CN116911445A - Subway pantograph replacement time prediction method based on limited samples - Google Patents
Subway pantograph replacement time prediction method based on limited samples Download PDFInfo
- Publication number
- CN116911445A CN116911445A CN202310788272.1A CN202310788272A CN116911445A CN 116911445 A CN116911445 A CN 116911445A CN 202310788272 A CN202310788272 A CN 202310788272A CN 116911445 A CN116911445 A CN 116911445A
- Authority
- CN
- China
- Prior art keywords
- data
- pantograph
- replacement
- time
- sampling
- 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
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000005070 sampling Methods 0.000 claims abstract description 44
- 238000005299 abrasion Methods 0.000 claims abstract description 32
- 238000013528 artificial neural network Methods 0.000 claims abstract description 14
- 238000004140 cleaning Methods 0.000 claims abstract description 7
- 238000001514 detection method Methods 0.000 claims abstract description 6
- 238000012549 training Methods 0.000 claims description 8
- 125000004122 cyclic group Chemical group 0.000 claims description 5
- 230000035772 mutation Effects 0.000 claims description 5
- 230000000306 recurrent effect Effects 0.000 claims description 4
- 239000013598 vector Substances 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 4
- 239000000463 material Substances 0.000 abstract description 2
- 230000003137 locomotive effect Effects 0.000 description 6
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007306 turnover Effects 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
- G06F18/15—Statistical pre-processing, e.g. techniques for normalisation or restoring missing data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Strategic Management (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Biology (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- Probability & Statistics with Applications (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Current-Collector Devices For Electrically Propelled Vehicles (AREA)
Abstract
The invention provides a subway pantograph replacement time prediction method based on a limited sample, which comprises the following steps: and acquiring material abrasion thickness data of the pantograph by a non-contact detection method. The obtained data are subjected to cleaning, dividing and rewriting processing to obtain processed data, the processed data are subjected to searching, least square fitting and random adoption to obtain slopes of sampling points and fitting lines, the relative days between the sampling point time and the adjacent two replacement point times are calculated, the obtained sampling points, slopes and relative days are constructed into a large number of four-dimensional characteristic data to train the GRU circulating neural network, and finally the days which are the current distance from the date to be replaced are predicted through the trained GRU circulating neural network, so that the approximate time when the abrasion of the pantograph reaches the replacement threshold value can be predicted well under the condition that samples are limited, and the influence on a prediction result caused by limited replacement times contained in the detected data is reduced.
Description
Technical Field
The invention belongs to the technical field of data prediction, and particularly relates to a subway pantograph replacement time prediction method based on a limited sample.
Background
With the rapid development of cities in China, the urban rail transit industry rapidly develops. 21 urban rail transit operation lines are newly increased in 2022 all year round, the operation mileage is newly increased by 847 km, the actual running of trains 3316 is carried out ten thousand times all year round, the passenger traffic is 194.0 hundred million people times, the arrival amount is 116.9 hundred million people times, and the passenger traffic turnover amount is 1560 hundred million people kilometers. The rail transit brings great convenience to the travel of people, and meanwhile, the characteristics of large mileage, multiple operation times and large traffic flow rate are achieved, so that the abrasion condition of the metro vehicle is accelerated. The pantograph is electrical equipment for obtaining electric energy by contacting the contact net with the locomotive, and the pantograph rubs with the contact net at the moment when the metro vehicle runs to obtain the electric energy required by running, so that the abrasion condition of the pantograph needs to be fully mastered.
At present, a non-contact method can be used for periodically detecting the abrasion condition of the pantograph of the metro vehicle, the abrasion degree of the pantograph in the actual condition is in nonlinear change, meanwhile, the pantograph replacement period is longer, and the number of times of replacement contained in detected data is limited, so that the metro pantograph replacement time prediction method based on limited samples is provided.
