CN116559977A - NPS-GRYANIK20 evaporation waveguide height prediction method based on BP neural network - Google Patents
NPS-GRYANIK20 evaporation waveguide height prediction method based on BP neural network Download PDFInfo
- Publication number
- CN116559977A CN116559977A CN202310539832.XA CN202310539832A CN116559977A CN 116559977 A CN116559977 A CN 116559977A CN 202310539832 A CN202310539832 A CN 202310539832A CN 116559977 A CN116559977 A CN 116559977A
- Authority
- CN
- China
- Prior art keywords
- evaporation waveguide
- height
- nps
- gryanik20
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000008020 evaporation Effects 0.000 title claims abstract description 117
- 238000001704 evaporation Methods 0.000 title claims abstract description 117
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 44
- 244000062645 predators Species 0.000 claims abstract description 62
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 34
- 238000012937 correction Methods 0.000 claims description 22
- 230000007613 environmental effect Effects 0.000 claims description 21
- 238000011088 calibration curve Methods 0.000 claims description 13
- 230000003746 surface roughness Effects 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 4
- 238000005259 measurement Methods 0.000 claims description 3
- 230000007935 neutral effect Effects 0.000 claims description 3
- 238000004891 communication Methods 0.000 abstract description 4
- 230000005540 biological transmission Effects 0.000 abstract description 3
- 239000011159 matrix material Substances 0.000 description 12
- 239000013598 vector Substances 0.000 description 12
- 244000062662 apex predator Species 0.000 description 10
- 230000002431 foraging effect Effects 0.000 description 8
- 230000033001 locomotion Effects 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 230000005653 Brownian motion process Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000005537 brownian motion Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 239000013256 coordination polymer Substances 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 238000009413 insulation Methods 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000004083 survival effect Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004379 similarity theory Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/10—Devices for predicting weather conditions
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Environmental & Geological Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Atmospheric Sciences (AREA)
- Biodiversity & Conservation Biology (AREA)
- Ecology (AREA)
- Environmental Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a BP neural network-based NPS-GRYANIK20 evaporation waveguide height prediction method, which is characterized in that an NPS-GRYANIK20 evaporation waveguide model-based environment interval division method is adopted, BP neural network prediction models based on predator algorithms are built in different environment intervals, and accurate prediction of evaporation waveguide height is realized, and the method specifically comprises the following steps: the method comprises the steps of obtaining a true value of the height of an evaporation waveguide, dividing an environment interval based on an NPS-GRYANIK20 evaporation waveguide model, predicting the height of the evaporation waveguide based on an improved BP neural network prediction model and the like. Compared with the prior art, the method has the advantages of improving the accuracy and reliability of the height prediction of the evaporation waveguide, better solving the problem of inaccurate height prediction of the evaporation waveguide, providing high-precision model support for the fine depiction of the transmission path of the ocean electromagnetic wave, and having important practical application value for the field of ocean communication.
Description
Technical Field
The invention relates to the technical field of evaporation waveguide calculation, in particular to an NPS-GRYANIK20 evaporation waveguide height prediction method based on a BP neural network.
Background
Evaporation waveguides are a common atmospheric waveguide phenomenon that occurs within the convective layer and is related to various weather factors of the sea surface atmosphere environment. The sea evaporation can cause the interaction of sea gas, and the phenomenon that the refractive curvature is smaller than the earth curvature occurs at a certain altitude, so that the phenomenon of normal atmospheric refractive index is generated, and the beyond-the-horizon propagation of electromagnetic waves on the sea surface can be realized. The evaporation waveguide has important influence on marine communication, marine development, marine rescue, military investigation and the like, and how to accurately predict the evaporation waveguide height becomes an important subject in marine evaporation waveguide research.
