CN114692969B - Switch machine fault prediction method, device, electronic equipment and storage medium - Google Patents
Switch machine fault prediction method, device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN114692969B CN114692969B CN202210320814.8A CN202210320814A CN114692969B CN 114692969 B CN114692969 B CN 114692969B CN 202210320814 A CN202210320814 A CN 202210320814A CN 114692969 B CN114692969 B CN 114692969B
- Authority
- CN
- China
- Prior art keywords
- distribution
- switch machine
- time window
- scoring information
- historical time
- 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.)
- Active
Links
Images
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
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M99/00—Subject matter not provided for in other groups of this subclass
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The application provides a method, a device, electronic equipment and a storage medium for predicting a fault of a point machine, wherein the method for predicting the fault of the point machine comprises the following steps: obtaining a plurality of scoring information of the switch machine, wherein the scoring information is used for indicating the degree of abnormality when the switch machine executes corresponding actions; performing distribution fitting on at least two scoring information corresponding to at least two actions executed by the switch machine in the same historical time window to obtain first distribution of the scoring information corresponding to at least two historical time windows; determining a first probability of failure of the switch machine in each historical time window according to a first distribution of the scoring information corresponding to the historical time window; and predicting a second probability of the switch machine to fail in a target time window according to the first probability of the switch machine to fail in each historical time window. The scheme can improve the accuracy of fault prediction of the switch machine.
Description
Technical Field
The application relates to the technical field of rail transit, in particular to a fault prediction method and device for a point switch, electronic equipment and a storage medium.
Background
The switch machine is an actuator of a switch control system for switching and locking switch points or rails, indicating the position and status of the switch points or rails in a supervision and interlock area. The switch machine is an important component of the switch control system, if the switch machine fails, the normal operation of the switch control system is affected, the switch machine is required to be subjected to failure prediction, and maintenance is performed in advance, so that the switch machine is prevented from being failed to affect the normal operation of the switch control system.
At present, operation data of all parts in the switch machine are collected, the collected operation data are input into a pre-trained fault prediction model, and whether all the parts in the switch machine fail or not is predicted through the fault prediction model.
However, three types of faults of the switch machine are initial faults, sudden faults and intermittent faults, and the operation data of each component in the switch machine cannot reflect the sudden faults and the intermittent faults possibly occurring in the switch machine, so that the fault prediction method for analyzing the operation data of each component in the switch machine through the fault prediction model cannot accurately predict the sudden faults and the intermittent faults of the switch machine, and the accuracy of fault prediction on the switch machine is low.
Disclosure of Invention
In view of this, the method, the device, the electronic equipment and the storage medium for predicting the faults of the switch machine can improve the accuracy of predicting the faults of the switch machine.
According to a first aspect of an embodiment of the present application, there is provided a fault prediction method for a switch machine, including: obtaining a plurality of scoring information of the switch machine, wherein the scoring information is used for indicating the degree of abnormality when the switch machine executes corresponding actions; performing distribution fitting on at least two scoring information corresponding to at least two actions executed by the switch machine in the same historical time window to obtain first distribution of the scoring information corresponding to at least two historical time windows; determining a first probability of failure of the switch machine in each historical time window according to a first distribution of the scoring information corresponding to the historical time window; and predicting a second probability of the switch machine to fail in a target time window according to the first probability of the switch machine to fail in each historical time window.
According to a second aspect of the embodiments of the present application, there is provided a fault prediction apparatus for a switch machine, including: an acquisition unit configured to acquire a plurality of pieces of scoring information of a switch machine, wherein the scoring information is used to indicate an abnormality degree when the switch machine performs a corresponding action; the fitting unit is used for carrying out distribution fitting on at least two scoring information corresponding to at least two actions executed by the switch machine in the same historical time window, and obtaining first distribution of the scoring information corresponding to at least two historical time windows; a calculation unit, configured to determine a first probability of failure of the switch machine within each historical time window according to a first distribution of the scoring information corresponding to the historical time window; and the prediction unit is used for predicting the second probability of the fault of the switch machine in the target interval window according to the first probability of the fault of the switch machine in each historical time window.
According to a third aspect of embodiments of the present application, there is provided an electronic device, including: the device comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the memory and the communication interface complete mutual communication through the communication bus; the memory is used for storing at least one executable instruction, and the executable instruction enables the processor to execute the operation corresponding to the fault prediction method of the switch machine provided in the first aspect.
According to a fourth aspect of embodiments of the present application, there is provided a computer readable storage medium having stored thereon computer instructions that, when executed by a processor, cause the processor to perform operations corresponding to the switch machine fault prediction method provided in the first aspect above.
According to a fifth aspect of embodiments of the present application, there is provided a computer program product tangibly stored on a computer-readable medium and comprising computer-executable instructions which, when executed, cause at least one processor to perform the switch machine fault prediction method as provided by the above-mentioned first aspect or any one of the possible implementations of the first aspect.
According to the technical scheme, after the scoring information corresponding to the actions executed by the switch machine in different historical time windows is obtained, the scoring information corresponding to each historical time window is subjected to distribution fitting, the first distribution of the scoring information corresponding to each historical time window is obtained, then the first probability of failure of the switch machine in each historical time window is respectively determined according to the first distribution corresponding to each historical time window, and then the second probability of failure of the switch machine in the target time window is predicted according to the first probability of failure of the switch machine in each historical time window. By determining a plurality of time windows, the running data of the switch machine in each time window has correlation before the switch machine has different types of faults, so that the probability of the fault of the switch machine in the target time window can be predicted, the probability represents the comprehensive probability of various types of faults of the switch machine, and the accuracy of fault prediction of the switch machine can be improved.