Disclosure of Invention
The invention aims to provide a subway pantograph replacement time prediction method based on a limited sample, which can better predict the approximate time of the abrasion of a pantograph reaching a replacement threshold under the condition of limited sample;
in order to achieve the above purpose, the present invention adopts the following technical scheme: a subway pantograph replacement time prediction method based on a limited sample comprises the following steps:
step one: acquiring original data through a non-contact detection method, cleaning the original data, and dividing and rewriting the cleaned data;
step two: fitting the head and tail part data of the abrasion allowance of the pantograph in the processed data, and estimating the front and back replacement points [ t ] c-1 ,s c-1 ]And [ t ] c1 ,s c1 ]；
Step three: searching the processed data for the data of the new pantograph after the abrasion of the pantograph reaches the threshold value, and replacing the new pantograph after the abrasion of the pantograph reaches the threshold valueThe data of the pantograph is expressed as a set of vectors of time and pantograph thickness t c0 ,s c0 ]I.e. the specific replacement point;
step four: selecting adjacent replacement points according to the search result of the step three, and randomly sampling between the selected adjacent replacement points to obtain sampling points [ t ] p ,s p ]And the slope k of the fit line of the point to the front-to-back data f And k l ；
Step five: calculate the sampling point t p ,s p ]The relative days between the time and the time of two adjacent replacement points are respectively expressed as d f And d l ；
Step six: will d f 、k f 、s p K l Make up four features [ d f ,k f ,s p ,k l ]，d l Training GRU cyclic neural network as label to obtain d from data to be tested f 、k f 、s p K l Make up four features [ d f ,k f ,s p ,k l ]The trained GRU recurrent neural network is input to predict the number of days from the current date to be replaced.
Further, the step one of cleaning the obtained data specifically comprises the following steps:
and checking the obtained data one by one, and deleting the hollow white data, the messy code data and the data which do not accord with the actual physical meaning in the obtained data.
Further, in the step one, the specific steps of dividing and rewriting the cleaned data are as follows:
classifying and dividing the cleaned data, enabling the data to correspond to the corresponding subway vehicles one by one, and rewriting the classified and divided data into the abrasion condition of a certain pantograph in a period of time, wherein the abrasion condition is expressed as [ t ] i ,s i ]，i∈[t sart ,t end ]Wherein t is start And t end Indicating the start and end dates, respectively, of sampling the pantograph.
In the second step, the head and tail data of the abrasion allowance of the pantograph are fitted into straight by least squareWhen the numerical value of two ends of the straight line meets the replacement threshold value, the corresponding time coordinates of the two ends of the straight line are the front and back two replacement points [ t ] c-1 ,s c-1 ]And [ t ] c1 ,s c1 ]。
Further, calculating an average value of the total pantograph wear allowance in the processed data, sliding on the pantograph data with a step distance of one data at a time by using a window of three data, differencing two adjacent data in three continuous data in the window to obtain two differences, and if one difference is larger than a maximum value in a preset parameter, the other difference is smaller than a minimum value in the preset parameter and both the differences are larger than the average value of the total pantograph wear allowance, finding that a clear replacement point exists, wherein the clear replacement point is represented as [ t ] c0 ,s c0 ]Otherwise, the searching result is that no explicit replacement point exists.
Further, when the two difference values are both larger than the maximum value in the preset parameters, the second data is mutation data, and the mutation data is removed.
Further, the third specific step is as follows:
if the search result is that there is an explicit replacement point, then at [ t ] respectively c-1 ,s c-1 ]，[t c0 ,s c0 ]And [ t ] c0 ,s c0 ]，[t c1 ,s c1 ]Sampling randomly to obtain sampling point set, otherwise, at [ t ] c-1 ,s c-1 ]，[t c1 ,s c1 ]Randomly sampling to obtain a sampling point set, and selecting a sampling point [ t ] from the obtained sampling point set p ,s p ]And several data before and after it, sampling point [ t ] p ,s p ]Fitting the data before and after the data into a straight line through least square, and then calculating the slope k of the fitting line f And k l 。
The beneficial effects are that: the material abrasion thickness data of the pantograph is obtained through a non-contact detection method, the obtained data is subjected to cleaning, dividing and rewriting processing to obtain processed data, and the processed data is subjected to searching, least square fitting and random sampling to obtain a sampling point [ t ] p ,s p ]And slope k of the fit line f And k l Then calculate the sampling point t p ,s p ]Relative days d between time and time of two adjacent replacement points f And d l Next, d f 、k f 、s p K f Make up four features [ d f ,k f ,s p ,k l ]And d l As a label, limited data are constructed into a large number of four-dimensional characteristic data, and the GRU cyclic neural network is trained, so that the approximate time for the abrasion of the pantograph to reach the replacement threshold can be well predicted under the condition of limited samples, and the influence on a prediction result caused by limited replacement times contained in the detected data is reduced.
