A surge event identification method and identification system based on multi-parameter learning

Abstract

The present invention discloses a surge event recognition method and recognition system based on multi-parameter learning, belonging to the field of electronic digital data processing. The present invention marks and labels the data required for collection according to known surge event data and non-surge event data to serve as training data for supervised learning, builds a surge possibility prediction model through the training data for supervised learning, obtains real-time parameter data to predict the surge possibility, obtains the real-time parameter data and substitutes it into a surge event hazard value analysis strategy to perform surge event hazard analysis, substitutes the obtained surge event possibility prediction result and the corresponding surge event hazard analysis result into an abnormal warning strategy to perform abnormal warning, analyzes the surge historical events and then accurately recognizes the surge event, thereby improving the accuracy of surge event judgment.

Description

Technical Field

[0001] The invention belongs to the technical field of electronic digital data processing, in particular to a surge event recognition method and recognition system based on multi-parameter learning. Background Art

[0002] A surge protector is a device used to protect electrical equipment and systems from power surges. Power surges, also known as surges, refer to a sudden increase in voltage or current in a short period of time, which may cause damage to electrical equipment. Surge protectors can detect this sudden increase in voltage or current and take prompt action upon detecting such a change, directing the overvoltage or overcurrent to the ground. Existing technologies cannot accurately identify surge events by analyzing historical surge events, resulting in a low accuracy rate in determining surge events.

[0003] The above problems exist in the prior art.

[0004] For example, a Chinese patent with authorization announcement number CN117761444B discloses a life monitoring method and system for a surge protector. The method includes: installing sensors in the power system, monitoring power data in real time through the sensors, and monitoring key parameters of the surge protector through the sensors; sending the collected power data and key parameters to the edge device; preprocessing the received power data and key parameters through the edge device to obtain preprocessed monitoring data; transmitting the preprocessed monitoring data to the cloud platform; after receiving the monitoring data, the cloud platform stores the monitoring data in segments, and processes the segmented stored monitoring data through a deep learning algorithm. By installing sensors in the power system, the key parameters and power data of the surge protector can be monitored in real time, and the operating status of the surge protector can be understood in a timely manner;

[0005] At the same time, for example, a Chinese patent with publication number CN117929902A provides a composite surge protector detection method, surge protector and storage medium, including: in response to the initial sampling signal of the target surge protector, obtaining the initial resistance information of all gap units on at least two discharge branches; according to the preset inspection cycle or in response to the inspection sampling signal of the target surge protector, obtaining the current resistance information of all gap units; respectively comparing the relationship between the current resistance information and the initial resistance information and the preset resistance threshold, and judging the resistance state information of each gap unit; according to the resistance state information of each gap unit on each discharge branch, calculating the branch state information of each discharge branch; according to the branch state information of all discharge branches, determining the overall lightning protection performance of the target surge protector. This solution realizes the accurate rating of the surge protection performance of the composite surge protector, makes the application of surge protectors tend to be intelligent, ensures the safety of the power circuit, and avoids the waste of compliant surge protectors.

[0006] The above patents all have the problems raised by this background technology: the existing technology cannot accurately identify surge events by analyzing historical surge events, resulting in a low accuracy rate in surge event judgment. In order to solve these problems, this application designs a surge event identification method and identification system based on multi-parameter learning. Summary of the invention

[0007] In view of the shortcomings of the prior art, the present invention proposes a surge event identification method and identification system based on multi-parameter learning. The present invention obtains parameter data related to surge events, imports the parameter data into a data screening strategy to screen out required data from the parameter data, and marks and annotates the required data for collection according to known surge event data and non-surge event data to serve as training data for supervised learning. A surge possibility prediction model is constructed through the training data of supervised learning, and real-time parameter data is obtained to predict the surge possibility. The obtained real-time parameter data is substituted into a surge event hazard value analysis strategy to perform surge event hazard analysis, and the obtained surge event possibility prediction results and the corresponding surge event hazard analysis results are substituted into an abnormal warning strategy to perform abnormal warning. The surge event is accurately identified by analyzing historical surge events, thereby improving the accuracy of surge event judgment.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A method for identifying surge events based on multi-parameter learning includes the following specific steps:

[0010] S1. Obtain parameter data related to the surge event, import the parameter data into a data screening strategy, and screen out required data from the parameter data;

[0011] S2. Based on known surge event data and non-surge event data, mark and annotate the data required for collection to serve as training data for supervised learning;

[0012] S3. Build a surge possibility prediction model through supervised learning training data, and obtain real-time parameter data to predict the surge possibility;

[0013] S4, obtaining real-time parameter data and substituting it into the surge event hazard value analysis strategy to perform surge event hazard analysis;

[0014] S5. Substituting the obtained surge event possibility prediction result and the corresponding surge event hazard analysis result into the abnormal warning strategy for abnormal warning.

