Method for early warning brandish of transmission wire based on improved Bayes-Adaboost algorithm

Patent metadata

Field	Value
Publication number	US-12536455-B2
Application number	US-202217747036-A
Country	US
Kind code	B2
Filing date	May 18, 2022
Priority date	May 18, 2021
Publication date	Jan 27, 2026
Grant date	Jan 27, 2026

Abstract

The present application discloses a method for early warning brandish of a transmission wire based on an improved Bayes-Adaboost algorithm, including: forming a classifier by training a historical brandish fault training set, and by using an Adaboost ensemble learning method, and obtaining an early warning result of the brandish of the transmission wire via the classifier according to real-time forecast meteorological information and information of different parameters of the transmission wire. The present invention can realize calculation and processing of forecast information of meteorological characteristic factors of the brandish of the transmission wire, structural parameters of the transmission wire and other related data, and obtain an early warning analysis result of a brandish disaster of the transmission wire in a region.

First claim

What is claimed is:

1 . A method for early warning of brandish of a transmission wire based on an Bayes-Adaboost algorithm, comprising:
Step 1: classifying and combining transmission wires according to internal factors that affect brandish excitation of the transmission wire, that is, parameters of the wire, to form a plurality of wire combinations; Step 2: statistically classifying historical brandish meteorological characteristic factors of the transmission wire that affect brandish excitation external factors of the transmission wire, to obtain meteorological characteristic factors of the brandish of the transmission wire; Step 3: obtaining historical brandish meteorological characteristic factor data sets of the plurality of wire combinations; Step 4: obtaining the historical brandish meteorological characteristic factor data sets of the transmission wire according to wire parameters of a prediction wire; Step 5: forming a strong classification learner with an Adaboost ensemble learning algorithm by using the data sets of Step 4; Step 6: obtaining a real-time meteorological forecast data x of the brandish meteorological characteristic factors of the transmission wire; Step 7: according to the real-time meteorological forecast data x in Step 6, obtaining an output of a brandish early warning model of the transmission wire via the strong classification learner in Step 5, comprising a prediction result y of the strong classification learner and a confidence margin(x, y); Step 8: according to an output result of the early warning model in Step 7, determining that an early warning level of a transmission wire brandish risk is obtained; Step 9: for a newly added brandish fault sample, correcting a brandish early warning model pair according to change rates of temperature and humidity, optimizing the output of the brandish early warning model according to Bayes, and re-predicting a brandish probability, wherein:
the brandish early warning model is corrected according to the change rates of the temperature and the humidity, and the brandish probability is re-predicted according to the Bayes on the output of the brandish early warning model, specifically:
at a current temperature, a probability of whether the transmission wire brandishes is a conditional probability of the brandish under a current meteorological-related characteristic parameters, and a formula is as follows:
P ( s|X )=margin( x,y )= C·ΠP ( x i ·ROTC i ·HOTC i |s )= C ·Π( P ( x i |s ))· P (ROTC i |s )· P (HOTC i |s )  (15)

s is a state of the transmission wire, that is, whether the transmission wire brandishes;
X is characteristic parameters related to a current meteorological condition;
P(ROTC i |s) is a change rate of a temperature of a brandish fault in the current meteorological condition;
P(HOTC i | s) is a change rate of a humidity of the brandish fault in the current meteorological condition;
C is a total probability of the brandish so far;
a weak classifier C t (X) is as follows:

C
t

(
X
)

=

P
⁡
(

x
i

|
s

)

=

Count
xi

Count
s

(
17
)

P(x i |s) is a weight of a single feature parameter x i ;
Count si is the number of the brandishes that occur under the condition of an i-th characteristic parameter, and Count s is the number of all brandishes;
for the newly added brandish fault sample, after each brandish event occurs, the weight of the single feature parameter is dynamically updated, and an update formula (17) is as follows:

C
~

t

(
X
)

=

P
⁡
(

x
inew

|
s

)

=

Count
xi

+
i

Count
s

+
i

(
18
)

finally, a final brandish probability is obtained by substituting formula (18) into formula (15) to replace a result of a single weak classifier C t (X);

Step 10: obtaining an early warning probability and an early warning level of brandish of a final transmission wire according to an aerial meteorological condition and a predicted probability of Step 9.

2 . The method for early warning of the brandish of the transmission wire based on the Bayes-Adaboost algorithm according to claim 1 , wherein:
in Step 1, three wire parameters are selected, namely a wire structure, a wire section area, and a span; the wire structure is divided into a single wire and a split wire; the wire area is divided into three tap positions, an area not exceeding a first set area is a first wire area tap position, an area greater than the first wire set area and not exceeding a second set area is a second wire area tap position, and an area greater than the second set area is a third wire tap position; the span is divided into three tap positions, a span not exceeding a first set span is a first span tap position, a span greater than a first wire set span and not exceeding a second set span is a second wire tap position, and a span greater than the second set span is a third wire tap position.

3 . The method for early warning of the brandish of the transmission wire based on

Assignees

• State Grid Henan Electric Power Company Electric Power Science Res Institute
• State Grid Henan Electric Power Co
• State Grid Corp China

Inventors

• Yao Degui
• Zhou Ning
• Guo Zhimin
• Wang Lei
• Lu Ming
• Li Zhe
• Zheng Wei
• Liu Shanfeng
• Li Xiaolei
• Wang Jinyu
• Liang Yun
• Wang Chao
• Zhang Bo
• Zhang Yupeng
• Yuan Shaoguang
• Tian Yangyang
• Mao Wandeng
• Zhang Xiaofei

Classifications

• G01W2203/00

  Real-time site-specific personalized weather information, e.g. nowcasting · CPC title

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G01W1/10 Primary

  Devices for predicting weather conditions (computers per se G06; display devices G09) · CPC title

• G06N7/01 Primary

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title