Drawings
Fig. 1 is a flowchart of a fault prediction method for a switch machine according to an embodiment of the present application;
FIG. 2 is a flowchart of a score information distribution fitting method according to an embodiment of the present disclosure;
Fig. 3 is a schematic diagram of a fault prediction device for a switch machine according to an embodiment of the present application;
fig. 4 is a schematic diagram of an electronic device according to a fourth embodiment of the present application.
List of reference numerals:
101: obtaining multiple scoring information for a switch machine
102: performing distribution fitting on a plurality of scoring information corresponding to the same historical time window to obtain a first distribution of scoring information
103: determining, from each first distribution, a first probability of failure of the switch machine within a corresponding historical time window
104: predicting a second probability of failure of the switch machine within the target time window according to the first probabilities
201: fitting different types of distribution to scoring information corresponding to the same historical time window to obtain a plurality of second distributions
202: calculating the matching degree of each second distribution and the scoring information
203: determining the second distribution with the largest corresponding matching degree as the first distribution of scoring information corresponding to the historical time window
100: the fault prediction method 200 of the switch machine comprises the following steps: score information distribution fitting method 300: fault prediction device for switch machine
301: the acquisition unit 302: fitting unit 303: calculation unit
304: prediction unit 400: electronic device 402: processor and method for controlling the same
404: communication interface 406: memory 408: communication bus
410: program
Detailed Description
As mentioned above, the switch machine is an important component of the switch control system, and failure of the switch machine can affect normal operation of the switch control system, so that failure prediction needs to be performed on the switch machine, and maintenance is performed on the switch machine in advance, so that failure of the switch machine in the operation process of the switch control system is avoided, and driving safety is affected. At present, a fault prediction model of the switch machine is trained in advance, operation data of all parts in the switch machine are collected, the operation data are input into the fault prediction model, and whether the switch machine can fail or not is predicted through the fault prediction model. However, the sudden failure and intermittent failure may occur in the switch machine, and the operation data of each component in the switch machine may not reflect the sudden failure and intermittent failure which are about to occur in the switch machine, so that the failure prediction method for analyzing the operation data of each component in the switch machine by the failure detection model may not accurately predict the sudden failure and intermittent failure of the switch machine, and thus the accuracy of failure prediction for the switch machine is low.
In the embodiment of the application, the scoring information for indicating the degree of abnormality of the action executed by the switch machine is obtained, the distribution of a plurality of scoring information corresponding to a plurality of actions executed by the switch machine in each time window is respectively determined, the first probability of failure of the switch machine in the corresponding historical time window is determined according to each distribution of the scoring information, and then the second probability of failure of the switch machine in the target time window is predicted according to the first probability corresponding to each historical time window, so that the prediction of the failure of the switch machine is realized. According to the embodiment of the application, the occurrence of the specific faults is not predicted, but the probability of the faults of the switch machine in a specific time period is predicted, so that various types of faults possibly occurring in the switch machine can be predicted, and the accuracy of fault prediction of the switch machine can be improved.
The following describes in detail a method, an apparatus and an electronic device for predicting a fault of a switch machine according to embodiments of the present application with reference to the accompanying drawings.
Fig. 1 is a flowchart of a method 100 for predicting a failure of a switch machine according to an embodiment of the present application. As shown in fig. 1, the switch machine fault prediction method 100 includes the steps of:
and 101, acquiring a plurality of grading information of the switch machine.
The scoring information is used for indicating the degree of abnormality of the switch machine when the switch machine executes corresponding actions, and each action of the switch machine corresponds to one scoring information. The switch machine can drive the switch to move each time to complete the position conversion of the switch, wherein the switch is equipment for branching one track into two or more tracks to enable rolling stock to be switched from one line to the other.
The scoring information of the switch machine may be an anomaly score of the switch machine, the higher the anomaly score, the more serious the anomaly of the switch machine, or the lower the anomaly score, the more serious the anomaly of the switch machine.
And 102, performing distribution fitting on at least two pieces of scoring information corresponding to at least two actions executed by the switch machine in the same historical time window, and obtaining first distribution of scoring information corresponding to at least two historical time windows.
A plurality of historical time windows are determined based on the execution time of the actions performed by the switch machine such that the switch machine performs at least two actions within each historical time window. And performing distribution fitting on scoring information corresponding to a plurality of actions executed by the switch machine in the same time window to obtain first distribution of scoring information corresponding to each historical time window. In the same history time window, the health condition of the switch machine is relatively stable, so that a plurality of scoring information corresponding to the same history time window can be distributed in a relatively smaller range, and therefore the distribution of a plurality of scoring information corresponding to the same history time window can be fitted, and the first distribution of scoring information corresponding to each history time window can be obtained.
It should be understood that the time window refers to a continuous period of time and the historical time window refers to a past period of time. The different historical time windows may include overlapping time periods, and the same scoring information may correspond to different historical time windows, but different historical time windows may correspond to at least one different scoring information, e.g., a historical time windowT 1 Corresponding to the scoring information 1 to 30, historical time window T 2 Corresponding to the scoring information 2 to the scoring information 31, the scoring information 1 and the scoring information 31 are the historical time window T 1 And a history time window T 2 Corresponding different scoring information.
Alternatively, the different historical time windows have the same time length, for example, each historical time window has a time length of 24 hours, and the number of scoring information corresponding to each time window may be the same or different. Alternatively, different historical time windows correspond to the same amount of scoring information, for example, each historical time window corresponds to N scoring information, and N is greater than or equal to 30 and less than or equal to 100, where the time length of each time window may be the same or different.
The first distribution of the scoring information corresponding to the historical time window characterizes the probability of the scoring information corresponding to different values of the historical time window, the scoring information can indicate the abnormal degree of the switch machine when the switch machine executes the action, the abnormal degree of the switch machine when the switch machine executes the action can reflect the health condition of the switch machine, the higher the abnormal degree is, the higher the possibility of the fault of the switch machine is, so the probability of the fault of the switch machine in the historical time window can be determined according to the first distribution of the scoring information corresponding to one historical time window.
