A method for predicting average wind speed profile based on building wind load extreme value equivalence

Abstract

The application provides a prediction method of average wind speed profile based on building wind load extreme value equivalence. The first step is average wind speed data processing; the second step is average wind speed profile calculation model optimization; the third step is target building wind load extreme value analysis; and the fourth step is equivalent average wind speed profile characteristic parameter prediction. According to the calculation principle of wind load in the building wind load standard, the method establishes the corresponding relationship between the average wind speed profile and the target wind load extreme value, has clear physical meaning and solid theoretical basis, so that the one-to-one corresponding relationship between the average wind speed profile which originally does not have the concept of return period and the target wind load extreme value of different return periods is established, and the deficiency that the existing average wind speed profile prediction method lacks theoretical basis is overcome.

Description

TECHNICAL FIELD

[0001] The present application relates to the civil engineering technical field, and particularly to a kind of average wind speed profile prediction method based on building wind load extreme value equivalence. BACKGROUND

[0002] Average wind speed profile is the curve describing the law that average wind speed changes with height in atmospheric boundary layer, is used to calculate the average wind speed at different heights of building structure surface, and is one of important basic parameters for determining structural wind load. The average wind speed profile calculation model adopted by current world major wind load standards includes exponential law model, logarithmic law model and D-H model. At present, based on the average wind speed profile prediction of measured wind speed data, the statistical mean or envelope value of average wind speed profile characteristic parameter sequence is used as the recommended value of characteristic parameter, which lacks sufficient theoretical basis and deviates from the original intention of average wind speed profile in structural wind load standard, i.e. average wind speed profile characteristic parameter is a basic parameter for determining structural wind load extreme value. SUMMARY

[0003] Therefore, the purpose of the present application is to provide an average wind speed profile prediction method based on building wind load extreme value equivalence, establish the corresponding relationship between average wind speed profile and target wind load extreme value, so that one-to-one correspondence is established between average wind speed profile without recurrence period concept and target wind load extreme value of different recurrence periods, and overcome the deficiency of lack of theoretical basis in existing average wind speed profile prediction method.

[0004] To achieve the above purpose, the present application adopts the following technical solution: an average wind speed profile prediction method based on building wind load extreme value equivalence, characterized in that it comprises the following steps:

[0005] Step S1, pre-process the measured original wind speed data Data draw at different heights of wind measurement tower, eliminate outliers, and classify wind speed data according to climate type to obtain measured wind speed samples Data under different climates in the region;

[0006] Step S2, take measured wind speed samples Data as analysis object, respectively adopt different average wind speed profile calculation models, i.e. exponential law model, logarithmic law model and Deaves-Harris model, for fitting to obtain average wind speed profile characteristic parameters of ground roughness category in the region where the wind measurement tower is located; test the fitting goodness of average wind speed profile prediction model to obtain optimal model and take it as recommended model Model best for average wind speed profile prediction in the region; adopt the optimal average wind speed profile recommended calculation model Model best to fit measured wind speed data sequence Data to obtain corresponding average wind speed profile characteristic parameter sample sequence α;

[0007] Step S3, the average wind speed profile characteristic parameter sample sequence a is substituted into the selected building wind load standard, the wind load-shear force, bending moment, axial force acting on the target building is calculated, and the corresponding wind load sample sequence F is obtained wind ; the wind load sample sequence F of the target building wind is subjected to extreme value analysis, and the wind load extreme values of different return periods are obtained

[0008]

[0009] Step S4, according to the wind load extreme value equivalent principle of the target building, the wind load extreme values of different return periods are reversely deduced to obtain the equivalent average wind speed profile characteristic parameter prediction value of the target building through the selected building wind load standard

[0010] In a preferred embodiment, the average wind speed data processing includes measured data preprocessing, specifically including removing random codes, removing dead values, and removing outliers; the removing dead values specifically refers to defining the data with the number of continuously appearing certain measured values exceeding 10 as a dead value; the removing principle is: finding the difference between the dead value and the first data different from the dead value, and if the difference is greater than 0.5, the data is removed; the removing outliers specifically refers to: first, according to the data discrimination code of the ultrasonic anemometer, automatically discriminating invalid data generated due to rainfall, removing the invalid data, and then using linear interpolation method for interpolation;

[0011] The multiple truncated variance method is adopted to determine whether the wind speed data is an outlier, and 3 times the variance is used as the determination standard; the specific method is as follows:

[0012] First, the wind speed time history data sequence: u(i) (i = 1, 2, …, n) is constructed into its difference array sequence du(i):

[0013] du(i) = u(i+2) - u(i), i = (1, 2, …, n-2) (1)

[0014] The average values of the difference sequence du(i) and du(i) 2 are respectively and as follows:

[0015]

[0016]

[0017] The truncated variance calculation

[0018]

[0019] The criterion for determining the outlier is that u(i+2) is an outlier when |du(i)|>3σ or |du(i+2)|>3σ;

[0020] The value of the outlier is replaced by interpolation of the previous relevant wind speed sequence as follows:

[0021] u(i) = u(i-1)R m +(1+R m )U m (5)

[0022] where u(i-1) is the value at the previous time of the outlier, R m and U m are the autocorrelation coefficient and the average value of the previous 10 data samples, and m=10 is taken in the analysis.

[0023] In a preferred embodiment, the average wind speed data processing further comprises classifying the wind speed data according to climate types; the climate types specifically include typhoon climate, good-state wind climate, and mixed wind climate; the selection of the average wind speed samples in the typhoon climate is based on the following three points: the 10-min average wind speed at the 10m height of the wind tower in the influence range of the typhoon 7-grade wind circle needs to be greater than 13.9m / s, and the roughness of the underlying surface of the wind speed samples in the windward direction is consistent; the selection of the average wind speed samples in the good-state wind climate is based on the following three points: the wind speed samples during the typhoon are removed, the 10-min average wind speed at the 10m height of the wind tower is greater than 10.8m / s, and the roughness of the underlying surface of the wind speed samples in the windward direction is consistent; the selection of the average wind speed samples in the mixed wind climate is based on the following two points: the 10-min average wind speed at the 10m height of the wind tower is greater than 10.8m / s, and the roughness of the underlying surface of the wind speed samples in the windward direction is consistent.

[0024] In a preferred embodiment, the exponential law model is:

[0025]

[0026] where z b , are the reference height and the average wind speed at the height, respectively; z are the arbitrary height and the average wind speed at the height, respectively; and α is the average wind speed profile index, which is related to the roughness of the ground surface.

[0027] The logarithmic law model is:

[0028]

[0029] where k is the Karman constant, taken as 0.4; u* is the friction velocity; and z0 is the roughness length of the ground surface, used to represent the roughness of the ground surface. ​

[0030] The Deaves-Harris model is:

[0031]

[0032] wherein, is the atmospheric boundary layer height, B' is an empirical parameter, taken as 6, and f is the Coriolis parameter, taken as 7.554 x 10-5s-1.

