CN109583593A - A kind of low-level wind shear recognition methods based on automatic weather station - Google Patents
A kind of low-level wind shear recognition methods based on automatic weather station Download PDFInfo
- Publication number
- CN109583593A CN109583593A CN201811285720.1A CN201811285720A CN109583593A CN 109583593 A CN109583593 A CN 109583593A CN 201811285720 A CN201811285720 A CN 201811285720A CN 109583593 A CN109583593 A CN 109583593A
- Authority
- CN
- China
- Prior art keywords
- wind
- data
- shear
- wind shear
- automatic weather
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000012544 monitoring process Methods 0.000 claims abstract description 28
- 238000010801 machine learning Methods 0.000 claims abstract description 9
- 238000003908 quality control method Methods 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims description 33
- 238000007689 inspection Methods 0.000 claims description 27
- 239000013598 vector Substances 0.000 claims description 24
- 230000008859 change Effects 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 6
- 230000005856 abnormality Effects 0.000 claims description 5
- 238000003066 decision tree Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 238000012706 support-vector machine Methods 0.000 claims description 4
- 238000002790 cross-validation Methods 0.000 claims description 3
- 238000007477 logistic regression Methods 0.000 claims description 3
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 claims description 2
- 230000002159 abnormal effect Effects 0.000 claims 1
- 230000008901 benefit Effects 0.000 abstract description 4
- 238000013479 data entry Methods 0.000 description 3
- 241001269238 Data Species 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000002459 sustained effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01P—MEASURING LINEAR OR ANGULAR SPEED, ACCELERATION, DECELERATION, OR SHOCK; INDICATING PRESENCE, ABSENCE, OR DIRECTION, OF MOVEMENT
- G01P13/00—Indicating or recording presence, absence, or direction, of movement
- G01P13/02—Indicating direction only, e.g. by weather vane
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01P—MEASURING LINEAR OR ANGULAR SPEED, ACCELERATION, DECELERATION, OR SHOCK; INDICATING PRESENCE, ABSENCE, OR DIRECTION, OF MOVEMENT
- G01P5/00—Measuring speed of fluids, e.g. of air stream; Measuring speed of bodies relative to fluids, e.g. of ship, of aircraft
Landscapes
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Educational Administration (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Aerodynamic Tests, Hydrodynamic Tests, Wind Tunnels, And Water Tanks (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The low-level wind shear recognition methods based on automatic weather station that the present invention relates to a kind of includes the following steps: the historical data for obtaining each wind element in monitoring region；Quality control and time synchronization are carried out to historical data, obtain data list；The valid data for actually occurring wind shear are chosen, the shear value of each wind element is calculated；Distance between standing two-by-two by automatic weather station scales standard value, and compared with wind shear discrimination threshold, the label of wind shear whether is occurred；Data set is constituted as a result using each wind element as input, the label for whether occurring wind shear, and SMOTE algorithm is used to make two kinds of tag types quantity balances；It is trained using machine learning method, and exports identification model；The latest data of each automatic weather station is read, identification model is inputted, obtains recognition result.Method of the invention has the advantages that accuracy of identification is high, local adaptability is good, can effectively be identified to runway horizontal wind shear near the ground, improve the guarantee efficiency of Meteorological Services.
Description
Technical field
The present invention relates to Data of Automatic Weather analysis fields, in particular relate to a kind of low latitude based on automatic weather station
Wind shear recognition methods.
Background technique
Wind shear is a kind of common meteor, refer in sustained height or different height short distance wind direction (or) wind speed
Quick variation.According to the difference of wind vector, wind shear can be divided into the horizontal shear of horizontal wind, horizontal vertical wind shear
With vertical wind shear.It is referred to as low wherein at 600 meters of flying height or less, mainly in the wind shear of takeoff and landing stage generation
Empty wind shear.It is counted according to World Meteorological Organization and International Civil Aviation Organization, low-level wind shear is to takeoff and landing stage flight safety
Maximum weather phenomenon is threatened, many great air crashes are caused by low-level wind shear.And in aircraft landing, for resolution
Wind field situation highly below is the most important thing of observation.