After the switch machine is abnormal, the degree of abnormality of the switch machine can be increased along with the increase of the running time of the switch machine, namely the probability of the switch machine to be failed is correspondingly increased, namely the probability of the switch machine to be failed is related to the running time of the switch machine, so that the second probability of the switch machine to be failed in a target time window can be predicted according to the first probability of the switch machine to be failed in a series of historical time windows, and further, after the second probability is larger than a certain threshold value, the switch machine can be maintained in advance, and the switch machine is prevented from being failed to influence the normal running of a switch control system.
It should be appreciated that the target time window is a continuous time period that includes a future time period, and may or may not include a past time period. For example, the current time is T k The target time window is the time point T k To time point T k+1 Time period, time point T k+1 At time point T k Thereafter, the target time window does not include the past time period at this time. Alternatively, the current time is T k The target time window is the time point T k-1 To time point T k+1 Time period, time point T k+1 At time point T k-1 Thereafter, time point T k At time point T k-1 And point in time T k+1 Between them.
It should be further understood that when different historical time windows have the same time length, the target time window and the historical time window have the same time length, and when different historical time windows correspond to the same amount of scoring information, the number of times the switch machine performs actions in the target time window is the same as the amount of scoring information corresponding to the historical time window, that is, the historical time window corresponds to the same amount of scoring information as the target time window, so that the number of variables in the fault prediction process is controlled, and the accuracy of the prediction result is ensured.
After the target time window is relative to the current time and the scoring information corresponding to each execution of the action of the switch machine in the target time window is obtained, the target time window is used as a historical time window to predict the probability of failure of the switch machine in the next target time window, so that the failure prediction of the switch machine is continuously performed.
In the embodiment of the application, after scoring information corresponding to actions executed by the switch machine in different historical time windows is obtained, distribution fitting is performed on a plurality of scoring information corresponding to each historical time window, a first distribution of the scoring information corresponding to each historical time window is obtained, then according to the first distribution corresponding to each historical time window, the first probability of failure of the switch machine in each historical time window is respectively determined, and then according to the first probability of failure of the switch machine in each historical time window, the second probability of failure of the switch machine in a target time window is predicted. By determining a plurality of time windows, the running data of the switch machine in each time window has correlation before the switch machine has different types of faults, so that the probability of the fault of the switch machine in the target time window can be predicted, the probability represents the comprehensive probability of various types of faults of the switch machine, and the accuracy of fault prediction of the switch machine can be improved.
In one possible implementation manner, when obtaining multiple scoring information of the switch machine, operation data of the switch machine when each action is executed can be obtained, and the operation data of the switch machine when each action is executed is input into a pre-trained scoring calculation model to obtain the scoring information of the switch machine when the corresponding action is executed. The operation data of the switch machine includes at least one of current data, voltage data, temperature data, humidity data, vibration data, and driving force data.
When the switch is abnormal, the operation data of the switch when the switch is operated is changed, and the degree of abnormality of the switch is different, and the operation data of the switch when the switch is operated is also different, compared with the normal operation of the switch, so that the scoring information for indicating the degree of abnormality of the switch can be determined according to the operation data of the switch when the switch is operated.
Parameters such as current, voltage, temperature, environmental humidity, vibration and driving force of the switch machine can correspondingly change along with different running states of the switch machine, so that current data, voltage data, temperature data, environmental humidity data, vibration data, driving force data and the like can be collected in the running process of the switch machine, and scoring information of the switch machine can be determined according to the collected running data.
The scoring calculation model is trained in advance, and scoring information representing the degree of abnormality of the switch machine can be calculated based on the operation data of the switch machine. The scoring calculation model can be obtained through training of an isolated Forest (Isolation Forest) anomaly detection algorithm, a support vector machine (One-class SVM) or a Gaussian Mixture Model (GMM) and the like.
In the embodiment of the application, when the switch machine is abnormal to different degrees, the running data such as the current data, the voltage data, the temperature data, the humidity data, the vibration data and the driving force data are correspondingly changed, so that the running data can be input into a pre-trained grading calculation model, the grading information of the switch machine is determined through the grading calculation model, the calculated grading information can be ensured to accurately reflect the abnormal degree of the switch machine, automatic calculation of the grading information can be realized, and the labor intensity of a user in the fault prediction process of the switch machine is reduced.
In one possible implementation manner, when the distribution fitting is performed on the plurality of pieces of scoring information corresponding to the same historical time window, the fitting may be performed on the plurality of pieces of scoring information corresponding to the same historical time window through different types of distributions, and then one distribution with a good matching degree with the scoring information is selected from the fitted plurality of distributions to be used as the first distribution. The following describes in detail the procedure of fitting the distribution of the score information.
Fig. 2 is a flowchart of a score information distribution fitting method 200 according to an embodiment of the present application. As shown in fig. 2, the score information distribution fitting method 200 includes the steps of:
For each historical time window, the switch machine performs a plurality of actions within the historical time window, each action performed by the switch machine corresponds to one piece of scoring information, and thus the historical time window corresponds to a plurality of pieces of scoring information. And fitting the plurality of scoring information corresponding to the historical time window by adopting different types of distribution to obtain a plurality of second distributions of the scoring information, wherein the different second distributions correspond to different distribution types.
When the distribution fitting is performed on the plurality of scoring information corresponding to the same historical time window, a plurality of distributions of different types such as normal distribution, poisson distribution, laplace distribution, binomial distribution and the like can be adopted, and the fitting is performed on the plurality of scoring information corresponding to the same historical time window, so that a second distribution corresponding to each distribution type is obtained.
