A structural temperature gradient simulation model and load characteristic value calculation method

Abstract

The present invention discloses a structural temperature gradient simulation model and a load characteristic value calculation method, which are used for non-repetitive simulation of structural temperature gradient annual cycle and solution of load characteristic representative value, including: obtaining structural temperature data of a complete annual cycle; analyzing the structural temperature data of a complete annual cycle, obtaining the daily maximum equivalent linear temperature gradient, and constructing a simulation model of the structural equivalent linear temperature gradient for an annual cycle; generating the daily maximum linear temperature gradient of a complete annual cycle based on the simulation model of the structural equivalent linear temperature gradient for an annual cycle, and using the Monte Carlo simulation sampling method and generalized extreme value distribution fitting to obtain the structural temperature gradient load characteristic value of once in m years. The present invention solves the problem that it is difficult to obtain enough annual temperature extreme value samples in engineering practice, has high calculation accuracy, and is applicable to different climatic conditions and structures, and can provide a reference for the design and maintenance of engineering structures.

Description

Technical Field

[0001] The invention belongs to the technical field of structural load characteristic representative value analysis and calculation, and relates to solving structural temperature gradient load characteristic representative value. Specifically, a structural temperature gradient simulation model and a load characteristic value calculation method are proposed. Background Art

[0002] The internal forces of super-high, super-long and large-span structures are more sensitive to temperature changes. Temperature effect is one of their main control conditions and cannot be ignored. For example, the vertical deformation caused by the temperature gradient of the large-span main beam of the high-speed railway will cause the track to be uneven, which will affect the safe operation of the train. Super-long steel-concrete structures such as terminal buildings, railway stations, and gymnasiums have temperature stress, and the deformation caused by temperature needs to be accurately calculated and controlled. The temperature stress of large-volume concrete dams in water conservancy hubs is the main factor causing uneven cracking of the dam.

[0003] Therefore, choosing the appropriate temperature load characteristic value during the structural design stage is crucial to the service life and operational safety of the structure. However, the specifications often do not have provisions for the temperature gradient load characteristic values ​​of special structures such as super-high, super-long and large-span structures. For special engineering structures, the temperature gradient load characteristic value needs to be calculated separately, and the conventional load characteristic value calculation method needs to be based on a large number of annual temperature extreme value samples, but it is difficult to obtain enough annual temperature extreme value samples in engineering practice. Therefore, it is urgent to propose a structural temperature gradient simulation model and load characteristic value calculation method based on a limited number of structural temperature data samples. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a structural temperature gradient simulation model and a load characteristic value calculation method which can accurately evaluate the temperature gradient load characteristic value of a structural component based on a short-term finite structural temperature data sample.

[0005] The present invention provides a method for constructing a structural temperature gradient simulation model, comprising the following steps:

[0006] S1. According to the structural characteristics and the measured data of the structural temperature over a complete annual cycle, the uniform temperature of the structure is obtained, which is calculated as follows:

[0007]

[0008] Among them, T i is the temperature of measuring point i, A i is the area represented by measuring point i, and A is the cross-sectional area;

[0009] S2. Obtaining the cross-sectional temperature gradient T(A) of the structure. The specific method includes: converting the measured temperature value T of each measuring point of the cross section i Subtract the uniform temperature of the cross section T uni, get the temperature difference T of the measuring point di ;

[0010] The point with the smallest temperature difference is taken as the reference point, and the temperature difference of the remaining points is subtracted from the reference point to obtain the final cross-sectional temperature gradient T(A);

[0011] S3. Obtain vertical linear temperature gradient ΔT z , the specific method includes: According to Hooke's law and the law of thermal expansion, the structural temperature self-stress is expressed as:

[0012] σ=EαΔT (2)

[0013] Where, σ is stress; E is elastic modulus; α is linear expansion coefficient; ΔT is uniform temperature change;