Currently, relying primarily on space wind field in relevant equipment search coverage for the detection identification of wind shear both at home and abroad
Distribution and situation of change.Automatic weather station is generally deployed in runway two sides, close automatically horizontal nearby for intuitively obtaining runway
Wind field, therefore in traditional wind shear recognition methods, it can use Data of Automatic Weather and calculate the horizontal shear of surface wind.Its
In, more mature efficient system is the LLWAS system in the U.S., but its to ground survey wind devices deployment quantity and position have compared with
High requirement, and it is insufficient to the specific aim of local landform, climatic characteristic, device characteristics.Current many medium and small airports are in construction
Wind devices are surveyed on the ground for being not equipped with the quantity as required by LLWAS system and installation site, and LLWAS system wind shear algorithm is in local
Adaptability it is also invalidated reliable, the judgement of forecaster is relied primarily in practical business, accuracy and coverage still have promotion
Space.
Summary of the invention
According to problems of the prior art, the present invention provides a kind of, and the low-level wind shear based on automatic weather station is known
Other method has the advantages that accuracy of identification is high, local adaptability is good, can carry out to runway horizontal wind shear near the ground effective
Identification, improve the guarantee efficiency of Meteorological Services.
The invention adopts the following technical scheme:
A kind of low-level wind shear recognition methods based on automatic weather station, which comprises the steps of:
S1 obtains the historical data of all collected each wind elements of automatic weather station in monitoring region；
S2 carries out quality control to historical data with rejecting abnormalities data, obtains the valid data of each wind element, and according to
The data moment of setting carries out time synchronization to valid data, obtains the corresponding data list of valid data；
S3 chooses the valid data in the time range that front and back is set at the time of actually occurring wind shear in data list,
The shear at all automatic weather stations being calculated in the monitoring region corresponding each wind element each data moment between standing two-by-two
Value；
S4, by it is described monitoring region in automatic weather station stand two-by-two between apart from bi-directional scaling standard value, will scale
Standard value afterwards obtains whether each data moment occurs wind shear compared with wind shear discrimination threshold, and provides each
Whether the data moment occurs the label of wind shear；
S5 is constituted as a result using each wind element of each time data as input, the label for whether occurring wind shear
Data set, and SMOTE algorithm EDS extended data set is used to make two kinds of tag types quantity balance that wind shear whether occurs, it obtains
Data set after balance；
S6 is trained the data set after balance using machine learning method, chooses the best training result of recall ratio,
And corresponding model is exported, using this model as identification model；
S7 reads the latest data of each automatic weather station in monitoring region, using the method for step S2 to newest number in real time
According to being pre-processed, and will treated that data input the identification model, and then obtain recognition result.
Preferably, in step S1, the wind element includes instantaneous wind speed, instantaneous wind direction, gustiness, fitful wind wind direction, two
Minute mean wind speed, two minutes mean wind directions, two minutes maximum wind velocities, two minutes maximum wind directions, ten minutes mean wind speeds, very
Clock mean wind direction, ten minutes maximum wind velocities, ten minutes maximum wind directions.
It is further preferred that carrying out quality control to historical data includes data normalization inspection processing, gas in step S2
Wait educational circles's limit value inspection processing, the processing of extreme value range check, internal consistency inspection processing, time consistency inspection processing；Institute
It states data normalization inspection processing to refer to and delete the repeated data occurred in historical data, and corresponding to historical data
Filename, the date, the time, automatic weather station site information carry out improve and uniform format；The climatology boundary value inspection
Processing refers to the local historical climate feature according to locating for Meteorological Field relevant criterion and monitoring region, goes through to what can not be occurred
History data are rejected；The extreme value range check processing refers to the history monthly average wind speed for choosing each automatic weather station acquisition
For basic unit, calculate the standard deviation of the annual monthly average wind speed, using this month history maximum wind velocity plus twice of standard deviation as
Maximum cuts twice of standard deviation as minimum using this month history minimum windspeed, and rejecting is more than that maximum value is less than minimum
Historical data；The internal consistency inspection processing specifically refers to carry out inspection processing to the physical link each wind element, i.e.,
Then wind speed when according to ten minutes maximum wind velocity >=two minute maximum wind velocities, two minutes maximum wind velocity >=instantaneous wind speeds, wind direction being 0 °
These judgment criterias of≤0.2m/s, rejecting abnormalities data；Time consistency inspection processing specifically refer to it is a certain current when
It is checked with the change rate of five minutes corresponding historical datas before this moment at quarter, upper limit threshold is greater than to change rate
Corresponding historical data is rejected, while to a certain current time and one hour corresponding historical data before this moment
Change rate checked, to change rate be less than lower threshold corresponding historical data reject, the upper limit threshold and
Lower threshold local historical climate feature according to locating for monitoring region determines.