And respectively calculating the matching degree of the scoring information corresponding to the historical time window of each second distribution after obtaining the plurality of second distributions of the scoring information corresponding to the historical time window aiming at the same historical time window. The matching degree of the second distribution and the scoring information is used for indicating the fitting goodness of the corresponding second distribution, namely the regression degree of the regression line or curve of the second distribution to the corresponding scoring information, and representing the accuracy of the change rule of the scoring information indicated by the second distribution.
And 203, determining the second distribution with the largest corresponding matching degree as the first distribution of the scoring information corresponding to the historical time window.
For the same historical time window, after calculating the matching degree of a plurality of second distributions and the scoring information, determining one second distribution with the largest matching degree with the scoring information as the first distribution of the scoring information corresponding to the historical time window according to the matching degree of each second distribution and the scoring information.
For example, the second distribution of the scoring information includes a normal distribution, a poisson distribution, a laplace distribution, and a binomial distribution, and after matching degrees of the normal distribution, the poisson distribution, the laplace distribution, and the binomial distribution with the scoring information are calculated, the matching degree of the normal distribution with the scoring information is the largest, and the normal distribution is determined as the first distribution of the scoring information.
It should be understood that, since the first distributions are determined for the different historical time windows respectively, the first distributions corresponding to the different historical time windows may be different types of distributions, and of course, may be the same type of distribution, so the first distributions of the scoring information corresponding to the respective historical time windows may be one or more of a normal distribution, a poisson distribution, a laplacian distribution, and a binomial distribution. For example, the first distribution of the scoring information corresponding to a part of the historical time window is a normal distribution, and the first distribution of the scoring information corresponding to the other historical time windows is a laplace distribution.
In the embodiment of the application, a plurality of scoring information corresponding to the historical time window is fitted through different types of distribution, a plurality of second distribution corresponding to different types of distribution is obtained, then the matching degree of each second distribution and the scoring information is calculated, one second distribution with the largest matching degree with the scoring information is determined to be the first distribution of the scoring information, the determined first distribution can more accurately reflect the distribution rule of the scoring information in the corresponding historical time window, and further the accuracy of fault prediction of the switch machine based on the first distribution is guaranteed.
In one possible implementation, step 201 fits the second distribution of scoring information, and maximum likelihood estimates (Maximum Likelihood Estimate, MLE) may be used to determine the parameter values for the corresponding distribution. The optimized parameter value combination can be determined through maximum likelihood estimation, so that the log likelihood of the corresponding distribution is maximum. When determining the parameter value combinations of the respective distributions based on the maximum likelihood, the parameter value combinations of the respective distributions may be determined specifically by the following formula (1).
In the embodiment of the application, when different types of distribution such as normal distribution, poisson distribution, laplacian distribution and binomial distribution are adopted to fit a plurality of scoring information corresponding to the same historical time window, the values of parameters of the corresponding types of distribution are adjusted based on each scoring information corresponding to the historical time window, so that the log likelihood of the corresponding types of distribution is maximum, the distribution defined by the values of each parameter at the moment is determined to be the second distribution of the scoring information, the determined second distribution can be ensured to accurately reflect the distribution condition of each scoring information in the same historical time window, the speed of fitting the second distribution can be improved, and the timeliness of fault prediction of the switch machine is further improved.
It should be noted that, since the different types of distributions, such as the normal distribution, the poisson distribution, the laplace distribution, and the binomial distribution, all include a plurality of parameters, when the score information in the same historical time window is fitted through the different types of distributions, such as the normal distribution, the poisson distribution, the laplace distribution, and the binomial distribution, each parameter of the corresponding type of distribution needs to be adjusted to obtain a better fitting result
Is a vector formed by combining the values of a plurality of parameters distributed correspondingly.
In one possible implementation, when calculating the matching degree between each second distribution and the scoring information in step 202, the log-likelihood value of the second distribution may be calculated according to the scoring information, where the log-likelihood value may indicate the goodness of fit of the second distribution, and the larger the log-likelihood value, the higher the matching degree between the second distribution and the scoring information. Specifically, when calculating the matching degree of a second distribution and scoring information, the probability density function of the second distribution is first obtained, and then the matching degree of the second distribution and scoring information is calculated by the following formula (2).
ln (L (θ)) is used to characterize the degree of matching of the second distribution with at least two scoring information, p (x) i I θ) is used to characterize the probability density function of the second distribution, n is used to characterize the number of scoring information corresponding to the same historical time window, x i For characterizing the ith scoring information in the n scoring information, θ is for characterizing the information represented by the ith scoring informationThe two distributions include a vector of parameters.
In this embodiment of the present application, for each second distribution, after obtaining a probability density function of the second distribution, according to the obtained probability density function and a plurality of scoring information corresponding to the second distribution, the matching degree between the second distribution and the scoring information is calculated by using the above formula (2), so as to ensure that the calculated matching degree can accurately reflect the fitting degree of the second distribution to the scoring information.
In one possible implementation, when calculating the matching degree between each second distribution and the scoring information in step 202, the matching degree between the second distribution and the scoring information may be determined according to the Anderson-Darling statistic, where the Anderson-Darling statistic may represent the goodness of fit of the second distribution, and the smaller the Anderson-Darling statistic, the better the matching degree between the second distribution and the scoring information. Specifically, when calculating the matching degree of a second distribution and scoring information, the probability density function of the second distribution is first obtained, and then the matching degree of the second distribution and scoring information is calculated by the following formula (3).
Z is used for representing the matching degree of the second distribution and at least two scoring information; f (x) is used to characterize the second distribution; n is used for representing the number of scoring information corresponding to the same historical time window; w (x) is used for representing a preset weight function, and w (x) is more than or equal to 0; p (x is used to characterize the probability density function of the second distribution; k is used to characterize the empirical coefficients determined from the at least two scoring information.