[0033] In a preferred embodiment, the regression analysis of the mean wind speed profile includes two steps:

[0034] 1) A linear equation is obtained by changing the nonlinear mean wind speed profile equation, and the fitting parameters of the mean wind speed profile model are preliminarily obtained by using linear least squares regression;

[0035] 2) Nonlinear least squares regression is directly performed based on the nonlinear mean wind speed profile model, and the Newton iteration method is used to solve the related parameters, and the iteration initial value is selected as the fitting value preliminarily obtained in step 1); the regression process of each mean wind speed profile model is as follows:

[0036] Exponential law model:

[0037] ① Calculation of the mean wind speed profile index initial value α0

[0038] Taking the logarithm of both sides of equation (6) gives,

[0039]

[0040] wherein i is the sample number of the measured wind speed data of the wind tower; Z i is the sample point height at the number i, is the measured wind speed at the height Z i ; Z R is the reference sample point height, is the measured wind speed at the height Z R , that is, the reference wind speed;

[0041] Let The linear least squares method is applied to linearly fit the mean wind speed profile index initial value α0, and when the deviation square sum of the fitted wind speed and the measured wind speed is the smallest, the calculation formula of α0 is:

[0042]

[0043] wherein n is the number of the measured wind speed data samples of the wind tower

[0044] ② Nonlinear regression of the mean wind speed profile index α

[0045] When the regression analysis is directly performed on the mean wind speed profile model, the residual square sum is The calculation is as follows:

[0046]

[0047] Make The necessary condition for the minimum is:

[0048]

[0049] The equation is obtained:

[0050]

[0051] Equation (7) is a nonlinear equation, and it is difficult to obtain an analytical solution directly. Newton's iterative method is used to obtain its approximate solution. The Newton iteration formula for the exponential law model is:

[0052]

[0053] The initial value of iteration is the average wind speed profile index initial value α0 obtained in the previous step;

[0054] Logarithmic law model:

[0055] ① Calculation of initial values of characteristic parameters of logarithmic law model

[0056] According to equation (7), we have:

[0057]

[0058] Let y i ′ = ln(z i ), B = ln(z0), and apply linear least squares method to linearly fit the measured average wind speed at each height. When the deviation square sum of the fitted wind speed and the measured wind speed tends to zero, the calculation formula of A and B can be obtained as:

[0059]

[0060]

[0061] In the formula: is the average value of x i ′, is the average value of y i ′; After obtaining the fitting parameters A and B, the initial values of friction velocity and ground roughness z0' can be obtained.

[0062] ② Nonlinear regression of logarithmic law model

[0063] Based on the nonlinear square difference The minimum principle, the undetermined parameters in the nonlinear least square regression analysis are z0 and According to the extreme value theorem of multivariate function, the nonlinear square difference The necessary condition for the minimum is:

[0064]

[0065] That is

[0066]

[0067] This is a nonlinear equation set, it is difficult to solve directly by using Newton iteration method, by eliminating the parameters The equation is constructed as:

[0068]

[0069] Based on equation (16), the parameter z0 is obtained by using Newton iteration formula, the initial value is selected as the ground roughness length z0' obtained in step 1; after obtaining z0, the friction velocity u* can be obtained according to the following equation

[0070]

[0071] Deaves-Harris model:

[0072] Through the logarithmic law regression calculation, the ground roughness z0 and the friction velocity u* have been obtained On this basis, the average wind speed profile gradient wind height z is calculated by the following equation G At this time, the three unknowns of the Deaves-Harris model are obtained

[0073]

[0074] In the formula: B' is an empirical parameter, taking 6; f represents the Coriolis parameter, taking 7.554×10 -5 s -1 .

[0075] In a preferred embodiment, according to the minimum principle of the square sum of fitting residuals, the goodness of fit of the curve is uniformly evaluated by using the goodness of fit test index RNL, and the calculation formula is:

[0076]

[0077] In the formula: is the fitting calculation wind speed corresponding to the height z i

[0078] R NL The square sum of residuals and the relative error are combined together, obviously R​NL ≤1, R NL The closer to 1, the better the curve fitting degree is;

[0079] In the use of the coefficient of determination R NL On the basis of the fitting degree test of the wind speed profile curve, the fitting relative error calculation process is as follows:

[0080] Calculate the residual sum of squares of the fitting calculation wind speed and the measured wind speed at each height

[0081]

[0082] In the formula: is the measured value, is the fitting calculation wind speed, N is the number of sample points of a mean wind speed profile; Calculate the residual sum of squares of the fitting calculation wind speed and the measured wind speed at each height

[0083]

[0084] Calculate the relative error of the fitting mean wind speed profile

[0085] E n = RMSE / U n (26)

[0086] In the formula: U n is the reference wind speed.

[0087] In a preferred embodiment, the mean wind speed profile characteristic parameter sample sequence obtained in step S2 is substituted into the current building wind load standard, i.e. "Building Structure Load Standard" GB50009-2012, "Building Wind Resistance Design Standard and Explanation" 2015 edition, American standard ASCE7-16, Japanese standard AIJ 2004, Canadian standard National Building Code of Canada 2015, European standard En 1991-1-4:2005, Australian / New Zealand standard AS NZA 1170.2-2011 or International Standardization Association standard ISO 4351-2009, to calculate the wind load-shear force, bending moment acting on the target building, and obtain the corresponding wind load sample sequence.

[0088] In a preferred embodiment, on the basis of obtaining the target building wind load sample sequence, the extreme value analysis is used to obtain the target building wind load extreme value; the extreme value analysis mainly includes sampling, probability regression, fitting degree test and extreme value prediction, wherein:

[0089] Sampling, in extreme value analysis, the sampling method mainly includes interval maximum value method, i.e. BMM method and threshold method, i.e. POT, wherein the interval maximum value method is to assume X1, X2, …, Xn The maximum value in each interval is taken as an interval maximum value sample, and the threshold value method is to take all data greater than a fixed value in the sample, and the fixed value is called a threshold value.

[0090] Probability regression analysis includes two steps of selection of a probability distribution model and estimation of parameters, commonly used probability models for wind load extreme value prediction include a generalized Pareto distribution, a generalized extreme value distribution, a generalized normal distribution, a generalized Logistic distribution and a P-III type distribution, and commonly used parameter estimation methods include a moment method, a maximum likelihood method, a linear moment method, a least square method and a probability weight method.

[0091] Fitting degree test, for samples obtained by using the same sampling method, fitting effects of different probability distribution models are different, and there is an optimal probability distribution model; meanwhile, for the same probability model, model parameters obtained by using different parameter estimation methods are generally different, and there is a most suitable parameter estimation method; therefore, in order to obtain target building wind load samples, different sampling methods + different probability models + different parameter estimation methods are used to perform regression analysis on the samples, fitting degree test is used to compare and select different combinations of sampling methods + probability models + parameter estimation methods, and the combination of the sampling method + the probability model + the parameter estimation method with the best fitting effect is obtained; the fitting degree test method commonly includes Q-Q plot test, Kolmogorov-Smirnov test, Anderson-Darling test and chi-square test.