[0014] If the cross-sectional temperature gradient T(A) is nonlinear, the cross-sectional bending moment calculation formula is:

[0015]

[0016] Where M y is the bending moment; T(A) is the nonlinear temperature gradient; z is the vertical coordinate of the cross section; A is the cross-sectional area; if the cross-sectional temperature gradient is linear, then the calculation formula for the cross-sectional bending moment is:

[0017]

[0018] In the formula, ΔT z is the vertical linear temperature gradient; I y is the section moment of inertia; h z is the section height;

[0019] For an Euler beam, the curvature is related to the bending moment by:

[0020]

[0021] In the formula, κ is the curvature. From formula (5), it can be seen that no matter it is a nonlinear temperature gradient or a linear temperature gradient, as long as the bending moment generated in the cross section is the same, the structural deformation caused is also the same. Therefore, the simultaneous formulas (3) and (4) are:

[0022]

[0023] The conversion formula between the nonlinear temperature gradient T(A) and the linear temperature gradient ΔT is obtained as follows:

[0024]

[0025] For the discrete temperature field obtained through experimental monitoring, the temperature gradient between the measuring points changes linearly, and then the section bending moment is obtained by piecewise integration and summation. The conversion formula for discrete nonlinear temperature gradient and linear temperature gradient is:

[0026]

[0027] Where n is the number of temperature sensors; T i is the temperature gradient function of the ith region; z i A is the height coordinate of the cross section of the i-th region; i is the area of ​​the ith region;

[0028] S4. Count and draw the daily maximum linear temperature gradient change curve of the structure within a complete annual cycle;

[0029] S5. Obtain the structural temperature gradient simulation model, and calculate as follows:

[0030] The annual cycle trend term f(t) of the daily maximum linear temperature gradient is fitted using a trigonometric function:

[0031]

[0032] Where, T av is the annual average of the maximum daily average temperature; T am is the annual fluctuation range of the daily maximum average temperature; T is the fluctuation period of the daily maximum average temperature, in days; It is the ordinal number of the day when the daily maximum average temperature reaches the annual minimum value;

[0033] Separate the periodic trend term from the daily maximum linear temperature gradient data to obtain the random fluctuation term R(t) of the daily maximum linear temperature gradient;

[0034] According to the histogram of the random fluctuation term, searching for the distribution function with the highest histogram fitting accuracy, and using maximum likelihood estimation to obtain parameter estimates of the distribution function;

[0035] Based on the estimated parameter values, the KS test and the chi-square test are used to perform a goodness of fit test on the distribution function to determine the random fluctuation term distribution function;

[0036] The structural temperature gradient simulation model is obtained by combining the periodic trend function and the random fluctuation term distribution function, which is expressed as follows:

[0037]

[0038] Among them, Tav is the annual average value of the daily maximum average temperature; Tam is the annual fluctuation range of the daily maximum average temperature; T is the fluctuation period of the daily maximum average temperature, in days; is the day number when the daily maximum average temperature reaches the annual minimum value; μ is the mean of the Gaussian distribution function of the random fluctuation term, and σ is the standard deviation of the Gaussian distribution function of the random fluctuation term.

[0039] On the other hand, to achieve the above-mentioned purpose, the present invention provides a method for calculating load characteristic values ​​based on a structural temperature gradient simulation model, comprising:

[0040] Obtain structural temperature data for a complete annual cycle;

[0041] Analyze the structural temperature data of the complete annual cycle to obtain the daily maximum equivalent linear temperature gradient and build a structural temperature gradient simulation model;

[0042] Based on the structural temperature gradient simulation model, the maximum daily linear temperature gradient of several complete annual cycles is cyclically generated, and the Monte Carlo simulation sampling method and generalized extreme value distribution fitting are used to obtain the characteristic value of the structural temperature gradient load once in m years; based on the structural temperature gradient simulation model, the maximum daily linear temperature gradient of a complete annual cycle is generated, and the calculation is as follows:

[0043] T Z_ymax =Max{ΔT ZM (t)},t∈(1,365) (11)

[0044] Where, T Z_ymax is the annual linear temperature gradient extreme value; T ZM is the daily maximum linear temperature gradient; t is the day number;

[0045] The method for obtaining the characteristic value of the structural temperature gradient load with a return period of m years by using the Monte Carlo simulation sampling method and the generalized extreme value distribution fitting according to the daily maximum linear temperature gradient of the complete annual cycle includes:

[0046] The Monte Carlo simulation sampling method is used to simulate the daily maximum linear temperature gradient change process of n complete annual cycles using the structural daily maximum linear temperature gradient simulation model;

[0047] According to n annual temperature data curves, n linear temperature gradient annual extreme value samples are obtained respectively;

[0048] The generalized extreme value distribution is used to fit the cumulative distribution function of the annual extreme value sample data, and the KS test and chi-square test are used to confirm the fitting parameters. The quantile value corresponding to the exceedance probability is selected in the cumulative distribution function to obtain the characteristic value of the structural temperature gradient load with a return period of once in m years.

[0049] Technical effect of the invention: The invention discloses a structural temperature gradient simulation model and a load characteristic value calculation method, which solves the problem that it is difficult to obtain sufficient annual temperature extreme value samples in engineering practice; by introducing the idea of ​​Monte Carlo simulation method, the conventional temperature load characteristic value calculation method is improved, and a calculation method for obtaining long-term temperature load characteristic values ​​based on short-term temperature monitoring data is proposed; the method has high calculation accuracy and is suitable for different climatic conditions and structures, and can provide a reference for the design and maintenance of engineering structures. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A schematic diagram of a flow chart of a structural temperature gradient simulation model and a load characteristic value calculation method proposed by the present invention;

[0051] Figure 2 A diagram showing a substructure research object to be measured and a measurement point arrangement diagram of a truss structure model according to an embodiment of the present invention;

[0052] Figure 3 It is a time history curve diagram of daily maximum value and period term of linear temperature gradient of steel box girder in an embodiment of the present invention;

[0053] Figure 4 This is a schematic diagram of a time history curve of a random fluctuation term of a linear temperature gradient of a steel box girder according to an embodiment of the present invention;

[0054] Figure 5 Schematic diagram of the calculation results of the linear temperature gradient load characteristic value of the steel box girder according to the embodiment of the present invention, wherein Figure 5 (a) is the maximum value histogram, Figure 5 (b) is the extreme value cumulative distribution function. DETAILED DESCRIPTION

[0055] like Figure 1 As shown, in this embodiment, a structural temperature gradient simulation model and a load characteristic value calculation method are provided. The simulation model is suitable for describing the periodic transformation law of the daily maximum linear temperature gradient of the engineering structure. The temperature load characteristic value calculation method is suitable for calculating the temperature gradient load characteristic value of the engineering structure. The specific process is as follows:

[0056] Step 1: Arrange temperature monitoring points on the corresponding structural sections and install temperature sensors to test the real-time temperature changes of each typical structural section. Use a data acquisition instrument to collect and store the temperature data of each component at a certain frequency, use a wireless transmission module to achieve long-distance real-time transmission of temperature sensor data, and transmit the collected multi-source sensor data to the cloud platform for real-time download and analysis.

[0057] Step 2: Download the original data of structural temperature collected on the cloud platform in real time for analysis, and calculate the uniform temperature and temperature gradient of the structure. Establish a periodic simulation model of the structural equivalent linear temperature gradient. The specific process is as follows:

[0058] 1) Obtain the measured data of the structural temperature over a complete annual cycle, and obtain the uniform temperature of the structure based on the structural characteristics and the original temperature data. For the discrete temperature field, the calculation formula for the uniform temperature of the structural section is:

[0059]

[0060] Where: T i is the temperature of measuring point i, A i is the area of ​​the region represented by measuring point i, and A is the cross-sectional area.

[0061] The structural cross section is divided into several regions from top to bottom according to the depth of the measuring point. The temperature of each region is equal to the temperature of the measuring point corresponding to its height. Then, the uniform temperature of the cross section is calculated using formula (1). For structures with small cross-sectional areas, the temperatures of the measuring points can be directly averaged as the uniform temperature of the cross section.

[0062] 2) Obtain the cross-sectional temperature gradient T(A) of the structure:

[0063] The measured temperature value T of each measuring point on the cross section i Subtract the uniform temperature of the cross section T uni , get the temperature difference T of the measuring point di ; Then take the point with the smallest temperature difference as the reference point, and subtract the reference point from the temperature differences of the remaining points to obtain the final cross-sectional temperature gradient T(A).

[0064] 3) Obtain the equivalent linear temperature gradient ΔT of the structure z :

[0065] According to Hooke's law and the law of thermal expansion, the structural temperature self-stress is expressed as:

[0066] σ=EαΔT (2)

[0067] Where σ is stress; E is elastic modulus; α is linear expansion coefficient; ΔT is uniform temperature change.

[0068] If the cross-sectional temperature gradient T(A) is nonlinear, the cross-sectional bending moment calculation formula is:

[0069]

[0070] Where: M y is the bending moment; T(A) is the nonlinear temperature gradient; z is the vertical coordinate of the cross section; A is the cross section area. If the cross section temperature gradient is linear, then the calculation formula for the cross section bending moment is:

[0071]

[0072] In the formula, ΔT z is the vertical linear temperature gradient; I y is the section moment of inertia; h z is the section height;

[0073] For an Euler beam, the curvature is related to the bending moment by:

[0074]

[0075] In the formula, κ is the curvature. From formula (5), it can be seen that no matter it is a nonlinear temperature gradient or a linear temperature gradient, as long as the bending moment generated in the cross section is the same, the structural deformation caused is also the same, so the simultaneous formulas (3) and (4) are:

[0076]

[0077] The conversion formula between the nonlinear temperature gradient T(A) and the linear temperature gradient ΔT is:

[0078]

[0079] For the discrete temperature field obtained through experimental monitoring, it is approximately assumed that the temperature gradient between the measuring points changes linearly, and then the section bending moment is obtained by piecewise integration and summation. The conversion formula for discrete nonlinear temperature gradient and linear temperature gradient is:

[0080]

[0081] Where n is the number of temperature sensors; T i is the temperature gradient function of the ith region; z i A is the height coordinate of the cross section of the i-th region; i is the area of ​​the ith region.

[0082] 4) Statistics and plotting of the daily maximum linear temperature gradient curve of the structure within a complete annual cycle:

[0083] Using the above equivalent method, the nonlinear temperature gradient corresponding to the temperature measurement point of the structural section is converted into a linear temperature gradient, and then the maximum positive (negative) linear temperature gradient ΔT of the structure is extracted. z-dmax ,The curve of the daily maximum linear temperature gradient change of the structure within a complete annual cycle is drawn in units of days.

[0084] 5) Solution of the simulation model of the maximum linear temperature gradient of the structure during the day:

[0085] The annual cycle trend term f(t) of the daily maximum linear temperature gradient is fitted using a trigonometric function:

[0086]

[0087] Where, T av is the annual average of the maximum daily average temperature; T am is the annual fluctuation range of the daily maximum average temperature; T is the fluctuation period of the daily maximum average temperature, in days; It is the ordinal number of the day when the daily maximum average temperature reaches the annual minimum value.

[0088] Then, the periodic trend term is separated from the daily maximum linear temperature gradient data to obtain the random fluctuation term R(t) of the daily maximum linear temperature gradient. The random fluctuation term R(t) is expressed as a random variable that obeys a certain distribution function, such as Gaussian distribution N(μ,σ 2 ). According to the histogram of the random fluctuation term, the distribution function with the highest fitting accuracy is searched, and the parameter estimation value of the distribution function is obtained by maximum likelihood estimation. Then, the KS test and chi-square test are used to test the goodness of fit of the distribution function, and finally the distribution function of the random fluctuation term is determined.

[0089] Combining the above periodic trend function and random fluctuation function can obtain the simulation model of the maximum linear temperature gradient of the structure:

[0090]

[0091] Step 3: Use the above structure daily maximum linear temperature gradient simulation model to arbitrarily generate the daily maximum linear temperature gradient ΔT under the conditions of a complete annual cycle. ZM , the annual extreme value of the daily maximum linear temperature gradient is regarded as a problem to be solved and expressed as:

[0092] T Z_ymax =Max{ΔT ZM (t)},t∈(1,365) (11)

[0093] Where, T Z_ymax is the annual linear temperature gradient extreme value; T ZM is the daily maximum linear temperature gradient; t is the day number.

[0094] The Monte Carlo method is adopted to simulate the daily maximum linear temperature gradient change process of n (n≥10000) years using the simulation model, and then n (n≥10000) linear temperature gradient annual extreme value samples are obtained from the n (n≥10000) years of temperature data, and the generalized extreme value distribution (GEV) is used to fit the cumulative distribution function of the annual extreme value sample data, and the KS test and the chi-square test are used to confirm the fitting parameters, and then the quantile value corresponding to the exceedance probability is selected in the cumulative distribution function, which is the characteristic value of the structural temperature gradient load once in m years; the present invention can obtain the characteristic representative value of the temperature load through short-term temperature monitoring data, overcomes the disadvantage of insufficient annual extreme value samples of structural temperature monitoring data in actual engineering, is simple to calculate and has high calculation accuracy, and is suitable for different climatic conditions and structures, and can provide a reference for the design and operation and maintenance of engineering structures.

[0095] Specific application examples of the present invention:

[0096] 1) Cable-stayed bridge structure

[0097] A cable-stayed bridge with a span of (60+120+324+120+60)m is taken as the research object. The bridge adopts a steel box truss composite structure, with a truss at the top and a steel box girder at the bottom. The truss is 14.0m wide, 12m high, with a section spacing of 12.0m, and a total of 57 sections in the whole bridge. The lower steel box girder is 2.5m high, 19.87m wide, 17.64m wide on the top plate, and 9.4m wide on the bottom plate. Both the upper and lower top plates are orthotropic plate structures.

[0098] 2) Temperature measurement point arrangement and real-time signal acquisition

[0099] like Figure 2 As shown in the figure, temperature monitoring points are arranged on the main beam structure of the cable-stayed bridge, and temperature sensors are installed to test the real-time changes in the temperature of the measuring points on the typical structural section of the main beam. The data acquisition instrument is used to collect and store the temperature data of each component at a certain frequency, and the wireless transmission module is used to realize the long-distance real-time transmission of the temperature sensor data. The collected multi-temperature measurement point data is transmitted to the cloud platform for real-time download and analysis.

[0100] 3) Verification of simulation model and calculation method

[0101] Meteorological data for Boston, USA from 1893 to 2019 were obtained from the National Centers for Environmental Information (NOAA). The meteorological data included the annual maximum temperature from 1893 to 2018 and the daily maximum temperature from 2006 to 2019.

[0102] According to the daily maximum temperature in Boston from 2006 to 2019, an annual maximum temperature simulation model was established, and then the Monte Carlo method was used to simulate the change process of the daily maximum temperature in 10,000 years to obtain 10,000 annual maximum temperature samples. Then, according to the distribution characteristics of the samples, the annual maximum temperature characteristic values ​​of the corresponding return period were calculated. Finally, the calculated once-in-50-year and once-in-100-year temperature characteristic values ​​were compared with the measured once-in-50-year and once-in-100-year maximum temperatures to verify the accuracy of the calculation method in this paper. The specific calculation steps are as follows:

[0103] First, the periodic fluctuation law f(t) of the daily maximum temperature data is obtained, as shown in formula (12).

[0104]

[0105] Then separate the periodic law and obtain the random law R(t). According to the histogram of R(t), search for the distribution function with the highest fitting accuracy, and finally obtain the Johnson SU distribution with the highest fitting accuracy. The probability density function expression of Johnson SU is (13):

[0106]

[0107] Among them, a and b are shape parameters.

[0108] The probability density function of R(t) is obtained by using the maximum likelihood estimation. The distribution function is tested by KS and chi-square hypothesis, and the p-value is much larger than the significance coefficient of 0.05. Therefore, it can be considered that the random fluctuation component of temperature R(t) obeys the Johnson SU distribution shown in the formula, as shown in formula (14).

[0109]

[0110] So far, the mathematical model of Boston's daily maximum temperature is obtained as follows:

[0111]

[0112] The simulation model is used to simulate the daily maximum temperature fluctuation curve for 10,000 years, thereby obtaining 10,000 annual maximum temperature samples.

[0113] The generalized extreme value distribution is used to fit the cumulative distribution function of the annual maximum temperature, and then the quantile corresponding to the exceedance probability p = 0.01 is selected in the cumulative distribution function, which is the characteristic value of Boston's maximum temperature once in a hundred years; the quantile corresponding to the exceedance probability p = 0.02 is taken, which is the characteristic value of Boston's maximum temperature once in fifty years. The calculation results are shown in Table 1. The results show that the calculation error of the load characteristic value obtained based on one year of monitoring data is 5.5%, and the calculation error of the load characteristic value obtained based on 14 years of monitoring data is 2.4%. It can be seen that the temperature load characteristic value obtained by the calculation method proposed in the present invention is very close to the measured value, and the larger the amount of original data, the higher the calculation accuracy. In summary, this method can be used to calculate the characteristic value of structural temperature load.

[0114] Table 1 Verification and comparison results of simulation model and calculation method

[0115]

[0116] 4) Practical Application

[0117] The daily maximum value of the linear temperature gradient of the steel box girder structure is obtained, such as Figure 3 As shown in Figure 2, it can be seen that the linear temperature gradient also has a seasonal fluctuation trend.

[0118] The periodic variation model for extracting the daily maximum of the linear temperature gradient is:

[0119] Daily maximum of linear temperature gradient:

[0120]

[0121] After obtaining the deterministic periodic term, the random fluctuation term is separated as Figure 4 shown.

[0122] According to the distribution characteristics of random fluctuation items, Gaussian distribution is used to fit the distribution of random fluctuation items of daily maximum value of temperature gradient. The probability density function expression of Gaussian distribution is:

[0123]

[0124] Among them, μ is the mean and σ is the standard deviation.

[0125] The maximum likelihood estimation is used to estimate the parameters of the formula, and the parameters and goodness of fit test results are shown in Table 2:

[0126] Table 2 Random term function goodness of fit test results

[0127]

[0128] The results of KS test and chi-square test show that the p-value is greater than the significant level of 0.05. Therefore, it can be considered that the random term of the daily maximum value of the linear temperature gradient of the steel box girder obeys a Gaussian distribution with a mean of 0 and a standard deviation of 4.17, and the random term of the daily minimum value of the linear temperature gradient of the steel box girder structure obeys an exponential Weibull distribution with k = 10.81, α = 1.04, and λ = 7.24.

[0129] So far, the simulation model of the daily maximum value of the linear temperature gradient of the steel box girder structure is obtained as follows:

[0130]

[0131] After obtaining the simulation model of the daily extreme value of the linear temperature gradient, the Monte Carlo method is used to simulate the time course curve of the daily extreme value of the linear temperature gradient for 10,000 years. Then, 10,000 samples of the annual extreme value of the linear temperature gradient are obtained from the 10,000-year data, and the histogram of the annual maximum value of the linear temperature gradient of the steel box girder is obtained as shown in the figure. Figure 5 (a) shown.

[0132] The generalized extreme value distribution is used to fit the distribution of the uniform temperature annual extreme value. The probability density function expression of the generalized extreme value distribution is:

[0133]

[0134] Where: k is the shape parameter; σ is the scale parameter; μ is the location parameter.

[0135] The fitting results and goodness-of-fit test results are shown in Table 3.

[0136] Table 3 Goodness-of-fit test results of generalized extreme value distribution

[0137]

[0138] It can be seen that the p-value of the KS test and the chi-square test are both greater than the significant level of 0.05. Therefore, it can be considered that the distribution function in Table 3 is the distribution function of the annual extreme value of the linear temperature gradient of the steel box girder. After obtaining the cumulative distribution function, the cumulative distribution function is made equal to the exceedance probability corresponding to the target return period, such as Figure 5 As shown in (b), the corresponding temperature load characteristic value can be obtained.

[0139] Finally, the maximum linear temperature gradient load characteristic value of the steel box girder structure of the embodiment once in fifty years is 25.01°C, and the minimum linear temperature gradient load characteristic value is -5.61°C; the maximum linear temperature gradient load characteristic value of the steel box girder structure once in a hundred years is 25.74°C, and the minimum linear temperature gradient load characteristic value is -5.80°C, as shown in Table 4.

[0140] Table 4 Characteristic values ​​of linear temperature gradient load

[0141]

[0142] 5) Conclusion

[0143] The above results show that the invented structural temperature gradient simulation model and load eigenvalue calculation method are scientific and reasonable, and innovatively express the daily temperature gradient of the structure as the sum of a periodic trend term (quasi-trigonometric function) and a random fluctuation term (random fluctuation function). It can accurately simulate the variation law and numerical value of the daily temperature gradient of the structure. The invention can be well used to solve the equivalent linear temperature gradient load eigenvalue of engineering structures.

Claims

1. A method for constructing a structural temperature gradient simulation model, characterized in that: The following steps are involved: S1. According to the structural characteristics and the measured data of the structural temperature over a complete annual cycle, the uniform temperature of the structure is obtained, which is calculated as follows: Among them, T i is the temperature of measuring point i, A i is the area represented by measuring point i, and A is the cross-sectional area; S2. Obtaining the cross-sectional temperature gradient T(A) of the structure. The specific method includes: converting the measured temperature value T of each measuring point of the cross section i Subtract the uniform temperature of the cross section T uni , get the temperature difference T of the measuring point di ; The point with the smallest temperature difference is taken as the reference point, and the temperature difference of the remaining points is subtracted from the reference point to obtain the final cross-sectional temperature gradient T(A); S3. Obtain vertical linear temperature gradient ΔT z , the specific method includes: According to Hooke's law and the law of thermal expansion, the structural temperature self-stress is expressed as: σ=EαΔT (2) Where, σ is stress; E is elastic modulus; α is linear expansion coefficient; ΔT is uniform temperature change; If the cross-sectional temperature gradient T(A) is nonlinear, the cross-sectional bending moment calculation formula is: Where M y is the bending moment; T(A) is the nonlinear temperature gradient; z is the vertical coordinate of the cross section; A is the cross-sectional area; if the cross-sectional temperature gradient is linear, then the calculation formula for the cross-sectional bending moment is: In the formula, ΔT z is the vertical linear temperature gradient; I y is the section moment of inertia; h z is the section height; For an Euler beam, the curvature is related to the bending moment by: In the formula, κ is the curvature. From formula (5), it can be seen that no matter it is a nonlinear temperature gradient or a linear temperature gradient, as long as the bending moment generated in the cross section is the same, the structural deformation caused is also the same. Therefore, the simultaneous formulas (3) and (4) are: The conversion formula between the nonlinear temperature gradient T(A) and the linear temperature gradient ΔT is obtained as follows: For the discrete temperature field obtained through experimental monitoring, the temperature gradient between the measuring points changes linearly, and then the section bending moment is obtained by piecewise integration and summation. The conversion formula for discrete nonlinear temperature gradient and linear temperature gradient is: Where n is the number of temperature sensors; T i is the temperature gradient function of the ith region; z i A is the height coordinate of the cross section of the i-th region; i is the area of ​​the ith region; S4. Count and draw the daily maximum linear temperature gradient change curve of the structure within a complete annual cycle; S5. Obtain the structural temperature gradient simulation model, and calculate as follows: The annual cycle trend term f(t) of the daily maximum linear temperature gradient is fitted using a trigonometric function: Where, T av is the annual average of the maximum daily average temperature; T am is the annual fluctuation range of the daily maximum average temperature; T is the fluctuation period of the daily maximum average temperature, in days; It is the ordinal number of the day when the daily maximum average temperature reaches the annual minimum value; Separate the periodic trend term from the daily maximum linear temperature gradient data to obtain the random fluctuation term R(t) of the daily maximum linear temperature gradient; According to the histogram of the random fluctuation term, searching for the distribution function with the highest histogram fitting accuracy, and using maximum likelihood estimation to obtain parameter estimates of the distribution function; Based on the estimated parameter values, the KS test and the chi-square test are used to perform a goodness of fit test on the distribution function to determine the random fluctuation term distribution function; The structural temperature gradient simulation model is obtained by combining the periodic trend function and the random fluctuation term distribution function, which is expressed as follows: Among them, T av is the annual average of the maximum daily average temperature; T am is the annual fluctuation range of the daily maximum average temperature; T is the fluctuation period of the daily maximum average temperature, in days; is the day number when the daily maximum average temperature reaches the annual minimum value; μ is the mean of the Gaussian distribution function of the random fluctuation term, and σ is the standard deviation of the Gaussian distribution function of the random fluctuation term.

2. A method for calculating load characteristic values ​​using a structural temperature gradient simulation model according to claim 1, characterized in that: include: Obtain structural temperature data for a complete annual cycle; Analyze the structural temperature data of the complete annual cycle to obtain the daily maximum equivalent linear temperature gradient and build a structural temperature gradient simulation model; Based on the structural temperature gradient simulation model, the maximum daily linear temperature gradient of several complete annual cycles is cyclically generated, and the Monte Carlo simulation sampling method and generalized extreme value distribution fitting are used to obtain the characteristic value of the structural temperature gradient load once in m years; based on the structural temperature gradient simulation model, the maximum daily linear temperature gradient of a complete annual cycle is generated, and the calculation is as follows: T Z_ymax =Max{ΔT ZM (t)},t∈(1,365) (11) Where, T Z_ymax is the annual linear temperature gradient extreme value; T ZM is the daily maximum linear temperature gradient; t is the day number; The method of obtaining the characteristic value of the structural temperature gradient load with a return period of m years by using the Monte Carlo simulation sampling method according to the daily maximum linear temperature gradient of the complete annual cycle includes: The Monte Carlo simulation sampling method is used to simulate the daily maximum linear temperature gradient change process of n complete annual cycles using the structural daily maximum linear temperature gradient simulation model; According to n annual temperature data curves, n annual extreme value samples of linear temperature gradient are obtained respectively; The generalized extreme value distribution is used to fit the cumulative distribution function of the annual extreme value sample data, and the KS test and chi-square test are used to confirm the fitting parameters. The quantile value corresponding to the exceedance probability is selected in the cumulative distribution function to obtain the characteristic value of the structural temperature gradient load with a return period of once in m years.

Images