In this embodiment of the present application, after obtaining the probability density function of the second distribution for each second distribution, the function F (x) used for characterizing the second distribution, the preset weight function w (x), the probability density function p (x) of the second distribution, and each piece of scoring information corresponding to the second distribution are brought into the above formula (3) to calculate the matching degree between the second distribution and the scoring information, where the calculated matching degree is equal to the reciprocal of the Anderson-Darling statistic, and the greater the calculated matching degree, the better the matching degree between the second distribution and the scoring information, so as to ensure that the calculated matching degree can accurately reflect the fitting degree between the second distribution and the scoring information, and further ensure the accuracy of the fault prediction on the switch machine based on the second distribution.
In one possible implementation manner, step 203 determines a second distribution with the largest matching degree as the first distribution of the scoring information corresponding to the corresponding historical time window, and since the matching degree of the second distribution and the scoring information may be calculated in multiple manners, when determining the first distribution corresponding to the historical time window from the multiple second distributions corresponding to the historical time window, the second distribution with the largest matching degree may be determined from the multiple second distributions as the first distribution according to the matching degree calculated in the same manner, for example, the matching degree calculated by the above formula (2) or the formula (3), and the first distribution may be determined from the second distributions based on the matching degree calculated in multiple manners, for example, the weighted average value of the matching degree calculated in multiple manners, and the second distribution with the largest weighted average value of the matching degree may be determined from the multiple second distributions as the first distribution.
In one possible implementation manner, when determining the first probability of the failure of the switch machine in the historical time window according to the first distribution of the scoring information corresponding to the historical time window, a probability density function of the first distribution corresponding to the historical time window may be obtained, and then the first probability of the failure of the switch machine in the historical time window is calculated according to the obtained probability density function through the following formula (4).
F T The method comprises the steps of representing first probability of faults of the switch machine in a historical time window corresponding to first distribution, wherein t is used for representing a preset grading information threshold value, and p (x) is used for representing a probability density function of the first distribution.
In the embodiment of the present application, the probability density function of the first distribution is used to indicate the probability distribution obeyed by describing the continuous random variable scoring information, the scoring information threshold value can be determined based on experience, the scoring information greater than the scoring information threshold value indicates that the switch machine is about to fail, through the above formula (4), the probability density function p (x) of the first distribution is integrated within the [ t, ] range, the integrated result is the first probability representing that the switch machine fails in the corresponding historical time window, the calculated first probability is ensured to accurately reflect the failure occurrence condition of the switch machine in the corresponding historical time window, and the accuracy of failure prediction of the switch machine based on the first probability is ensured.
In one possible implementation manner, after the first probabilities that the switch machine fails in the plurality of historical time windows are obtained, the first probabilities corresponding to each historical time window can be combined according to the time sequence corresponding to the historical time window to obtain a probability sequence, and then the probability sequence is input into a pre-trained failure prediction model to obtain the second probability that the switch machine output by the failure prediction model fails in the target time window.
In the embodiment of the application, the running process of the switch machine is divided into a plurality of time windows, before various types of faults occur in the switch machine, the running data of the switch machine in each time window can generate corresponding changes, and the changes of the running data have correlation in the time dimension, so that the first probabilities corresponding to different historical time windows can be combined into probability sequences according to time sequences, the change rule of the probability sequences is identified through a pre-trained fault prediction model, the probability of faults of the switch machine in a target time window is determined, and therefore the occurrence of various types of faults of the switch machine can be predicted, the occurrence of a certain specific fault is not predicted, and the comprehensiveness of fault prediction on the switch machine is ensured.
Fig. 3 is a schematic diagram of a fault prediction device 300 for a switch machine according to an embodiment of the present application. As shown in fig. 3, the switch machine failure prediction apparatus 300 includes:
an obtaining unit 301, configured to obtain a plurality of scoring information of the switch machine, where the scoring information is used to indicate an abnormality degree when the switch machine performs a corresponding action;
the fitting unit 302 is configured to perform distribution fitting on at least two pieces of scoring information corresponding to at least two actions executed by the switch machine in the same historical time window, so as to obtain a first distribution of scoring information corresponding to at least two historical time windows;
a calculating unit 303, configured to determine a first probability of a failure of the switch machine within each historical time window according to a first distribution of scoring information corresponding to the historical time window;
and a prediction unit 304, configured to predict a second probability of failure of the switch machine in the inter-target window according to the first probability of failure of the switch machine in each historical time window.
In this embodiment of the present application, the obtaining unit 301 may be used to perform step 101 in the above method embodiment, the fitting unit 302 may be used to perform step 102 in the above method embodiment, the calculating unit 303 may be used to perform step 103 in the above method embodiment, and the predicting unit 304 may be used to perform step 104 in the above method embodiment.
In one possible implementation manner, the obtaining unit 301 may obtain operation data of each time the switch machine performs an action, and input the operation data of each time the switch machine performs an action into a pre-trained score calculation model to obtain score information corresponding to the action performed by the switch machine, where the operation data includes at least one of current data, voltage data, temperature data, humidity data, vibration data, and driving force data of the switch machine.
In one possible implementation manner, the fitting unit 302 may perform fitting of different types of distributions on at least two scoring information corresponding to at least two actions performed by the switch machine in the same historical time window, obtain at least two second distributions of the scoring information, calculate a matching degree between each second distribution and at least two scoring information, and determine a second distribution with the largest matching degree as a first distribution of the scoring information corresponding to the corresponding historical time window.
In one possible implementation, the fitting unit 302 may obtain a probability density function of each second distribution, and calculate the probability density function of each second distribution according to the formula
Calculating the matching degree of the second distribution and at least two scoring information, wherein ln (L (theta)) is used for representing the matching degree of the second distribution and the at least two scoring information, and p (x) i I θ) is used to characterize the probability density function of the second distribution, n is used to characterize the number of scoring information corresponding to the same historical time window, x i For characterizing the ith scoring information in the n scoring information, θ is for characterizing a vector of parameters included in the second distribution.
