CN-121997065-A - Breaker service life prediction system and method based on Internet of things

CN121997065ACN 121997065 ACN121997065 ACN 121997065ACN-121997065-A

Abstract

The invention discloses a breaker service life prediction system and method based on the Internet of things, and relates to the technical field of breakers, wherein the system comprises an Internet of things data acquisition module, a scene grouping module, a data quantization module, a key factor identification module, a service life prediction module and an early warning priority output module; the system comprises an Internet of things data acquisition module, a scene grouping module, a data quantization module, a key factor identification module, a life prediction module and an early warning priority output module, wherein the Internet of things data acquisition module is used for acquiring environment data and working condition data, the scene grouping module is used for grouping scenes of the circuit breakers from two dimensions of environment conditions and working condition loads, the data quantization module is used for analyzing comprehensive life loss of each circuit breaker, the key factor identification module is used for outputting key factors of a corresponding scene group, the life prediction module is used for constructing a scene life prediction model, the key factor data acquired in real time by the Internet of things are substituted into the prediction model to calculate residual life, and the early warning priority output module is used for carrying out early warning response when the model output result is smaller than the residual life early warning time limit and carrying out priority ranking on all label factors contained in the same scene group.

Inventors

LI XIAOBIN
GUI SHENGQING
ZHAN CHUANXIANG
Xin Haomiao
GUAN HUASHEN
LIU YILONG
QI HAN
SUN QIANG
LV JIAWEI
NI HUIHAO

Assignees

广东电网有限责任公司江门供电局

Dates

Publication Date: 20260508
Application Date: 20251231

Claims (8)