At present, a prediction model method is a main stream method for obtaining the height of an evaporation waveguide, and the method is based on a Moning-Oibuf similarity theory, and by substituting the atmospheric temperature, the sea surface temperature, the atmospheric pressure, the relative humidity and the wind speed of a certain height on the sea into the evaporation waveguide prediction model, combining an atmospheric correction refractive index formula, and carrying out iterative computation for several times to obtain an atmospheric correction refractive profile, wherein the height value corresponding to the lowest point of the correction refractive index at the profile is the evaporation waveguide height. However, this approach lacks flexibility in the selection of the generic function, stability correction function, and scale parameters under non-uniform, strongly stable atmospheric conditions, resulting in large deviations in waveguide height. Therefore, there is a need to improve the accuracy of prediction of ocean evaporation waveguide height by predictive modeling.
Disclosure of Invention
The invention aims to provide a method for predicting the height of an NPS-GRYANIK20 evaporation waveguide based on a BP neural network, which aims at the defects of the prior art, adopts a method for dividing the environment interval based on the NPS-GRYANIK20 evaporation waveguide model, builds a BP neural network prediction model based on a predator algorithm in different environment intervals, realizes accurate prediction of the height of the evaporation waveguide, analyzes the dependence of different environment parameters on the NPS-GRYANIK20 evaporation waveguide height model, divides different environment intervals, optimizes BP neural network parameters by introducing a predator algorithm, builds a BP neural network prediction model based on the predator algorithm in different environment intervals, greatly improves the accuracy and reliability of the evaporation waveguide height prediction by optimizing the BP neural network parameters, has the advantages of scientific and reasonable method, strong applicability and good effect, effectively solves the problem of inaccurate prediction of the existing evaporation waveguide height prediction model, and has higher practical application value.
The purpose of the invention is realized in the following way: the NPS-GRYANIK20 evaporation waveguide height prediction method based on the BP neural network is characterized in that an environmental interval division method based on an NPS-GRYANIK20 evaporation waveguide model is adopted, BP neural network prediction models based on predator algorithms are built in different environmental intervals, and accurate prediction of the evaporation waveguide height is achieved, and the method specifically comprises the following steps:
(1) And (3) obtaining the true value of the height of the evaporation waveguide: and acquiring data of atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity and wind speed at different heights by an offshore vertical gradient observation method, and correcting the data to obtain the true occurrence height of the evaporation waveguide, namely the true value of the evaporation waveguide height.
The data correction is a data processing method and mainly comprises two aspects of sensor calibration curve correction and tide level data correction. In actual observation, the sensor output signal is often affected by various factors, such as temperature, humidity, noise, etc., which may cause a deviation between the sensor output signal and the actual measured value. To eliminate these deviations, sensor calibration curve corrections are required. In particular, the sensor calibration curve is obtained by laboratory calibration of the sensor, describing the relationship between the sensor output signal and the actual measured value. During the data processing process, the sensor calibration curve is used for revising the observed data, and the observed data is revised according to different parts of the sensor calibration curve, so that a more accurate measurement result is obtained. The correction of the tide level data refers to correcting the tide level data in actual observation to eliminate errors caused by various factors. In the observation of the tide level, there may be a deviation in the tide level data due to the complexity of the marine environment, the limitation of the equipment, and the like. To eliminate these deviations, it is necessary to employ different correction methods, such as correction of tidal parameters obtained by tidal analysis, or correction by comparison with reference station data. By correcting the tide level data, more accurate tide data can be obtained, and more accurate data support is provided for monitoring and researching marine environments.
(2) Environmental interval division based on NPS-GRYANIK20 evaporation waveguide model: based on the GRYANIK20 similarity function, the traditional simple linear similarity function is updated, and an NPS evaporation waveguide model based on the GRYANIK20 similarity function, which is simply called as an NPS-GRYANIK20 evaporation waveguide model, is formed. And (3) obtaining the evaporation waveguide height by adopting an NPS-GRYANIK20 evaporation waveguide model on the basis of obtaining the atmospheric temperature, the sea surface temperature, the atmospheric pressure, the relative humidity and the wind speed of the meteorological site in the step (1).