[0015] Specifically, S1 includes the following specific steps:

[0016] S11, obtaining the equipment voltage change curve and current change curve during the occurrence of historical surge events, and simultaneously obtaining lightning strike data, power equipment operation data and electromagnetic interference data before the surge event occurs, wherein the lightning strike data includes lightning current amplitude and lightning current duration data, the power equipment operation data includes power equipment circuit short-circuit time data, and the electromagnetic interference data includes electromagnetic interference intensity data;

[0017] S12, obtaining the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data before the historical surge event occurs, obtaining the equipment voltage change curve, equipment current change curve and power equipment circuit short-circuit time data during the equipment operation process, and importing them into the equipment impact probability calculation formula to calculate the equipment impact probability, wherein the equipment impact probability calculation formula in the i-th historical surge event is: Where m is the number of equipment operation data, T is the recording time of the surge event, Atj is the maximum slope of the jth data change curve of the equipment during the surge time, ts is the short-circuit time of the power equipment circuit, x ij t is the data at time t of the jth data change curve of the i-th surge event, where x jmax is the maximum value of the safety range of the jth data change curve, x j min is the minimum value of the safe range of the jth data change curve, |x ijt -x j max | indicates that when x ijt -x j max When it is less than 0, |x ijt -x j max | Take 0, if x ijt -x j max is greater than or equal to 0, then |x ijt-x j max |Get x ijt -x j max ;

[0018] S13, importing the lightning current amplitude and lightning current duration data of the historical surge event into the lightning current impact probability calculation formula to calculate the lightning current impact probability, wherein the lightning current impact probability calculation formula in the i-th historical surge event is: Among them, Si is the lightning current duration data of the i-th surge event in history, T1 is the set time standard value, R i is the magnitude of lightning current in the i-th surge event in history, R max is the maximum value of the lightning current safety range, R min It is the minimum value of the lightning current safety range;

[0019] S14, importing the electromagnetic interference intensity data before the historical surge event into the electromagnetic interference intensity probability calculation formula to calculate the electromagnetic interference intensity probability, wherein the electromagnetic interference intensity probability calculation formula in the i-th historical surge event is: L i is the electromagnetic interference intensity in the i-th historical surge event, L max is the maximum value of the electromagnetic interference intensity safety range, L min It is the minimum value of the safe range of electromagnetic interference intensity;

[0020] S15. Compare the acquired probability of equipment impact in historical surge events, the probability of lightning current impact in historical surge events, and the probability of electromagnetic interference intensity in historical surge events with the set probability threshold. If any one of the probability of equipment impact in historical surge events, the probability of lightning current impact in historical surge events, and the probability of electromagnetic interference intensity in historical surge events is greater than the set probability threshold, it is set as an abnormal event. If the probability of equipment impact in historical surge events, the probability of lightning current impact in historical surge events, and the probability of electromagnetic interference intensity in historical surge events are all less than the set probability threshold, it is set as a normal event.

[0021] Specifically, the step of marking and labeling the data required for collection according to the known surge event data and non-surge event data in S2 to serve as training data for supervised learning includes the following specific steps:

[0022] S21. Obtaining equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data corresponding to the abnormal event;

[0023] S22. Mark the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data corresponding to the collected abnormal events as training data for supervised learning.

[0024] Specifically, the method of constructing a surge possibility prediction model through supervised learning training data and obtaining real-time parameter data to predict the surge possibility includes the following specific steps:

[0025] S31. Obtain the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data, electromagnetic interference intensity data and surge event judgment data corresponding to the historical abnormal events, and construct a supervised learning neural network model with the input of equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data, and the output of surge event judgment data, i.e., surge possibility prediction model;

[0026] S32, the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short circuit time data, electromagnetic interference intensity data and surge event judgment data corresponding to the extracted historical abnormal events are used as 70% of the parameter training set and 30% of the parameter test set; the supervised learning neural network model is trained with the 70% parameter training set to obtain an initial supervised learning neural network model; the initial supervised learning neural network model is tested with the 30% parameter test set, and the initial supervised learning neural network model that meets the preset maximum historical surge event judgment accuracy is output as the supervised learning neural network model, wherein the output strategy formula of the specific neurons in the supervised learning neural network model is: in is the output of the s-term neuron in the x+1 layer, is the connection weight between the x-th layer neuron d and the x+1 layer s-item neuron, represents the output of neuron d in the xth layer, represents the bias of the linear relationship between the x-th layer neuron d and the x+1 layer s-term neuron, σ() represents the Sigmoid activation function, and D is the number of neurons in the x-th layer;

[0027] S33, importing the acquired real-time equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data, and electromagnetic interference intensity data into the constructed supervised learning neural network model to output the probability of surge event occurrence.

[0028] Specifically, the surge event hazard value analysis strategy in S4 includes the following specific steps:

[0029] S41, obtaining a real-time device voltage change curve, a device current change curve, and a device temperature change curve during device operation;

[0030] S42, substituting the obtained real-time device voltage change curve, device current change curve and device temperature change curve into the surge event hazard value calculation formula to calculate the surge event hazard value, wherein the surge event hazard value calculation formula is: Among them, E is the number of types of device data obtained, is the average value of the e-th device data type, T em The closest to the security range of the e-th device data type The value of T emax is the maximum value of the safety range of the e-th device data type, T emin It is the minimum value of the safety range of the e-th device data type.

[0031] Specifically, the specific steps of the abnormal warning strategy in S5 are as follows:

[0032] S51, obtaining the calculated surge event hazard value and surge event occurrence probability;

[0033] S52, substituting the obtained surge event hazard value and surge event occurrence probability of the customer into the surge event management value calculation formula to calculate the surge event management value, wherein the surge event management value calculation formula is: Y=a1M+(1-a1)W, wherein W is the output surge event occurrence probability, and a1 is the surge event hazard value ratio coefficient;

[0034] S53, if Y is greater than 0 and less than or equal to 0.2, no surge event warning is performed;

[0035] If Y is greater than 0.2 and less than or equal to 0.4, a level 3 surge event warning is issued;

[0036] If Y is greater than 0.4 and less than or equal to 0.6, a level 2 surge event warning is issued;

[0037] If Y is greater than 0.6, a level 1 surge event warning is issued.

[0038] It should be noted here that the surge event hazard value ratio coefficient and probability threshold are obtained as follows: obtain at least five thousand sets of equipment voltage change curves, equipment current change curves, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data, calculate and judge the losses caused by the surge event, substitute the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data into the surge event control value calculation formula to calculate the surge event control value, import the calculated surge event control value and the loss caused by the surge event into the fitting software, and output the surge event hazard value ratio coefficient and probability threshold that meet the highest judgment accuracy.

[0039] Specifically, a surge event identification system based on multi-parameter learning is implemented based on the above-mentioned surge event identification method based on multi-parameter learning, and specifically includes: a data acquisition module, a data screening module, a possibility prediction module, a surge event hazard analysis module, an abnormal warning module and a control module, wherein the data acquisition module is used to obtain parameter data related to surge events, import the parameter data into a data screening strategy to filter out required data from the parameter data, the data screening module is used to mark and annotate the data required for collection according to known surge event data and non-surge event data, so as to serve as training data for supervised learning, and the possibility prediction module is used to construct a surge possibility prediction model through the training data of supervised learning, and obtain real-time parameter data to predict the surge possibility.

[0040] Specifically, the surge event hazard analysis module is used to obtain real-time parameter data and substitute it into the surge event hazard value analysis strategy to perform surge event hazard analysis, and the abnormal warning module is used to substitute the obtained surge event possibility prediction results and the corresponding surge event hazard analysis results into the abnormal warning strategy to perform abnormal warning.

[0041] Specifically, the control module is used to control the operation of the data acquisition module, the data screening module, the possibility prediction module, the surge event hazard analysis module and the abnormal warning module.

[0042] Specifically, a surge event identification surge protector based on multi-parameter learning includes: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0043] The processor executes the above-mentioned surge event identification method based on multi-parameter learning by calling the computer program stored in the memory.

[0044] Specifically, a computer-readable storage medium stores instructions, and when the instructions are executed on a computer, the computer executes a surge event identification method based on multi-parameter learning as described above.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] The present invention obtains parameter data related to surge events, imports the parameter data into a data screening strategy to screen out required data from the parameter data, marks and annotates the data required for collection according to known surge event data and non-surge event data to serve as training data for supervised learning, builds a surge possibility prediction model through the training data for supervised learning, obtains real-time parameter data to predict surge possibility, obtains real-time parameter data and substitutes it into a surge event hazard value analysis strategy to perform surge event hazard analysis, substitutes the obtained surge event possibility prediction result and the corresponding surge event hazard analysis result into an abnormal warning strategy to perform abnormal warning, analyzes historical surge events and then accurately identifies surge events, thereby improving the accuracy of surge event judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A schematic diagram of a flow chart of a surge event identification method based on multi-parameter learning according to the present invention;

[0048] Figure 2 This is a schematic diagram of a specific flow chart of step S1 of a surge event identification method based on multi-parameter learning of the present invention;

[0049] Figure 3 This is a schematic diagram of a specific flow of step S2 of a surge event identification method based on multi-parameter learning of the present invention;

[0050] Figure 4 The present invention is a schematic diagram of the architecture of a surge event identification system based on multi-parameter learning. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0052] Example 1

[0053] See also Figure 1-Figure 3 The present invention provides an embodiment: a surge event recognition method based on multi-parameter learning, which comprises the following specific steps:

[0054] S1. Obtain parameter data related to the surge event, import the parameter data into a data screening strategy, and screen out required data from the parameter data;

[0055] S2. Based on known surge event data and non-surge event data, mark and annotate the data required for collection to serve as training data for supervised learning;

[0056] S3. Build a surge possibility prediction model through supervised learning training data, and obtain real-time parameter data to predict the surge possibility;

[0057] S4, obtaining real-time parameter data and substituting it into the surge event hazard value analysis strategy to perform surge event hazard analysis;

[0058] S5, substituting the obtained surge event possibility prediction result and the corresponding surge event hazard analysis result into the abnormal warning strategy for abnormal warning;

[0059] It should be noted that in this embodiment, S1 includes the following specific steps:

[0060] S11. Obtain the equipment voltage change curve and current change curve during the historical surge event, and simultaneously obtain the lightning strike data, power equipment operation data and electromagnetic interference data before the surge event occurs, wherein the lightning strike data includes lightning current amplitude and lightning current duration data, the power equipment operation data includes power equipment circuit short-circuit time data, and the electromagnetic interference data includes electromagnetic interference intensity data;

[0061] S12, obtaining the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data before the historical surge event occurs, obtaining the equipment voltage change curve, equipment current change curve and power equipment circuit short-circuit time data during the equipment operation process, and importing them into the equipment impact probability calculation formula to calculate the equipment impact probability, wherein the equipment impact probability calculation formula in the i-th historical surge event is: Where m is the number of equipment operation data, T is the recording time of the surge event, Atj is the maximum slope of the jth data change curve of the equipment during the surge time, ts is the short-circuit time of the power equipment circuit, x ijt is the data at time t of the jth data change curve of the i-th surge event, where x j max is the maximum value of the safety range of the jth data change curve, x j min is the minimum value of the safe range of the jth data change curve, |x ijt -x j max | indicates that when x ijt -x j max When it is less than 0, |x ijt -x j max | Take 0, if x ijt -x j maxis greater than or equal to 0, then |x ijt -x j max |Get x ijt -x j max ;

[0062] S13, importing the lightning current amplitude and lightning current duration data of the historical surge event into the lightning current impact probability calculation formula to calculate the lightning current impact probability, wherein the lightning current impact probability calculation formula in the i-th historical surge event is: Among them, Si is the lightning current duration data of the i-th surge event in history, T1 is the set time standard value, R i is the magnitude of lightning current in the i-th surge event in history, R max is the maximum value of the lightning current safety range, R min It is the minimum value of the lightning current safety range;

[0063] S14, importing the electromagnetic interference intensity data before the historical surge event into the electromagnetic interference intensity probability calculation formula to calculate the electromagnetic interference intensity probability, wherein the electromagnetic interference intensity probability calculation formula in the i-th historical surge event is: L i is the electromagnetic interference intensity in the i-th historical surge event, L max is the maximum value of the electromagnetic interference intensity safety range, L min It is the minimum value of the safe range of electromagnetic interference intensity;

[0064] S15. Compare the acquired probability of equipment impact in historical surge events, the probability of lightning current impact in historical surge events, and the probability of electromagnetic interference intensity in historical surge events with the set probability threshold. If any one of the probability of equipment impact in historical surge events, the probability of lightning current impact in historical surge events, and the probability of electromagnetic interference intensity in historical surge events is greater than the set probability threshold, it is set as an abnormal event. If the probability of equipment impact in historical surge events, the probability of lightning current impact in historical surge events, and the probability of electromagnetic interference intensity in historical surge events are all less than the set probability threshold, it is set as a normal event.

[0065] Specifically, in S2, based on known surge event data and non-surge event data, marking and labeling the data required for collection to serve as training data for supervised learning includes the following specific steps:

[0066] S21. Obtaining equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data corresponding to the abnormal event;

[0067] S22, marking the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short circuit time data and electromagnetic interference intensity data corresponding to the collected abnormal events as training data for supervised learning;

[0068] It should be noted that in this embodiment, constructing a surge possibility prediction model through supervised learning training data and obtaining real-time parameter data to predict the surge possibility includes the following specific steps:

[0069] S31. Obtain the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data, electromagnetic interference intensity data and surge event judgment data corresponding to the historical abnormal events, and construct a supervised learning neural network model with the input of equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data, and the output of surge event judgment data, i.e., surge possibility prediction model;

[0070] S32, the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short circuit time data, electromagnetic interference intensity data and surge event judgment data corresponding to the extracted historical abnormal events are used as 70% of the parameter training set and 30% of the parameter test set; the supervised learning neural network model is trained with the 70% parameter training set to obtain an initial supervised learning neural network model; the initial supervised learning neural network model is tested with the 30% parameter test set, and the initial supervised learning neural network model that meets the preset maximum historical surge event judgment accuracy is output as the supervised learning neural network model, wherein the output strategy formula of the specific neurons in the supervised learning neural network model is: in is the output of the s-term neuron in the x+1 layer, is the connection weight between the x-th layer neuron d and the x+1 layer s-item neuron, represents the output of neuron d in the xth layer, represents the bias of the linear relationship between the x-th layer neuron d and the x+1 layer s-term neuron, σ() represents the Sigmoid activation function, and D is the number of neurons in the x-th layer;

[0071] S33, importing the acquired real-time equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data, and electromagnetic interference intensity data into the constructed supervised learning neural network model to output the probability of surge event occurrence;

[0072] It should be noted that in this embodiment, the surge event risk value analysis strategy in S4 includes the following specific steps:

[0073] S41, obtaining a real-time device voltage change curve, a device current change curve, and a device temperature change curve during device operation;

[0074] S42, substituting the obtained real-time device voltage change curve, device current change curve and device temperature change curve into the surge event hazard value calculation formula to calculate the surge event hazard value, wherein the surge event hazard value calculation formula is: Among them, E is the number of types of device data obtained, is the average value of the e-th device data type, T em The closest to the security range of the e-th device data type The value of T emax is the maximum value of the safety range of the e-th device data type, T emin is the minimum value of the safety range of the e-th device data type;

[0075] It should be noted that in this embodiment, the specific steps of the abnormal warning strategy in S5 are as follows:

[0076] S51, obtaining the calculated surge event hazard value and surge event occurrence probability;

[0077] S52, substituting the obtained surge event hazard value and surge event occurrence probability of the customer into the surge event management value calculation formula to calculate the surge event management value, wherein the surge event management value calculation formula is: Y=a1M+(1-a1)W, wherein W is the output surge event occurrence probability, and a1 is the surge event hazard value ratio coefficient;