Drawings
FIG. 1 is a flow chart of an algorithm of the present invention;
FIG. 2 is a graph of pantograph data with explicit change points in an embodiment;
FIG. 3 is a graph of pantograph data without explicit change points in an embodiment;
FIG. 4 is a diagram showing the prediction results in the embodiment.
Detailed Description
The invention is further explained below with reference to the drawings.
As shown in fig. 1, the invention provides a subway pantograph replacement time prediction method based on a limited sample, which comprises the following steps:
step one: and acquiring original data by a non-contact detection method, cleaning the original data, and dividing and rewriting the cleaned data.
Step two: fitting the head and tail part data of the abrasion allowance of the pantograph in the processed data, and estimating the front and back replacement points [ t ] c-1 ,s c-1 ]And [ t ] c1 ,s c1 ]。
Step three: searching the processed data for the data of the new pantograph after the abrasion of the pantograph reaches the threshold value, wherein the data of the new pantograph after the abrasion of the pantograph reaches the threshold value is expressed as a set of vectors [ t ] of time and thickness of the pantograph c0 ,s c0 ]I.e. the specific replacement point.
Step four: according to the steps ofThirdly, selecting adjacent replacement points according to the search result, and randomly sampling between the selected adjacent replacement points to obtain sampling points [ t ] p ,s p ]And the slope k of the fit line of the point to the front-to-back data f And k l 。
Step five: calculate the sampling point t p ,s p ]The relative days between the time and the time of two adjacent replacement points are respectively expressed as d f And d l 。
Step six: will d f 、k f 、s p K l Make up four features [ d f ,k f ,s p ,k l ]，d l Training GRU cyclic neural network as label to obtain d from data to be tested f 、k f 、s p K l Make up four features [ d f ,k f ,s p ,k l ]The trained GRU recurrent neural network is input to predict the number of days from the current date to be replaced.
In step one, original data is obtained by a non-contact detection method, the original data includes thickness, time and car number of a pantograph collected at a certain time, data samples in the original data are shown in table 1, and data samples in the obtained data include sample 1, sample 2, sample 3 and sample 4. Wherein sample 1 is normal data and consists of three parts, namely date, locomotive pantograph number and pantograph abrasion allowance. Sample 2 is scrambled, blank data, and related data cannot be recorded correctly during the acquisition process. Sample 3 is data which does not conform to physical significance, the maximum margin of the newly replaced pantograph is 19.5 units thick, and the data collected in sample 3 are all higher than the data. Sample 4 is identifiable but recorded with erroneous data, and the number of the subway locomotive is missing, but the missing part can be deduced from the existing number, and the wear allowance recording error can be ignored. Table 1 is as follows:
| date of day | Locomotive pantograph number | Pantograph wear allowance | |
| Sample 1 | 2022-08-1722:43:57 | S7020A,S7020B,S7019B,S7019A | 12.7,12.7,13.8,13.6 |
| Sample 2 | 2021-11-2202:35:07 | 55118A | |
| Sample 3 | 2021-11-2123:35:25 | S7018A,S7018B,S7017B,S7017A | 29.4,25.4,27.4,26.4 |
| Sample 4 | 2021-11-2116:51:35 | S7001A,S7001B,S7002A | 9.5,10.0,10.2,6.999999 |
And after the data sample is cleaned, dividing and rewriting are carried out, so that the serial numbers of the pantographs of the locomotives are in one-to-one correspondence with the abrasion allowance of the pantographs. And not all pantographs are detected at the same time, so that a certain blank exists. As shown in table 2.