The vertical profile of the near-formation temperature T and specific humidity R in the NPS-GRYANIK20 evaporation waveguide model is represented by the following formulas (a) - (b):
1)
2)
wherein T (z) and R (z) are the air temperature and specific humidity at the height z, respectively; t (T) 0 And R is 0 Sea surface temperature and specific humidity respectively; θ and R are characteristic dimensions of the bit temperature θ and specific humidity R, respectively; k is a Kalman constant; z 0t Is the temperature roughness height; psi h Is a temperature universal function; Γ -shaped structure d Decreasing the rate for dry insulation; l is a similar length.
The wind speed and temperature stability function in the form of GRYANIK20 in the NPS-GRYANIK20 evaporation waveguide model is represented by the following formulas (c) to (d):
3)
4)
wherein a is m ，a h ，b m And b h Is an empirical constant; pr (Pr) 0 Is defined as Pr, the neutral limit being the Plantnumber 0 ＝0.98，a m ＝5.0，a h ＝5.0，b m ＝0.3，b h ＝0.4。
The pressure profile in the NPS-GRYANIK20 evaporation waveguide model is determined by the following equation (e):
5)
wherein p (z) 1 ) And p (z) 2 ) Respectively the measured heights z 1 And z 2 Air pressure at the location; c is the dry air gas constant; g is gravity acceleration; t (T) v At a height z 1 And z 2 The virtual temperature average value at the point.
The water vapor pressure profile is determined from the following equation (f) based on the specific humidity R as a function of the water vapor pressure e:
6)
where f is a constant of 0.622.
The atmospheric refractive index profile can be obtained according to the temperature, air pressure and water vapor pressure profiles determined by the formulas (a), (b) and (f), and the height corresponding to the minimum position of the profile is the evaporation waveguide height.
According to the calculation method of the NPS-GRYANIK20 evaporation waveguide model, the dependence degree of different environmental parameters on the NPS-GRYANIK20 evaporation waveguide model is analyzed by setting the environmental parameter change ranges such as atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity and wind speed, and further different environmental intervals are divided.
(3) Building a BP neural network prediction model: and (3) introducing a predator algorithm to optimize the weight and the bias parameters of the BP neural network to obtain an optimal value. And (3) according to the analysis result of the step (2), inputting the obtained atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity, wind speed, rainfall attenuation, sea surface roughness and evaporation waveguide model height true value in each interval, and taking the error of the evaporation waveguide model height predicted value and the evaporation waveguide model height true value as a judgment basis. And stopping training when the error meets a set threshold value to obtain a nonlinear mapping relation between input and output, so as to construct a BP neural network prediction model based on a predator algorithm to predict the height of the evaporation waveguide.
The BP neural network prediction model based on the predator algorithm is as follows: in the predator algorithm, according to the theory of survival of the fittest, the most prey individuals in the predator population are called top predators. Using top predators to construct a matrix named ellite, each array in the matrix providing the predators with information on the location of the currently found prey in the next foraging, the ellite matrix represented by the following formula (g):
7)
where N is the number of predators, D is the predators individual dimension, and x is the top predator vector.
After each iteration, if predators with better predators appear, the predators with better predators replace the original top predators to become new top predators, the Elite matrix is updated, and the updated matrix is expressed as a Prey matrix by the following formula (h):
8)
wherein X is ij Represents the j dimension of the i-th prey.
The optimization of the predator algorithm is mainly performed around the two matrixes, and the optimization process can be divided into the following steps according to different speed ratios of predators and prey: three phases of high, equal and low speed ratios, briefly described as follows:
1) In high ratio scenarios, predators move slower than prey. This stage belongs to the detection stage of the algorithm, and the mathematical formula is expressed by the following formulas (i) - (j):
9)
10)
in the method, in the process of the invention,is a random vector containing a normal distribution, representing brownian motion. />A step vector representing the next movement of the ith predator. P is a constant for adjusting the movement step of the predator,/->Is a random vector in the range of 0 to 1.
2) In the constant velocity ratio scenario, the prey and predator speeds are equivalent, and at this stage, the algorithm gradually transitions from exploration to exploitation, and the formula is expressed by the following formulas (k) - (1):
11)
12)
in the method, in the process of the invention,is a set of random vectors that obey the Levy distribution, CP is an adaptive control variable for controlling predator movement steps, T is the current iteration number, and T is the maximum iteration number.
