[0083] Extreme weather scenarios: Time: Dust storm event in March 2024; Meteorological conditions: PM10 peak value 1500 μg / m³, lasting for 18 hours; Power output plummeted: photovoltaic power dropped to 15% of its rated capacity, and wind power to 30%.

[0084] Comparison of verification results: Table 3 Comparison of Validation Results As shown in Table 3 and Figure 2 Based on actual verification data from the sandstorm event in March 2024 involving the Gansu 10-million-kilowatt-level wind and solar power reservoir cluster, the MDP method of this invention demonstrates a comprehensive performance leap compared to traditional static methods: In predicting the probability of insufficient power supply, traditional static methods only predict 8.5% LOLP, severely underestimating the risk of load loss under extreme weather conditions, with a deviation of 3.6 percentage points from the actual 12.1% LOLP. In contrast, the MDP method of this invention accurately predicts 12.3%, differing from the actual value by only 0.2 percentage points, reducing the prediction error to 1.6% and improving accuracy by over 35%, effectively capturing the risk of a sharp drop in renewable energy output caused by sandstorms.

[0085] In assessing expected power shortages, traditional methods predict a shortage of 520 MWh, a significant underestimation of 342 MWh compared to the actual shortage of 862 MWh, resulting in an error rate as high as 40%. The MDP method, through dynamic state evolution modeling, optimizes the prediction to 845 MWh, differing from the actual value by only 17 MWh, reducing the error to within 2%, and improving the assessment accuracy by nearly 20 times. This provides a reliable basis for dispatching departments to accurately grasp the scale of the energy shortage.

[0086] In terms of conditional value at risk (VaR) forecasting, traditional methods estimated the economic loss at 1.85 million yuan, far lower than the actual loss of 3.05 million yuan, underestimating it by 1.2 million yuan (nearly 40%). The MDP method, by integrating power outage costs and frequency deviation penalties, accurately predicted the loss to be 3.12 million yuan, with a deviation of only 70,000 yuan (2.3%) from the actual loss. This enabled the risk assessment in the economic dimension to reach engineering-usable accuracy for the first time, laying a quantitative foundation for cost-benefit decision-making.

[0087] Regarding the timeliness of early warnings, traditional static methods, based on current cross-section assessments, can only issue warnings 30 minutes in advance, which is insufficient to meet the activation sequence requirements of slow-reserve resources. The MDP method, with its risk evolution projection capabilities, extends the warning time to 75 minutes, gaining an additional 45-minute golden intervention window. This provides ample preparation time for key actions such as cross-provincial support coordination, thermal power plant ramp-up, and orderly power consumption notifications, which is the core guarantee for avoiding power shortages.

[0088] Regarding the accuracy of supply guarantee mode switching, traditional methods suffer from insufficient assessment precision, resulting in only 2 correct switchings out of 5 attempts, with an accuracy rate of only 40%. Frequent erroneous or missed switchings severely impact system stability. The MDP method, through mapping the RSI comprehensive index to a dynamic threshold, achieves 100% accuracy across all 5 switching attempts, realizing zero misjudgments and completely solving the decision-making challenges of "when to switch, how much to switch, and how to switch" under extreme weather conditions.

[0089] In summary, this invention, through dynamic modeling using MDP, improves the overall accuracy of risk prediction by 35%, advances the warning time by 45 minutes, and increases the accuracy of power supply mode switching from 40% to 100%, successfully avoiding power shortage losses of approximately 8 million yuan. It fills the technical gap in real-time risk quantification assessment of GW-level wind, solar and energy storage base clusters under extreme weather conditions.