Table 2 is as follows:
| S7001A | S7001B | S7002A | … | S7022A | S7022B | |
| 2021-11-1122:47:34 | 15.7 | 13.2 | 15.2 | … | ||
| 2021-11-1122:34:57 | … | 8.6 | 9.4 | |||
| 2021-11-1222:46:50 | 12.5 | 12.7 | 14.5 | … |
in the second step, the data of the new pantograph is searched for after the abrasion of the pantograph reaches the threshold value in the processed data, namely, the specific replacement point is obtained, and the obtained search results are two cases, namely, the specific replacement point exists and the specific replacement point does not exist, as shown in fig. 2-3. Under the condition that a clear replacement point exists, the abrasion allowance of the pantograph has obvious numerical jump, so that the replacement point can be found out, and meanwhile, the mutation point which does not accord with the general change trend of the allowance can be removed. The pseudo code is as follows:
the search result is obtained by the specific steps of: calculating the average value of the total pantograph abrasion allowance in the processed data, sliding on the pantograph data with a step distance of one data at a time by using a window of three data, differencing two adjacent data in three continuous data in the window to obtain two differences, if one difference is larger than the maximum value in a preset parameter, the other difference is smaller than the minimum value in the preset parameter, and the two differences are both larger than the average value of the total pantograph abrasion allowance, the searching result is that a clear replacement point exists, and the clear replacement point is represented as [ t ] c0 ,s c0 ]Otherwise, the searching result is that no explicit replacement point exists.
In the third step, the abrasion allowance of the pantograph is one tenth of the total head and tailThe data of (2) are fitted into a straight line, a point which accords with a replacement threshold value is searched on an extension line of the straight line, and the point is used as a predicted replacement point, so that a front replacement point and a rear replacement point [ t ] are obtained c-1 ,s c-1 ]And [ t ] c1 ,s c1 ]. The pseudo code is as follows:
in the self embodiment, the fitting method adopts a least square method to fit a straight line, and the fitting line is [9,10]The first point in between is identified as [ t ] c-1 ,s c-1 ]In [4,5]The first point in between is identified as [ t ] c1 ,s c1 ]。
In the fourth step, according to the search result of the third step, selecting adjacent replacement points, and randomly sampling between the selected adjacent replacement points to obtain sampling points [ t ] p ,s p ]And the slope k of the fit line of the point to the front-to-back data f And k l . The method comprises the following specific steps: if the search result is that there is an explicit replacement point, then at [ t ] respectively c-1 ,s c-1 ]，[t c0 ,s c0 ]And [ t ] c0 ,s c0 ]，[t c1 ,s c1 ]Sampling randomly to obtain sampling point set, otherwise, at [ t ] c-1 ,s c-1 ]，[t c1 ,s c1 ]Randomly sampling to obtain a sampling point set, and selecting a sampling point [ t ] from the obtained sampling point set p ,s p ]And several data before and after it, sampling point [ t ] p ,s p ]Fitting the data before and after the data into a straight line through least square, and then calculating the slope k of the fitting line f And k l . The pseudo code is as follows:
further, a sampling point [ t ] is selected p ,s p ]And the number of the data before and after the sampling point set is not more than one tenth of the number of the data of the sampling point set.
In step fiveCalculate the sampling point t p ,s p ]The relative days between the time and the time of two adjacent replacement points are respectively expressed as d f And d l And can be used d f And d l Replacing the date in the data after the processing of steps one to four. Wherein d is calculated f And d l The formula of (2) is:
d f ＝t p -t c-1
d l ＝t c1 -t p 。
in step six, d f 、k f 、s p 、k l D l Make up four features [ d f ,k f ,s p ,k l ]，d l Training GRU cyclic neural network as label to obtain d from data to be tested f 、k f 、s p K l Make up four features [ d f ,k f ,s p ,k l ]The trained GRU recurrent neural network is input to predict the number of days from the current date to be replaced.