3) In a low ratio scenario, where predators are faster than prey, the algorithm is transformed from exploration to development, and the best foraging strategy for predators is Levy movement, the mathematical formula of which is expressed by the following (m):
13)
14)
in nature, environmental changes can also affect predators' foraging activities, in predator algorithms, f is considered locally optimal, and its mathematical model is defined by the following equation (o):
15)
where f=0.2 denotes the probability of foraging activity occurring under the influence of environmental changes, r is a randomly generated random number in the range of 0 to 1,is a binary vector comprising 0 or 1, ">And->The upper and lower limits of the decision variables, respectively.
Modeling a predator algorithm model, taking a Mean Square Error (MSE) of the BP neural network as an fitness function, wherein the MSE is used for evaluating an error between a predicted result and an actual result of the BP neural network and is expressed by the following formula (p):
16)
where N is the number of samples, y i Is the actual value of the i-th sample,the predicted value of the ith sample is that the smaller the MSE is, the smaller the error between the predicted result and the actual result of the BP neural network is, and the better the prediction performance is.
(4) Evaporation waveguide height prediction: and (3) acquiring new data such as atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity, wind speed and the like, inputting the data into the BP neural network prediction model based on the predator algorithm established in the step (3), and carrying out the evaporation waveguide model height prediction. In addition, after the prediction model is built in the step (3), the model is evaluated, new data such as atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity, wind speed, rainfall attenuation, sea surface roughness and the like are substituted into the nonlinear mapping relation built in the step (3), the evaporation waveguide height predicted value at the current moment is obtained, and the evaluation is performed with the evaporation waveguide height true value, wherein the evaluation mode comprises absolute error, variance, correlation and the like between the evaporation waveguide height predicted value and the evaporation waveguide height true value.
Compared with the prior art, the method has the advantages of improving the accuracy and reliability of the height prediction of the evaporation waveguide, better solving the problem of inaccurate height prediction of the evaporation waveguide, providing high-precision model support for the fine depiction of the transmission path of the ocean electromagnetic wave, and having important practical application value for the field of ocean communication.
Drawings
FIG. 1 is a schematic flow chart of the present invention;
fig. 2 is a flow chart of a BP neural network based on predator algorithm.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
Example 1
Referring to fig. 1, the present invention achieves evaporation waveguide height prediction as follows:
step one, obtaining the true value of the height of the evaporation waveguide
And acquiring data of atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity and wind speed at different heights by an offshore vertical gradient observation method, and correcting a sensor calibration curve and tidal level data on the data to obtain the true occurrence height of the evaporation waveguide, namely the true value of the evaporation waveguide height.
The data correction is a data processing method and mainly comprises two aspects of sensor calibration curve correction and tide level data correction. In actual observation, the sensor output signal is often affected by various factors, such as temperature, humidity, noise, etc., which may cause a deviation between the sensor output signal and the actual measured value. To eliminate these deviations, sensor calibration curve corrections are required. In particular, the sensor calibration curve is obtained by laboratory calibration of the sensor, describing the relationship between the sensor output signal and the actual measured value. During the data processing process, the sensor calibration curve is used for revising the observed data, and the observed data is revised according to different parts of the sensor calibration curve, so that a more accurate measurement result is obtained. The correction of the tide level data refers to correcting the tide level data in actual observation to eliminate errors caused by various factors. In the observation of the tide level, there may be a deviation in the tide level data due to the complexity of the marine environment, the limitation of the equipment, and the like. To eliminate these deviations, it is necessary to employ different correction methods, such as correction of tidal parameters obtained by tidal analysis, or correction by comparison with reference station data. By correcting the tide level data, more accurate tide data can be obtained, and more accurate data support is provided for monitoring and researching marine environments.