Still more preferably, in step S2, it is specific that time synchronization is carried out to valid data according to the data moment of setting
Refer to the data list that valid data are established to valid data using per half a minute as the data moment, i.e., by each of valid data
Entry collects in that the hithermost moment half a minute corresponding data list of its original moment.
Still more preferably, in step S3, front and back each 12 at the time of actually occurring wind shear in data list is chosen
Hour in valid data, according in valid data instantaneous wind speed and instantaneous wind direction obtain instantaneous vector wind, according to fitful wind wind
Speed and fitful wind wind direction obtain vector fitful wind, obtain two minutes mean vectors according to two minutes mean wind speeds and two minutes mean wind directions
Wind obtains ten minutes average vector winds according to ten minutes mean wind speeds and ten minutes mean wind directions, calculates in the monitoring region
All automatic weather stations stand two-by-two between the phasor difference of instantaneous vector wind, the phasor difference of vector fitful wind, two minutes average vector winds
Phasor difference, ten minutes average vector winds phasor difference, and then obtain all automatic weather stations stand two-by-two between each data moment
Shear value.
Still more preferably, in step S4, according to the international judgment criteria of wind shear, i.e. between two automatic weather stations
Spacing when being 4km, when the wind shear value between two automatic weather stations is greater than 15 section, then be judged to that wind shear has occurred,
International judgment criteria is floated downward 10% to 20%, in this, as wind shear discrimination threshold.
Still more preferably, in step S4, the distance between the automatic weather station in the monitoring region is stood two-by-two is pressed
Scaling standard value, by the standard value after scaling compared with wind shear discrimination threshold, if the instantaneous wind at a certain data moment
Shear value is greater than wind shear discrimination threshold or battle array wind shear value is greater than wind shear discrimination threshold, average wind shear in two minutes
Value is greater than wind shear discrimination threshold and instantaneous wind shear value and wind shear discrimination threshold difference are within 20%, then when this data
Have occurred wind shear quarter, it is on the contrary then wind shear does not occur；And whether wind shear is occurred to each data moment and provides corresponding mark
Label, that is, have occurred the label that wind shear provides "Yes", the label that wind shear provides "No" do not occur.
Still more preferably, in step S5, two kinds of tag types quantity balance that wind shear whether occurs is specific
Refer to that data that label is "Yes" and label are the data bulk difference of "No" less than 5%.
Still more preferably, step S6 specifically refers to that support vector machines, logic is respectively adopted using machine learning method
It returns and three kinds of models of decision tree, the data set after input balances is trained, and fold into capable intersection using 10 in training process
Verifying chooses the best training result of recall ratio, and exports corresponding model, using this model as identification model.
The advantages and beneficial effects of the present invention are:
1) present invention first acquisition monitoring region in automatic weather station each wind element historical data, to historical data into
The control of row quality simultaneously carries out time synchronization to valid data according to the data moment of setting, obtains the corresponding data column of valid data
Table；The valid data in the time range that front and back is set at the time of actually occurring wind shear in data list are chosen, are calculated
The shear value at each wind element each data moment；Again by it is described monitoring region in automatic weather station stand two-by-two between distance by than
Example scaling standard value, by the standard value after scaling compared with wind shear discrimination threshold, and then obtains whether each data moment is sent out
Raw wind shear, and provide the label whether each data moment occur wind shear；Using each wind element of each time data as
Input, the label for whether occurring wind shear constitute data set as a result, and use SMOTE algorithm EDS extended data set make whether
Two kinds of tag types quantity balance of wind shear, the data set after being balanced occurs；Then, using machine learning method to flat
Data set after weighing apparatus is trained, and chooses the best training result of recall ratio, and export corresponding model, using this model as knowledge
Other model；Finally, the latest data of each automatic weather station in monitoring region, the method pair controlled using the quality are read in real time
Latest data is pre-processed, and by treated, data input the identification model, and then obtain recognition result.The present invention adopts
It is analyzed and is trained with based on Data of Automatic Weather, effective identification can be made to low latitude horizontal wind shear, compare traditional base
In the wind shear identification technology of Data of Automatic Weather, the history number of each wind element is analyzed in innovation by the way of machine learning
According to containing special to the monitoring local landform in region, climate characteristic, each site location feature and each measuring device data
Point etc. be relatively difficult to directly give amendment or it is existing study considering for not yet specific influence factor, so that larger improve Low level wind
The accuracy of shear identification, provides strong support for Meteorological safeguard and service.