In this example, table 2 was converted into table 3 by the procedure described above. And then, inputting the data in the table 3 as training data into the GRU circulating neural network, training the GRU circulating neural network, processing the data to be detected into the data in the same format in the table 3, and inputting the data into the trained GRU circulating neural network to predict the number of days when the current distance is the date to be replaced. Table 3 is as follows:
2200 training data sets are obtained through 22 groups of subway locomotive pantograph data, training results are shown in fig. 4, errors of most prediction results are smaller than 5 days, and the prediction accuracy is 94% under the condition that the prediction errors are allowed. Therefore, the subway pantograph replacement time prediction method based on the limited samples can predict the approximate time of the pantograph abrasion reaching the replacement threshold value with higher accuracy under the condition of limited samples.
The foregoing is merely a preferred embodiment of the present invention and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present invention, which are intended to be comprehended within the scope of the present invention.
Claims (7)
1. A subway pantograph replacement time prediction method based on a limited sample is characterized by comprising the following steps:
step one: acquiring original data through a non-contact detection method, cleaning the original data, and dividing and rewriting the cleaned data;
step two: fitting the head and tail part data of the abrasion allowance of the pantograph in the processed data, and estimating the front and back replacement points [ t ] c-1 ,s c-1 ]And [ t ] c1 ,s c1 ]；
Step three: searching the processed data for the data of the new pantograph after the abrasion of the pantograph reaches the threshold value, wherein the data of the new pantograph after the abrasion of the pantograph reaches the threshold value is expressed as a set of vectors [ t ] of time and thickness of the pantograph c0 ,s c0 ]I.e. the specific replacement point;
step four: selecting adjacent replacement points according to the search result of the step three, and randomly sampling between the selected adjacent replacement points to obtain sampling points [ t ] p ,s p ]And the slope k of the fit line of the point to the front-to-back data f And k l ；
Step five: calculate the sampling point t p ,s p ]The relative days between the time and the time of two adjacent replacement points are respectively expressed as d f And d l ；
Step six: will d f 、k f 、s p K l Make up four features [ d f ,k f ,s p ,k l ]，d l Training GRU cyclic neural network as label to obtain d from data to be tested f 、k f 、s p K l Make up four features [ d f ,k f ,s p ,k l ]Input of trained GRU recurrent neural network to predict current distanceDays of day to be replaced.
2. The method for predicting the replacement time of a subway pantograph based on a limited sample according to claim 1, wherein the cleaning the obtained data in the step one specifically comprises the following steps:
and checking the obtained data one by one, and deleting the hollow white data, the messy code data and the data which do not accord with the actual physical meaning in the obtained data.
3. The method for predicting the replacement time of a subway pantograph based on a limited sample according to claim 1, wherein the step one of dividing and rewriting the cleaned data comprises the specific steps of:
classifying and dividing the cleaned data, enabling the data to correspond to the corresponding subway vehicles one by one, and rewriting the classified and divided data into the abrasion condition of a certain pantograph in a period of time, wherein the abrasion condition is expressed as [ t ] i ,s i ]，i∈[t start ,t end ]Wherein t is start And t end Indicating the start and end dates, respectively, of sampling the pantograph.
4. The method for predicting the replacement time of a subway pantograph based on a limited sample according to claim 1, wherein in the second step, the head and tail part data of the wear allowance of the pantograph are fitted into a straight line by least square, and when the values of two ends of the straight line meet the replacement threshold value, the corresponding time coordinates of the two ends of the straight line are the front and rear replacement points [ t ] c-1 ,s c-1 ]And [ t ] c1 ,s c1 ]。
5. The method for predicting the change time of the subway pantograph based on the limited samples according to claim 1, wherein the three specific steps are as follows:
calculating average value of total pantograph wear allowance in processed data, sliding on the pantograph data with step distance of one data at a time by using window of three data, and comparing two adjacent numbers in three continuous data in the windowObtaining two differences according to the difference, if one difference is larger than the maximum value in the preset parameters and the other difference is smaller than the minimum value in the preset parameters and both the two differences are larger than the average value of the total pantograph abrasion allowance, the finding result is that a clear replacement point exists, and the clear replacement point is expressed as [ t ] c0 ,s c0 ]Otherwise, the searching result is that no explicit replacement point exists.
6. The method for predicting the replacement time of the subway pantograph based on the limited samples according to claim 5, wherein when the two difference values are larger than the maximum value in the preset parameters, the second data is mutation data, and the mutation data is removed.