Step two, environmental interval division based on NPS-GRYANIK20 evaporation waveguide model
The invention realizes the update of the traditional simple linear similarity function based on the GRYANIK20 similarity function, and forms an NPS evaporation waveguide model based on the GRYANIK20 similarity function, which is simply called as an NPS-GRYANIK20 evaporation waveguide model. And on the basis of acquiring the atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity and wind speed of the meteorological site, acquiring the evaporation waveguide height by adopting an NPS-GRYANIK20 evaporation waveguide model. In the NPS-GRYANIK20 evaporation waveguide model, the vertical profile of the near-formation temperature T and specific humidity R is expressed by the following formulas (a) to (b):
17)
18)
wherein T (z) and R (z) are the air temperature and specific humidity at the height z, respectively; t (T) 0 And R is 0 Sea surface temperature and specific humidity respectively; θ and R are characteristic dimensions of the bit temperature θ and specific humidity R, respectively; k is a Kalman constant; zo (zo) t Is the temperature roughness height; psi h Is a temperature universal function; Γ -shaped structure d Decreasing the rate for dry insulation; l is a similar length.
The wind speed and temperature stability function in the form of GRYANIK20 in the NPS-GRYANIK20 evaporation waveguide model is represented by the following formulas (c) to (d):
19)
wherein a is m ，a h ，b m And b h Is an empirical constant; pr (Pr) 0 Is defined as Pr, the neutral limit being the Plantnumber 0 ＝0.98，a m ＝5.0，a h ＝5.0，b m ＝0.3，b h ＝0.4。
The pressure profile in the NPS-GRYANIK20 evaporation waveguide model is determined by the following equation (e):
20)
wherein p (z) 1 ) And p (z) 2 ) Respectively the measured heights z 1 And z 2 Air pressure at the location; c is the dry air gas constant; g is gravity acceleration; t (T) v At a height z 1 And z 2 The virtual temperature average value at the point.
The water vapor pressure profile is determined from the following equation (f) based on the specific humidity R as a function of the water vapor pressure e:
21)
where f is a constant of 0.622.
The atmospheric refractive index profile can be obtained according to the temperature, air pressure and water vapor pressure profiles determined by the formulas (a), (b) and (f), and the height corresponding to the minimum position of the profile is the evaporation waveguide height.
According to the calculation method of the NPS-GRYANIK20 evaporation waveguide model, the dependence degree of different environmental parameters on the NPS-GRYANIK20 evaporation waveguide model is analyzed by setting the environmental parameter change ranges such as atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity and wind speed, and further different environmental intervals are divided.
(3) BP neural network prediction model-based construction
And (3) introducing a predator algorithm to optimize the weight and the bias parameters of the BP neural network to obtain an optimal value. And (3) according to the analysis result of the step (2), inputting the obtained atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity, wind speed, rainfall attenuation, sea surface roughness and evaporation waveguide model height true value in each interval, and taking the error of the evaporation waveguide model height predicted value and the evaporation waveguide model height true value as a judgment basis. And stopping training when the error meets a set threshold value to obtain a nonlinear mapping relation between input and output, so as to construct a BP neural network prediction model based on a predator algorithm to predict the height of the evaporation waveguide.
Referring to fig. 2, the BP neural network prediction model based on the predator algorithm is as follows: in the predator algorithm, according to the theory of survival of the fittest, the most prey individuals in the predator population are called top predators. Using top predators to construct a matrix named ellite, each array in the matrix providing the predators with information on the location of the currently found prey in the next foraging, the ellite matrix represented by the following formula (g):
22)
where N is the number of predators, D is the predators individual dimension, and x is the top predator vector.
After each iteration, if predators with better predators appear, the predators with better predators replace the original top predators to become new top predators, the Elite matrix is updated, and the updated matrix is expressed as a Prey matrix by the following formula (h):
23)
wherein X is ij Represents the j dimension of the i-th prey.