[0090] In some publications, this application provides a Markov-based wind-solar-storage reservoir cluster system for extreme weather operation, including: The input unit is used to obtain the current weather status, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources in the province in the target area. The weather status, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources in the province are discretized to construct a four-dimensional finite state set. The processing unit is used to obtain the corresponding meteorological state transition probability, power generation output transition probability, load state transition probability and flexibility resource availability transition probability based on the four-dimensional finite state set, and to obtain the meteorological state transition probability under meteorological conditions by multiplying the meteorological state transition probability, power generation output transition probability, load state transition probability and flexibility resource availability transition probability. Steady-state probability is obtained by the meteorological state transition probability of meteorological conditions, and risk indicators are calculated based on the steady-state probability. The risk indicators include power shortage probability, expected power shortage, and conditional risk value. The risk severity index is obtained by normalizing and weighting the power shortage probability, expected power shortage, and conditional risk value. A weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. The output unit is used to classify risk levels based on the real-time comprehensive risk index and dynamically adjusted risk thresholds, and to switch the operation strategy of the wind, solar and energy storage base cluster system according to the risk level classification results.

[0091] Furthermore, obtain the current weather conditions, power output status of the wind-solar-storage base system, load demand status, and available capacity status of flexible resources within the province in the target area, including: Weather conditions at time t for Among them, Normal: normal weather conditions; Light dust storm; Severe sandstorm; Light smog; Severe smog; Extreme: extreme complex weather conditions; Power output status of the wind-solar-storage base system at time t for ,in, ; ,in, For the i-th output level, The lower limit of the probability of power transfer for wind-solar-storage base cluster systems. N represents the upper limit of the power output transfer probability of the wind-solar-storage base cluster system, where N is a constant. Load demand status at time t for ,in, The load demand status is discretized into M levels corresponding to the status; Intra-provincial flexible resource availability status at time t , ,in, The load demand status is discretized into K levels corresponding to the status.

[0092] Furthermore, the probability of weather state transition, power generation transition, load state transition, and flexibility resource availability transition corresponding to weather conditions, wind-solar-storage base system output conditions, load demand conditions, and available capacity of provincial flexibility resources are obtained, including: Weather state transition probability The specific calculation formula is: in, Indicates the current weather conditions. Indicates the weather conditions at the next moment; Power generation output transfer probability The specific calculation formula is: ,in, : Expected change in power generation output under weather condition w; Standard deviation of output variation; This represents the i-th output level; This represents the j-th output level; Load state transition probability The specific calculation formula is: ; This represents the current load output level. The load output level for the next moment; Active resource availability transfer probability The specific calculation formula is as follows: , and These represent the i-th and j-th flexibility resource states, respectively. Margin The specific calculation formula is as follows: , Let t represent the load demand state. Let t represent the available capacity of flexible resources within the province. The power output status of the wind-solar-storage base system at time t.

[0093] Furthermore, a risk severity index is obtained by normalizing and weighting the probability of insufficient power supply, the expected power shortage, and the conditional risk value. ,include: in, This represents the probability of insufficient power supply. In anticipation of power shortage, Conditional risk value, To assess the time window, =1, This represents the largest power shortage in history under extreme circumstances. Maximum conditional risk value in extreme historical scenarios; Let be the weighting coefficient, satisfying .

[0094] Furthermore, a weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. ,include: in, The risk severity index is calculated over a 15-minute period. This represents the risk severity index over one hour. This represents the risk severity index over a 24-hour period. Time scale weights; .

[0095] Furthermore, risk levels are classified based on real-time comprehensive risk indices and dynamically adjusted risk thresholds, including: Risk level The specific formula for determination is: in, i 1 represents the low-to-medium risk threshold; i 2 represents the medium-to-high risk threshold; risk severity index Less than i 1， Risk level Corresponding to low risk Risk Severity Index In i 1 and i Risk level between 2 and 3 Corresponding to medium risk Risk Severity Index Greater than i 2. Risk Level Corresponding to high risk .

[0096] Furthermore, dynamically adjusted risk thresholds are used to classify risk levels, including: Update the low-to-medium risk threshold or the medium-to-high risk threshold according to the trend of risk changes; The update formulas for low-to-medium risk thresholds or medium-to-high risk thresholds are as follows: in, The static baseline threshold is k, which is an adjustment coefficient. For risk change trends; For low-to-medium risk thresholds, The updated low-to-medium risk thresholds are based on changing risk trends; for medium-to-high risk thresholds, The updated medium-to-high risk thresholds are based on changing risk trends. in, This is a static baseline threshold, targeting low- to medium-risk thresholds. Equal to the previous version i 1; For medium- to high-risk areas, Equal to the previous version i 2; k is an adjustment factor; This represents the trend of risk changes.

