Random near-extreme day weather data simulation method, equipment and medium

Abstract

The invention discloses a random near-extreme day weather data simulation method, equipment and a medium, which are used for screening near-extreme day weather data from historical weather data of previous years by taking daily highest dry-bulb temperature exceeding threshold temperature as a standard, and extracting corresponding hour dry-bulb temperature, wet-bulb temperature and solar radiation data; according to the method, four groups of characteristic weather parameters are identified by quantifying probability characteristics of the characteristic weather parameters, and a typical variation coefficient representing a daily variation mode of each weather element is obtained by performing normalization processing and clustering analysis on the daily variation mode of historical near-extreme weather data; and finally, in combination with characteristic meteorological parameters and typical variable coefficients obtained by random sampling, constructing random near-extreme day weather data, and meanwhile, considering uncertainty and a coupling relationship of near-extreme weather parameters, obtaining probability characteristics of a near-extreme load in daily load random simulation, so that the simulation efficiency is improved, and the accuracy of a simulation result is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of load data simulation, and in particular to a method, device and medium for simulating random near-extreme day weather data. Background Art

[0002] The constant change of weather directly affects the use of power loads such as air conditioners. In the hot and humid summer, a large number of refrigeration air conditioners will be put into operation to eliminate the stuffiness, and as the proportion of cooling load usage increases, the impact of temperature on load becomes more obvious; under cold conditions, heating demand increases significantly, which will put great pressure on the heating system, especially in extremely cold areas. In recent years, my country's meteorological science has developed rapidly, and the impact of urban micrometeorology on urban power loads and air conditioners has attracted widespread attention. A large number of literatures have studied the relationship between weather-sensitive loads and weather. Many studies have shown that among all weather factors, temperature has the greatest and most regular impact on weather-sensitive loads. When the weather is extremely cold or hot, a large number of air conditioners and heating loads will be put into operation to maintain the most comfortable temperature for the human body. From a lot of life experience, it can be seen that in the winter with low temperature, as the temperature decreases, the power load increases, which is mainly caused by heating loads such as heating; in the summer with high temperature, as the temperature rises, the power load increases, which is mainly caused by cooling loads such as air conditioners. Research shows that the power load is lowest at 14-18°C because this is the most comfortable temperature for the human body, requiring neither cooling nor heating.

[0003] A significant shortcoming of existing air conditioning load forecasting methods is the lack of consideration of air conditioning load changes under extreme weather conditions. Extreme weather, such as continuous high temperatures or extremely low temperatures, can cause dramatic fluctuations in air conditioning loads, which is crucial for determining the design load of air conditioning systems because it is directly related to the capacity and performance of air conditioning systems. At the same time, air conditioning loads under extreme weather conditions also have a significant impact on load forecasting of power systems because it may lead to a sudden increase in power demand, posing a challenge to the stability and reliability of the power grid. However, due to the relatively small number of samples of extreme weather events in historical data, traditional data-driven forecasting models often fail to fully simulate these extreme scenarios during the generation process. This results in reduced accuracy and reliability of the model's forecasts when faced with actual extreme weather events. For example, traditional multivariate linear regression models, gray forecasting models, and forecasting methods based on BP neural networks, while performing well under conventional weather conditions, have limited forecasting capabilities under extreme weather conditions.

[0004] In order to improve the prediction ability of prediction models under extreme weather conditions, researchers have begun to explore new prediction methods. For example, prediction models based on long short-term memory networks (LSTM), especially multi-channel multi-layer LSTM networks (MM-LSTM), have shown superior performance in air-conditioning load prediction for non-industrial buildings. Through appropriate feature engineering, these models can more accurately capture the relationship between air-conditioning load and meteorological factors, thereby improving prediction accuracy. In addition, prediction methods based on tensor low-rank completion algorithms have also shown potential in short-term load prediction under extreme weather conditions. This method can improve the model's ability to respond to extreme events by completing missing load data.