1. The method for predicting the service life of the circuit breaker based on the Internet of things is characterized by comprising the following steps of: Step 100, acquiring environment data and working condition data by using an Internet of things sensor deployed on the circuit breakers, and grouping the circuit breakers in a scene from two dimensions of the environment condition and the working condition load; Step 200, analyzing the comprehensive life loss of each circuit breaker by taking the switching-on and switching-off katon frequency as a representation and the abnormal frequency of the temperature of the head as a representation of the mechanical life of each circuit breaker in the same scene group; Step 300, carrying out standardization processing on the label factors corresponding to the environment data and the working condition data, analyzing the correlation coefficient of each label factor and the comprehensive life loss in the same scene group, and outputting the key factors of the corresponding scene group; Step 400, constructing a scene life prediction model by taking key factors as input variables and comprehensive life loss as output variables, and substituting key factor data acquired in real time by the Internet of things into the prediction model to calculate the residual life; And S500, presetting a residual life early warning time limit, performing early warning response when the output result of the model is smaller than the residual life early warning time limit, re-checking the residual life when the operation working condition of the circuit breaker is suddenly changed, and performing priority ranking on all label factors contained in the same scene group based on the mathematical relationship between the residual life after the sudden change and the residual life before the sudden change.
2. The method for predicting the service life of a circuit breaker based on the Internet of things as set forth in claim 1, wherein the step S100 comprises the following specific processes: the environmental data comprise salt fog concentration, average daily temperature, dust concentration and relative humidity; The working condition data comprise monthly operation times, a single opening and closing current peak value, harmonic content and continuous operation time length; the sensor of the Internet of things sets acquisition frequency, acquires corresponding data by taking environment data and working condition data as acquisition targets in an acquisition period, and performs assignment of a recorded scene label on each breaker; The method comprises the steps of carrying out unified quantization on each label of the circuit breaker completing scene label assignment, wherein the unified quantity refers to that assignment of each scene label record is converted into quantization scores within 0-10 according to corresponding quantization rules, calculating total label scores Q, Q-sigma Q of each circuit breaker, Q represents the quantization scores of each scene label of the circuit breaker record, and calculating the label similarity of any two circuit breakers by utilizing a formula S= [ 1-I Q A -Q B /80 ] ×100%, wherein Q A represents the total label score of the circuit breaker A, and Q B represents the total label score of the circuit breaker B.
3. The method for predicting service life of circuit breakers based on Internet of things according to claim 1, wherein analyzing the comprehensive service life loss of each circuit breaker comprises the following steps: Obtaining mechanical life loss L Machinery ,L Machinery = switching-off and switching-on stuck frequency/4 recorded by each circuit breaker, and electric life loss L Electrical appliance ,L Electrical appliance = abnormal temperature frequency/2 recorded by the circuit breaker; And calculating the comprehensive service life loss L Total (S) of each circuit breaker by using a formula L Total (S) =0.4*L Machinery +0.6*L Electrical appliance .
4. The method for predicting the service life of a circuit breaker based on the Internet of things as set forth in claim 1, wherein the step S300 comprises the following specific steps: Step S310, taking a daily average value of salt fog concentration, temperature and dust concentration in environmental factors, taking a month sum of month average operation times in working condition factors, taking a single average value of current peaks, taking a total harmonic distortion rate of harmonic content, and normalizing data of all factors to a range of 0-1; Step S320, extracting sample data of all the circuit breakers in the same scene group in a historical acquisition period, wherein each group of sample data comprises a comprehensive life loss and eight factor values, calculating a correlation coefficient r of each label factor and the corresponding comprehensive life loss, traversing all label factors to generate the corresponding correlation coefficient, and screening label factors of |r| > correlation coefficient threshold r0 as key factors.
5. The method for predicting the service life of a circuit breaker based on the Internet of things as set forth in claim 1, wherein the step S400 comprises the following specific steps: in step S410, the step of constructing the scene life prediction model is to construct an LSTM scene life prediction model fusing the attention mechanism, and the model formula is as follows: L Total (S) =σ(W 1 ·Attention(X 1 ,X 2 ,…,Xn)+W 2 ·LSTM(X 1 ,X 2 ,…,Xn)+b); Wherein sigma is Sigmoid activation function to limit output value in 0-1 interval, attention ) LSTM for the attention mechanism layer ) For the long-term and short-term memory network layer, W 1 、W 2 is a weight matrix, and b is a bias term, wherein X 1 ,X 2 , xn represents the key factors after the 1 st, 2 nd, n standardized in the same scene group; Attention(X)=Σ 1 n α i X i ,α i =exp(s(X i ))/Σ 1 n exp(s(X j )), wherein s (X i ) is a scoring function, s (X j ) represents scoring other key factors than the key factor corresponding to X i ; Step S420, collecting real-time key factor data for standardization, substituting the real-time key factor data into the attention mechanism layer for output, obtaining the supplementary loss captured by the long-short-period memory network layer, calculating to obtain the predicted comprehensive life loss L Total (S) , and the residual life T Residual of =(1-L Total (S) )/a, wherein a represents the annual loss rate of historical calculation.
6. The method for predicting the service life of a circuit breaker based on the Internet of things as set forth in claim 1, wherein the step S500 comprises the following steps: the judgment basis for the abrupt change of the operating condition of the circuit breaker is as follows: Collecting a tag factor value Xt monitored by an Internet of things sensor in real time and a past 12-hour average value X 0,t-12 , and calculating a mutation rate P, P= | (X t -X 0,t-12 )/X max |, wherein X max represents a rated maximum value corresponding to a tag factor) by using a formula; setting a mutation rate threshold P 0 , outputting a corresponding label factor to generate mutation when P > P 0 , returning to the scene life prediction model to increase the weight of the corresponding mutation label factor, adjusting the time step parameter of the LSTM layer, substituting the label factor value to predict the comprehensive life loss again and calculating the residual life T Projection(s) .
7. The method for predicting the service life of a circuit breaker based on the Internet of things as set forth in claim 1, wherein the step S500 further comprises the following steps: Traversing all operation condition mutation events, taking each label factor as a classification target, generating mutation event sets corresponding to the same classification target, extracting predicted residual life T Residual of before mutation corresponding to the mutation event sets and residual life T Projection(s) calculated after mutation, and calculating a difference index Z, Z=T Residual of -T Projection(s) ; Calculating mutation influence degree V of each difference index set by using a formula of V=D 1 /D 0 , wherein D 1 represents the number of mutation events with the difference index larger than a difference index threshold value, and D 0 represents the total number of events in the mutation event set; and sequencing the tag factors from large to small according to the numerical value of the mutation influence degree to generate a priority sequence.
8. The breaker service life prediction system based on the Internet of things, which uses the breaker service life prediction method based on the Internet of things as claimed in any one of claims 1 to 6, is characterized by comprising an Internet of things data acquisition module, a scene grouping module, a data quantization module, a key factor identification module, a service life prediction module and an early warning priority output module; The data acquisition module of the Internet of things is used for acquiring environmental data and working condition data; The scene grouping module is used for grouping scenes of the circuit breaker from two dimensions of environmental conditions and working condition loads; The data quantization module is used for analyzing the comprehensive life loss of each circuit breaker by taking the switching-on and switching-off clamp frequency as a representation and the abnormal frequency of the temperature of the head as a representation of the mechanical life of each circuit breaker in the same scene group; The key factor identification module is used for analyzing the correlation coefficient of each tag factor and the comprehensive life loss in the same scene group and outputting the key factor of the corresponding scene group; The life prediction module is used for constructing a scene life prediction model, substituting key factor data acquired in real time by the Internet of things into the prediction model to calculate the residual life; the early warning priority output module is used for presetting a residual life early warning time limit, carrying out early warning response when the model output result is smaller than the residual life early warning time limit, rechecking the residual life when the operation working condition of the circuit breaker is suddenly changed, and carrying out priority ranking on all label factors contained in the same scene group based on the mathematical relationship between the residual life after the sudden change and the residual life before the sudden change.