In one possible implementation, the fitting unit 302 may obtain a probability density function of each second distribution, and calculate the probability density function of each second distribution according to the formula
Calculating the matching degree of the second distribution and at least two scoring information, wherein Z is used for representing the matching degree of the second distribution and the at least two scoring information; f (x) is used to characterize the second distribution; n is used for representing the number of scoring information corresponding to the same historical time window; w (x) is used for representing a preset weight function, and w (x) is more than or equal to 0; p (x) is used to characterize the probability density function of the second distribution; k is used to characterize the empirical coefficients determined from the at least two scoring information.
In one possible implementation, the computing unit 303 may obtain a probability density function of each first distribution and calculate a probability density function of each first distribution according to the probability density function of the first distribution by the formula
Calculating a first probability of failure of the switch machine within a historical time window corresponding to the first distribution, wherein F T The method comprises the steps of representing first probability of faults of the switch machine in a historical time window corresponding to first distribution, wherein t is used for representing a preset grading information threshold value, and p (x) is used for representing a probability density function of the first distribution.
In one possible implementation, the prediction unit 304 may combine the first probabilities according to a time sequence of the corresponding historical time window to obtain a probability sequence, and input the probability sequence into a pre-trained failure prediction model to obtain a second probability that the switch machine fails within the target time window.
It should be noted that, because the content of information interaction and execution process between each unit in the switch machine fault prediction device is based on the same concept as the switch machine fault prediction method embodiment, specific content can be referred to the description in the switch machine fault prediction method embodiment, and will not be repeated here.
Fig. 4 is a schematic diagram of an electronic device provided in an embodiment of the present application, and the embodiment of the present application is not limited to a specific implementation of the electronic device. Referring to fig. 4, an electronic device 400 provided in an embodiment of the present application includes: a processor 402, a communication interface (Communications Interface) 404, a memory 406, and a communication bus 408. Wherein:
A communication interface 404 for communicating with other electronic devices or servers.
The processor 402 is configured to execute the process 410, and may specifically perform the relevant steps in any of the foregoing embodiments of the machine fault prediction method.
In particular, program 410 may include program code including computer-operating instructions.
The processor 402 may be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement embodiments of the present application. The one or more processors comprised by the smart device may be the same type of processor, such as one or more CPUs; but may also be different types of processors such as one or more CPUs and one or more ASICs.
The program 410 is particularly useful for causing the processor 402 to perform the switch machine fault prediction method of any of the foregoing embodiments.
The specific implementation of each step in the procedure 410 may refer to corresponding descriptions in the corresponding steps and units in any of the foregoing embodiments of the fault prediction method for a point machine, which are not repeated herein. It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the apparatus and modules described above may refer to corresponding procedure descriptions in the foregoing method embodiments, which are not repeated herein.
Through the electronic equipment of the embodiment of the application, after the scoring information corresponding to the actions executed by the switch machine in different historical time windows is obtained, distribution fitting is carried out on a plurality of scoring information corresponding to each historical time window, the first distribution of the scoring information corresponding to each historical time window is obtained, then the first probability of failure of the switch machine in each historical time window is respectively determined according to the first distribution corresponding to each historical time window, and then the second probability of failure of the switch machine in the target time window is predicted according to the first probability of failure of the switch machine in each historical time window. By determining a plurality of time windows, the running data of the switch machine in each time window has correlation before the switch machine has different types of faults, so that the probability of the fault of the switch machine in the target time window can be predicted, the probability represents the comprehensive probability of various types of faults of the switch machine, and the accuracy of fault prediction of the switch machine can be improved.
Embodiments also provide a computer readable storage medium storing instructions for causing a machine to perform a switch machine failure prediction method as described herein. Specifically, a system or apparatus provided with a storage medium on which a software program code realizing the functions of any of the above embodiments is stored, and a computer (or CPU or MPU) of the system or apparatus may be caused to read out and execute the program code stored in the storage medium.
In this case, the program code itself read from the storage medium may realize the functions of any of the above-described embodiments, and thus the program code and the storage medium storing the program code form part of the present application.
Examples of the storage medium for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (e.g., CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a nonvolatile memory card, and a ROM. Alternatively, the program code may be downloaded from a server computer by a communication network.
Further, it should be apparent that the functions of any of the above-described embodiments may be implemented not only by executing the program code read out by the computer, but also by causing an operating system or the like operating on the computer to perform part or all of the actual operations based on the instructions of the program code.
Further, it is understood that the program code read out by the storage medium is written into a memory provided in an expansion board inserted into a computer or into a memory provided in an expansion module connected to the computer, and then a CPU or the like mounted on the expansion board or the expansion module is caused to perform part and all of actual operations based on instructions of the program code, thereby realizing the functions of any of the above embodiments.
Embodiments of the present application also provide a computer program product tangibly stored on a computer-readable medium and comprising computer-executable instructions that, when executed, cause at least one processor to perform the switch machine fault prediction method provided by the above embodiments. It should be understood that each solution in this embodiment has the corresponding technical effects in the foregoing method embodiments, which are not repeated herein.
It should be noted that not all the steps and modules in the above flowcharts and the system configuration diagrams are necessary, and some steps or modules may be omitted according to actual needs. The execution sequence of the steps is not fixed and can be adjusted as required. The system structure described in the above embodiments may be a physical structure or a logical structure, that is, some modules may be implemented by the same physical entity, or some modules may be implemented by multiple physical entities, or may be implemented jointly by some components in multiple independent devices.