[0005] Although current air conditioning load forecasting models have shown a certain degree of accuracy under standard climate conditions, their forecasting performance under extreme meteorological conditions has not yet reached the expected level. The main reason for this phenomenon is that the sample size of extreme meteorological events in historical datasets is insufficient, and the data collection of these extreme meteorological events has only begun to become normalized and systematized in recent years. Therefore, the limitations of existing datasets restrict the performance of different forecasting models in dealing with load forecasting problems under extreme meteorological conditions. To overcome this challenge, scientific simulation and expansion of sample data to build a more comprehensive dataset is a key strategy to improve the model's forecasting ability. By integrating this approach into forecasting models and algorithms, the forecasting accuracy of air conditioning load under extreme meteorological conditions can be significantly improved, thereby providing more accurate data support for the design optimization of air conditioning systems and power load forecasting. Summary of the invention

[0006] The technical problem to be solved by the present invention is that the prediction performance under extreme meteorological conditions has not yet reached the expected level, and the prediction accuracy is low. The purpose is to provide a random near-extreme daily weather data simulation method, equipment and medium. Taking the daily maximum dry-bulb temperature exceeding the threshold temperature as the standard, near-extreme daily weather data are screened out from the historical weather data of previous years, and the corresponding hourly dry-bulb temperature, wet-bulb temperature and solar radiation data are extracted. By quantifying the probabilistic characteristics of the characteristic weather parameters, four groups of characteristic weather parameters are identified. By normalizing and clustering the daily variation pattern of the historical near-extreme weather data, the typical coefficient of variation representing the daily variation pattern of each weather element is obtained. Finally, the characteristic meteorological parameters and the typical coefficient of variation obtained by random sampling are combined to construct random near-extreme daily weather data. At the same time, the uncertainty and coupling relationship of the near-extreme weather parameters are considered to obtain the probabilistic characteristics of the near-extreme load in the random simulation of daily load, thereby improving the simulation efficiency and ensuring the accuracy of the simulation results.

[0007] The present invention is achieved through the following technical solutions:

[0008] A first aspect of the present invention provides a method for simulating random near-extreme daily weather data, comprising the following specific steps:

[0009] For obtaining historical weather data, filter out historical near-extreme day weather data;

[0010] Quantitative analysis of historical extreme daily weather data was performed to obtain conditional probability distributions of multiple sets of characteristic weather parameters;

[0011] Based on the conditional probability distribution of multiple groups of characteristic weather parameters, random characteristic weather parameters with coupling relationships are generated through sampling methods;

[0012] The daily variation patterns of historical near-extreme day weather data were normalized to obtain the coefficient of variation of characteristic weather parameters;

[0013] K-Means clustering was used to analyze the coefficient of variation of characteristic weather parameters, and the centroid of the cluster area in the largest cluster was selected as the typical coefficient of variation;

[0014] The typical coefficient of variation is randomly sampled to obtain random near-extreme daily weather data.

[0015] Furthermore, the uncertainty of historical near-extreme day weather data is quantified, specifically including:

[0016] Based on the daily maximum dry-bulb temperature > T0, extract the corresponding hourly dry-bulb temperature, wet-bulb temperature and solar radiation;

[0017] The probability characteristics of dry-bulb temperature, wet-bulb temperature and solar radiation data are quantified to obtain the conditional probabilities of dry-bulb temperature, wet-bulb temperature and solar radiation data, where the conditional probabilities include:

[0018] The conditional probability of daily dry-bulb temperature peak and the conditional probability of daily dry-bulb temperature mean for dry-bulb temperature;

[0019] The conditional probability of the daily wet-bulb temperature peak and the conditional probability of the daily wet-bulb temperature mean for wet-bulb temperature;

[0020] Conditional probability of daily peak solar radiation for solar radiation.