[0092] In a preferred embodiment, according to the target building wind load extreme value equivalence principle, based on the wind load extreme values of different return periods obtained in step S3, the equivalent average wind speed profile characteristic parameter prediction values of the target building of different return periods are obtained by reverse derivation.

[0093] Compared with the prior art, the present application has the following beneficial effects:

[0094] The present application is based on that the average wind speed profile characteristic parameter is the basic parameter for determining the extreme value of the wind load of a building structure, and through the measured average wind speed profile characteristic parameter sample sequence at different time, the corresponding wind load sample sequence acting on the structure is obtained, and then through extreme value analysis, the wind load extreme value of different return periods acting on the building structure is obtained, and then according to the equivalent principle of the building wind load extreme value, the corresponding equivalent average wind speed profile characteristic parameter prediction value of different return periods is obtained through the selected building wind load standard. According to the calculation principle of the wind load in the building wind load standard, the corresponding relationship between the average wind speed profile and the target wind load extreme value is established, which has clear physical meaning and solid theoretical basis, so that the one-to-one correspondence between the average wind speed profile which originally does not have the concept of return period and the target wind load extreme value of different return periods is established, and the deficiency of the existing average wind speed profile prediction method lacking theoretical basis is overcome. BRIEF DESCRIPTION OF DRAWINGS

[0095] Figure 1 The workflow diagram of the preferred embodiment of the present application is shown in the figure;

[0096] Figure 2 The front elevation view of the 5# residence (unit: m) of the preferred embodiment of the present application is shown in the figure;

[0097] Figure 3 The side elevation view of the 5# residence (unit: m) of the preferred embodiment of the present application is shown in the figure;

[0098] Figure 4 The plan view of the 5# residence (unit: m) of the preferred embodiment of the present application is shown in the figure;

[0099] Figure 5 The goodness-of-fit test result of the average wind speed profile under the good-state wind climate of the preferred embodiment of the present application is shown in the figure;

[0100] Figure 6 The goodness-of-fit test result of the average wind speed profile under the typhoon climate of the preferred embodiment of the present application is shown in the figure;

[0101] Figure 7 The goodness-of-fit test result of the average wind speed profile under the mixed wind climate of the preferred embodiment of the present application is shown in the figure.

[0102] In the figure: Figure 5 (a) is the goodness-of-fit coefficient R of the average wind speed profile under the good-state wind climate NL , Figure 5 (b) is the relative error of the average wind speed profile fitting under the good-state wind climate; Figure 6 (a) is the goodness-of-fit coefficient R of the average wind speed profile under the typhoon climate NL , Figure 6 (b) is the relative error of the average wind speed profile fitting under the good-state wind climate; Figure 7(a) the determination coefficient R for fitting the average wind speed profile in mixed wind climate NL , Figure 7 (b) the relative error for fitting the average wind speed profile in good wind climate. DETAILED DESCRIPTION

[0103] The application will be further described below in conjunction with the accompanying drawings and embodiments.

[0104] It should be noted that the following detailed description is illustrative only and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0105] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application; as used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise, it will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, steps, operations, elements, components, and / or groups thereof, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof.

[0106] As shown in Figure 1 The present application is a method for predicting the average wind speed profile based on the equivalent of the extreme value of building wind load, which comprises the following four steps: first, average wind speed original data processing to obtain wind speed samples under different climate types; second, regression analysis of the measured data using the exponential law, logarithmic rate and D-H model, and obtaining the optimal model of the average wind speed profile through goodness-of-fit test, and then obtaining the sample sequence of the average wind speed profile characteristic parameters; third, selecting a suitable building wind load standard to convert the sample sequence of the average wind speed profile characteristic parameters into the target building wind load sample sequence, and obtaining the extreme value of the target building wind load under different return periods through extreme value analysis; fourth, obtaining the predicted value of the average wind speed profile characteristic parameters under different return periods according to the equivalent principle of the target building wind load extreme value. In an embodiment of the present application, the method is implemented as follows:

[0107] Step S1. Average wind speed original data processing

[0108] (1) Preprocessing of measured data

[0109] Eliminate garbled data, as the data missing or garbled data caused by signal transmission or instrument failure, so first need to eliminate these garbled data.

[0110] Eliminate "stiff value", define the data whose number of continuous appearance of a certain measured value exceeds 10 as stiff value. Elimination principle: find the difference between the stiff value and the first data different from it, which is greater than 0.5, then eliminate the data

[0111] The invalid data caused by rainfall is removed. Since the typhoon landing process is often accompanied by heavy rainfall, the reliability of the data of the ultrasonic observation instrument may be affected by the rainfall. The reliability discrimination and quality control process of the sample wind speed data are as follows: ①First, according to the data discrimination code of the ultrasonic anemometer, the invalid data caused by rainfall is automatically discriminated and removed, and then the linear interpolation method is used for interpolation. ②The multiple truncated variance method is used to judge whether the wind speed data is a wild point. In this paper, 3 times variance is used as the judgment standard. The specific method is as follows:

[0112] First, the wind speed time series data sequence: u(i) (i=1, 2, …, n) is constructed into its difference array sequence du(i):

[0113] du(i)=u(i+2)-u(i)i=(1,2,...,n-2) (1)

[0114] The average values of the difference sequence du(i) and du(i) 2 are as follows:

[0115]

[0116]

[0117] The truncated variance calculation

[0118]

[0119] The wild point standard is: when |du(i)|﹥3σ or |du(i+2)|﹥3σ, u(i+2) is a wild point.

[0120] ③The value of the wild point is replaced by the interpolation of the previous related wind speed sequence as follows:

[0121] u(i)=u(i-1)R m +(1+R m )U m (5)

[0122] In the formula, u(i-1) is the value of the previous time of the wild point, R m , U m are the autocorrelation coefficient and average value of the previous 10 data samples, respectively, and m=10 is taken in the analysis.

[0123] (2) Classification of climate categories

[0124] Typhoon climate, in order to accurately reflect the characteristics of the wind field of typhoon, need to exclude the influence of typhoon peripheral circulation on the data sample, this paper based on the following 3 points: in the typhoon 7 level wind circle influence range, 10m height of the wind tower 10min average wind speed need to be greater than 13.9m / s(7 level wind), wind speed sample windward direction of the underlying surface roughness is consistent.

[0125] Good state wind climate, the average wind speed sample selection under good state wind climate based on the following 3 points: eliminate typhoon wind speed sample, 10m height of the wind tower 10min average wind speed greater than 10.8m / s(6 level wind speed), wind speed sample windward direction of the underlying surface roughness is consistent.

[0126] Mixed wind climate, mixed wind climate does not distinguish the climate category, at this time the average wind speed sample selection based on the following 2 points: 10m height of the wind tower 10min average wind speed greater than 10.8m / s(6 level wind speed), wind speed sample windward direction of the underlying surface roughness is consistent.