Detailed description of the invention
Fig. 1 is the flow chart of method of the invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
As shown in Figure 1, a kind of low-level wind shear recognition methods based on automatic weather station, includes the following steps:
S1 obtains the historical data of all collected each wind elements of automatic weather station in monitoring region；
Specifically, the wind element includes instantaneous wind speed, instantaneous wind direction, gustiness, fitful wind wind direction, two minutes average winds
Speed, two minutes mean wind directions, two minutes maximum wind velocities, two minutes maximum wind directions, ten minutes mean wind speeds, ten minutes average winds
To, ten minutes maximum wind velocities, ten minutes maximum wind directions.
S2 carries out quality control to historical data with rejecting abnormalities data, obtains the valid data of each wind element, and according to
The data moment of setting carries out time synchronization to valid data, obtains the corresponding data list of valid data；
Specifically, carrying out quality control to historical data includes data normalization inspection processing, climatology boundary value inspection
Processing, the processing of extreme value range check, internal consistency inspection processing, time consistency inspection processing；The data normalization inspection
Reason is investigated and prosecuted to refer to and deletes the repeated data occurred in historical data, and to the corresponding filename of historical data, the date, when
Between, the site information of automatic weather station improve and uniform format；The climatology boundary value inspection processing refers to according to gas
The local historical climate feature as locating for industry relevant criterion and monitoring region, picks the historical data that can not occur
It removes；The extreme value range check processing refers to that choosing the history monthly average wind speed of each automatic weather station acquisition is basic unit,
The standard deviation for calculating the annual monthly average wind speed, using this month history maximum wind velocity plus twice of standard deviation as maximum, with this
The moon, history minimum windspeed cut twice of standard deviation as minimum, and rejecting is more than historical data of the maximum value less than minimum；Institute
It states internal consistency inspection processing to specifically refer to carry out inspection processing to the physical link each wind element, i.e., most according to ten minutes
When big wind speed >=two minute maximum wind velocity, two minutes maximum wind velocity >=instantaneous wind speeds, wind direction are 0 ° then wind speed≤0.2m/s these
Judgment criteria, rejecting abnormalities data；Time consistency inspection processing specifically refer to a certain current time and this moment it
The change rate of five minutes preceding corresponding historical datas is checked, the corresponding history number of upper limit threshold is greater than to change rate
It is carried out according to being rejected, while to the change rate at a certain current time and one hour corresponding historical data before this moment
Check, the corresponding historical data for being less than lower threshold to change rate is rejected, the upper limit threshold and lower threshold according to
Local historical climate feature locating for region is monitored to determine.
Time synchronization is carried out to valid data according to the data moment of setting to specifically refer to valid data with per half a minute
The data list of valid data is established for the data moment, i.e., by each entry in valid data collect its original moment near
In that close moment half a minute corresponding data list.
S3 chooses the valid data in the time range that front and back is set at the time of actually occurring wind shear in data list,
The shear at all automatic weather stations being calculated in the monitoring region corresponding each wind element each data moment between standing two-by-two
Value；
Specifically, the valid data before and after choosing at the time of actually occurring wind shear in data list in each 12 hours,
According in valid data instantaneous wind speed and instantaneous wind direction obtain instantaneous vector wind, sweared according to gustiness and fitful wind wind direction
Fitful wind is measured, obtains two minutes average vector winds according to two minutes mean wind speeds and two minutes mean wind directions, it is average according to ten minutes
Wind speed and ten minutes mean wind directions obtain ten minutes average vector winds, calculate all automatic weather stations in the monitoring region two-by-two
The phasor difference of instantaneous vector wind between standing, the phasor difference of vector fitful wind, the phasor difference of two minutes average vector winds, ten minutes it is average
The phasor difference of vector wind, so obtain all automatic weather stations stand two-by-two between each data moment shear value.
S4, by it is described monitoring region in automatic weather station stand two-by-two between apart from bi-directional scaling standard value, will scale
Standard value afterwards obtains whether each data moment occurs wind shear compared with wind shear discrimination threshold, and provides each
Whether the data moment occurs the label of wind shear；
Specifically, according to the international judgment criteria of wind shear, i.e. when spacing between two automatic weather stations is 4km, two
When wind shear value between a automatic weather station is greater than 15 section, then it is judged to that wind shear has occurred, it will be under international judgment criteria
Floating 10% to 20%, in this, as wind shear discrimination threshold.
By it is described monitoring region in automatic weather station stand two-by-two between apart from bi-directional scaling standard value, after scaling
Standard value is compared with wind shear discrimination threshold, if the instantaneous wind shear value at a certain data moment is greater than wind shear discrimination threshold,
Or battle array wind shear value is greater than wind shear discrimination threshold, two minutes average wind shear values are greater than wind shear discrimination threshold and instantaneous
Wind shear value and wind shear discrimination threshold difference are within 20%, then this data moment has occurred wind shear, on the contrary then do not send out
Raw wind shear；And whether wind shear is occurred to each data moment and provides corresponding label, that is, wind shear has occurred and provides "Yes"
Label, the label that wind shear provides "No" does not occur.
S5 is constituted as a result using each wind element of each time data as input, the label for whether occurring wind shear
Data set, and SMOTE algorithm EDS extended data set is used to make two kinds of tag types quantity balance that wind shear whether occurs, it obtains
Data set after balance；
Specifically, two kinds of tag types quantity balance that wind shear whether occurs specifically refers to the number that label is "Yes"
According to label be "No" data bulk difference less than 5%.
S6 is trained the data set after balance using machine learning method, chooses the best training result of recall ratio,
And corresponding model is exported, using this model as identification model；
Specifically, three kinds of support vector machines, logistic regression and decision tree models is respectively adopted using machine learning method,
Data set after input balance is trained, and folds into row cross validation using 10 in training process, and it is best to choose recall ratio
Training result, and corresponding model is exported, using this model as identification model.
S7 reads the latest data of each automatic weather station in monitoring region, using the method for step S2 to newest number in real time
According to being pre-processed, and will treated that data input the identification model, and then obtain recognition result.
Method of the invention is illustrated below with reference to embodiment.
8 automatic meteorological station equipments are now deployed with around Hefei City new bridge International airport runway, first to be located at runway north and south
Illustrate the specific embodiment of the method for the present invention for the automatic weather station A and automatic weather station B at both ends.
It is about 2.7km at a distance between two station automatic weather station A and automatic weather station B, acquires automatic weather station A and automatic gas
As the historical data that station two station B is built a station so far certainly, the method according to the invention first pre-processes historical data, rejects not
Trust data, and historical data being synchronized in time, by data available naturalization to synchronization to utilize.
Concern report actually occurs 12 hours before and after wind shear moment periods, extracts the wind element letter wherein paid close attention to
Breath, including instantaneous wind speed, wind direction, gustiness, wind direction, two minutes mean wind speeds, wind directions, ten minutes wind speed, wind directions.With the world
The universal standard (4km range inscribe becomes larger in 15 sections (about 7.7m/s)) is zoomed in and out by distance, and floats downward 10%, i.e. threshold value W=
7.7*(2.7/4)*0.9≈4.6m/s.The shear value of each wind element at two stations is calculated, and using above-mentioned threshold value W as automatic meteorological
Stand A and two station automatic weather station B each wind element shear discrimination standard.
During differentiation, in conjunction with forecaster's experience, instantaneous wind shear value, fitful wind are successively considered from big to small by weight
Shear value, two minutes average wind shear values, the shear value of ten minutes average winds, instantaneous wind shear value sentences greater than wind shear
Other threshold value or battle array wind shear value are greater than wind shear discrimination threshold, two minutes average wind shear values are greater than wind shear and differentiate threshold
Value and instantaneously wind shear value and wind shear discrimination threshold difference, then labeled as wind shear occurs, otherwise mark within 20%
For wind shear does not occur.Front and back 12 hours automatic weather station A and automatic gas at the time of actually occurring wind shear to report
As the wind shear label of each wind element in station two station B and label constitutes data set.
Using SMOTE algorithm EDS extended data set, it will expand to and be labeled as labeled as the data entry that wind shear does not occur
The data entry of wind shear occurs unanimously or relatively (the two quantity difference controls within 5%).
The data set expanded is inputted into support vector machines, three kinds of models of logistic regression and decision tree are trained, use
10 folding cross validations take recall ratio as preferential, output optimal result model, using this model as identification model.
