Method for predicting high-dimensional temperature of aircraft in wide-area high-altitude task environment

Abstract

The invention provides a high-dimensional temperature prediction method for an aircraft in a wide-area high-altitude task environment. The method comprises the following steps: step 1, carrying out temperature-time first-time prediction based on regression analysis; step 2, temperature-time secondary prediction based on nonlinear fitting and error random representation; 3, temperature-height prediction based on an empirical formula and dynamic interpolation; 4, temperature-longitude and latitude prediction based on dynamic interpolation; and 5, carrying out fusion prediction on the high-dimensional temperature. According to the method, mathematical statistical methods and empirical formulas such as regression analysis, nonlinear fitting and linear (nonlinear) interpolation are comprehensively utilized, a prediction model of four-dimensional variables including time, height, longitude and latitude of temperature factors is constructed, the purpose of outputting a reasonable long-term temperature prediction value when any four-dimensional coordinate in the aircraft task process is given is achieved, and the method is suitable for being applied to the aircraft. Therefore, guidance is provided for the task process.

Description

Technical Field

[0001] The present invention provides a method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment, which relates to a long-term prediction technology of high-dimensional temperature environmental factors based on comprehensive empirical models and statistical models such as regression analysis, nonlinear fitting, and dynamic interpolation, and belongs to the field of equipment environmental engineering. Background Art

[0002] Temperature, as one of the most typical natural environmental factors in the aircraft mission process, can have a continuous impact on the performance and function realization of the aircraft. Taking missiles as an example, temperature can have a significant impact on the thrust, aerodynamic drag, flight lift, guidance accuracy and warhead material of the missile engine, thus directly affecting the missile's launch decision, performance level and function realization. If the temperature conditions during the aircraft mission can be predicted in advance, it can not only make a preliminary assessment of its reliability, safety and mission success during the flight, but also provide guidance on its mission start decision and the execution of compensation measures during the mission based on the assessment. Therefore, reasonable and accurate temperature prediction has important guiding significance for the aircraft's own performance and mission success. Specifically, in terms of the temperature characteristics during the aircraft mission, it is usually in a four-dimensional "space" where time, altitude, longitude and latitude are intertwined, and the high-dimensional space-time coupling effect brings major challenges to the reasonable and accurate prediction of temperature.

[0003] At present, the temperature prediction methods mainly include mathematical statistics and intelligent prediction. The existing methods mainly focus on the prediction of temperature with respect to a single variable such as altitude or time, and mainly focus on short-term prediction. There is a lack of long-term temperature prediction research considering longitude and latitude and high-dimensional space-time coupling. Therefore, it is urgent to study the prediction method of high-dimensional space-time temperature factors. Summary of the invention

[0004] (1) Purpose of the present invention: The present invention utilizes mathematical statistics methods such as regression analysis, nonlinear fitting, linear (nonlinear) interpolation, and empirical formulas to construct a prediction model of temperature factors with respect to four-dimensional variables including time, altitude, longitude, and latitude, so as to achieve the purpose of outputting a reasonable long-term temperature prediction value for any four-dimensional coordinate during a given aircraft mission, thereby providing guidance for the mission process.

[0005] (2) Technical solution:

[0006] The technical solution of the present invention is as follows: A method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment comprises the following steps:

[0007] Step 1: First prediction of temperature-time based on regression analysis; specifically includes the following steps:

[0008] a) Determination of the 4-dimensional spatial range and resolution of the aircraft mission environment, including time, altitude, longitude and latitude:

[0009] First, the 4-dimensional spatial range that the aircraft may experience during the mission is clarified; secondly, the data resolution of the four dimensions is set respectively: in the present invention, the time resolution is set to 6h / data point, the altitude resolution is set to 1km / data point, and the resolutions of longitude and latitude are both set to 0.25 degrees / data point.

[0010] b) Design and train a temperature-time linear regression model based on adjacent data; specifically including the selection of adjacent data and the design and training of a linear regression model:

[0011] First, the data of the latest year in the historical data is selected as the benchmark. The nearby data in the present invention is the historical temperature information of 2022.

[0012] Secondly, the linear regression model was designed and trained: the inter-annual increment of the monthly average temperature was used as the dependent variable and the year was used as the independent variable. The linear regression models Δy of the dependent variables of January, February, March, …, December with respect to the year were constructed respectively. i (x) = k i (x-x0)+b i In this formula, i represents month, x represents year, x0 represents the initial year of historical data, Δy i (x) represents the increment of the average temperature of the i-th month of the x-th year compared with the same period of the previous year, k i and b i They represent the slope and intercept of the linear regression model in month i. Further, the linear regression model training is carried out using MATLAB software to obtain the coefficients of the linear regression model.

[0013] Step 2: Temperature-time secondary prediction based on nonlinear fitting and random error characterization; specifically includes the following steps:

[0014] a) Temperature-time prediction based on nonlinear fitting:

[0015] After the linear regression prediction in "Step 1", there is still a certain deviation between the predicted value and the true value of the temperature increment (due to factors such as dimensionality compression and long-term prediction). i (x) represents the difference between the predicted value and the true value of the average temperature level increment of the i-th month of the x-th year after "step 1". Through statistical analysis and fitting of historical data, the present invention uses a trigonometric function to characterize the deviation between the predicted value and the true value. The specific expression is:

[0016] ξ i (x) = A i sin(B i(x-x0)+C i ), (1)

[0017] In the formula, A i , B i and C i They represent the unknown parameters in the nonlinear model of the i-th month, which can be obtained by fitting historical data through MATLAB software; x and x0 represent the year to be calculated and the initial year of historical data, respectively.

[0018] b) Temperature-time prediction based on random characterization of errors:

[0019] After the prediction of regression analysis and nonlinear fitting, the predicted value is close to the true value, but due to the limitation of the number of training samples, there is still an error between the two. In this regard, the present invention introduces a random value method to characterize the error ξ2 in this case i (x), that is, ξ2 i (x) = rand(min i ,max i ). In this formula, 'rand' represents a random value function, min i Indicates the minimum error of this stage in the training sample set, and max i Indicates the corresponding maximum error value.

[0020] c) Temperature-time comprehensive prediction model:

[0021] By superimposing linear regression, nonlinear regression and error random characterization models and using formula (2), it is possible to achieve long-term temperature prediction using the historical temperature data of the region under given latitude, longitude and altitude.

[0022] y i,j (x) = y i,j (2022)+k i (x-2018)+b i +A i sin(B i (x-2018)+C i )+rand(min i ,max i )(2)

[0023] In this formula, i represents the month, j represents the specific time of the predicted point in the month, and y i,j (x) and y i,j (2022) represent the value of temperature at specific time j in any future year and in the i-th month of 2022 respectively.

[0024] Step 3: Temperature-height prediction based on empirical formula and dynamic interpolation; specifically includes the following steps:

[0025] a) Tropospheric temperature-altitude prediction based on empirical formula:

[0026] The main heat source of the troposphere is the ground. Therefore, the farther away from the ground, the less energy is obtained, which causes the atmospheric temperature to decrease with increasing altitude. In the troposphere, the relationship between temperature and altitude can be approximately characterized by a generally accepted empirical formula:

[0027] T=T0-6×Δy, (3)

[0028] In the formula, T represents the temperature at the current height; T0 represents the temperature at a known height; Δy represents the height difference, in km. The height range of the troposphere varies with latitude. The average height range in low-latitude areas is 17-18 km, the average range in mid-latitude areas is 10-12 km, and the average height in the polar regions is 8-9 km. Therefore, when using the empirical formula to carry out tropospheric temperature-height prediction, it is necessary to flexibly adjust the application range of the formula according to the latitude of the target area.

[0029] b) Stratospheric temperature-altitude prediction based on dynamic interpolation:

[0030] The stratosphere is mainly between the top of the troposphere and 50 km. At the bottom of the stratosphere (about 25 km below), the temperature changes very slowly with altitude, and the rate of change is approximately 0; within the range of 25 km to 50 km, due to the dominant position of solar radiation, the temperature increases with increasing altitude. Therefore, within the stratosphere, the present invention conducts temperature-altitude prediction in stages, determines the altitude envelope interval of the point to be predicted at the resolution, and estimates the temperature value of the point to be predicted by interpolating the temperature at the endpoint of the interval.

[0031] Specifically, in the case of the height resolution (1 km / data point) used in the present invention, the height h to be predicted is selected as X The envelope interval, h X ∈(h1,h2), where h1 and h2 represent the two endpoints of the height envelope. The temperature T at the height to be predicted X Satisfaction: T X =(h2-h X )T1+(h X -h1)T2, where T1 and T2 represent the temperatures at the end points of the envelope respectively.

[0032] The implementation diagram of this step is as follows Figure 2 shown.

[0033] Step 4: Temperature-latitude and longitude prediction based on dynamic interpolation;

[0034] The change of temperature with longitude and latitude is affected by complex and difficult-to-quantify factors such as topography. Therefore, in the two-dimensional space of longitude and latitude, the two-dimensional limited space is divided into several grids based on the resolution of historical data. When a point to be predicted is given, its envelope grid is determined, and the distance d1, d2, d3, d4 between the point to be predicted and the boundary point of the envelope grid is calculated (see Appendix Figure 3 ), respectively using the formula and Calculate the two predicted values ​​of the point to be predicted; where T i (i=1, 2, 3, 4) represents the temperature value at the endpoint of the envelope grid, and T' and T" represent the temperature prediction values ​​under the diagonal envelope twice. Further, the temperature value of the point to be predicted is calculated using the formula T=(T'+T") / 2.

[0035] Step 5: Fusion prediction of high-dimensional temperature; specifically includes the following steps:

[0036] a) Determine the four-dimensional envelope range of the point to be predicted:

[0037] For any point to be predicted, its 4-dimensional envelope range is determined at a given resolution. The number of envelope points at the latitude and longitude level is usually 4, the number of envelope points at the altitude level is usually 2, and a point that meets the resolution and is closest to the time is selected at the time level.

[0038] b) Determine the predicted temperature values ​​at several points within the envelope:

[0039] At a given resolution, the temperature prediction values ​​of several points within the envelope are calculated through "Step 1". Specifically, the first temperature-time prediction based on regression analysis is carried out, and the second temperature-time prediction is further carried out based on nonlinear fitting and random characterization of errors. Through this link, the prediction of the temperature of several points within the envelope at the time level is completed.

[0040] c) High-dimensional temperature fusion prediction by priority:

[0041] According to the priority of altitude, longitude and latitude, firstly calculate the temperature values ​​of several points within the altitude envelope to be predicted through "step three"; specifically, according to the different situations of the predicted point in the troposphere or stratosphere, select the corresponding temperature prediction model with respect to altitude. In the troposphere, the empirical formula is used, while in the stratosphere, the dynamic interpolation method is used.

[0042] Then use "Step 4" to determine the envelope grid at the latitude and longitude level, and calculate two prediction values ​​T' and T", and use T = (T' + T") / 2 to calculate the final temperature prediction value of the point to be predicted.

[0043] (3) Advantages and effects:

[0044] The present invention provides a method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment, which has the following advantages:

[0045] ① The present invention conducts prediction research on high-dimensional temperature environmental factors during the aircraft mission from the perspectives of time, altitude, longitude and latitude, which can significantly improve the rationality and accuracy of long-term temperature prediction;

[0046] ② The present invention combines regression analysis, nonlinear interpolation and error random characterization methods to construct a temperature-time prediction model, which can significantly improve the prediction accuracy of temperature with respect to the time dimension;

[0047] ③ The empirical formula and dynamic interpolation method used in the present invention provide ideas for solving the prediction problem of temperature at the altitude and longitude and latitude levels, and promote the prediction research of high-dimensional environmental factors;

[0048] ④ The method of the present invention is scientific, has good processability, and has broad promotion and application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 The present invention is a flow chart of the method.

[0050] Figure 2 Schematic diagram of the temperature-altitude prediction method in the present invention.

[0051] Figure 3 It is a schematic diagram of the temperature-longitude and latitude prediction method in the present invention. DETAILED DESCRIPTION

[0052] In order to make the above-mentioned objects and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments;

[0053] The present invention provides a method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment, and its implementation flow chart is as follows: Figure 1 As shown, the specific implementation method is as follows;

[0054] Step 1: First prediction of temperature-time based on regression analysis; specifically divided into the following steps:

[0055] a) Determination of the 4-dimensional spatial range and resolution of the aircraft mission environment: time, altitude, longitude, and latitude

[0056] The latitude and longitude range of the aircraft's mission space is set to 115-121 degrees (east longitude), 23-29 degrees (north latitude); the altitude range is set to 0-50 km. The resolution of the time layer is set to 6 hours / data point; the resolution of longitude and latitude is 0.25 degrees / data point; the resolution of altitude is 1 km / data point. Three points to be predicted are selected, and their four-dimensional coordinates are shown in Table 1.

[0057] Table 1 Four-dimensional coordinates of the points to be predicted

[0058]

[0059] b) Design and train a temperature-time regression model based on nearby data

[0060] The temperature data of the whole year of 2022 in this area is selected as the adjacent data, and the linear regression model is used to model the interannual incremental change process of the monthly average temperature level, that is, the formula Δy i =k i Δx+b i in, Δy i represents the change in the average temperature level of the i-th month between two consecutive years, k i and b i Represents the parameters of the incremental linear regression model of the ith month. The regression model was trained using all the temperature historical data of the region from 2018 to 2022 in MATLAB software, and the linear regression model parameters for 12 months were obtained, as shown in Table 2.

[0061] Table 2 Parameters of the linear regression model of monthly mean temperature annual increment-time

[0062]

[0063] Step 2: Temperature-time secondary prediction based on nonlinear fitting and random error characterization; specifically includes the following steps:

[0064] a) Temperature-time prediction based on nonlinear fitting

[0065] After the linear regression calculation in "step 1", there is still a deviation between the actual value of the annual increment of the monthly average temperature and the predicted value, so a nonlinear fitting method is chosen to further fit the deviation. The nonlinear fitting formula used in the present invention is:

[0066] ξ i (x) = A i sin(B i (x-x0)+C i ). For any month, the “cftool” toolbox in MATLAB software is used to fit the nonlinear model and determine the three parameters to be determined. The results are shown in Table 3.

[0067] Table 3 Parameters of the nonlinear fitting model of monthly mean temperature annual increment-time

[0068]

[0069] b) Temperature-time prediction based on random characterization of errors

[0070] Using the formula ξ2 i (x) = rand(min i ,max i ) Calculate the random error term ξ2 i (x), the results are shown in Table 4.

[0071] Table 4. Parameters of the random characterization model for the error of the interannual increment of monthly mean temperature

[0072]

[0073] c) Temperature-time comprehensive prediction

[0074] By superimposing linear regression, nonlinear regression and error random characterization models, the temperature value of the point to be predicted is calculated using formula (2) under the condition of fixed longitude, latitude and altitude.

[0075] Step 3: Temperature-height prediction based on empirical formula and dynamic interpolation;

[0076] a) Tropospheric temperature-altitude prediction based on empirical formula

[0077] Determine the envelope interval that includes the height of the point to be predicted, and use the temperature values ​​at the end points of the interval and the formula T = T0-6×Δy to calculate the temperature of the point to be predicted. It is not difficult to find that "point to be predicted 1" is in the troposphere, so the empirical formula of this link is used to carry out its prediction in the height dimension.

[0078] b) Stratospheric temperature-altitude prediction based on dynamic interpolation

[0079] It is not difficult to find that the "point to be predicted 2" is in the layer in the stratosphere where the temperature changes very slowly with altitude. Therefore, under the data resolution set by the present invention, the endpoint value of the altitude interval containing the "point to be predicted 2" is used to replace the temperature value of the point.

[0080] For "point 3 to be predicted", it is in the range where the temperature changes significantly with altitude. Therefore, at a given resolution, the temperature T1 and T2 of the endpoints of the altitude interval containing "point 2 to be predicted" are used to calculate the temperature of the point 3 through the formula T X =(h2-h X )T1+(h X -h1)T2 calculates the temperature of the point.

[0081] Step 4: Temperature-latitude and longitude prediction based on dynamic interpolation;

[0082] It is not difficult to find that, at the latitude and longitude level, both "point 2 to be predicted" and "point 3 to be predicted" are at the data resolution set by the present invention, so there is no need to perform interpolation of this step for these two points.

[0083] For "point 1 to be predicted", in the two-dimensional space of longitude and latitude, calculate the distances from this point to the four endpoints of the longitude and latitude envelope grid: d1=0.224, d2=0.25, d1=0.158, d1=0.112;

[0084] Use the formulas and Calculate two predicted values ​​of "point 1 to be predicted";

[0085] The final temperature prediction value of "point to be predicted 1" is calculated using the formula T=(T'+T') / 2.

[0086] Step 5: Fusion prediction of high-dimensional temperature; specifically includes the following steps:

[0087] a) Determine the four-dimensional envelope range of the point to be predicted

[0088] For any predicted point that is not on the resolution grid, determine its 4-dimensional envelope point. The number of envelope points in the latitude and longitude level is usually 4, the number of envelope points in the altitude level is usually 2, and one approximate point is selected in the time level. The envelope ranges of the three predicted points are shown in Table 5.

[0089] b) Determine the predicted temperature values ​​at several points within the envelope;

[0090] At a given resolution, the temperature prediction values ​​of several envelope points are calculated through “Step 1”, and the results are shown in Table 5.

[0091] Table 5 Temperature prediction values ​​of coordinate points within the envelope of the three predicted points

[0092]

[0093] a) High-dimensional temperature fusion prediction by priority

[0094] According to the priority of altitude, longitude and latitude, firstly, the temperature values ​​of several points within the altitude envelope are calculated by "step three"; finally, "step four" is used to fuse the temperature values ​​of several points within the longitude and latitude envelope to obtain the final temperature value of the predicted point, as shown in Table 6.

[0095] Table 6 Long-term temperature prediction results of three predicted points

[0096]

[0097] It is not difficult to conclude that the high-dimensional temperature long-term prediction method proposed by the present invention can obtain a more reasonable temperature value. And by comparing the prediction effects of the three points to be predicted, it can be found that as the number of dimensions of historical data resolution that the points to be predicted meet increases, the temperature prediction value becomes more accurate.

[0098] In summary, the present invention relates to a method for high-dimensional temperature prediction of aircraft in a wide-range high-altitude mission environment, which relates to a long-term prediction technology for high-dimensional temperature environmental factors based on regression analysis, nonlinear fitting and dynamic interpolation. The specific steps of the method are: 1. The first prediction of temperature-time based on regression analysis; 2. The second prediction of temperature-time based on nonlinear fitting and random characterization of errors; 3. The prediction of temperature-altitude based on empirical formula and dynamic interpolation; 4. The prediction of temperature-latitude and longitude based on dynamic interpolation; 5. The fusion prediction of high-dimensional temperature. The present invention is suitable for the study of high-dimensional spatiotemporal temperature prediction of aircraft with wide-range high-altitude mission environment characteristics, which can significantly improve the rationality and accuracy of long-term prediction and provide effective information input for its mission decision-making.

Claims

1. A method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment, characterized in that: The following steps are involved: Step 1: First prediction of temperature-time based on regression analysis; including: Step 1.1: Determine the 4-dimensional spatial range and resolution of the aircraft mission environment, including time, altitude, longitude, and latitude; Step 1.2: Based on the adjacent data, design and train the temperature-time linear regression model; Step 2: Temperature-time secondary prediction based on nonlinear fitting and random characterization of errors; including: Step 2.1: Temperature-time prediction based on nonlinear fitting; Step 2.2: Temperature-time prediction based on stochastic characterization of errors; Step 2.3: Temperature-time integrated prediction model; Step 3: Temperature-height prediction based on empirical formula and dynamic interpolation; including: Step 3.1: Tropospheric temperature-height prediction based on empirical formula; Step 3.2: Stratospheric temperature-altitude prediction based on dynamic interpolation; Step 4: Temperature-latitude and longitude prediction based on dynamic interpolation; In the two-dimensional space of longitude and latitude, the two-dimensional limited space is divided into several grids based on the resolution of historical data; Step 5: Fusion prediction of high-dimensional temperature; including: Step 5.1: Determine the four-dimensional envelope range of the point to be predicted; Step 5.2: Determine the predicted temperature values ​​at several points within the envelope; Step 5.3: Fusion of high-dimensional temperature predictions in order of priority.

2. The method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment according to claim 1, characterized in that: In step 1.1, first clarify the 4-dimensional spatial range that the aircraft may experience during the mission; secondly, set the data resolution of the four dimensions respectively: the time resolution is set to 6h / data point, the altitude resolution is set to 1km / data point, and the longitude and latitude resolutions are both set to 0.25 degrees / data point.

3. The method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment according to claim 1 or 2, characterized in that: In step 1.2, it specifically includes the selection of adjacent data and the design and training of the linear regression model: First, the data of the latest year in the historical data is selected as the benchmark, and the data selected is the historical temperature information of 2022; Secondly, the linear regression model was designed and trained: the inter-annual increment of the monthly average temperature was used as the dependent variable and the year was used as the independent variable. The linear regression models Δy of the dependent variables of January, February, March, …, December with respect to the year were constructed respectively. i (x) = k i (x-x0)+b i ; In this formula, i represents month, x represents year, x0 represents the initial year of historical data, Δy i (x) represents the increment of the average temperature of month i in year x compared with the same period of the previous year, k i and b i They represent the slope and intercept of the linear regression model in the i-th month respectively; further, the linear regression model is trained using MATLAB software to obtain the coefficients of the linear regression model.

4. The method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment according to claim 1, characterized in that: In step 2.1, there is still a certain deviation between the predicted value and the true value of the temperature increment; ξ i (x) represents the difference between the predicted value and the true value of the average temperature level increment of the i-th month of the x-th year after step 1. Through statistical analysis and fitting of historical data, a trigonometric function is selected to characterize the deviation between the predicted value and the true value. The specific expression is: x i (x)=A i sin(B i (x-x0)+C i ), (1) In the formula, A i , B i and C i They represent the unknown parameters in the nonlinear model of the i-th month, which are obtained by fitting historical data using MATLAB software; x and x0 represent the year to be calculated and the initial year of the historical data, respectively.

5. The method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment according to claim 1 or 4, characterized in that: In step 2.2, the error ξ2 in this case is characterized by introducing a random value method i (x), that is, ξ2 i (x) = rand(min i ,max i ); In this formula, 'rand' represents a random value function, min i Indicates the minimum error of this stage in the training sample set, and max i Indicates the corresponding maximum error value.

6. The method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment according to claim 1, 4 or 5, characterized in that: In step 2.3, linear regression, nonlinear regression and error random characterization models are superimposed, and the historical temperature data of the region are used to achieve long-term temperature prediction under given latitude, longitude and altitude using formula (2); y i,j (x)=y i,j (2022)+k i (x-2018)+b i +A i sin(B i (x-2018)+C i )+rand(min i ,max i ) (2) In this formula, i represents the month, j represents the specific time of the predicted point in the month, and y i,j (x) and y i,j (2022) represent the value of temperature at specific time j in any future year and in the i-th month of 2022 respectively.

7. The method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment according to claim 1, characterized in that: In step 3.1, in the troposphere, the relationship between temperature and altitude can be approximated by the generally accepted empirical formula: T=T0-6×Δy, (3) Where T is the temperature at the current altitude; T0 is the temperature at a known altitude; Δy is the altitude difference in km; the altitude range of the troposphere varies with latitude, with an average altitude range of 17 to 18 km in low latitudes, 10 to 12 km in mid-latitudes, and 8 to 9 km in the polar regions.

8. The method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment according to claim 1 or 7, characterized in that: In step 3.2, in the flat process, the temperature-height prediction is carried out in stages, the height envelope interval of the predicted point is determined at the resolution, and the temperature value of the predicted point is estimated by the temperature of the endpoint of the interpolation interval; Select the height h to be predicted X The envelope interval, h X ∈(h1,h2), where h1 and h2 represent the two endpoints of the height envelope; the temperature T at the height to be predicted X Satisfaction: T X =(h2-h X )T1+(h X -h1)T2, where T1 and T2 represent the temperatures at the end points of the envelope respectively.

9. The method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment according to claim 1, characterized in that: In step 4, when a point to be predicted is given, its envelope grid is determined, and the distances d1, d2, d3, and d4 between the point to be predicted and the boundary points of the envelope grid are calculated using the formula and Calculate the two predicted values ​​of the point to be predicted; where T i (i=1,2,3,4) represents the temperature value at the endpoint of the envelope grid, T' and T" represent the temperature prediction values ​​under the diagonal envelope twice respectively; further, the temperature value of the point to be predicted is calculated using the formula T=(T'+T") / 2.

10. The method for predicting high-dimensional temperature of an aircraft in a wide-range high-altitude mission environment according to claim 1, characterized in that: In step 5.1, specifically: for any point to be predicted, under a given resolution, determine its 4-dimensional envelope range; the number of envelope points at the latitude and longitude level is 4, the number of envelope points at the altitude level is 2, and a point at the time level that meets the resolution and is closest to the time is selected; In step 5.2, the temperature prediction values ​​of several points within the envelope are calculated at a given resolution; first, the temperature-time first prediction based on regression analysis is carried out, and then the temperature-time second prediction is carried out based on nonlinear fitting and random characterization of errors; through this link, the temperature prediction of several points within the envelope at the time level is completed; In step 5.3, specifically: first, calculate the temperature values ​​of several points within the envelope of the altitude to be predicted according to the priority order of altitude, latitude and longitude; select the corresponding temperature prediction model with respect to altitude according to the different situations that the point to be predicted is in the troposphere or stratosphere; select the empirical formula in the troposphere and the dynamic interpolation method in the stratosphere; Then, the envelope grid at the latitude and longitude level is used to calculate two prediction values ​​T' and T", and the final temperature prediction value of the point to be predicted is calculated using T = (T' + T") / 2.

Images