Description

Breaker service life prediction system and method based on Internet of things Technical Field The invention relates to the technical field, in particular to a breaker service life prediction system and method based on the Internet of things. Background Along with the development of an electric power system to an intelligent and distributed direction, a circuit breaker is used as core protection equipment for guaranteeing the safety of electric power supply, and the accuracy of service life prediction of the circuit breaker is directly related to the operation and maintenance efficiency and the power supply reliability of a power grid. Under different application scenes, the environmental conditions and working condition loads of the circuit breaker are obviously different, so that the life loss mechanism of the circuit breaker presents obvious scene specificity, for example, the circuit breaker in coastal areas is easy to accelerate insulation aging due to salt spray corrosion, the circuit breaker in a metallurgical plant aggravates contact ablation due to impact current, and the circuit breaker in a data center causes mechanical structure fatigue due to high-frequency switching-on and switching-off. The service life of a circuit breaker as a core protection device of an electric power system is directly related to the stability and safety of electric power supply. Under different application scenes, the environmental conditions and working condition loads of the circuit breaker are obviously different, so that life loss mechanisms are different. The existing breaker service life prediction method based on the Internet of things depends on a general data model, and has two major core problems that firstly, analysis logic is not designed aiming at scene differentiation characteristics, special influence factors of scenes are ignored, so that prediction deviation is large, secondly, the depth processing of data collected by the Internet of things is lacked, key factor locking efficiency is low, core loss factors in different scenes cannot be rapidly positioned, and further service life prediction accuracy is influenced. Thirdly, the early warning mechanism lacks priority management, and when a plurality of circuit breakers in the same scene group trigger early warning simultaneously, factors and equipment with larger influence on service life cannot be accurately identified, so that operation and maintenance resource allocation is unreasonable, and intervention efficiency is influenced. The problems cause that the prior art is difficult to realize accurate prediction of the service life of the circuit breaker under different differentiation scenes, so that untimely or excessive maintenance is easy to cause, the operation and maintenance cost of a power grid is increased, and even equipment faults are caused to cause power interruption. Disclosure of Invention The invention aims to provide a breaker service life prediction system and method based on the Internet of things, so as to solve the problems in the prior art. In order to achieve the purpose, the invention provides the technical scheme that the method for predicting the service life of the circuit breaker based on the Internet of things comprises the following steps: Step 100, acquiring environment data and working condition data by using an Internet of things sensor deployed on the circuit breakers, and grouping the circuit breakers in a scene from two dimensions of the environment condition and the working condition load; Step 200, analyzing the comprehensive life loss of each circuit breaker by taking the switching-on and switching-off katon frequency as a representation and the abnormal frequency of the temperature of the head as a representation of the mechanical life of each circuit breaker in the same scene group; Step 300, carrying out standardization processing on the label factors corresponding to the environment data and the working condition data, analyzing the correlation coefficient of each label factor and the comprehensive life loss in the same scene group, and outputting the key factors of the corresponding scene group; Step 400, constructing a scene life prediction model by taking key factors as input variables and comprehensive life loss as output variables, and substituting key factor data acquired in real time by the Internet of things into the prediction model to calculate the residual life; And S500, presetting a residual life early warning time limit, performing early warning response when the output result of the model is smaller than the residual life early warning time limit, re-checking the residual life when the operation working condition of the circuit breaker is suddenly changed, and performing priority ranking on all label factors contained in the same scene group based on the mathematical relationship between the residual life after the sudden change and the residual life before the sudden change. The priority ordering