In the above embodiments, the hardware module may be mechanically or electrically implemented. For example, a hardware module may include permanently dedicated circuitry or logic (e.g., a dedicated processor, FPGA, or ASIC) to perform the corresponding operations. The hardware modules may also include programmable logic or circuitry (e.g., a general-purpose processor or other programmable processor) that may be temporarily configured by software to perform the corresponding operations. The particular implementation (mechanical, or dedicated permanent, or temporarily set) may be determined based on cost and time considerations.
While the present application has been illustrated and described in detail in the drawings and in the preferred embodiments, the present application is not limited to the disclosed embodiments, and it will be appreciated by those skilled in the art that the code audits of the various embodiments described above may be combined to obtain further embodiments of the present application, which are also within the scope of the present application.
Claims (10)
1. A method (100) of predicting a failure of a switch machine, comprising:
obtaining a plurality of scoring information of the switch machine, wherein the scoring information is used for indicating the degree of abnormality when the switch machine executes corresponding actions;
Performing distribution fitting on at least two scoring information corresponding to at least two actions executed by the switch machine in the same historical time window to obtain first distribution of the scoring information corresponding to each historical time window;
determining a first probability of failure of the switch machine in each historical time window according to a first distribution of the scoring information corresponding to the historical time window;
and predicting a second probability of the switch machine to fail in a target time window according to the first probability of the switch machine to fail in each historical time window.
2. The method of claim 1, wherein the obtaining a plurality of scoring information for the switch machine comprises:
acquiring operation data of the switch machine when the switch machine executes actions each time, wherein the operation data comprises at least one of current data, voltage data, temperature data, humidity data, vibration data and driving force data of the switch machine;
and inputting the operation data of the switch machine when each action is executed into a pre-trained score calculation model to obtain the score information corresponding to the action executed by the switch machine.
3. The method of claim 1, wherein said performing a distribution fit on at least two of said scoring information corresponding to at least two actions performed by said switch machine within a same historical time window to obtain a first distribution of said scoring information within each said historical time window comprises:
Fitting different types of distribution to at least two scoring information corresponding to at least two actions executed by the switch machine in the same historical time window to obtain at least two second distributions of the scoring information;
calculating the matching degree of each second distribution and the at least two scoring information;
and determining the second distribution corresponding to the maximum matching degree as a first distribution of the scoring information corresponding to the corresponding historical time window.
4. A method according to claim 3, wherein said calculating the degree of matching of each of said second distributions to said at least two scoring information comprises:
acquiring a probability density function of each second distribution;
calculating the matching degree of the second distribution and the at least two scoring information according to the probability density function of the second distribution through the following formula;
ln (L (θ)) is used to characterize the degree of matching of the second distribution with the at least two scoring information, p (x) i I theta) is used to characterize the probability density function of the second distribution, n is used to characterize the number of scoring information corresponding to the same historical time window, x i And (c) representing the ith scoring information in the n scoring information, wherein θ represents a vector composed of parameters included in the second distribution.
5. A method according to claim 3, wherein said calculating the degree of matching of each of said second distributions to said at least two scoring information comprises:
acquiring a probability density function of each second distribution;
calculating the matching degree of the second distribution and the at least two scoring information according to the probability density function of the second distribution through the following formula;
z is used for representing the matching degree of the second distribution and the at least two scoring information; f (x) is used to characterize the second distribution; n is used for representing the quantity of the scoring information corresponding to the same historical time window; w (x) is used for representing a preset weight function, and w (x) is more than or equal to 0; p (x) is used to characterize the probability density function of the second distribution; k is used to characterize the empirical coefficients determined from the at least two scoring information.
6. The method of claim 1, wherein said determining a first probability of said switch machine failing within each of said historical time windows based on a first distribution of said scoring information corresponding to that historical time window comprises:
acquiring a probability density function of each first distribution;
according to the probability density function of the first distribution, calculating a first probability of the fault of the switch machine in the historical time window corresponding to the first distribution through the following formula;
F T The probability density function is used for representing the first probability of the fault of the switch machine in the historical time window corresponding to the first distribution, t is used for representing a preset grading information threshold value, and p (x) is used for representing a probability density function of the first distribution.
7. The method of any of claims 1-6, wherein predicting a second probability of failure of the switch machine within a target time window based on a first probability of failure of the switch machine within each of the historical time windows, comprises:
combining the first probabilities according to the time sequence corresponding to the historical time window to obtain a probability sequence;
and inputting the probability sequence into a pre-trained fault prediction model to obtain a second probability of the fault of the switch machine in a target time window.
8. A point machine failure prediction apparatus (300), comprising:
an acquisition unit (301) configured to acquire a plurality of scoring information of a switch machine, wherein the scoring information is used to indicate an abnormality degree when the switch machine performs a corresponding action;
a fitting unit (302) configured to perform distribution fitting on at least two score information corresponding to at least two actions executed by the switch machine in the same historical time window, so as to obtain a first distribution of the score information corresponding to each historical time window;
A calculation unit (303) for determining a first probability of failure of the switch machine within each historical time window according to a first distribution of the scoring information corresponding to the historical time window;
and the predicting unit (304) is used for predicting the second probability of the fault of the switch machine in the target time window according to the first probability of the fault of the switch machine in each historical time window.
9. An electronic device (400), characterized by comprising: -a processor (402), a communication interface (404), a memory (406) and a communication bus (408), said processor (402), said memory (406) and said communication interface (404) completing communication with each other via said communication bus (408);
the memory (406) is configured to store at least one executable instruction that causes the processor (402) to perform operations corresponding to the switch machine fault prediction method (100) of any one of claims 1-7.