[0127] Step S2. Average wind speed profile calculation model is preferred

[0128] The average wind speed profile model adopted by the world's major wind load standards includes: exponential law model, logarithmic law model and D-H model, among them

[0129] Exponential law model:

[0130]

[0131] In the formula, z b , respectively, the reference height and the average wind speed at the height; z、 respectively, the arbitrary height and the average wind speed at the height; α is the average wind speed profile index, which is related to the roughness of the ground.

[0132] Logarithmic law model:

[0133]

[0134] In the formula: k is Karman constant, generally take 0.4; is the friction velocity; z0 is the ground roughness length, which is used to represent the roughness of the ground.

[0135] D-H model:

[0136]

[0137] In the formula, is the atmospheric boundary layer height, B' is the empirical parameter, generally take 6, f is the Coriolis parameter, take 7.554×10 -5 s -1.

[0138] For the obtained wind speed samples at different heights and different times, firstly, regression analysis is carried out based on the average wind speed profile models adopted by the main existing wind load standards in the world, to obtain the fitting parameter sequences of the average wind speed profile models. Then, the average wind speed profile models are compared and selected by using the goodness-of-fit test, to obtain the optimal average wind speed profile calculation model. Among them,

[0139] (1) Regression analysis of average wind speed profile

[0140] In order to accurately obtain the fitting parameters of the average wind speed profile model, the regression process of the average wind speed profile mainly includes two steps: ① the nonlinear average wind speed profile formula is changed to a linear formula, and the fitting parameters of the average wind speed profile model are preliminarily obtained by using linear least squares regression; ② nonlinear least squares regression is directly carried out based on the nonlinear average wind speed profile model, and Newton iteration method is used to solve the related parameters, and the iteration initial value is selected as the fitting value preliminarily obtained in step ①. The regression process of each average wind speed profile model is as follows:

[0141] Exponential law model:

[0142] ① Calculation of average wind speed profile index initial value α0

[0143] Taking the logarithm of both sides of formula (6) gives

[0144]

[0145] In the formula, i is the sample number of the measured wind speed data of the wind tower; Z i is the height of the sample point numbered i, is the measured wind speed at the height Z i ; Z R is the reference sample point height, is the measured wind speed at the height Z R , that is, the reference wind speed;

[0146] Let The linear least squares method is applied to linearly fit the average wind speed profile index initial value α0, and when the sum of squared deviations between the fitted wind speed and the measured wind speed is minimized, the calculation formula of α0 is:

[0147]

[0148] In the formula, n is the number of measured wind speed data samples of the wind tower

[0149] ② Nonlinear regression of average wind speed profile index α

[0150] When the average wind speed profile model is directly subjected to regression analysis, the residual sum of squares The calculation is performed as follows:

[0151]

[0152] Make The minimum necessary condition is:

[0153]

[0154] The equation is obtained:

[0155]

[0156] Formula (7) is a nonlinear equation, and it is difficult to obtain an analytical solution directly. Generally, Newton iteration method can be used to obtain an approximate solution. According to Newton iteration method, the Newton iteration formula of the exponential law model is:

[0157]

[0158] The iteration initial value is the average wind speed profile index initial value α0 obtained in the previous step

[0159] Logarithmic law model:

[0160] ① Calculation of initial values of characteristic parameters of logarithmic law model

[0161] According to formula (7), we have:

[0162]

[0163] Let y i ′ = ln(z i ), B = ln(z0), and apply linear least squares method to linearly fit the measured average wind speed at each height. When the deviation square sum of the fitted calculated wind speed and the measured wind speed tends to zero, the calculation formula of A and B can be obtained as:

[0164]

[0165]

[0166] In the formula: is the average value of x i ′, is the average value of y i ′; after obtaining the fitting parameters A and B, the initial values of friction velocity and ground roughness and z0' can be obtained.

[0167] ② Nonlinear regression of logarithmic law model

[0168] Based on nonlinear square difference Following the principle of minimization, the undetermined parameters in the logarithmic law model regression analysis are z0 and According to the extrema theorem for multivariable functions, the nonlinear difference of squares... The minimum necessary condition is:

[0169]

[0170] Right now

[0171]

[0172] This is a system of nonlinear equations, which is quite complex and difficult to solve directly using Newton's iteration method. We can solve it by eliminating parameters. The constructible equation is:

[0173]

[0174] Based on equation (16), the parameter z0 can be obtained using Newton's iterative formula, with the initial value being the ground roughness length z0' obtained in step ①. After obtaining z0, the friction velocity can be calculated using the following formula.

[0175]

[0176] DH model:

[0177] The surface roughness z0 and friction velocity have been obtained through logarithmic law regression calculation. Based on this, the average wind speed profile gradient wind height z can be calculated using the following formula. G At this point, all three unknowns of the DH model can be obtained.

[0178]

[0179] In the formula: B' is an empirical parameter, generally taken as 6; f represents the Coriolis force parameter, taken as 7.554 × 10⁻⁶. -5 s -1 .

[0180] (2) Goodness-of-fit test

[0181] Because nonlinear least squares regression is used, the conventional linear regression goodness-of-fit test method is no longer valid. Instead, based on the principle of minimizing the sum of squared residuals, the goodness-of-fit index RNL can be used to uniformly evaluate the curve's goodness of fit. Its calculation formula is as follows:

[0182]

[0183] In the formula: For height z i The wind speed is calculated by fitting the corresponding location.

[0184] R NL By organically combining the sum of squared residuals with the relative error, it is clear that R0 NL ≤1, R NL The closer a value is to 1, the better the curve fit.

[0185] Using the determination coefficient R NL Based on the goodness-of-fit test of the wind speed profile curves, and to facilitate comparison of the fitting errors of each model, the relative fitting error of the average wind speed profile curve is also given. The smaller the relative error, the better the model's fitting effect. The calculation process of the relative fitting error is as follows:

[0186] Calculate the sum of squares of the residuals between the fitted wind speed and the measured wind speed at each height.

[0187]

[0188] In the formula: These are measured values. To fit the calculated wind speed, N is the number of sample points for an average wind speed profile.

[0189] Calculate the root mean square error of the sum of squares of the residuals between the fitted wind speed and the measured wind speed at each height.

[0190]

[0191] Calculate the relative error of the fitted mean wind speed profile

[0192] E n =RMSE / U n (26)

[0193] In the formula: U n For reference wind speed.

[0194] (3) Calculation of the sample sequence of characteristic parameters of average wind speed profile

[0195] After obtaining the optimal wind profile model through goodness-of-fit test, the model is used to perform regression calculations on the measured wind speed samples to obtain the sample sequence of average wind speed profile characteristic parameters under different climates.

[0196] Step S3. Extreme Value Analysis of Target Building Wind Load

[0197] (1) Calculation of target building wind load sample sequence

[0198] The characteristic parameter sample sequence of the average wind speed profile obtained in step S2 is substituted into the current building wind load standard, i.e., the Code of Load for Building Structures (GB50009-2012), the Code for Wind Resistance Design of Buildings and Its Explanation (2015 edition), the American standard (ASCE7-16), the Japanese standard (AIJ 2004), the Canadian standard (National Building Code of Canada 2015), the European standard (En 1991-1-4:2005), the Australian / New Zealand standard (AS NZA 1170.2-2011) or the International Standardization Association standard (ISO 4351-2009), to calculate the wind load, i.e., shear force and bending moment, acting on the target building and obtain the corresponding wind load sample sequence

[0199] (2) Extreme value analysis of the target building wind load

[0200] On the basis of obtaining the target building wind load sample sequence, the extreme value of the target building wind load is obtained by using the extreme value analysis. The extreme value analysis mainly includes four steps, i.e., sampling, probability regression, goodness-of-fit test and extreme value prediction, in which:

[0201] ① Sampling:

[0202] In the extreme value analysis, the sampling methods mainly include the block maximum method (BMM method) and the threshold method (POT). In the block maximum method, it is assumed that X1, X2, …, X n are mutually independent and subject to the distribution F(x), and they are divided into several intervals with appropriate lengths, and the maximum value in each interval is taken to form a sample, which is the block maximum sample; in the threshold method, all the data greater than a fixed value in the sample are taken, and the fixed value is called the threshold. For a sample with a short observation time, the threshold method is suggested to be used for sampling.

[0203] ② Probability regression:

[0204] The probability regression analysis includes two steps, i.e., selection of the probability distribution model and estimation of the parameters. In the wind load extreme value analysis, the probability model can be selected from the generalized Pareto distribution, the generalized extreme value distribution (GEV), the generalized normal distribution, the generalized Logistic distribution and the P-III type distribution, and the parameter estimation method can be selected from the moment method, the maximum likelihood method, the linear moment method, the least square method and the probability weight method; in which:

[0205] Generalized Pareto distribution:

[0206]

[0207] In the formula, u is the threshold value, which is also the location parameter of the distribution; σ is the scale parameter; and γ is the shape parameter.

[0208] Generalized Extreme Value (GEV) distribution:

[0209] Extreme Value Type I (Gumbel distribution)

[0210]

[0211] Extreme Value Type II (Frechet distribution)

[0212]

[0213] Extreme Value Type III (Weibull distribution)

[0214]

[0215] where μ is the location parameter, σ is the scale parameter, and γ is the shape parameter.

[0216] Generalized Normal distribution:

[0217]

[0218] Generalized Logistic distribution:

[0219]

[0220] where, when k≠0, when k=0,

[0221] Further simplifying equation (32) gives:

[0222]

[0223] P-III Type distribution:

[0224]

[0225] where Γ(α) is the gamma function,

[0226] 3. Goodness-of-fit test:

[0227] In the probability regression analysis, for the samples obtained by using the same sampling method, the fitting effects of different probability distribution models are different, and there is an optimal probability distribution model; at the same time, for the same probability model, the model parameters obtained by using different parameter estimation methods are generally different, and there is the most suitable parameter estimation method. Therefore, in obtaining the target building wind load sample, the sample needs to be analyzed by using "different sampling methods + different probability models + different parameter estimation methods", and the combination of "different sampling methods + probability models + parameter estimation methods" is compared and selected through the goodness-of-fit test, so as to obtain the combination of "sampling method + probability model + parameter estimation method" with the best fitting effect for the sample. The goodness-of-fit test method commonly includes the commonly used goodness-of-fit test methods such as Q-Q plot test method, Kolmogorov-Smirnov test method (K-S test), Anderson-Darling test method (A-D test) and chi-square test

[0228] Extreme value prediction:

[0229] The optimal "sampling method + probability model + parameter estimation method" combination obtained in the above step is used for extreme value analysis of the target building wind load sample sequence, so as to obtain the wind load extreme values of different return periods.

[0230] Step S4. Prediction of equivalent average wind speed profile characteristic parameters

[0231] According to the equivalent principle of the target building wind load extreme value, the equivalent average wind speed profile characteristic parameters of the target building are obtained by reverse derivation based on the wind load extreme values of different return periods obtained in step S3. The following is a specific embodiment of the present application, which is based on the measured wind speed and direction data of a certain wind tower in Pingtan, and takes the #5 high-rise building of Xinyuan community in Pingtan comprehensive experimental zone (two shores standard common pilot project) as an example, and the detailed description is as follows, but the present application is not limited thereto.

[0232] Table 1 Comparison of parameter estimation methods of each probability distribution model under POT extreme value sample in good wind climate

[0233]

[0234] Table 2 Comparison of parameter estimation methods of each probability distribution model under POT extreme value sample in typhoon

[0235]

[0236]

[0237] Table 3 Comparison of parameter estimation methods of each probability distribution model under POT extreme value sample in mixed climate

[0238]

[0239] Table 4. Predicted results of base shear (kN) under different wind climates and return periods

[0240]

[0241] Table 5. Predicted results of average wind speed profile index (kN) under mixed wind climate and different return periods

[0242]

[0243] Example 1

[0244] Basic information of the target building:

[0245] Xinyuan on both sides of Pingtan is the first demonstration community mainly built for Taiwan compatriots, and is also one of the key construction projects in the experimental zone in 2020. The project aims to build a new industrial zone, a high-end service zone and a living area for the promotion of life, and to serve as a window for cooperation between Fujian and Taiwan, and to build a common home for both sides. The project is located at the intersection of Xinggang Road and Shunyi Road in the center of Jinjing District on Pingtan Island. The total land area of the project is 20081 m2, and the total construction area is 65376.34 m2, including four high-rise residential buildings and commercial buildings. The four high-rise residential buildings in the community have basically the same height and shape, and are not lost in general. In this example, the #5 residential building is taken as the target building for illustration. The total height of the 5# residential building is 83.1 m, with 26 floors. Except that the floor height of the 1st floor is 3.1 m and the floor height of the top floor is 4.5 m, the floor height of the remaining floors is 3.0 m. The width of the front face of the residential building is 34.8 m, and the width of the side face is 16.2 m. The front elevation, side elevation and plan of the 5# residential building are shown in Figure 2 、 Figure 3 and Figure 4 respectively.

[0246] Basic information of the measured data:

[0247] The wind observation tower is located in Fuguan Village, Jinjing Town, Jinjiang City, Fujian Province, with longitude and latitude of 118°37′49.6″ and 24°32′54.2″ respectively. The surrounding topographic features are coastal landforms. The wind observation tower is 70 m high, and wind speed meters are installed at 10 m, 30 m, 50 m and 70 m from the ground respectively. The wind speed meters at each height of the wind observation tower output 10 min average wind speed and corresponding wind direction (no wind direction data at 30 m) data every 10 min. The recording time is from April 2009 to February 2011.

[0248] According to the method steps S1 described in the invention content, the obtained measured data is preprocessed, and abnormal data is removed. According to the selection criteria of different climate samples, the data is screened.