[0078] S53, if Y is greater than 0 and less than or equal to 0.2, no surge event warning is performed;

[0079] If Y is greater than 0.2 and less than or equal to 0.4, a level 3 surge event warning is issued;

[0080] If Y is greater than 0.4 and less than or equal to 0.6, a level 2 surge event warning is issued;

[0081] If Y is greater than 0.6, a level 1 surge event warning is issued;

[0082] It should be noted that in this embodiment, the surge event hazard value ratio coefficient and probability threshold are obtained as follows: at least five thousand sets of equipment voltage change curves, equipment current change curves, lightning current amplitudes, lightning current duration data, power equipment circuit short-circuit time data, and electromagnetic interference intensity data are obtained, and the losses caused by the surge event are calculated and judged, and the equipment voltage change curves, equipment current change curves, lightning current amplitudes, lightning current duration data, power equipment circuit short-circuit time data, and electromagnetic interference intensity data are substituted into the surge event control value calculation formula to calculate the surge event control value, and the calculated surge event control value and the loss caused by the surge event are imported into the fitting software, and the surge event hazard value ratio coefficient and probability threshold that meet the highest judgment accuracy are output;

[0083] Through this embodiment, it is possible to achieve the following: obtaining parameter data related to surge events, importing the parameter data into a data screening strategy to screen out required data from the parameter data, marking and annotating the data required for collection based on known surge event data and non-surge event data to serve as training data for supervised learning, building a surge possibility prediction model through the training data for supervised learning, obtaining real-time parameter data to predict surge possibility, obtaining real-time parameter data and substituting it into a surge event hazard value analysis strategy to perform surge event hazard analysis, substituting the obtained surge event possibility prediction results and the corresponding surge event hazard analysis results into an abnormal warning strategy to perform abnormal warning, and accurately identifying surge events through analysis of historical surge events, thereby improving the accuracy of surge event judgment.

[0084] Example 2

[0085] like Figure 4As shown, a surge event recognition system based on multi-parameter learning is implemented based on the above-mentioned surge event recognition method based on multi-parameter learning, and specifically includes: a data acquisition module, a data screening module, a possibility prediction module, a surge event risk analysis module, an abnormal warning module and a control module, wherein the data acquisition module is used to obtain parameter data related to the surge event, import the parameter data into the data screening strategy to filter out the required data from the parameter data, the data screening module is used to mark and annotate the data required for collection according to the known surge event data and non-surge event data, so as to serve as the training data for supervised learning, the possibility prediction module is used to construct a surge possibility prediction model through the training data of supervised learning, and obtain real-time parameter data to predict the surge possibility; the surge event risk analysis module is used to obtain real-time parameter data and substitute it into the surge event risk value analysis strategy to perform surge event risk analysis, and the abnormal warning module is used to substitute the obtained surge event possibility prediction result and the corresponding surge event risk analysis result into the abnormal warning strategy to perform abnormal warning; the control module is used to control the operation of the data acquisition module, the data screening module, the possibility prediction module, the surge event risk analysis module and the abnormal warning module.

[0086] Example 3

[0087] This embodiment provides a surge event identification surge protector based on multi-parameter learning, comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0088] The processor executes the above-mentioned surge event identification method based on multi-parameter learning by calling the computer program stored in the memory.

[0089] The surge event recognition surge protector based on multi-parameter learning may have relatively large differences due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) and one or more memories, wherein at least one computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement a surge event recognition method based on multi-parameter learning provided in the above method embodiment. The surge event recognition surge protector based on multi-parameter learning can also include other components for realizing the functions of the device, for example, the surge event recognition surge protector based on multi-parameter learning can also have components such as wired or wireless network interfaces and input and output interfaces to input and output data. This embodiment will not be described in detail here.

[0090] Example 4

[0091] This embodiment provides a computer-readable storage medium having a rewritable computer program stored thereon;

[0092] When the computer program runs on a computer device, the computer device is enabled to execute the above-mentioned surge event identification method based on multi-parameter learning.

[0093] For example, the computer readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0094] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0095] It should be understood that determining B based on A does not mean determining B only based on A. B can also be determined based on A and / or other information.

[0096] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When a computer instruction or computer program is loaded or executed on a computer, a process or function according to an embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center through a wired network or / and a wireless network. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0097] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0098] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

Claims

1. A surge event recognition method based on multi-parameter learning, characterized in that: It includes the following specific steps: S1. Obtain parameter data related to the surge event, import the parameter data into a data screening strategy, and screen out required data from the parameter data; S2. Based on known surge event data and non-surge event data, mark and annotate the data required for collection to serve as training data for supervised learning; S3. Build a surge possibility prediction model through supervised learning training data, and obtain real-time parameter data to predict the surge possibility; S4, obtaining real-time parameter data and substituting it into the surge event hazard value analysis strategy to perform surge event hazard analysis; S5, substituting the obtained surge event possibility prediction result and the corresponding surge event hazard analysis result into the abnormal warning strategy for abnormal warning; S1 includes the following specific steps: S11, obtaining the equipment voltage change curve and current change curve during the occurrence of historical surge events, and simultaneously obtaining lightning strike data, power equipment operation data and electromagnetic interference data before the surge event occurs, wherein the lightning strike data includes lightning current amplitude and lightning current duration data, the power equipment operation data includes power equipment circuit short-circuit time data, and the electromagnetic interference data includes electromagnetic interference intensity data; S12, obtaining the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data before the historical surge event occurs, obtaining the equipment voltage change curve, equipment current change curve and power equipment circuit short-circuit time data during the equipment operation process, and importing them into the equipment impact probability calculation formula to calculate the equipment impact probability, wherein the equipment impact probability calculation formula in the i-th historical surge event is: Where m is the number of equipment operation data, T is the recording time of the surge event, Atj is the maximum slope of the jth data change curve of the equipment during the surge time, ts is the short-circuit time of the power equipment circuit, x ijt is the data at time t of the jth data change curve of the i-th surge event, where x jmax is the maximum value of the safety range of the jth data change curve, x jmin is the minimum value of the safe range of the jth data change curve, |x ijt -x jmax | indicates that when x ijt -x jmax When it is less than 0, |x ijt -x jmax | Take 0, if x ijt -x jmax is greater than or equal to 0, then |x ijt -x jmax |Get x ijt -x jmax ; S13, importing the lightning current amplitude and lightning current duration data of the historical surge event into the lightning current impact probability calculation formula to calculate the lightning current impact probability, wherein the lightning current impact probability calculation formula in the i-th historical surge event is: Among them, Si is the lightning current duration data of the i-th surge event in history, T1 is the set time standard value, R i is the magnitude of lightning current in the i-th surge event in history, R max is the maximum value of the lightning current safety range, R min It is the minimum value of the lightning current safety range; S14, importing the electromagnetic interference intensity data before the historical surge event into the electromagnetic interference intensity probability calculation formula to calculate the electromagnetic interference intensity probability, wherein the electromagnetic interference intensity probability calculation formula in the i-th historical surge event is: L i is the electromagnetic interference intensity in the i-th historical surge event, L max is the maximum value of the electromagnetic interference intensity safety range, L min It is the minimum value of the safe range of electromagnetic interference intensity; S15, compare the acquired probability of equipment impact in historical surge events, the probability of lightning current impact in historical surge events, and the probability of electromagnetic interference intensity in historical surge events with the set probability threshold; if any one of the probability of equipment impact in historical surge events, the probability of lightning current impact in historical surge events, and the probability of electromagnetic interference intensity in historical surge events is greater than the set probability threshold, it is set as an abnormal event; if the probability of equipment impact in historical surge events, the probability of lightning current impact in historical surge events, and the probability of electromagnetic interference intensity in historical surge events are all less than the set probability threshold, it is set as a normal event; the step of marking and labeling the data required for collection according to the known surge event data and non-surge event data in S2 as training data for supervised learning includes the following specific steps: S21. Obtaining equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data corresponding to the abnormal event; S22, marking the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short circuit time data and electromagnetic interference intensity data corresponding to the collected abnormal events as training data for supervised learning; said constructing a surge possibility prediction model through the supervised learning training data, and obtaining real-time parameter data to predict the surge possibility includes the following specific steps: S31. Obtain the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data, electromagnetic interference intensity data and surge event judgment data corresponding to the historical abnormal events, and construct a supervised learning neural network model with the input of equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short-circuit time data and electromagnetic interference intensity data, and the output of surge event judgment data, i.e., surge possibility prediction model; S32, the equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short circuit time data, electromagnetic interference intensity data and surge event judgment data corresponding to the extracted historical abnormal events are used as 70% of the parameter training set and 30% of the parameter test set; the supervised learning neural network model is trained with the 70% parameter training set to obtain an initial supervised learning neural network model; the initial supervised learning neural network model is tested with the 30% parameter test set, and the initial supervised learning neural network model that meets the preset maximum historical surge event judgment accuracy is output as the supervised learning neural network model, wherein the output strategy formula of the specific neurons in the supervised learning neural network model is: in is the output of the s-term neuron in the x+1 layer, is the connection weight between the x-th layer neuron d and the x+1 layer s-item neuron, represents the output of neuron d in the xth layer, represents the bias of the linear relationship between the x-th layer neuron d and the x+1 layer s-term neuron, σ() represents the Sigmoid activation function, and D is the number of neurons in the x-th layer; S33, importing the acquired real-time equipment voltage change curve, equipment current change curve, lightning current amplitude, lightning current duration data, power equipment circuit short circuit time data, and electromagnetic interference intensity data into the constructed supervised learning neural network model to output the probability of surge event occurrence; the surge event hazard value analysis strategy in S4 includes the following specific steps: S41, obtaining a real-time device voltage change curve, a device current change curve, and a device temperature change curve during device operation; S42, substituting the obtained real-time device voltage change curve, device current change curve and device temperature change curve into the surge event hazard value calculation formula to calculate the surge event hazard value, wherein the surge event hazard value calculation formula is: Among them, E is the number of types of device data obtained, is the average value of the e-th device data type, T em The closest to the security range of the e-th device data type The value of T emax is the maximum value of the safety range of the e-th device data type, T emin is the minimum value of the safety range of the e-th device data type; the specific steps of the abnormal warning strategy in S5 are as follows: S51, obtaining the calculated surge event hazard value and surge event occurrence probability; S52, substituting the obtained surge event hazard value and surge event occurrence probability of the customer into the surge event management value calculation formula to calculate the surge event management value, wherein the surge event management value calculation formula is: Y=a1M+(1-a1)W, wherein W is the output surge event occurrence probability, and a1 is the surge event hazard value ratio coefficient; S53, if Y is greater than 0 and less than or equal to 0.2, no surge event warning is performed; If Y is greater than 0.2 and less than or equal to 0.4, a level 3 surge event warning is issued; If Y is greater than 0.4 and less than or equal to 0.6, a level 2 surge event warning is issued; If Y is greater than 0.6, a level 1 surge event warning is issued.

2. A surge event recognition system based on multi-parameter learning, which is implemented based on a surge event recognition method based on multi-parameter learning as claimed in claim 1, characterized in that: It specifically includes: a data acquisition module, a data screening module, a possibility prediction module, a surge event hazard analysis module, an abnormal warning module and a control module, wherein the data acquisition module is used to obtain parameter data related to surge events, import the parameter data into a data screening strategy to filter out required data from the parameter data, the data screening module is used to mark and annotate the data required for collection according to known surge event data and non-surge event data, so as to serve as training data for supervised learning, and the possibility prediction module is used to construct a surge possibility prediction model through the training data of supervised learning, and obtain real-time parameter data to predict the surge possibility.

3. A surge event identification system based on multi-parameter learning as described in claim 2, characterized in that: The surge event hazard analysis module is used to obtain real-time parameter data and substitute it into the surge event hazard value analysis strategy to perform surge event hazard analysis. The abnormal warning module is used to substitute the obtained surge event possibility prediction results and the corresponding surge event hazard analysis results into the abnormal warning strategy to perform abnormal warning.

4. A surge event identification system based on multi-parameter learning as claimed in claim 3, characterized in that: The control module is used to control the operation of the data acquisition module, the data screening module, the possibility prediction module, the surge event hazard analysis module and the abnormal warning module.

5. A computer-readable storage medium, characterized in that: Instructions are stored, and when the instructions are executed on a computer, the computer is caused to execute the surge event identification method based on multi-parameter learning as claimed in claim 1.

Images