[0021] Furthermore, the conditional probability distribution based on multiple groups of characteristic weather parameters generates random characteristic weather parameters with coupling relationship through sampling method, which specifically includes:

[0022] According to the daily dry-bulb temperature peak conditional probability and the daily dry-bulb temperature average conditional probability of the dry-bulb temperature, the daily wet-bulb temperature peak conditional probability and the daily wet-bulb temperature average conditional probability of the wet-bulb temperature, and the daily solar radiation peak conditional probability of the solar radiation;

[0023] generating the joint probability of daily peak dry-bulb temperature and daily average dry-bulb temperature, the joint probability of daily peak dry-bulb temperature and daily peak wet-bulb temperature, the joint probability of daily peak dry-bulb temperature, daily peak wet-bulb temperature and daily average wet-bulb temperature, and the joint probability of daily peak dry-bulb temperature and daily peak solar radiation;

[0024] Generate random characteristic weather parameters based on joint probability.

[0025] Furthermore, the specific calculation steps of the joint probability include:

[0026] The joint probability of the daily peak dry-bulb temperature and the daily mean dry-bulb temperature is:

[0027] P(T dry,p T dry,m )=P(T dry,m |T dry,p )P(T dry,p )

[0028] The joint probability of a daily peak dry-bulb temperature and a daily peak wet-bulb temperature is:

[0029] P(T dry,p T wet,p )=P(T wet,p |T dry,p )P(T dry,p )

[0030] The joint probability of the daily peak dry-bulb temperature, the daily peak wet-bulb temperature, and the daily average wet-bulb temperature is:

[0031] P(T dry,p T wet,p T wet,m )=P(T wet,m |T wet,p )P(T wet,p |T dry,p )P(T dry,p )

[0032] The joint probability of daily peak dry-bulb temperature and daily peak solar radiation is:

[0033] P(T dry,p I sun,p )=P(T sun,p |T dry,p )P(T dry,p )

[0034] Among them, T dry,p represents the daily dry-bulb temperature peak, T dry,m represents the daily average dry-bulb temperature, T wet,p represents the daily peak wet-bulb temperature, T wet,mrepresents the daily average wet-bulb temperature, I sun,p Represents the peak daily solar radiation.

[0035] Furthermore, the daily variation pattern of the historical near-extreme day weather data is normalized to obtain the coefficient of variation of the characteristic weather parameters, specifically including:

[0036]

[0037]

[0038] In the formula, are the coefficients of change of dry-bulb temperature, wet-bulb temperature and solar radiation on the nth day; X d ,i n , are the hourly values ​​of the three weather elements on the nth day, i = 1, 2…, 24; is the daily peak; is the daily average.

[0039] Furthermore, the coefficient of variation of characteristic weather parameters is analyzed by K-Means clustering, and the centroid of the cluster area in the largest class is selected as the typical coefficient of variation, which specifically includes:

[0040] Construct a joint matrix based on the coefficient of variation;

[0041] Combined with the joint matrix, K-Means clustering was used to obtain the typical coefficient of variation;

[0042] In the process of clustering to obtain the typical coefficient of variation, the stability and non-guarantee rate of the load probability characteristics are also evaluated.

[0043] Furthermore, the specific steps of evaluating the stability and non-guaranteed rate of the load probability characteristic include:

[0044]

[0045] R′=ηR

[0046]

[0047] In the formula, Q i,j is the load value; Q dpeak,mean (N) is the average of all daily peak loads in the previous simulation N; N is the number of simulations; ΔN is the simulation interval; S(N) is the rate of change of the daily peak load average; n Qhourly>Qdesign Indicates the number of hours Q per hour of load hourly Exceeding the design load Q design; R is the non-guaranteed rate calculated based on the load sample under the recent extreme weather scenario; R′ is the corrected non-guaranteed rate under the weather scenario in previous years; η is the correction factor, N ext is the number of days with historical extreme weather data, and d is the number of years of previous years’ data calculated.

[0048] Furthermore, the characteristic meteorological parameters and typical coefficients of variation are randomly sampled to obtain random near-extreme day weather data, specifically including:

[0049] T dry,i =T dry,p -(T dry,p -T dry,m )*β′ d,i

[0050] T wet,i =T wet,p -(T wet,p -T wet,m )*β′ w,i

[0051] I sun,i =I sun,p *β′ s,i

[0052] Where, T dry,i 、T wet,i 、T sun,i are random near-extreme dry-bulb temperature, wet-bulb temperature and solar radiation, i=1,2…,24.

[0053] A second aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a method for simulating random near-extreme daily weather data is implemented.

[0054] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for simulating random near-extreme daily weather data.

[0055] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0056] By quantifying the uncertainty of near-extreme weather parameters, especially considering the coupling relationship between weather parameters, the probabilistic characteristics of near-extreme loads in random simulation of daily loads are directly obtained. First, the near-extreme daily weather data of previous years are screened out from the historical weather data with the daily maximum dry-bulb temperature exceeding the threshold temperature as the standard, and the corresponding hourly dry-bulb temperature, wet-bulb temperature and solar radiation data are extracted. Then, the probabilistic characteristics of these characteristic weather parameters are quantified, the four most likely related groups of characteristic weather parameters are identified, and the joint probability of each group of characteristic weather parameters is expressed using conditional probability. In addition, the typical coefficient of variation representing the daily variation pattern of each weather element is obtained by normalizing and clustering the daily variation pattern of historical near-extreme weather data. Finally, the random near-extreme daily weather data is constructed by combining the characteristic meteorological parameters and typical coefficient of variation obtained by random sampling. This method improves the simulation efficiency and ensures the accuracy of the simulation results by retaining the coupling relationship between weather parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0058] Figure 1 is a modeling flow chart in an embodiment of the present invention;

[0059] Figure 2 It is a temperature change image of part of the data in an embodiment of the present invention;

[0060] Figure 3 It is the data of the temperature of the near-extreme weather in the embodiment of the present invention;

[0061] Figure 4 It is a probability distribution diagram of the maximum dry-bulb temperature and the average temperature of the day in the embodiment of the present invention;

[0062] Figure 5 is a probability distribution diagram of the maximum dry-bulb temperature and the maximum wet-bulb temperature of the day in an embodiment of the present invention;

[0063] Figure 6 It is a probability distribution diagram of the maximum wet-bulb temperature and the average temperature of the day in the embodiment of the present invention;

[0064] Figure 7 is a probability distribution diagram of the maximum dry-bulb temperature and the radiation peak value of the day in an embodiment of the present invention;

[0065] Figure 8is a dry-bulb temperature variation coefficient image in an embodiment of the present invention;

[0066] Fig. 9 is a wet-bulb temperature variation coefficient image in an embodiment of the present invention;

[0067] Fig.10 is a solar radiation variation coefficient image in an embodiment of the present invention;

[0068] Fig.11 This is a meteorological data simulation image in an embodiment of the present invention. DETAILED DESCRIPTION

[0069] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0070] As a possible implementation, Figure 1 As shown, this embodiment provides a random near-extreme daily weather data simulation method. In this embodiment, in order to ensure that the probability load obtained by random simulation of daily load is similar to the probability load under the real near-extreme weather scenario, it is necessary to accurately quantify the probability and regular characteristics of near-extreme weather parameters while considering the coupling relationship between different weather parameters. Therefore, a refined random near-extreme daily weather data simulation model is proposed to improve the probability and calculation efficiency of the design load accuracy of the air-conditioning system.