The real time data for acquiring automatic weather station A and two station automatic weather station B, carries out the pre- of method same as described above to it
After processing, trained identification model is inputted, and then obtain recognition result.
In conclusion the present invention provides a kind of low-level wind shear recognition methods based on automatic weather station, has and knows
The advantage that other precision is high, local adaptability is good, can effectively identify runway horizontal wind shear near the ground, improve gas
As the guarantee efficiency of service.
Claims (9)
1. a kind of low-level wind shear recognition methods based on automatic weather station, which comprises the steps of:
S1 obtains the historical data of all collected each wind elements of automatic weather station in monitoring region；
S2 carries out quality control to historical data with rejecting abnormalities data, obtains the valid data of each wind element, and according to setting
The data moment to valid data carry out time synchronization, obtain the corresponding data list of valid data；
S3 chooses the valid data in the time range that front and back is set at the time of actually occurring wind shear in data list, calculates
Obtain all automatic weather stations in the monitoring region stand two-by-two between corresponding each wind element each data moment shear value；
S4, by it is described monitoring region in automatic weather station stand two-by-two between apart from bi-directional scaling standard value, after scaling
Standard value obtains whether each data moment occurs wind shear compared with wind shear discrimination threshold, and provides each data
Whether the moment occurs the label of wind shear；
Each wind element of each time data is constituted data as a result by S5
Collection, and SMOTE algorithm EDS extended data set is used to make two kinds of tag types quantity balance that wind shear whether occurs, it is balanced
Data set afterwards；
S6 is trained the data set after balance using machine learning method, chooses the best training result of recall ratio, and defeated
Corresponding model out, using this model as identification model；
S7, in real time read monitoring region each automatic weather station latest data, using step S2 method to latest data into
Row pretreatment, and by treated, data input the identification model, and then obtain recognition result.
2. a kind of low-level wind shear recognition methods based on automatic weather station according to claim 1, it is characterised in that: step
In rapid S1, the wind element includes instantaneous wind speed, instantaneous wind direction, gustiness, fitful wind wind direction, two minutes mean wind speeds, two points
Clock mean wind direction, two minutes maximum wind velocities, two minutes maximum wind directions, ten minutes mean wind speeds, ten minutes mean wind directions, ten minutes
Maximum wind velocity, ten minutes maximum wind directions.
3. a kind of low-level wind shear recognition methods based on automatic weather station according to claim 2, it is characterised in that；Step
In rapid S2, carrying out quality control to historical data includes data normalization inspection processing, climatology boundary value inspection processing, extreme value
Range check processing, internal consistency inspection processing, time consistency inspection processing；The data normalization inspection processing refers to
The repeated data occurred in historical data is deleted, and to the corresponding filename of historical data, date, time, automatic gas
As the site information at station carries out perfect and uniform format；The climatology boundary value inspection processing refers to according to Meteorological Field correlation
Local historical climate feature locating for standard and monitoring region, rejects the historical data that can not occur；The extreme value
Range check processing refers to that choosing the history monthly average wind speed of each automatic weather station acquisition is basic unit, calculates the annual moon
The standard deviation of mean wind speed, using this month history maximum wind velocity plus twice of standard deviation as maximum, with this month history minimum wind
Speed cuts twice of standard deviation as minimum, and rejecting is more than historical data of the maximum value less than minimum；The internal consistency
Inspection processing specifically refers to carry out inspection processing to the physical link each wind element, i.e., according to ten minutes maximum wind velocity >=two point
Then these judgment criterias of wind speed≤0.2m/s when clock maximum wind velocity, two minutes maximum wind velocity >=instantaneous wind speeds, wind direction are 0 °, are rejected
Abnormal data；The time consistency inspection processing was specifically referred to a certain current time and five minutes phases before this moment
The change rate for the historical data answered is checked that the corresponding historical data for being greater than upper limit threshold to change rate is rejected, together
When a certain current time is checked with the change rate of one hour corresponding historical data before this moment, to change rate
Corresponding historical data less than lower threshold is rejected, and the upper limit threshold and lower threshold are according to locating for monitoring region
Local historical climate feature determines.
4. a kind of low-level wind shear recognition methods based on automatic weather station according to claim 3, it is characterised in that: step
In rapid S2, according to the data moment set to valid data carry out time synchronization specifically refer to by valid data with per half a minute as
The data moment establishes the data list of valid data, i.e., by each entry in valid data collect its original moment near
That moment half a minute corresponding data list in.