[0097] The operation strategy of the wind, solar and energy storage base system is switched according to the risk level classification, including: Operational strategy The formula for determining it is: .

[0098] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0099] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure one One or more processes and / or boxes Figure one A device that provides the functions specified in one or more boxes.

[0100] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure one One or more processes and / or boxes Figure one The function specified in one or more boxes.

[0101] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure one One or more processes and / or boxes Figure one The steps of the function specified in one or more boxes.

[0102] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0103] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

[0104] In the description of this specification, the references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0105] The foregoing has shown and described the basic principles, main features, and advantages of this application. Those skilled in the art should understand that this application is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this application. Various changes and modifications can be made to this application without departing from the spirit and scope thereof, and all such changes and modifications fall within the scope of the claims of this application.

Claims

1. A method for extreme weather operation of a wind-solar-storage reservoir system based on Markovian principles, characterized in that: include: Obtain the current weather conditions, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources within the province in the target area. Discretize the weather conditions, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources within the province to construct a four-dimensional finite state set. The corresponding meteorological state transition probability, power generation output transition probability, load state transition probability, and flexibility resource availability transition probability are obtained from the four-dimensional finite state set. The meteorological state transition probability under meteorological conditions is obtained by multiplying the meteorological state transition probability, power generation output transition probability, load state transition probability, and flexibility resource availability transition probability. Steady-state probability is obtained by the meteorological state transition probability of meteorological conditions, and risk indicators are calculated based on the steady-state probability. The risk indicators include the probability of insufficient power supply, the expected power shortage, and the conditional risk value. The risk severity index is obtained by normalizing and weighting the probability of insufficient power supply, the expected power shortage, and the conditional risk value. A weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. Based on the real-time comprehensive risk index and dynamically adjusted risk thresholds, risk levels are classified, and the operation strategy of the wind, solar and energy storage base cluster system is switched according to the risk level classification results.

2. The extreme weather operation method for a wind-solar-storage reservoir system based on Markov as described in claim 1, characterized in that, Obtain the current weather conditions, power output status of the wind, solar and energy storage system clusters, load demand status, and available capacity of flexible resources within the province in the target area, including: Weather conditions at time t for Among them, Normal: normal weather conditions; Light dust storm; Severe sandstorm; Light smog; Severe smog; Extreme: extreme complex weather conditions; Power output status of the wind-solar-storage base system at time t for ,in, ; ,in, For the i-th output level, The lower limit of the probability of power transfer for wind-solar-storage base cluster systems. N represents the upper limit of the power output transfer probability of the wind-solar-storage base cluster system, where N is a constant. Load demand status at time t for ,in, The load demand status is discretized into M levels corresponding to the status; Intra-provincial flexible resource availability status at time t , ,in, The load demand status is discretized into K levels corresponding to the status.

3. The extreme weather operation method for a wind-solar-storage reservoir system based on Markov as described in claim 2, characterized in that, Acquire the meteorological state transition probability, power generation output transition probability, load state transition probability, and flexibility resource availability transition probability corresponding to the meteorological state, wind-solar-storage base system output state, load demand state, and available capacity of provincial flexibility resources, including: Weather state transition probability The specific calculation formula is: in, Indicates the current weather conditions. Indicates the weather conditions at the next moment; Power generation output transfer probability The specific calculation formula is: ,in, : Expected change in power generation output under weather condition w; Standard deviation of output variation; This represents the i-th output level; This represents the j-th output level; Load state transition probability The specific calculation formula is: ； This represents the current load output level. The load output level for the next moment; Active resource availability transfer probability The specific calculation formula is as follows: , and These represent the i-th and j-th flexibility resource states, respectively. Margin The specific calculation formula is as follows: , Let t represent the load demand state. Let t represent the available capacity of flexible resources within the province. The power output status of the wind-solar-storage base system at time t.

4. The extreme weather operation method for a wind-solar-storage reservoir system based on Markov as described in claim 1, characterized in that, The risk severity index is obtained by normalizing and weighting the probability of insufficient power supply, the expected power shortage, and the conditional risk value. ,include: in, This represents the probability of insufficient power supply. In anticipation of power shortage, Conditional risk value, To assess the time window, =1, This represents the largest power shortage in history under extreme circumstances. Maximum conditional risk value in extreme historical scenarios; Let be the weighting coefficient, satisfying .

5. The extreme weather operation method for a Markov-based wind-solar-storage reservoir system according to claim 4, characterized in that, A weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. ,include: in, The risk severity index is calculated over a 15-minute period. This represents the risk severity index over one hour. This represents the risk severity index over a 24-hour period. Time scale weights; .

6. The extreme weather operation method for a Markov-based wind-solar-storage reservoir system according to claim 1, characterized in that, Risk levels are classified based on real-time comprehensive risk indices and dynamically adjusted risk thresholds, including: Risk level The specific formula for determination is: in, θ 1 represents the low-to-medium risk threshold; θ 2 represents the medium-to-high risk threshold; risk severity index Less than θ 1， Risk level Corresponding to low risk Risk Severity Index In θ 1 and θ Risk level between 2 and 3 Corresponding to medium risk Risk Severity Index Greater than θ 2. Risk Level Corresponding to high risk .

7. The extreme weather operation method for a Markov-based wind-solar-storage reservoir system according to claim 6, characterized in that, Dynamically adjusted risk thresholds are used to classify risk levels, including: Update the low-to-medium risk threshold or the medium-to-high risk threshold according to the trend of risk changes; The update formulas for low-to-medium risk thresholds or medium-to-high risk thresholds are as follows: in, The static baseline threshold is k, which is an adjustment coefficient. For risk change trends; For low-to-medium risk thresholds, The updated low-to-medium risk thresholds are based on changing risk trends; for medium-to-high risk thresholds, The updated medium-to-high risk thresholds are based on changing risk trends. in, This is a static baseline threshold, targeting low- to medium-risk thresholds. Equal to the previous version θ 1; For medium- to high-risk areas, Equal to the previous version θ 2; The operation strategy of the wind, solar and energy storage base system is switched according to the risk level classification, including: Operational strategy The formula for determining it is: 。 8. A Markov-based wind-solar-storage base cluster system for extreme weather operation, characterized in that: include: The input unit is used to obtain the current weather status, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources in the province in the target area. The weather status, wind-solar-storage base system output status, load demand status, and available capacity status of flexible resources in the province are discretized to construct a four-dimensional finite state set. The processing unit is used to obtain the corresponding meteorological state transition probability, power generation output transition probability, load state transition probability and flexibility resource availability transition probability based on the four-dimensional finite state set, and to obtain the meteorological state transition probability under meteorological conditions by multiplying the meteorological state transition probability, power generation output transition probability, load state transition probability and flexibility resource availability transition probability. Steady-state probability is obtained by the meteorological state transition probability of meteorological conditions, and risk indicators are calculated based on the steady-state probability. The risk indicators include the probability of insufficient power supply, the expected power shortage, and the conditional risk value. The risk severity index is obtained by normalizing and weighting the probability of insufficient power supply, the expected power shortage, and the conditional risk value. A weighted fusion calculation of risk severity indices across multiple time scales yields a real-time comprehensive risk index. The output unit is used to classify risk levels based on the real-time comprehensive risk index and dynamically adjusted risk thresholds, and to switch the operation strategy of the wind, solar and energy storage base cluster system according to the risk level classification results.

9. A computer device, characterized in that, include: One or more processors; The processor is used to store one or more programs; When the one or more programs are executed by the one or more processors, the extreme weather operation method of the wind-solar-storage base cluster system based on Markov as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that, It contains a computer program, which, when executed, implements the extreme weather operation method of the Markov-based wind-solar-storage base cluster system as described in any one of claims 1 to 7.

Images