10. A computer readable storage medium having stored thereon computer instructions which, when executed by a processor, cause the processor to perform the method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210320814.8A CN114692969B (en) | 2022-03-29 | 2022-03-29 | Switch machine fault prediction method, device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210320814.8A CN114692969B (en) | 2022-03-29 | 2022-03-29 | Switch machine fault prediction method, device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114692969A CN114692969A (en) | 2022-07-01 |
| CN114692969B true CN114692969B (en) | 2023-05-12 |
Family
ID=82141223
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210320814.8A Active CN114692969B (en) | 2022-03-29 | 2022-03-29 | Switch machine fault prediction method, device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114692969B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107054410A (en) * | 2017-04-01 | 2017-08-18 | 广州地铁集团有限公司 | The intelligent diagnosis system and diagnostic method of point machine |
| CN111046583A (en) * | 2019-12-27 | 2020-04-21 | 中国铁道科学研究院集团有限公司通信信号研究所 | Switch machine fault diagnosis method based on DTW algorithm and ResNet network |
| CN111914320A (en) * | 2020-06-06 | 2020-11-10 | 同济大学 | No-sample turnout fault diagnosis method based on deep learning |
| CN112906230A (en) * | 2021-03-05 | 2021-06-04 | 北京全路通信信号研究设计院集团有限公司 | Turnout fault diagnosis method and device, electronic equipment and medium |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8694836B2 (en) * | 2010-01-27 | 2014-04-08 | Telcordia Technologies, Inc. | Fault diagnosis employing probabilistic models and statistical learning |
| CN107769972B (en) * | 2017-10-25 | 2019-12-10 | 武汉大学 | Power communication network equipment fault prediction method based on improved LSTM |
| JP6714626B2 (en) * | 2018-02-26 | 2020-06-24 | 株式会社京三製作所 | Railway equipment condition determination device and railway equipment condition determination method |
| CN109739214B (en) * | 2019-01-14 | 2020-08-25 | 中国石油大学(华东) | Method for detecting intermittent faults in industrial process |
| CN110441629B (en) * | 2019-07-29 | 2022-02-15 | 交控科技股份有限公司 | Fault diagnosis method and device based on turnout switch machine operating current curve |
| CN110435716A (en) * | 2019-08-20 | 2019-11-12 | 湖南磁浮交通发展股份有限公司 | A kind of changing points control circuit |
| CN110927497B (en) * | 2019-12-09 | 2022-03-08 | 交控科技股份有限公司 | Point switch fault detection method and device |
| CN112947357B (en) * | 2021-01-24 | 2022-03-01 | 西北工业大学 | Aircraft intermittent fault diagnosis method based on double observers |
| CN113516310B (en) * | 2021-07-12 | 2022-04-19 | 内蒙古电力(集团)有限责任公司内蒙古电力科学研究院分公司 | Transformer fault early warning method and system |
| CN113627496A (en) * | 2021-07-27 | 2021-11-09 | 交控科技股份有限公司 | Method, device, electronic equipment and readable storage medium for predicting fault of turnout switch machine |
-
2022
- 2022-03-29 CN CN202210320814.8A patent/CN114692969B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107054410A (en) * | 2017-04-01 | 2017-08-18 | 广州地铁集团有限公司 | The intelligent diagnosis system and diagnostic method of point machine |
| CN111046583A (en) * | 2019-12-27 | 2020-04-21 | 中国铁道科学研究院集团有限公司通信信号研究所 | Switch machine fault diagnosis method based on DTW algorithm and ResNet network |
| CN111914320A (en) * | 2020-06-06 | 2020-11-10 | 同济大学 | No-sample turnout fault diagnosis method based on deep learning |
| CN112906230A (en) * | 2021-03-05 | 2021-06-04 | 北京全路通信信号研究设计院集团有限公司 | Turnout fault diagnosis method and device, electronic equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114692969A (en) | 2022-07-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3439241B1 (en) | Using machine learning to monitor link quality and predict link faults | |
| US11165802B2 (en) | Network security assessment using a network traffic parameter | |
| US10311356B2 (en) | Unsupervised behavior learning system and method for predicting performance anomalies in distributed computing infrastructures | |
| CN106598029B (en) | Train control on board equipment Reliability Prediction Method based on performance degradation | |
| US7636051B2 (en) | Status monitor apparatus | |
| CN106327033B (en) | Power system cascading failure analysis method based on Markov process | |
| US10809695B2 (en) | Information processing apparatus, machine learning device and system | |
| Ntalampiras et al. | A fault diagnosis system for interdependent critical infrastructures based on HMMs | |
| CN112433896B (en) | Method, device, equipment and storage medium for predicting server disk faults | |
| EP3968479A1 (en) | Systems and methods for automatic power topology discovery | |
| US10731906B2 (en) | Controlling a compressor of a refrigeration system by predicting when a temperature alarm is triggered | |
| CN111460692A (en) | Equipment residual life prediction method and system considering degradation rate mutual influence | |
| CN114692969B (en) | Switch machine fault prediction method, device, electronic equipment and storage medium | |
| CN115453356A (en) | Power equipment running state monitoring and analyzing method, system, terminal and medium | |
| WO2023272520A1 (en) | Anomaly detection method and apparatus for industrial equipment, electronic device, and storage medium | |
| Abeygunawardane et al. | A realistic maintenance model based on a new state diagram | |
| US20210182739A1 (en) | Ensemble learning model to identify conditions of electronic devices | |
| CN111611097A (en) | Fault detection method, device, equipment and storage medium | |
| CN114060228B (en) | Method and device for detecting abnormal torque jump of wind generating set and controller | |
| Mebratu et al. | AI assisted fan failure prediction using workload fingerprinting | |
| CN116457796A (en) | Scheme for tuning memory interfaces based on reinforcement learning | |
| CN112231127A (en) | Electronic device and method for analyzing reliability of equipment | |
| CN101854225B (en) | Method and system for detecting change of target state from position data stream | |
| Alayil et al. | Prediction of power outage during cyclone using machine learning | |
| Papataxiarhis et al. | Event correlation and forecasting over multivariate streaming sensor data |
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 |