5. a kind of low-level wind shear recognition methods based on automatic weather station according to claim 4, it is characterised in that: step
In rapid S3, the valid data in each 12 hours of front and back at the time of actually occurring wind shear in data list are chosen, according to effective
Instantaneous wind speed and instantaneous wind direction in data obtain instantaneous vector wind, obtain vector fitful wind according to gustiness and fitful wind wind direction,
Two minutes average vector winds are obtained according to two minutes mean wind speeds and two minutes mean wind directions, according to ten minutes mean wind speeds and ten
Minute mean wind direction obtains ten minutes average vector winds, calculate all automatic weather stations in the monitoring region stand two-by-two between wink
When the phasor difference of vector wind, the phasor difference of vector fitful wind, the phasor difference of two minutes average vector winds, ten minutes average vector winds
Phasor difference, so obtain all automatic weather stations stand two-by-two between each data moment shear value.
6. a kind of low-level wind shear recognition methods based on automatic weather station according to claim 5, it is characterised in that: step
In rapid S4, according to the international judgment criteria of wind shear, i.e. when spacing between two automatic weather stations is 4km, two automatic gas
As station between wind shear value be greater than 15 section when, then be judged to that wind shear has occurred, by international judgment criteria float downward 10% to
20%, in this, as wind shear discrimination threshold.
7. a kind of low-level wind shear recognition methods based on automatic weather station according to claim 6, it is characterised in that: step
In rapid S4, between the automatic weather station in the monitoring region is stood two-by-two apart from bi-directional scaling standard value, after scaling
Standard value is compared with wind shear discrimination threshold, if the instantaneous wind shear value at a certain data moment is greater than wind shear discrimination threshold,
Or battle array wind shear value is greater than wind shear discrimination threshold, two minutes average wind shear values are greater than wind shear discrimination threshold and instantaneous
Wind shear value and wind shear discrimination threshold difference are within 20%, then this data moment has occurred wind shear, on the contrary then do not send out
Raw wind shear；And whether wind shear is occurred to each data moment and provides corresponding label, that is, wind shear has occurred and provides "Yes"
Label, the label that wind shear provides "No" does not occur.
8. a kind of low-level wind shear recognition methods based on automatic weather station according to claim 7, it is characterised in that: step
In rapid S5, two kinds of tag types quantity balance that wind shear whether occurs specifically refers to the data and label that label is "Yes"
For "No" data bulk difference less than 5%.
9. a kind of low-level wind shear recognition methods based on automatic weather station according to claim 8, it is characterised in that: step
Rapid S6 specifically refers to that three kinds of support vector machines, logistic regression and decision tree models is respectively adopted using machine learning method, inputs
Data set after balance is trained, and folds into row cross validation using 10 in training process, chooses the best training of recall ratio
As a result, and export corresponding model, using this model as identification model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811285720.1A CN109583593B (en) | 2018-10-31 | 2018-10-31 | Low-altitude wind shear identification method based on automatic meteorological station |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811285720.1A CN109583593B (en) | 2018-10-31 | 2018-10-31 | Low-altitude wind shear identification method based on automatic meteorological station |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109583593A true CN109583593A (en) | 2019-04-05 |
| CN109583593B CN109583593B (en) | 2020-11-13 |
Family
ID=65920911
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811285720.1A Active CN109583593B (en) | 2018-10-31 | 2018-10-31 | Low-altitude wind shear identification method based on automatic meteorological station |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109583593B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113962462A (en) * | 2021-10-25 | 2022-01-21 | 中国科学院空天信息创新研究院 | Wind field stability prediction method and system based on convolutional neural network |
| CN115169236A (en) * | 2022-07-22 | 2022-10-11 | 黄河水利委员会水文局 | Automatic identification method for cut line based on combination of data mining and weather analysis |