7. The method for predicting the change time of the subway pantograph based on the limited samples according to claim 1, wherein the three specific steps are as follows:
if the search result is that there is an explicit replacement point, then at [ t ] respectively c-1 ,s c-1 ]，[t c0 ,s c0 ]And [ t ] c0 ,s c0 ]，[t c1 ,s c1 ]Sampling randomly to obtain sampling point set, otherwise, at [ t ] c-1 ,s c-1 ]，[t c1 ,s c1 ]Randomly sampling to obtain a sampling point set, and selecting a sampling point [ t ] from the obtained sampling point set p ,s p ]And several data before and after it, sampling point [ t ] p ,s p ]Fitting the data before and after the data into a straight line through least square, and then calculating the slope k of the fitting line f And k l 。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310788272.1A CN116911445A (en) | 2023-06-30 | 2023-06-30 | Subway pantograph replacement time prediction method based on limited samples |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310788272.1A CN116911445A (en) | 2023-06-30 | 2023-06-30 | Subway pantograph replacement time prediction method based on limited samples |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116911445A true CN116911445A (en) | 2023-10-20 |
Family
ID=88352214
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310788272.1A Pending CN116911445A (en) | 2023-06-30 | 2023-06-30 | Subway pantograph replacement time prediction method based on limited samples |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116911445A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117647935A (en) * | 2024-01-29 | 2024-03-05 | 兰州交通大学 | Active control method for backstepping of pantograph |
-
2023
- 2023-06-30 CN CN202310788272.1A patent/CN116911445A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117647935A (en) * | 2024-01-29 | 2024-03-05 | 兰州交通大学 | Active control method for backstepping of pantograph |
| CN117647935B (en) * | 2024-01-29 | 2024-05-24 | 兰州交通大学 | Active control method for backstepping of pantograph |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116911445A (en) | Subway pantograph replacement time prediction method based on limited samples | |
| CN106682699B (en) | A kind of vehicle exhaust emission signature analysis processing method based on clustering | |
| CN109903555B (en) | Bus passenger getting-off data prediction method and system based on big data | |
| CN111680638B (en) | Passenger path identification method and passenger flow clearing method based on same | |
| CN116168356B (en) | Vehicle damage judging method based on computer vision | |
| CN109580260A (en) | A kind of inferior health diagnostic method of track vehicle door system | |
| CN115311858A (en) | Urban road section grading control method based on traffic flow toughness | |
| CN105976615A (en) | Method for realizing vehicle track flow data abnormity detection | |
| CN117115562A (en) | Inbound railway freight train carriage counting method based on deep learning | |
| CN115731261B (en) | Vehicle lane change behavior recognition method and system based on expressway radar data | |
| CN111461183A (en) | Flat wheel fault detection method for optimizing BP neural network by improving artificial fish swarm algorithm | |
| CN114501336A (en) | Road traffic volume measuring and calculating method and device, electronic equipment and storage medium | |
| CN112363415A (en) | Station real-time monitoring system based on wireless sensor network | |
| CN115376315B (en) | Multi-level bayonet quality control method for road network emission accounting | |
| CN116611864A (en) | Urban rail transit train speed and passenger flow prediction method based on RNN | |
| CN113011713B (en) | Information entropy-based driver driving stability assessment method | |
| CN114265955A (en) | Fault prediction method and device based on diagnosis fault code evolution map | |
| CN113033694B (en) | Data cleaning method based on deep learning | |
| Asplund et al. | Assessment of the data quality of wayside wheel profile measurements. | |
| Yuan et al. | A gray prediction method for economic loss of road traffic accidents based on Logistic model. | |
| CN111353515B (en) | Multi-scale classification-based train wheel set tread damage classification and identification method | |
| CN107832173B (en) | Urban rail transit vehicle real-time fault diagnosis method based on working condition detection | |
| CN113011712B (en) | Driver driving stability assessment method based on Euclidean distance and information entropy | |
| CN110889468A (en) | Multi-model uncertain reproduction result analysis method capable of eliminating error information | |
| Kraft et al. | An approach for the classification of track geometry defects |
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 |