The optimization of the predator algorithm is mainly performed around the two matrixes, and the optimization process can be divided into the following steps according to different speed ratios of predators and prey: three phases of high, equal and low speed ratios, briefly described as follows:
1) In high ratio scenarios, predators move slower than prey. This stage belongs to the detection stage of the algorithm, and the mathematical formula is expressed by the following formulas (i) - (j):
24)
25)
in the method, in the process of the invention,is a random vector containing a normal distribution, representing brownian motion. />A step vector representing the next movement of the ith predator. P is a constant for adjusting the movement step of the predator,/->Is a random vector in the range of 0 to 1.
2) In the constant velocity ratio scenario, the prey and predator speeds are equivalent, and at this stage, the algorithm gradually transitions from exploration to exploitation, and the formula is expressed by the following formulas (k) - (1):
26)
27)
in the method, in the process of the invention,is a set of random vectors that obey the Levy distribution, CP is an adaptive control variable for controlling predator movement steps, T is the current iteration number, and T is the maximum iteration number.
3) In a low ratio scenario, where predators are faster than prey, the algorithm is transformed from exploration to development, and the best foraging strategy for predators is Levy movement, the mathematical formula of which is expressed by the following (m):
28)
29)
in nature, environmental changes can also affect predators' foraging activities, in predator algorithms, f is considered locally optimal, and its mathematical model is defined by the following equation (o):
30)
where f=0.2 denotes the probability of foraging activity occurring under the influence of environmental changes, r is a randomly generated random number in the range of 0 to 1,is a binary vector comprising 0 or 1, ">And->The upper and lower limits of the decision variables, respectively.
Modeling a predator algorithm model, taking a Mean Square Error (MSE) of the BP neural network as an fitness function, wherein the MSE is used for evaluating an error between a predicted result and an actual result of the BP neural network and is expressed by the following formula (p):
31)
where N is the number of samples, y i Is the actual value of the i-th sample,the predicted value of the ith sample is that the smaller the MSE is, the smaller the error between the predicted result and the actual result of the BP neural network is, and the better the prediction performance is.
(4) Evaporation waveguide height prediction
And (3) acquiring new data such as atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity, wind speed and the like, inputting the data into the BP neural network prediction model based on the predator algorithm established in the step (3), and carrying out the evaporation waveguide model height prediction. In addition, after the prediction model is built in the step (3), the model is evaluated, new data such as atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity, wind speed, rainfall attenuation, sea surface roughness and the like are substituted into the nonlinear mapping relation built in the step (3), and the evaporation waveguide height prediction value at the current moment is obtained, and is evaluated with the evaporation waveguide height true value, wherein the evaluation mode comprises the following steps: absolute error, variance, correlation, etc. between the evaporation waveguide height prediction value and the evaporation waveguide height true value.
Aiming at the problem that the evaporation waveguide height prediction is inaccurate in the existing evaporation waveguide height prediction model, the invention provides an NPS-GRYANIK20 evaporation waveguide height prediction method based on a BP neural network. According to the method, the dependence of different environment parameters on the NPS-GRYANIK20 evaporation waveguide height model is analyzed, different environment intervals are further divided, the BP neural network parameters are optimized by introducing the predator algorithm, the BP neural network prediction model based on the predator algorithm is built in the different environment intervals, so that the accuracy and reliability of evaporation waveguide height prediction are improved, the method is scientific and reasonable, the applicability is high, the effect is good, high-precision model support can be provided for fine depiction of an ocean electromagnetic wave transmission path, and important practical application value is provided for the ocean communication field.
The invention is further described with reference to the following claims, which are not intended to limit the scope of the invention.
Claims (5)
1. The NPS-GRYANIK20 evaporation waveguide height prediction method based on the BP neural network is characterized in that an environmental interval division method based on an NPS-GRYANIK20 evaporation waveguide model is adopted, BP neural network prediction models based on predator algorithms are built in different environmental intervals, and accurate prediction of the evaporation waveguide height is achieved, and the method specifically comprises the following steps:
step one, obtaining the true value of the height of the evaporation waveguide
Acquiring atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity and wind speed data at different heights by an offshore vertical gradient observation method, and correcting a sensor calibration curve and tidal level data to obtain the true occurrence height of the evaporation waveguide, namely the true value of the evaporation waveguide height;
step two, environmental interval division based on NPS-GRYANIK20 evaporation waveguide model
On the basis of atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity and wind speed, an NPS-GRYANIK20 evaporation waveguide model is adopted to obtain an evaporation waveguide height predicted value;
third, building a BP neural network prediction model
According to the height of the evaporation waveguide, inputting the obtained atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity, wind speed, rainfall attenuation, sea surface roughness and the true value of the height of the evaporation waveguide in each interval, taking the error of the predicted value of the height of the evaporation waveguide and the true value of the height of the evaporation waveguide as a judgment basis, stopping training when the error meets a set threshold value, and obtaining a nonlinear mapping relation between the input and the output, thereby constructing a BP neural network prediction model based on a predator algorithm;
step four, predicting the height of the evaporation waveguide
Inputting the re-acquired atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity, wind speed, rainfall attenuation and sea surface roughness data into a BP neural network prediction model based on predator algorithm to obtain a predicted value of the evaporation waveguide height at the current moment;
step five, evaluating the performance of the prediction model
The performance of the evaporation waveguide height prediction model is evaluated by calculating the root mean square error and the average absolute error between the evaporation waveguide height prediction value and the evaporation waveguide height true value, if the error is small, the prediction precision is high, otherwise, the prediction precision is low.
2. The method for predicting the height of an NPS-GRYANIK20 evaporation waveguide based on a BP neural network according to claim 1, wherein the environmental section of the NPS-GRYANIK20 evaporation waveguide model is divided into NPS evaporation waveguide models based on GRYANIK20 similarity functions, the atmospheric temperature, the sea surface temperature, the atmospheric pressure, the relative humidity and the wind speed of a meteorological site are obtained by an offshore vertical gradient observation method, the evaporation waveguide height is obtained, the environmental parameter variation ranges of the atmospheric temperature, the sea surface temperature, the atmospheric pressure, the relative humidity and the wind speed are set, the dependence degree of different environmental parameters on the NPS-GRYANIK20 evaporation waveguide models is analyzed, and further, different environmental sections are divided, and the wind speed and the temperature stability function of the GRYANIK20 form in the NPS-GRYANIK20 evaporation waveguide models are expressed by the following formulas (c) - (d)
Wherein a is m ,a h ,b m And b h Is an empirical constant; pr (Pr) 0 Is defined as Pr, the neutral limit being the Plantnumber 0 ＝0.98，a m ＝5.0，a h ＝5.0，b m ＝0.3，b h ＝0.4。
3. The method for predicting the height of the evaporation waveguide of the NPS-GRYANIK20 based on the BP neural network according to claim 1, wherein the BP neural network prediction model is characterized in that a predator algorithm is introduced to optimize the weight and the bias parameters of the BP neural network, and then the optimal value is obtained, and then the obtained atmospheric temperature, sea surface temperature, atmospheric pressure, relative humidity, wind speed, rainfall attenuation, sea surface roughness and the evaporation waveguide model height true value are taken as input in each interval according to the divided environment interval, the error of the evaporation waveguide height prediction value and the evaporation waveguide height true value is taken as a decision basis, and training is stopped when the error meets a set threshold value, so that a nonlinear mapping relation between input and output is obtained, and the evaporation waveguide height prediction model is built.
4. The BP neural network-based NPS-greanik 20 evaporation waveguide height prediction method of claim 1, wherein the sensor calibration curve correction revises the observed data using a laboratory calibration-obtained curve, thereby obtaining a more accurate measurement result.
5. The BP neural network-based NPS-greanik 20 evaporation waveguide height prediction method of claim 1, wherein the tidal level data correction is corrected using tidal parameters or by comparison with reference station data to obtain more accurate tidal data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310539832.XA CN116559977A (en) | 2023-05-15 | 2023-05-15 | NPS-GRYANIK20 evaporation waveguide height prediction method based on BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310539832.XA CN116559977A (en) | 2023-05-15 | 2023-05-15 | NPS-GRYANIK20 evaporation waveguide height prediction method based on BP neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116559977A true CN116559977A (en) | 2023-08-08 |
Family
ID=87491255
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310539832.XA Pending CN116559977A (en) | 2023-05-15 | 2023-05-15 | NPS-GRYANIK20 evaporation waveguide height prediction method based on BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116559977A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117421601A (en) * | 2023-12-19 | 2024-01-19 | 山东省科学院海洋仪器仪表研究所 | Sea surface evaporation waveguide near-future rapid forecasting method |
| CN117540342A (en) * | 2024-01-05 | 2024-02-09 | 山东省科学院海洋仪器仪表研究所 | Short-term prediction method and system for fusion of evaporation waveguide |
-
2023
- 2023-05-15 CN CN202310539832.XA patent/CN116559977A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117421601A (en) * | 2023-12-19 | 2024-01-19 | 山东省科学院海洋仪器仪表研究所 | Sea surface evaporation waveguide near-future rapid forecasting method |
| CN117421601B (en) * | 2023-12-19 | 2024-03-01 | 山东省科学院海洋仪器仪表研究所 | Sea surface evaporation waveguide near-future rapid forecasting method |
| CN117540342A (en) * | 2024-01-05 | 2024-02-09 | 山东省科学院海洋仪器仪表研究所 | Short-term prediction method and system for fusion of evaporation waveguide |
| CN117540342B (en) * | 2024-01-05 | 2024-03-26 | 山东省科学院海洋仪器仪表研究所 | Short-term prediction method and system for fusion of evaporation waveguide |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116559977A (en) | NPS-GRYANIK20 evaporation waveguide height prediction method based on BP neural network | |
| CN111310889B (en) | Evaporation waveguide profile estimation method based on deep neural network | |
| WO2022194045A1 (en) | Three-dimensional temperature-salinity field analysis and forecasting method for offshore platform guarantee | |
| WO2018218832A1 (en) | Method and apparatus for correcting predicted wind speed of wind farm | |
| CN112711899B (en) | Fusion prediction method for height of evaporation waveguide | |
| CN106779137A (en) | A kind of method that ship oil consumption is predicted according to sea situation and operating condition | |
| CN112288164B (en) | Wind power combined prediction method considering spatial correlation and correcting numerical weather forecast | |
| CN113657028B (en) | Online aerosol optical thickness prediction method based on multi-source information | |
| CN113469426A (en) | Photovoltaic output power prediction method and system based on improved BP neural network | |
| CN115859789A (en) | Method for improving inversion accuracy of polar atmosphere temperature profile | |
| CN113240198A (en) | Port ship track prediction method based on TCN model | |
| CN113468799A (en) | Method and system for acquiring near-ground PM2.5 concentration in static meteorological satellite observation | |
| CN108663727B (en) | Method for estimating height of evaporation waveguide in world sea area by using evaporation rate | |
| CN115356702A (en) | Evaporation waveguide deep learning joint inversion method based on marine multi-source radar echo | |
| CN113609639B (en) | Evaporation waveguide correction method suitable for stable condition | |
| CN113095547B (en) | Short-term wind power prediction method based on GRA-LSTM-ICE model | |
| CN112800075B (en) | Ship manipulation prediction database updating method based on six-degree-of-freedom attitude data of real ship | |
| CN117172149A (en) | Evaporation waveguide prediction method based on data feature classification and neural network model | |
| CN113642249A (en) | Gyroscope zero-bias error compensation method | |
| CN114817837B (en) | Troposphere parameter profile inversion method based on buoy platform microwave radiometer | |
| CN116525135B (en) | Method for predicting epidemic situation development situation by space-time model based on meteorological factors | |
| Moghaddamnia et al. | Evaporation estimation using support vector machines technique | |
| Chang et al. | Neural network with multi-trend simulating transfer function for forecasting typhoon wave | |
| CN113311509B (en) | Sensitivity test method of MWHTS (metal wrap through) to sea surface air pressure based on neural network | |
| CN113536657A (en) | Ocean atmosphere refractive index prediction method based on machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |