CN114840616A - Dynamic atmospheric natural environment modeling method based on space-time interpolation - Google Patents

Abstract

The invention provides a dynamic atmospheric natural environment modeling method based on space-time interpolation, which comprises the following implementation steps: the method comprises the following steps: collecting typical atmospheric environment data; step two: modeling atmospheric environment data by time-space interpolation; step three: comprehensive statistical analysis of environmental data; through the steps, the invention provides a dynamic atmospheric natural environment modeling method based on space-time interpolation, which analyzes environmental factors aiming at typical atmospheric environmental factors and space-time positions, comprehensively considers the influence of time and geographic positions, establishes a space-time dynamic change model with geographic position precision reaching 0.5 degrees through a space-time interpolation method, provides atmospheric natural environment dynamic change data, and performs statistical analysis on the atmospheric natural environment data in a time dimension to form the dynamic atmospheric natural environment modeling method based on the space-time interpolation. The method is scientific, easy to realize and more in line with engineering practice.

Description

A Dynamic Atmospheric Natural Environment Modeling Method Based on Spatio-temporal Interpolation

technical field

The invention provides a dynamic atmospheric natural environment modeling method based on space-time interpolation. It is a dynamic atmospheric natural environment modeling method based on the collection of typical atmospheric environmental data, spatial-temporal interpolation modeling of atmospheric environmental data, and comprehensive statistical analysis of environmental data. Based on typical atmospheric environmental factors and temporal and spatial positions, it comprehensively considers the influence of time and geographic location, and establishes a temporal and spatial dynamic change model with a geographic location accuracy of 0.5 degrees by means of temporal and spatial interpolation. Comprehensive analysis is carried out to form a dynamic atmospheric natural environment modeling method. This patent applies to the technical field related to natural environment modeling.

Background technique

Under the atmospheric natural environment conditions, the natural environment conditions in which the products are located not only continuously change on the time scale, but also have different changing laws on the spatial scale, resulting in different environmental effects. Due to differences in environmental factors It is of great significance to establish a model that conforms to the temporal and spatial variation of atmospheric natural environment factors. In order to obtain the natural environment data outside the observation site, the statistical method is usually combined with the geographic information system, and the estimation is based on the observation value of the existing observation site, that is, the spatial interpolation of the natural environment data. One class is deterministic methods and the other is geostatistical methods. Deterministic interpolation methods are based on the similarity between information points or the smoothness of the entire surface to create a fitting surface, such as inverse distance weighted average interpolation method, trend surface method, spline function method, etc. Geostatistical interpolation methods are Using the statistical law of sample points, the spatial autocorrelation between sample points is quantified, so as to build a spatial structure model of sample points around the point to be predicted, such as kriging interpolation. The characteristic of the deterministic interpolation method is that the interpolation result at the sample point is basically consistent with the actual value of the original sample point. If the non-deterministic interpolation method is used, the interpolation result at the sample point and the measured value of the sample may not be consistent. Far from it. No matter which interpolation method is used, there is a distance attenuation effect, that is, the closer the points in space are, the more likely they have similar observed values; while the farther the points are, the less likely their eigenvalues are similar. In the atmospheric natural environment, the mutual influence effect between points far apart is small and can be ignored. By dividing the block according to the connectivity in the environmental area to be measured, the complex whole is decomposed into a series of local units, which is conducive to reducing error, improving the accuracy of interpolation and prediction. Therefore, in order to accurately describe the actual environmental conditions experienced by the product, it is necessary to describe its time and space variation laws at the same time, and to model the dynamic atmospheric natural environment. Based on this, the present invention proposes a dynamic atmospheric natural environment modeling method based on spatiotemporal interpolation.

SUMMARY OF THE INVENTION

(1) purpose of the present invention:

The purpose of the present invention is to provide a dynamic atmospheric natural environment modeling method based on space-time interpolation, which is aimed at a series of problems such as difficulty in collecting, querying, and analyzing atmospheric natural environment data, and provides a dynamic atmospheric natural environment modeling method. It is a dynamic atmospheric natural environment modeling method that includes typical atmospheric environment data collection, spatial-temporal interpolation modeling of atmospheric environment data, and comprehensive statistical analysis of environmental data. By taking into account the differences of environmental factors in different regions, typical atmospheric environmental factors in different time and space locations are collected, and through comprehensive statistical analysis of environmental factors, considering the influence of time and geographical location, the Kriging interpolation method is used to establish a geographical location accuracy of 0.5 The spatial and temporal dynamic change model of degree is used to model the dynamic atmospheric natural environment, and finally the dynamic atmospheric natural environment modeling method based on spatio-temporal interpolation is given.

(2) Technical solution: Based on the above theories and ideas, the present invention provides a dynamic atmospheric natural environment modeling method based on space-time interpolation, and its specific implementation steps are as follows:

Step 1: Collection of typical atmospheric environment data;

The atmospheric environmental factors mainly include temperature, humidity, solar radiation, rainfall, pollutants, etc.; the present invention is aimed at products working in the atmospheric environment, and the environmental parameters to be considered include temperature and relative humidity; temperature is an important factor affecting the corrosion failure of materials, It will affect the condensation of water vapor and the solubility of gases and salts in the water film. Where the average temperature is high, the atmospheric corrosion rate is larger; if there are pollutants in the atmospheric environment, such as chloride ions, they are deposited or dissolved in the material. When the surface liquid film is in, corrosion micro-battery or oxygen concentration cell can be formed on the surface of the material, resulting in corrosion failure on the surface of the material; the higher the relative humidity in the atmosphere, the easier the metal surface is to condense, and the time the electrolyte film on the surface exists. The longer it is, the faster the corrosion failure will increase, and the working performance of the product will also be affected; therefore, for the above typical environmental data, in order to meet the requirements of the final difference model geographic location accuracy of 0.5 degrees of latitude and longitude, it is necessary to Data collection sources are screened by environmental observatories:

① Screening and collecting environmental data according to different sources of environmental factor data (observation stations, literature, data networks);

② Write programs to process and organize scientific format file data;

Step 2: Spatial-temporal interpolation modeling of atmospheric environment data;

Since the collected and sorted environmental factor data is incomplete and missing, in order to cover and obtain the overall environmental factor information of the target area, it is necessary to perform spatial-temporal interpolation modeling of atmospheric environmental data on the environmental data of the existing geographical location to obtain Overall environmental factor data for the target area;

Kriging interpolation method is widely used in the interpolation calculation method of geostatistics, which has the characteristics of linearity, unbiasedness and minimum estimation variance; the idea of kriging interpolation method is to give the observation data of known sample points in space with corresponding weight values , used to estimate the corresponding data of the unknown point; due to the continuity of the atmospheric environment, it can be assumed that the environmental variables such as temperature and relative humidity satisfy the second-order stationary assumption and the eigen assumption, so the ordinary kriging interpolation method can be used for calculation; An environmental factor variable is Z(x), and its observed values at known geographic locations are: Z(x ₁ ), Z(x ₂ ), Z(x ₃ ), …, Z(x _n );

Then the estimated value of the variable at x ₀ is obtained by the following formula:

In the formula: Z ^* (x ₀ ) represents the estimated value of the variable at x ₀ , Z(x _i ) represents the observed value (i=1,...,n), λ _i is the weight coefficient, which can be expressed as the observation at each position The contribution of the value to the estimated value; therefore, the problem turns into the determination of λ _i . Since the expression already satisfies the linearity requirements, it can be solved by using the remaining two constraints: unbiasedness and minimum estimated variance; therefore, the output of Solving the system of equations is:

In the formula: μ is the Lagrangian number, c(x _i , x _j ) is the covariance function of x _i and x _j ; since the relationship between the variogram γ(·) and the covariance function c(·) approximates This trade-off can be solved by substituting the variogram into the equation system instead of the covariance function. The variogram model is used to represent the function that changes with distance in the space. Common variogram models are: Gaussian model, exponential model, spherical model Wait;

Said: "Gaussian model", means

Said: "Exponential Model", means

Said: "spherical model", means

Among them, Δx is x _j -x _i , j>i; e is the index; a is the variation range, representing the uniform spatial correlation range of the parameters, the speed of the spatial variation of the response parameters, the larger the a, the greater the spatial correlation range. Larger, indicating that the spatial change speed of the parameter is smaller; C ₀ is the nugget constant, which is caused by the parameter error and microstructure, and represents the part of the random change of the parameter; C ₀ +C is the base value, reflecting the parameter in the numerical value The maximum change in size; C is the arch height, which represents the structural change of the parameter;

Step 3: Comprehensive statistical analysis of environmental data;

For different regions, there are certain differences in the changes of environmental factors. For example, the temperature difference in different regions is large, and the drastic changes in temperature also affect the working conditions of the product. If the product encounters a temperature drop during the working process, the surface of the material will be When the temperature is lower than the surrounding atmospheric temperature, the condensation of water vapor in the atmosphere will condense on the surface of the material, which will accelerate corrosion; at the same time, the environmental factors at the same location at different times have certain differences, such as high temperature and drought in summer in some areas and low temperature in winter. Humidity will lead to significant changes in the working environment of the product, which will directly affect the working performance of the product; therefore, according to the characteristics of environmental factors in different regions, the environment is classified in time and space, and the time dimension accurate to the hour is used. Classification and screening at different latitudes and longitudes, and then through comprehensive statistical analysis of environmental data, a dynamic atmospheric natural environment modeling method based on spatiotemporal interpolation is formed;

Through the above steps, the effect of accurately estimating specific environmental factors of a given time information in a certain area through spatial information such as latitude and longitude is achieved, and the practical problems of environmental factor statistics, modeling, and prediction difficulties are solved.

(3) Advantages and efficacy:

The present invention is a dynamic atmospheric natural environment modeling method based on space-time interpolation, and has the following advantages:

① The present invention analyzes the environmental factors according to typical atmospheric environmental factors and space-time positions, comprehensively considers the influence of time and geographical position, and establishes a space-time dynamic change model with a geographical position accuracy of 0.5 degrees by means of Kriging interpolation, providing Atmospheric natural environment dynamic change data, and statistical analysis of atmospheric natural environment data in the time dimension, forming a dynamic atmospheric natural environment modeling method based on spatiotemporal interpolation.

② The method proposed by the present invention is simple to calculate, easy to implement, more in line with engineering practice, convenient for engineering and technical personnel to master and use, scientific method, good manufacturability, and convenient for application and popularization.

Description of drawings

Figure 1 is a flow chart of the method of the present invention.

Figure 2 is a map of the Pacific Ocean-Southeast Asia region.

Figure 3 shows the interpolation calculation results of the atmospheric temperature data in the Pacific Ocean.

Figure 4 shows the annual temperature and relative humidity changes in the South China Sea.

Figure 5 is the distribution map of monthly temperature probability density in Hainan.

Detailed ways

The present invention will be further described in detail below by taking part of the data in the atmospheric natural environment data of a certain sea area as an example and in conjunction with FIG. 1 . The present invention proposes a dynamic atmospheric natural environment modeling method based on space-time interpolation, as shown in Figure 1, and its specific implementation steps are as follows:

Step 1: Collection of typical atmospheric environment data

Select the Pacific Ocean-Southeast Asia region as the target sea area for analysis, as shown in Figure 2. The specific latitude and longitude are as follows:

Table 1 Longitude and latitude of sea area

In order to meet the requirement that the geographic location accuracy of the final interpolation model reaches 0.5 degrees of latitude and longitude, it is necessary to screen the environmental observation stations from which the data is collected. The sources of data on different environmental factors are shown in the table below:

Table 2 Data sources for environmental factors collection

The collected data is in the NetCDF scientific format, which cannot be read directly. After being converted by the programming program, it is organized into an EXCEL processable format, and the required environmental factors are screened;

Said: "NetCDF", refers to the network common data format (Network Common Data Form), which was developed by scientists of the University Corporation for Atmospheric Research (University Corporation for Atmospheric Research) for the characteristics of scientific data, is an array-oriented data format. It is a description and coding standard of data that is suitable for network sharing; it is widely used in atmospheric science, hydrology, oceanography, environmental simulation, geophysics and many other fields. Users can easily manage and manipulate this dataset in a variety of ways;

Step 2: Spatial-temporal interpolation modeling of atmospheric environment data

According to the completed environmental factor data, the latitude and longitude are gridded in units of 0.5 degrees in space, and the environmental factors are processed in the time dimension accurate to the hour. Based on the above data, interpolation calculation and natural environment construction are carried out. By removing the trend term, calculating the autocorrelation coefficient, solving the Kriging equations, and analyzing the errors, the spatial and temporal dynamic modeling of the atmospheric natural environment is carried out. Specifically, when performing kriging interpolation calculation and modeling, first The trend term needs to be removed from the original data, because this periodic change significantly affects the stationarity of the variable distribution, and at the same time causes an obvious hole effect in the fitting of the time variogram, which is not conducive to the modeling of the spatiotemporal variogram. Therefore, In the interpolation experiment, the seasonal factor should be removed first, and the corresponding seasonal item should be added after the interpolation; then the decay law of the autocorrelation coefficient of the sequence with the number of delay periods should be calculated, and the detrended item data should be judged by the autocorrelation graph inspection method. Then, the time-space variation function is constructed for the data after detrending, and fitting is performed to solve the Kriging equations and interpolate to calculate the spatio-temporal distribution data of environmental factors; Taking the atmospheric temperature and environmental data on August 1 as an example, the calculation results of the environmental profile are shown in Figure 3, and the units are in degrees Celsius; it can be seen from the figure that the temperature in January is lower, and the temperature in August is generally higher than that in January;

Step 3: Comprehensive statistical analysis of environmental data

Based on the big data of environmental factors given by the calculation results of spatio-temporal interpolation, the geographical location of environmental factors is classified spatially according to the latitude and longitude information, and the changes of environmental factors are classified in time according to different time information, and comprehensive statistical analysis of environmental data is carried out;

Since the Pacific Ocean covers a large area and includes different climate types, the statistical properties of environmental factors within this range vary greatly, so you can choose to narrow the target area to the South China Sea (105°E～118°E, 4°N～21° N); According to the atmospheric temperature and environmental data, the changes in the annual temperature and relative humidity in the South China Sea in the past 14 years are calculated and shown in Figure 4. It can be seen that the annual average temperature has been maintained at a relatively stable stage, and the annual change is small. , the variation range is only between 1 °C; at the same time, the annual average relative humidity and the average temperature have an opposite trend of rising and falling, because the increase in relative humidity may lead to a decrease in temperature, which is consistent with the actual situation;

For the study of the distribution of environmental factors, consider further reducing the target area to Hainan Province (108°E～111°E, 18°N～20°N), and calculate the monthly temperature probability density of Hainan Province according to the atmospheric temperature environment data The distribution is shown in Figure 5. It can be seen that the months from January to December show a typical normal distribution. At the same time, the temperature gradually increases from January to the highest in August, and then gradually shows a downward trend;

Through the above steps, a dynamic atmospheric natural environment modeling method based on spatiotemporal interpolation is formed;

To sum up, the present invention designs a dynamic atmospheric natural environment modeling method that considers the collection of typical atmospheric environment data, the spatial-temporal interpolation modeling of atmospheric environment data, and the comprehensive statistical analysis of environmental data; Screening of environmental factors, comprehensively considering the influence of time and geographical location, Kriging interpolation method is used to establish a time-space dynamic change model with a geographical location accuracy of 0.5 degrees, providing dynamic change data of the atmospheric natural environment, and analyzing the atmospheric natural environment data through the time dimension. Statistical analysis is carried out to form a dynamic atmospheric natural environment modeling method based on spatiotemporal interpolation.

Claims (2)

1. A dynamic atmospheric natural environment modeling method based on space-time interpolation is characterized in that: the implementation steps are as follows:

the method comprises the following steps: collecting typical atmospheric environment data;

atmospheric environmental factors include temperature, humidity, solar radiation, rainfall and pollutants; for products operating in atmospheric environments, the environmental parameters to be considered include temperature and relative humidity; the temperature is an important factor influencing the corrosion failure of the material, and can influence the condensation of water vapor and the solubility of gas and salts in a water film, and the atmospheric corrosion rate is higher in places with high average temperature; if pollutants exist in the atmospheric environment and are deposited and dissolved in a material surface liquid film, a corrosion micro-battery and an oxygen concentration difference battery can be formed on the surface of the material, so that the surface of the material is corroded and loses efficacy; the higher the relative humidity in the atmosphere is, the more the metal surface is easy to dewing, the longer the existence time of an electrolyte film on the surface is, the higher the corrosion failure speed is, and the influence on the working performance of a product is also generated; therefore, for the above typical environmental data, in order to meet the requirement that the accuracy of the final differential model geographic position reaches 0.5 degree longitude and latitude, an environmental observation station from which data is collected needs to be screened:

firstly, screening and collecting environmental data according to different environmental factor data sources;

writing a program to process and arrange the scientific format file data;

step two: modeling atmospheric environment data by time-space interpolation;

the collected and sorted environmental factor data are incomplete and missing, so that when the overall environmental factor information of the target region position is required to be obtained in a covering manner, the atmospheric environmental data spatial-temporal interpolation modeling is required to be carried out on the environmental data of the existing geographic position, and the overall environmental factor data of the target region is obtained;

the kriging interpolation method is to give corresponding weight values to observation data of spatial known sample points to estimate corresponding data of unknown points; because the continuity of the atmospheric environment can assume the environmental variables of temperature and relative humidity, which all satisfy the second-order stationary assumption and the intrinsic assumption, the calculation can be carried out by using a common kriging interpolation method; and setting a certain environment factor variable as Z (x), wherein the observed values of the environment factor variable at all known geographic positions are respectively as follows: z (x) ₁ ),Z(x ₂ ),Z(x ₃ )，…,Z(x _n )；

Then at x ₀ The estimated value of the variable is obtained by the following formula:

in the formula: z ^* (x ₀ ) Represents x ₀ At an estimate of this variable, Z (x) _i ) Representative of observed values, i ═ 1, …, n, λ _i Is a weight coefficient expressed as the degree of contribution of the observed value of each position to the estimated value; the problem thus turns into λ _i Since the expression already satisfies the linear requirement, the remaining two constraints are used to solve: unbiased and minimum estimated variance; thus, the output solution equation set is:

in the formula: μ is the Lagrangian constant, c (x) _i ,x _j ) Is x _i And x _j The covariance function of (a); since the relationship between the variation function γ (-) and the covariance function c (-) approximates this trade-off, the system of equations is solved by substituting the variation function into the equation system instead of the covariance function, and the function varying with distance in space is represented by a variation function model including: a gaussian model, an exponential model and a spherical model;

step three: comprehensive statistical analysis of environmental data;

for different regions, the change of environmental factors is different, the severe change of air temperature also influences the working condition of the product, if the temperature of the product is reduced in the working process, the temperature of the surface of the material is lower than the ambient atmospheric temperature, and the water vapor in the atmosphere is condensed and condensed on the surface of the material, so that the corrosion is accelerated; meanwhile, the environmental factors of the same place and different time have difference, which can cause the working environment of the product to change obviously and directly influence the working performance of the product; therefore, the environment is classified in time and space according to the characteristics of the environment factors in different areas, the environment factors are classified and screened in different longitudes and latitudes according to the time dimension accurate to the hour, and then the dynamic atmospheric natural environment modeling method based on the space-time interpolation is formed through comprehensive statistical analysis of the environment data.

2. The dynamic atmospheric natural environment modeling method based on spatio-temporal interpolation according to claim 1, characterized in that: in step two, the Gaussian model refers to

The index model refers to

The spherical model refers to

Wherein Δ x is x _j -x _i J > i; e is an index; a is a variation, represents a space correlation range uniformly possessed by the parameter, and reflects the speed of space change of the parameter, wherein the larger a is, the larger the space correlation range is, and the smaller the space change speed of the parameter is; c ₀ The gold block constant is caused by parameter errors and microstructures and represents a part of parameter random variation; c ₀ + C is a base station value and reflects the maximum variation amplitude of the parameter in numerical value; c is the arch height, representing the part of the parameter that varies structurally.

Images