[0071] The model described in this embodiment directly obtains the probability characteristics of near-extreme loads in random simulation of daily loads by quantifying the uncertainty of near-extreme weather parameters, especially considering the coupling relationship between weather parameters. First, the near-extreme daily weather data are screened out from the historical weather data of previous years, and the corresponding hourly dry-bulb temperature, wet-bulb temperature and solar radiation data are extracted with the daily maximum dry-bulb temperature exceeding 30°C as the standard. Then the probability characteristics of these characteristic weather parameters are quantified, the four most likely related groups of characteristic weather parameters are identified, and their joint probability is expressed using conditional probability. In addition, by normalizing and clustering the daily variation pattern of historical near-extreme weather data, the typical coefficient of variation representing the daily variation pattern of each weather element is obtained. Finally, the characteristic meteorological parameters and typical coefficient of variation obtained by random sampling are combined to construct random near-extreme daily weather data. According to the results, this method improves the simulation efficiency, and by retaining the coupling relationship between weather parameters, the accuracy of the simulation results is ensured, providing a new method for load calculation data reference for air conditioning system design under extreme weather.

[0072] Specifically, we first select the historical near-extreme daily weather data, and we need to quantify the uncertainty of the near-extreme weather parameters. Therefore, we need to first select the historical near-extreme daily weather data from the historical weather data of previous years. In order to meet the needs of air conditioning load calculation, all three weather factors (dry-bulb temperature, wet-bulb temperature and solar radiation) need to be included in the selected data, and the simultaneous occurrence of different factors needs to be retained. Therefore, we take "daily peak dry-bulb temperature>T0" as the standard, select the days that meet the requirements, and then select the corresponding hourly dry-bulb temperature, wet-bulb temperature and solar radiation data of these dates as the historical near-extreme daily weather data (dry-bulb thermometer temperature is the temperature measured by the thermometer freely exposed to the air, and it should avoid the interference of radiation and moisture, and is usually regarded as the actual temperature of the measured air; wet-bulb temperature refers to the temperature of the system when a large amount of water contacts limited wet air under adiabatic conditions, and the latent heat required for water evaporation comes entirely from the sensible heat released by the decrease in the temperature of the wet air.). Among them, T0 is related to the simulation accuracy. If T0 is too large, some historical near-extreme weather data that may affect the determination of the design load may not be selected, thereby reducing the simulation accuracy; if T0 is too small, a large amount of historical non-extreme weather data may be selected, reducing the simulation efficiency.

[0073] The peak values ​​and daily average values ​​of three weather elements were selected as characteristic weather parameters (table), and their form referred to the outdoor design parameter form in my country's standards. On this basis, the probability characteristics of weather parameters were quantified. The characteristic weather parameters are not independent of each other, and their appearance is coupled in some way. In order to consider this coupling relationship, four groups of characteristic weather parameters that are most likely to be related were determined according to the weather formation mechanism. On this basis, the joint probability of each group is expressed with the help of conditional probability. This joint probability can reflect the probability and coupling characteristics of the parameters. The details are as follows:

[0074] 1. Joint probability: T dry,p and T dry,m

[0075] P(T dry,p T dry,m )=P(T dry,m |T dry,p )P(T dry,p )

[0076] 2. Joint probability: T dry,p and T wet,p

[0077] P(T dry,p T wet,p )=P(T wet,p |T dry,p )P(T dry,p )

[0078] 3. Joint probability: T dry,p ,T wet,p and T wet,m

[0079] P(T dry,p T wet,p T wet,m )=P(T wet,m |T wet,p )P(T wet,p |T dry,p )P(T dry,p )

[0080] Assume that T dry,p and T wet,m Independent of each other.

[0081] 4. Joint probability: T dry,p and I sun,p

[0082] P(T dry,p I sun,p )=P(T sun,p |T dry,p )P(T dry,p )

[0083] Table 1 Characteristic weather parameters

[0084]

[0085]

[0086] As shown in Table 1, in order to generate random characteristic weather parameters, the conditional probabilities of four groups are obtained according to the selected historical near-extreme weather data statistics, and then the corresponding joint probabilities are expressed. Based on this sampling, it can be used to generate random characteristic weather parameters (T dry,p ,T dry,m ,T wet,p ,T wet,m ,I sun,p ) is replaced by a coupling relationship.

[0087] In order to facilitate the analysis of the daily variation patterns of the three weather elements, the selected historical near-extreme weather data (including the three weather elements) were normalized to a coefficient of variation less than 1. The specific formula is as follows:

[0088]

[0089] In the formula, are the coefficients of change of dry-bulb temperature, wet-bulb temperature and solar radiation on the nth day; X d ,i n , are the hourly values ​​of the three weather elements on the nth day, i = 1, 2…, 24h; is the daily peak; is the daily average.

[0090] In order to maintain the coupling relationship between the daily variation patterns of the three weather elements, a joint matrix M containing the variation coefficients of the three weather elements was established. On this basis, the K-Means clustering method was used to simultaneously analyze the variation coefficients of the three weather elements. The centroid of the cluster area in the largest class was selected as the typical variation coefficient (β′ d,i ,β′ w,i ,β′ s,i ) to represent the daily variation pattern of each weather element. The joint matrix is ​​constructed as follows:

[0091]

[0092] Where M is the joint matrix of the coefficients of variation of the three weather elements, and m indicates the total number of days in the selected historical near-extreme weather data. For the joint matrix M, it plays the role of jointly analyzing the coefficients of variation of the three weather elements (dry-bulb temperature, wet-bulb temperature, and solar radiation). The hourly values ​​of the three weather elements in the selected historical near-extreme weather data are standardized to obtain the coefficient of variation of each weather element. Then, these coefficients of variation are combined into a matrix in which each row represents the coefficient of variation of a specific weather element at different hours. M will be used in the K-Means cluster analysis and will not be shown in detail.

[0093] Combining the random characteristic meteorological parameters obtained by step sampling and the typical variation coefficient obtained by cluster analysis, random near-extreme daily weather data can be constructed. The specific equation is as follows:

[0094] T dry,i =T dry,p -(T dry,p -T dry,m )*β′ d,i

[0095] T wet,i =T wet,p -(T wet,p -T wet,m )*β′ w,i

[0096] I sun,i =I sun,p *β′ s,i

[0097] Where, T dry,i 、T wet,i 、Tsun,i are random near-extreme dry-bulb temperature, wet-bulb temperature and solar radiation, i=1,2…,24h.

[0098] In order to determine the number of load samples, it is necessary to ensure the accuracy of the obtained load probability distribution with the least simulation time. Therefore, in this study, the change rate of the daily peak load average value S(N) is used as an indicator of the stability of the load probability characteristics, and S(N) < 0.2% is used as the simulation convergence criterion to determine the minimum number of simulations N minutes. The specific equation is as follows:

[0099]

[0100] In the formula, Q i,j is the load value; Q dpeak,mean (N) is the average of all daily peak loads in the previous simulation N; N is the number of simulations; ΔN is the simulation interval, which is taken as 50; S(N) is the rate of change of the daily peak load average.

[0101] The non-guarantee rate is the probability that the selected design load exceeds the hourly load under all load conditions. The calculation steps are as follows: (1) The non-guarantee rate counts the corresponding design load according to the probability distribution of the load sample; (2) Since these load samples only contain loads under near-extreme weather scenarios, R is corrected so that it can represent the non-guarantee rate under weather scenarios in previous years. The specific equation is as follows:

[0102]

[0103] R′=ηR

[0104]

[0105] Where n Qhourly>Qdesign Indicates the number of hours Q per hour of load hourly Exceeding the design load Q design ; R is the non-guaranteed rate calculated based on the load sample under the recent extreme weather scenario; R′ is the corrected non-guaranteed rate under the weather scenario in previous years; η is the correction factor, N ext is the number of days with historical extreme weather data, and d is the number of years of previous years’ data calculated.

[0106] As a possible implementation method, this embodiment uses the historical meteorological time series data of a certain city in China as the original data to verify the data results achieved by the technical solution of this application. Figure 2 and Figure 3 As shown in the figure, the temperature of some near-extreme weather data is included in the data. The data contains three types of weather elements. The historical temperature changes (based on the daily dry-bulb peak temperature greater than 31°C (T0=31°C)) are comprehensively analyzed, and about 1,800 near-extreme weather daily meteorological data are selected.

[0107] Firstly, the selected near-extreme weather data were statistically analyzed, and the daily peak values ​​and daily average values ​​of the three weather elements were used as characteristic weather parameters to obtain the conditional probability distribution of four groups of characteristic weather parameters. The probability distribution results are shown in the figure below: Figures 4 to 7 shown.

[0108] Based on the three weather elements (dry-bulb temperature, wet-bulb temperature and solar radiation) in the selected historical near-extreme weather data, the daily average and normalization calculations are performed according to the following formula to obtain the coefficient of variation of the three weather elements.

[0109]

[0110] like Figures 8 to 10 As shown in the figure, for the calculation results of the coefficient of variation, cluster analysis is performed to obtain the largest cluster of the coefficient of variation of the three types of weather elements. The class center of the largest cluster is used as the typical coefficient of variation of the weather element (as shown by the lines of different grayscales in each figure), representing the typical change law characteristics of the weather element. It can be seen that for temperature, the intraday temperature fluctuations under near-extreme weather are large, showing the characteristics of low temperature in the morning and high temperature in the afternoon.

[0111] Combining the conditional probability distribution of the four groups of weather characteristic parameters generated above and the typical variation coefficient of weather elements obtained by clustering, near-extreme day weather data are simulated and generated through random sampling, such as Fig.11 shown.

[0112] As a possible implementation, this embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, a method for simulating random near-extreme day weather data is implemented.

[0113] As a possible implementation, this embodiment provides a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, a method for simulating random near-extreme daily weather data is implemented.

[0114] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for simulating random near-extreme daily weather data, characterized in that: The specific steps include: For obtaining historical weather data, filter out historical near-extreme day weather data; Quantitative analysis of historical extreme daily weather data was performed to obtain conditional probability distributions of multiple sets of characteristic weather parameters; Based on the conditional probability distribution of multiple groups of characteristic weather parameters, random characteristic weather parameters with coupling relationships are generated through sampling methods; The daily variation patterns of historical near-extreme day weather data were normalized to obtain the coefficient of variation of characteristic weather parameters; K-Means clustering was used to analyze the coefficient of variation of characteristic weather parameters, and the centroid of the cluster area in the largest cluster was selected as the typical coefficient of variation; The typical coefficient of variation is randomly sampled to obtain random near-extreme daily weather data.

2. The method for simulating random near-extreme day weather data according to claim 1, characterized in that: The uncertainty of historical extreme daily weather data is quantified, including: Based on the daily maximum dry-bulb temperature > T0, extract the corresponding hourly dry-bulb temperature, wet-bulb temperature and solar radiation; The probability characteristics of dry-bulb temperature, wet-bulb temperature and solar radiation data are quantified to obtain the conditional probabilities of dry-bulb temperature, wet-bulb temperature and solar radiation data, where the conditional probabilities include: The conditional probability of daily dry-bulb temperature peak and the conditional probability of daily dry-bulb temperature mean for dry-bulb temperature; The conditional probability of the daily wet-bulb temperature peak and the conditional probability of the daily wet-bulb temperature mean for wet-bulb temperature; Conditional probability of daily peak solar radiation for solar radiation.

3. The method for simulating random near-extreme day weather data according to claim 2, characterized in that: The conditional probability distribution based on multiple groups of characteristic weather parameters generates random characteristic weather parameters with coupling relationship through sampling method, specifically including: According to the daily dry-bulb temperature peak conditional probability and the daily dry-bulb temperature average conditional probability of the dry-bulb temperature, the daily wet-bulb temperature peak conditional probability and the daily wet-bulb temperature average conditional probability of the wet-bulb temperature, and the daily solar radiation peak conditional probability of the solar radiation; generating the joint probability of daily peak dry-bulb temperature and daily average dry-bulb temperature, the joint probability of daily peak dry-bulb temperature and daily peak wet-bulb temperature, the joint probability of daily peak dry-bulb temperature, daily peak wet-bulb temperature and daily average wet-bulb temperature, and the joint probability of daily peak dry-bulb temperature and daily peak solar radiation; Generate random characteristic weather parameters based on joint probability.

4. The method for simulating random near-extreme day weather data according to claim 3, characterized in that: The specific calculation steps of joint probability include: The joint probability of the daily peak dry-bulb temperature and the daily mean dry-bulb temperature is: P(T dry,p T dry,m )=P(T dry,m |T dry,p )P(T dry,p ) The joint probability of a daily peak dry-bulb temperature and a daily peak wet-bulb temperature is: P(T dry,p T wet,p )=P(T wet,p |T dry,p )P(T dry,p ) The joint probability of the daily peak dry-bulb temperature, the daily peak wet-bulb temperature, and the daily average wet-bulb temperature is: P(T dry,p T wet,p T wet,m )=P(T wet,m |T wet,p )P(T wet,p |T dry,p )P(T dry,p ) The joint probability of daily peak dry-bulb temperature and daily peak solar radiation is: P(T dry,p I sun,p )=P(T sun,p |T dry,p )P(T dry,p ) Among them, T dry,p represents the daily dry-bulb temperature peak, T dry,m represents the daily average dry-bulb temperature, T wet,p represents the daily peak wet-bulb temperature, T wet,m represents the daily average wet-bulb temperature, I sun,p Represents the peak daily solar radiation.

5. The method for simulating random near-extreme day weather data according to claim 1, characterized in that: The daily variation pattern of the historical near-extreme day weather data is normalized to obtain the coefficient of variation of the characteristic weather parameters, specifically including: In the formula, are the coefficients of change of dry-bulb temperature, wet-bulb temperature and solar radiation on the nth day; X d ,i n , are the hourly values ​​of the three weather elements on the nth day, i = 1, 2…, 24; is the daily peak; is the daily average.

6. The method for simulating random near-extreme day weather data according to claim 1, characterized in that: The coefficient of variation of characteristic weather parameters is analyzed by K-Means clustering, and the centroid of the cluster area in the largest class is selected as the typical coefficient of variation, which specifically includes: Construct a joint matrix based on the coefficient of variation; Combined with the joint matrix, K-Means clustering was used to obtain the typical coefficient of variation; In the process of clustering to obtain the typical coefficient of variation, the stability and non-guarantee rate of the load probability characteristics are also evaluated.

7. The method for simulating random near-extreme day weather data according to claim 6, characterized in that: The specific steps of evaluating the stability and non-guaranteed rate of the load probability characteristic include: R′=ηR In the formula, Q i,j is the load value; Q dpeak,mean (N) is the average of all daily peak loads in the previous simulation N; N is the number of simulations; ΔN is the simulation interval; S(N) is the rate of change of the daily peak load average; Indicates the number of hours Q per hour of load hourly Exceeding the design load Q design ; R is the non-guaranteed rate calculated based on the load sample under the recent extreme weather scenario; R′ is the corrected non-guaranteed rate under the weather scenario in previous years; η is the correction factor, N ext is the number of days with historical extreme weather data, and d is the number of years of previous years’ data calculated.

8. The method for simulating random near-extreme day weather data according to claim 1, characterized in that: The random sampling of characteristic meteorological parameters and typical coefficients of variation to obtain random near-extreme day weather data specifically includes: T dry,i =T dry,p -(T dry,p -T dry,m )*β′ d,i T wet,i =T wet,p -(T wet,p -T wet,m )*β′ w,i AND sun,i =I sun,p *β′ s,i Where, T dry,i , T wet,i , T sun,i are random near-extreme dry-bulb temperature, wet-bulb temperature and solar radiation, i=1,2…,24.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the random near-extreme day weather data simulation method as described in any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the random near-extreme daily weather data simulation method as described in any one of claims 1 to 8 is implemented.

Images