[0249] The mean wind speed profile calculation model is optimized according to step S2 of the method described in the invention. First, regression analysis is performed on the processed data based on exponential law, logarithmic law, and DH models to obtain the characteristic parameters of different mean wind speed profile models. Then, the fitting determination coefficient RNL and the mean of the fitting relative error of the three mean wind speed profile models are calculated respectively. The goodness-of-fit test results for benign wind, typhoon, and mixed wind climates are shown in the appendix. Figure 5 , 6 As shown in Figure 7. (From Appendix) Figure 5 (a) It can be seen that the fitting determination coefficient R of the three mean wind speed profiles under favorable wind climate conditions is... NL All values ​​are greater than 0.95 and close to 1, indicating that the three average wind speed profile models have good fitting effects. (See attached...) Figure 5 (b) It can be seen that the mean relative fitting error of the three average wind speed profile models is less than 3%, which also shows a good fitting effect. Comparing the relative fitting errors of the three models, it can be seen that the relative fitting errors of the exponential law and logarithmic law are relatively close, and both are much smaller than the relative error of the DH model; under typhoon climate, the attached... Figure 6 (a) It can be seen that for the five typhoons, regardless of the wind profile model used, the fitting determination coefficient R is consistent. NL All are greater than 0.95 and all are close to 1. Meanwhile, from the attached... Figure 6 (b) It can be seen that the relative fitting errors of the average wind speed profiles for the five typhoons, except for Typhoon Linfa, are all less than 3%, and the relative fitting error for Typhoon Linfa is also less than 5%, indicating that the fitting effects of the three average wind speed profile models under typhoon climate conditions are all good. Under typhoon climate conditions, comparing the relative fitting errors of the three models shows that, for any typhoon, the relative fitting errors of the exponential law and logarithmic law models are relatively close, and both are smaller than the relative error of the DH model; Appendix Figure 7 The results of the goodness-of-fit test for the mean wind speed profile under mixed wind climate are similar to those under benign wind climate, and will not be repeated here. In summary, the exponential law model and the logarithmic law model fit better than the DH model for benign wind, typhoon, and mixed wind climates. The fitting effects of the exponential law and logarithmic law models are close. Since the Fujian region adopts the Building Structure Load Code (GB 50009-2012) for wind-resistant design, and the code recommends using the exponential law model for wind load calculation, the exponential law model is selected as the optimal mean wind speed profile calculation model for this region in this example.

[0250] According to the method step S2 described in the invention, the optimal average wind speed profile calculation model, namely the exponential law model, is used to perform regression calculation on the wind speed measured samples processed under different climates to obtain the sample sequence of the average wind speed profile index α under different climates.

[0251] The method S3 according to the application is used for the target building of the 5th residence of the community on both sides of the fragrance garden, and based on the fitting average wind speed profile index sample sequence obtained in step S2, the wind load sequence of the target building is calculated according to the building structure load specification (GB 50009-2012). This example takes the base shear force of the building along the wind direction as an example for illustration.

[0252] The IWL method (inertial force load method) is used for calculating the wind load along the wind direction of the high-rise building in the building structure load specification, and the static equivalent load of the structure is calculated in the form of average wind pressure multiplied by wind vibration coefficient β z , wherein the standard value of the wind load on the main force structure surface shall be calculated according to the following formula:

[0253] w k =β z μ s μ z w0 (1)

[0254] In the formula, w k is the standard value of the wind load; β z is the wind vibration coefficient at the height z; μ s is the wind load shape coefficient; μ z is the wind pressure height variation coefficient; and w0 is the basic wind pressure. The above coefficients are respectively taken as follows.

[0255] (1) Basic wind pressure w0:

[0256] Since the basic wind pressure range given by the building structure load specification (GB 50009-2012) in Fujian is relatively rough, in order to more accurately obtain the basic wind pressure of the station location, the recommended value given in the Fujian building structure basic wind pressure specification is selected, and the basic wind pressure of the station location with a 50-year return period is taken as 0.75 kN / m 2 :

[0257] (2) Shape coefficient:

[0258] The cross section of the 5th residence is a rectangular cross section, the front width is 34.8 m, and the side width is 16.2 m. Since the building height is greater than 45 m, the residence shape coefficient needs to be taken according to the standard table 8.3.1 item 31. For the building depth-width ratio perpendicular to the front of the residence in the incoming flow direction, it is 16.2 / 34.8=0.466<1, so the shape coefficient of the incoming flow wind pressure in this direction is μ s =0.8-(-0.6)=1.4.

[0259] (3) Wind pressure height variation coefficient:

[0260] Formula (2) is a standard wind pressure height variation coefficient calculation formula, in which coefficient b is a basic wind pressure conversion coefficient under different landforms, and the product of this parameter and basic wind pressure w0 can obtain the basic wind pressure at 10m height under the landform of the building. The selected sample in this paper is under the standard A landform, and the value of A landform is taken as 1.284. Parameter a in the formula is the average wind speed profile index, which is selected according to the sample sequence of average wind speed profile index a under different climates.

[0261]

[0262] (4) Along-wind wind vibration coefficient:

[0263] Since the height of the residence is greater than 30m, and the height-width ratio is 2.388, which is greater than 1.5, and the basic natural vibration period T1 is 1.64s, which is greater than 0.25s, the influence of fluctuating wind on the structure should be considered. The along-wind wind vibration response calculation should be carried out according to the structure random vibration theory, and the wind vibration coefficient method can be used to calculate the along-wind wind load, and the wind vibration coefficient at height z can be calculated according to the following formula:

[0264]

[0265] In the formula: g is the peak factor, which is taken as 2.5; I 10 is the nominal turbulence intensity at 10m height, since the residence is surrounded by A landform, the value is taken as 0.12; R is the resonance component factor of fluctuating wind load, which is obtained from formula (4); B z is the background component factor of fluctuating wind load, which is obtained from formula (5).

[0266]

[0267]

[0268] In the formula: f1 is the first-order natural frequency of the structure (Hz), the first-order natural frequency of the residence is 0.61Hz; k w is the ground roughness correction coefficient, which is taken as 1.28 for A ground roughness; ζ1 is the structural damping ratio, which is taken as 0.05 for reinforced concrete and masonry structure.

[0269]

[0270] In the formula: φ1(z) is the first-order mode shape coefficient of the structure, which should be determined according to the dynamic calculation of the structure. Since the shape, mass and stiffness of the residence in this paper are relatively uniform, the calculation is carried out according to the recommended formula in the standard, i.e. formula (7).

[0271]

[0272] H is the total height of the structure, which is taken as 83.1 according to the actual height of the building; ρx with p z are the horizontal and vertical direction correlation coefficients respectively, calculated by equations (8) and (9); k and a1 are 0.944 and 0.670 respectively, taken from standard table 8.4.5-1.

[0273] The vertical direction correlation coefficient can be calculated as follows:

[0274]

[0275] The horizontal direction correlation coefficient can be calculated as follows:

[0276]

[0277] wherein B is the windward surface width of the structure.

[0278] (5) Along-wind load:

[0279] The wind load on each floor is calculated by equation (10) according to the building height along the height of the residence:

[0280] F i = β zi μ s μ zi w0Δh i B i (10)

[0281] The along-wind load on each floor is superimposed to obtain the base shear V of the target building. The base shear V of the target building is subjected to extreme value analysis according to the method described in the disclosure content step S3:

[0282] First, the base shear V of the target building is sampled. Since the data obtained in this example is relatively short in the measurement period, it is not convenient to divide groups, so the crossing threshold method is used for sampling. When the crossing threshold method is used for sampling, the average excess function diagram method and the estimation stability discrimination method are used to determine the optimal base shear threshold of the target building under different climates. The optimal threshold values under the climates of good wind, typhoon and mixed wind are 21700kN, 21500kN and 21700kN respectively.

[0283] According to the method steps in the content of the application, after obtaining the samples exceeding the threshold, the samples need to be subjected to probability regression analysis. In the present example, the probability models of generalized Pareto distribution, generalized extreme value distribution (GEV), generalized normal distribution, generalized Logistic distribution, P-III type distribution and the like are subjected to regression analysis. According to the selected probability model, different parameter estimation methods are selected to estimate the unknown parameters of the probability model, and the parameter estimation methods include moment method, maximum likelihood method, linear moment method, least square method and probability weight method. The parameter estimation results of each probability model under the climate of good wind, typhoon and mixed wind are shown in Tables 1, 2 and 3.

[0284] According to the method step S3 in the content of the application, after the regression analysis, the prediction combination of "sampling method + probability model + parameter estimation method" needs to be selected by using goodness-of-fit test. In the present example, the K-S test and A-D test are used for goodness-of-fit test, and the test results under the climate of good wind, typhoon and mixed wind are shown in Tables 1, 2 and 3. As shown in Table 1, for the target building base shear samples exceeding the threshold obtained by using POT sampling method under the climate of good wind, the K-S test value and A-D test value obtained by using "GPD distribution + moment method" combination for regression analysis are the smallest, so "POT sampling + GPD distribution + L-moment method" is the best prediction combination for the target building base shear samples under the mixed climate. Similarly, the best prediction combinations under the climate of typhoon and mixed wind are "POT sampling + GPD distribution + L-moment method" and "POT sampling + GLO distribution + L-moment method" respectively.

[0285] According to the method step S3 in the content of the application, after obtaining the best prediction combination, the extreme value analysis of the target building base shear is carried out by using the best prediction combination under each climate, and the extreme values of the target building base shear under different return periods are shown in Table 4.

[0286] According to the method step S4 in the content of the application, for the obtained extreme values of the target building wind load under different return periods, based on the equivalent principle of building wind load extreme value, the corresponding return period average wind speed profile index prediction value a is inversely solved according to the calculation process of the target building base shear in the wind direction, as shown in Table 5.

[0287] The above is the preferred embodiment of the present application, and any changes made according to the technical solutions of the present application, as long as the generated function does not exceed the scope of the technical solutions of the present application, are all within the protection scope of the present application.

Claims

1. A method for predicting the equivalent mean wind speed profile based on the extreme value of the wind load on a building, characterized in that The method comprises the following steps: Step S1, obtaining raw wind speed data at different heights of the wind measurement tower The raw wind speed data is preprocessed to eliminate abnormal values and classified according to climate types to obtain measured wind speed samples under different climates ; Step S2, fitting the measured wind speed sample For the analysis object, different average wind speed profile calculation models, namely exponential law model, logarithmic law model and Deaves-Harris model, are used respectively to perform fitting to obtain the average wind speed profile characteristic parameters of the ground roughness category of the area where the wind measurement tower is located; the fitting degree of the average wind speed profile prediction model is tested to obtain the optimal model and use it as the recommended model Model best for the average wind speed profile prediction of the area; the optimal average wind speed profile recommended calculation model Model best is used to perform fitting on the measured wind speed data sequence to obtain the corresponding average wind speed profile characteristic parameter sample sequence α; Step S3, the average wind speed profile characteristic parameter sample sequence α is substituted into the selected building wind load standard, the wind load-shear force, bending moment, axial force acting on the target building is calculated, and the corresponding wind load sample sequence is obtained ; the wind load sample sequence of the target building is subjected to extreme value analysis to obtain wind load extreme values of different return periods ; Step S4, according to the target building wind load extreme value equivalence principle, based on the wind load extreme value of different return period , the equivalent average wind speed profile characteristic parameter prediction value of the target building is obtained by reverse derivation of the selected building wind load standard .

2. The method according to claim 1, wherein, The average wind speed data processing comprises measured data preprocessing, specifically including removing garbled codes, removing dead values and removing outliers; The removing of the dead values specifically comprises defining data with a number of continuously appearing certain measured values exceeding 10 as the dead values; the removing principle is: finding out the difference between the dead value and the first data different from the dead value, and if the difference is greater than 0.5, the data is removed; The removing of the outliers specifically comprises: firstly, automatically judging invalid data generated due to rainfall influence according to the data discrimination code of the ultrasonic anemometer, and removing the invalid data; and then, using a linear interpolation method for interpolation; The multiple-truncated variance method is adopted to judge whether the wind speed data is an outlier, and 3 times variance is adopted as the judgment standard; the specific method is as follows: Firstly, a wind speed time series data sequence u(i) (i=1, 2,..., n) is constructed into a difference array sequence du(i): （1） The average of the difference sequence is calculated and respectively and is as follows: （2） （3） The truncated variance calculation thereof is: (4) The criterion for determining outliers is that when or then is an outlier; The value of the outlier is replaced by the interpolation of the previous related wind speed sequence as follows: （5） wherein is the value at the previous time of the wild point, and are the autocorrelation coefficient and the average value of the 10 data samples before the wild point, respectively.

3. The method of claim 1, wherein the method is characterized by: The average wind speed data processing further comprises classifying the wind speed data according to climate types; the climate types specifically comprise a typhoon climate, a good-state wind climate and a mixed wind climate; The average wind speed sample selection under the typhoon climate is based on the following three points: within the influence range of a typhoon 7-grade wind circle, the 10-minute average wind speed at the 10m height of the wind measurement tower is greater than 13.9 m / s, and the roughness of the underlying surface in the windward direction of the wind speed sample is consistent; The average wind speed sample selection under the good-state wind climate is based on the following three points: removing the wind speed sample during the typhoon, the 10-minute average wind speed at the 10m height of the wind measurement tower is greater than 10.8 m / s, and the roughness of the underlying surface in the windward direction of the wind speed sample is consistent; The average wind speed sample selection under the mixed wind climate is based on the following two points: the 10-minute average wind speed at the 10m height of the wind measurement tower is greater than 10.8 m / s, and the roughness of the underlying surface in the windward direction of the wind speed sample is consistent.

4. The method of claim 1, wherein the method is characterized by: The exponential law model is: (6) wherein , are the reference height and the mean wind speed at this height, respectively; , are an arbitrary height and the mean wind speed at this height, respectively; is the mean wind speed profile exponent, which is related to the roughness of the ground. The logarithmic law model is: (7) In the formula, k is Karman constant, and 0.4 is taken; is the friction velocity; is the ground roughness length, used to represent the roughness degree of the ground; The Deaves-Harris model is as follows: （8） wherein is the atmospheric boundary layer height, is an empirical parameter, taken as 6, f is the Coriolis parameter, taken as 7.554 x 10 -5 s -1 .

5. The average wind speed profile prediction method based on building wind load extreme value equivalence according to claim 4, characterized in that, The average wind speed profile regression analysis comprises two steps: 1) a linear formula is obtained by changing a nonlinear average wind speed profile formula, and fitting parameters of the average wind speed profile model are preliminarily obtained by using linear least square regression; 2) nonlinear least square regression is directly performed based on the nonlinear average wind speed profile model, and relevant parameters are solved by using Newton iteration method, and the iteration initial value is selected to be the fitting value preliminarily obtained in step 1); the regression process of each average wind speed profile model is as follows: Exponential law model: ① average wind speed profile exponential initial value α0 calculation Taking logarithm on both sides of equation (6), we have (9) wherein: i is the sample number of the measured wind speed data of the wind tower; Z i is the sample point height at the i-th, is the measured wind speed at the height Z i ; Z R is the sample point height at the i-th, is the measured wind speed at the height Z R , i.e. the reference wind speed; Set , , the average wind speed profile index initial value a0 is linearly fitted by using linear least square method, and the calculation formula of a0 is obtained when the deviation square sum of fitted calculation wind speed and measured wind speed is minimum: (10) In the formula, n is the total number of wind speed data samples of the wind measurement tower; ② average wind speed profile nonlinear regression of exponential law When regression analysis is directly performed on the average wind speed profile model, the residual sum of squares The calculation is performed by the following equation: (11) To make The minimum necessary conditions are: (12) The equation is obtained: (13) Formula (7) is a nonlinear equation, and it is difficult to directly obtain an analytical solution, and the Newton iteration method is used to obtain an approximate solution; according to the Newton iteration method, the Newton iteration formula of the exponential law model is as follows: （14） The iteration initial value is selected to be the average wind speed profile exponential initial value α0 obtained in the previous step; Logarithmic law model: ① calculation of logarithmic law model characteristic parameter initial value From equation (7) we have: (15) Set , , , , the linear least square method is applied to linear fitting of each height measured average wind speed, and when the deviation square sum of fitting calculation wind speed and measured wind speed tends to zero, the calculation formula of A and B can be obtained. （16） （17） wherein: is the average value of is the average value of is the average value of is the average value of After getting the fitting parameters A, B, the friction velocity can be calculated and the initial value of the ground roughness ; ② nonlinear regression of logarithmic law model Based on the nonlinear square difference The principle of the minimum, the undetermined parameter in the regression analysis of the logarithmic law model is and According to the extreme value theorem of multivariate function, the nonlinear square difference The necessary condition for the minimum is: , (18) That is (19) This is a nonlinear equation set, directly using Newton iteration method for solving complex difficult, by eliminating parameters The equation is constructed as: (20) Based on formula (16) using Newton iteration formula, i.e. to obtain the parameters , the initial value is selected from the ground rough length obtained in step ① ; after obtaining , the friction velocity can be obtained according to the following formula: (21) Deaves-Harris model: The ground roughness and the friction velocity The average wind speed profile gradient wind height is then calculated from (22) wherein: is an empirical parameter, taken as 6; denotes the Curie parameter, taken as .

6. The method of predicting the equivalent mean wind velocity profile based on the extreme value of the wind load of a building according to claim 5, wherein, According to the principle of least square sum of fitting residual, the goodness of fit index R NL The goodness of fit of the curve is evaluated uniformly, and the calculation formula is: （23） In the formula: is the height z i corresponds to the fitted calculated wind speed; The residual sum of squares is combined with the relative error to give a figure of merit , The closer to 1, the better the curve fit. In the use of the determination coefficient On the basis of the goodness-of-fit test of wind speed profile curve, the fitting relative error calculation process is shown as follows: Calculate the sum of squares of the residuals between the fitted wind speed and the measured wind speed at each height. (24); Calculate the root mean square error of the sum of squares of the residuals between the fitted wind speed and the measured wind speed at each height. (25); Calculate the relative error of the fitted mean wind speed profile. (26) In the formula: is the reference wind speed.

7. The method of predicting the equivalent mean wind velocity profile based on the extreme value of the wind load of a building according to claim 6, wherein, The average wind speed profile characteristic parameter sample sequence obtained in step S2 is substituted into the current building wind load standard to calculate the wind load-shear force, bending moment acting on the target building, and a corresponding wind load sample sequence is obtained.

8. The method of predicting the equivalent mean wind velocity profile based on the extreme value of the wind load of a building according to claim 7, wherein, On the basis of obtaining the target building wind load sample sequence, the extreme value of the target building wind load is obtained by extreme value analysis; the extreme value analysis includes sampling, probability regression, goodness-of-fit test and extreme value prediction, wherein: Sampling, in the extreme value analysis, the sampling method includes the interval maximum method, namely BMM method and threshold method, namely POT, wherein the interval maximum method is to assume X1, X2, …, X n The threshold method is to take all data greater than a fixed value in the sample, and the fixed value is called threshold. The probability regression analysis includes two steps of selection of a probability distribution model and estimation of parameters; the probability model for wind load extreme value prediction includes generalized Pareto distribution, generalized extreme value distribution, generalized normal distribution, generalized Logistic distribution and P-III type distribution, and the parameter estimation method includes moment method, maximum likelihood method, linear moment method, least square method and probability weight method; The goodness-of-fit test is used to obtain the target building wind load sample, and different sampling methods+different probability models+different parameter estimation methods are used for regression analysis of the sample; the combination of different sampling methods+probability models+parameter estimation methods is compared and selected through the goodness-of-fit test, and the combination of the sampling method+probability model+parameter estimation method with the best fitting effect for the sample is obtained; the goodness-of-fit test method commonly includes Q-Q plot test method, Kolmogorov-Smirnov test method (K-S test), Anderson-Darling test method (A-D test) and chi-square test; The extreme value prediction is performed on the target building wind load sample sequence by using the best sampling method+probability model+parameter estimation method combination obtained in the previous step, and the wind load extreme value of different return periods is obtained.

9. The method of claim 8, wherein the method is characterized by: According to the equivalent principle of the target building wind load extreme value, based on the wind load extreme value of different return periods obtained in step S3, the equivalent average wind speed profile characteristic parameter prediction value of the target building of different return periods is obtained by reverse deduction.

Images