| CN115830380A (en) * | 2022-12-06 | 2023-03-21 | 中科三清科技有限公司 | Wind shear line identification method and device based on artificial intelligence, storage medium and terminal |
| CN115861811A (en) * | 2022-12-06 | 2023-03-28 | 中科三清科技有限公司 | Wind shear region identification method and device, storage medium and terminal |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105607063A (en) * | 2016-01-05 | 2016-05-25 | 北京无线电测量研究所 | Detection method and system for low-altitude wind shear at airport |
| CN106772387A (en) * | 2016-12-21 | 2017-05-31 | 中国航空工业集团公司雷华电子技术研究所 | A kind of wind shear recognition methods |
-
2018
- 2018-10-31 CN CN201811285720.1A patent/CN109583593B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105607063A (en) * | 2016-01-05 | 2016-05-25 | 北京无线电测量研究所 | Detection method and system for low-altitude wind shear at airport |
| CN106772387A (en) * | 2016-12-21 | 2017-05-31 | 中国航空工业集团公司雷华电子技术研究所 | A kind of wind shear recognition methods |
Non-Patent Citations (4)
| Title |
|---|
| JAMES N. K. LIU ET AL.: "Chaotic Oscillatory-Based Neural Network for Wind Shear and Turbulence Forecast With LiDAR Data", 《IEEE》 * |
| 蒋立辉: "基于仿真雷达图像的低空风切变类型识别研究", 《激光与红外》 * |
| 迟继峰 等: "复杂地形多测风塔综合地貌及风切变拟合修正的风资源评估方法研究", 《华电技术》 * |
| 郭忠立: "双流机场风切变预警系统方案设计及LLWAS算法研究", 《气象水文海洋仪器》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113962462A (en) * | 2021-10-25 | 2022-01-21 | 中国科学院空天信息创新研究院 | Wind field stability prediction method and system based on convolutional neural network |
| CN115169236A (en) * | 2022-07-22 | 2022-10-11 | 黄河水利委员会水文局 | Automatic identification method for cut line based on combination of data mining and weather analysis |
| CN115830380A (en) * | 2022-12-06 | 2023-03-21 | 中科三清科技有限公司 | Wind shear line identification method and device based on artificial intelligence, storage medium and terminal |
| CN115861811A (en) * | 2022-12-06 | 2023-03-28 | 中科三清科技有限公司 | Wind shear region identification method and device, storage medium and terminal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109583593B (en) | 2020-11-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109583593A (en) | A kind of low-level wind shear recognition methods based on automatic weather station | |
| Rasmussen et al. | Verification of the origins of rotation in tornadoes experiment: VORTEX | |
| Lang et al. | The severe thunderstorm electrification and precipitation study | |
| CN113189615B (en) | Method for inspecting power transmission line by using vertical take-off and landing fixed wing unmanned aerial vehicle | |
| CN103779808A (en) | Power transmission line intelligent inspection system based on LiDAR | |
| CN203084193U (en) | A measuring system used for measuring the distance between a wire and ground | |
| CN107818382A (en) | Flight arrival time Forecasting Methodology | |
| Gál et al. | Urban heat island patterns and their dynamics based on an urban climate measurement network | |
| CN110196962A (en) | A kind of air speed abnormality recognition method based on Density Estimator | |
| CN104200082A (en) | Typhoon landing prediction method | |
| CN109407177A (en) | Dense fog identifying system and methods for using them based on machine learning and conventional meteorological observation | |
| Kato | A statistical method for separating urban effect trends from observed temperature data and its application to Japanese temperature records | |
| Eilts et al. | Oklahoma downbursts and their asymmetry | |
| Hagelin et al. | Nowcasting with the AROME model: First results from the high-resolution AROME airport | |
| CN205750554U (en) | A kind of electromagnetic field positioner of overhead transmission line line walking unmanned plane | |
| Gurke et al. | The development of the wake vortices warning system for Frankfurt airport: Theory and implementation | |
| Bartok et al. | Data mining for fog prediction and low clouds detection | |
| Rubnich et al. | An algorithm to identify robust convective weather avoidance polygons in en route airspace | |
| CN112097743A (en) | Building earthquake-resistant characteristic analysis method based on unmanned aerial vehicle technology | |
| CN103424782B (en) | The automatic identifying method of the radial convergence in a kind of middle level | |
| Drüe et al. | Comparison of boundary-layer profiles and layer detection by AMDAR and WTR/RASS at Frankfurt airport | |
| CN105488582A (en) | Mountain talus landslide prediction method and device | |
| CN111967653B (en) | Method for constructing airport runway wind forecast model and forecast method and system | |
| Zhou et al. | The wind and temperature information of AMDAR data applying to the analysis of severe weather nowcasting of airport | |
| Mandel | An early look at the development of an unmanned automated surface aviation weather observation system |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |