Fault arc intelligent detection method, device and system based on closed environment

Abstract

The invention discloses a fault arc intelligent detection method, device and system based on a closed environment, and relates to the field of electrical safety monitoring, and the method comprises the steps: obtaining an ultraviolet light signal, an ultrasonic signal and an electromagnetic wave signal; preprocessing the ultraviolet light signal to extract an ultraviolet light characteristic signal, and identifying a fault arc according to the ultraviolet light characteristic signal to generate an ultraviolet light identification result; preprocessing the ultrasonic signal to extract an ultrasonic characteristic signal, and identifying a fault arc according to the ultrasonic characteristic signal to generate an ultrasonic identification result; preprocessing the electromagnetic wave signal to extract an electromagnetic wave characteristic signal, and identifying a fault arc according to the electromagnetic wave characteristic signal to generate an electromagnetic wave identification result; and generating a fault arc detection result according to the ultraviolet light identification result, the ultrasonic wave identification result and the electromagnetic wave identification result. According to the invention, the fault arc can be detected by fusing the characteristics of three high-frequency signals of ultraviolet light, ultrasonic wave and electromagnetic wave in a closed environment, so that the problem of leak detection of the fault arc is solved, and the loss of power distribution infrastructures caused by the fault arc is avoided.

Description

Technical Field

[0001] The present invention relates to the field of electrical safety monitoring, and in particular to a method, device and system for intelligent detection of arc faults based on a closed environment. Background Art

[0002] Inside the distribution box, due to poor maintenance, loose machine screws, poor contact, electrical aging, improper assembly and other reasons, local fault arcs are likely to occur and gradually develop into electrical fires, leading to the burning of substation equipment and distribution facilities, and even casualties.

[0003] At present, the core technology of the technical solutions and products for detecting fault arcs is mainly based on analyzing faults based on electrical signals. However, in a complex circuit environment, sometimes the current and voltage of the arc are similar to the electrical signals under normal working conditions, which can easily lead to missed detection. At the same time, the detection method based on electrical signals is difficult to detect weak arcs, which can also lead to missed detection. Therefore, analyzing faults based on electrical signals often leads to missed detection, which causes electrical faults to spread.

[0004] Based on the currently disclosed electrical accident analysis and fault arc detection solutions, relying solely on electrical signals to detect electrical anomalies cannot achieve good fault detection results and is prone to missed detections. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a method, device and system for intelligent detection of arc faults based on a closed environment, which can effectively solve the problem of missed detection of arc faults and avoid losses of power distribution infrastructure caused by arc faults.

[0006] In order to solve the above technical problems, the present invention provides an intelligent fault arc detection method based on a closed environment, including: obtaining ultraviolet light signals, ultrasonic signals and electromagnetic wave signals; preprocessing the ultraviolet light signals to extract ultraviolet light characteristic signals, and identifying the fault arc according to the ultraviolet light characteristic signals to generate ultraviolet light identification results; preprocessing the ultrasonic signals to extract ultrasonic characteristic signals, and identifying the fault arc according to the ultrasonic characteristic signals to generate ultrasonic identification results; preprocessing the electromagnetic wave signals to extract electromagnetic wave characteristic signals, and identifying the fault arc according to the electromagnetic wave characteristic signals to generate electromagnetic wave identification results; generating a fault arc detection result according to the ultraviolet light identification results, ultrasonic identification results and electromagnetic wave identification results.

[0007] As an improvement of the above-mentioned scheme, the step of preprocessing the ultraviolet light signal to extract the ultraviolet light characteristic signal includes: dividing the ultraviolet light signal according to a preset segmentation period, counting the number of ultraviolet photons in each segmentation period respectively, and taking the number of ultraviolet photons as the ultraviolet light characteristic signal; the step of preprocessing the ultrasonic signal to extract the ultrasonic characteristic signal includes: extracting the signal peak and point information from the ultrasonic signal as the ultrasonic characteristic signal; the step of preprocessing the electromagnetic wave signal to extract the electromagnetic wave characteristic signal includes: normalizing and standardizing the electromagnetic wave signal, and extracting the mean information, peak information and kurtosis information from the normalized and standardized electromagnetic wave signal as the electromagnetic wave characteristic signal.

[0008] As an improvement of the above-mentioned scheme, the step of identifying the fault arc according to the ultraviolet light characteristic signal to generate an ultraviolet light identification result includes: comparing each group of ultraviolet light characteristic signals within the preset time with the preset photon number interval to determine whether an arc occurs. If the ultraviolet light characteristic signal is within the photon number interval, it indicates that an arc occurs; if the ultraviolet light characteristic signal is outside the photon number interval, it indicates that no arc occurs; comparing the number of arc occurrences within the preset time with the preset arc number interval. If the number of arc occurrences within the preset time is greater than the arc number interval, the ultraviolet light identification result is "fault arc"; if the number of arc occurrences within the preset time is within the arc number interval, the ultraviolet light identification result is "suspected fault arc"; if the number of arc occurrences within the preset time is within the arc number interval, the ultraviolet light identification result is "suspected fault arc"; if the number of arc occurrences within the preset time is less than the arc number interval, the ultraviolet light identification result is "non-fault arc".

[0009] As an improvement of the above scheme, the step of identifying the fault arc according to the ultrasonic characteristic signal to generate an ultrasonic identification result includes: comparing the peak value of each group of ultrasonic characteristic signals within the preset time with the preset peak threshold value to determine whether an arc occurs. If the peak value of the ultrasonic characteristic signal is greater than the peak threshold value, it indicates that an arc occurs; if the peak value of the ultrasonic characteristic signal is less than or equal to the peak threshold value, it indicates that no arc occurs; comparing the number of arc occurrences within the preset time with the preset arc number interval. If the number of arc occurrences within the preset time is greater than the arc number interval, the ultrasonic identification result is "fault arc"; if the number of arc occurrences within the preset time is within the arc number interval, the ultrasonic identification result is "suspected fault arc"; if the number of arc occurrences within the preset time is within the arc number interval, the ultrasonic identification result is "suspected fault arc"; if the number of arc occurrences within the preset time is less than the arc number interval, the ultrasonic identification result is "non-fault arc".

[0010] As an improvement of the above solution, the step of identifying the fault arc according to the electromagnetic wave characteristic signal to generate an electromagnetic wave identification result includes: inputting the electromagnetic wave characteristic signal into a pre-trained arc electromagnetic wave identification model to generate an electromagnetic wave identification result.

[0011] As an improvement of the above-mentioned scheme, the training method of the arc electromagnetic wave recognition model includes: obtaining historical electromagnetic wave signals; extracting feature information from the historical electromagnetic wave signals, the feature information including mean information, peak information and kurtosis information; constructing feature samples according to the feature information and historical fault categories; dividing the feature samples into a training set and a test set; inputting the training set into a SVM model to generate an arc electromagnetic wave recognition model; and inputting the test set into the arc electromagnetic wave recognition model to optimize the arc electromagnetic wave recognition model.

[0012] As an improvement of the above-mentioned scheme, the step of generating a fault arc detection result based on the ultraviolet light recognition result, the ultrasonic recognition result and the electromagnetic wave recognition result includes: when any two or three of the ultraviolet light recognition result, the ultrasonic recognition result and the electromagnetic wave recognition result are "fault arc", the fault arc detection result is "fault arc"; otherwise, the fault arc detection result is "non-fault arc".

[0013] As an improvement of the above-mentioned scheme, the intelligent detection method of fault arc based on a closed environment also includes: storing the intelligent recognition results in the detection process, the intelligent recognition results including the ultraviolet light recognition results, ultrasonic recognition results, electromagnetic wave recognition results and fault arc detection results; acquiring and storing electrical system operation data; sending the intelligent recognition results and electrical system operation data to a host computer.

[0014] Correspondingly, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein the processor implements the steps of the above-mentioned closed environment-based intelligent fault arc detection method when executing the computer program.

[0015] Correspondingly, the present invention also provides an intelligent fault arc detection system based on a closed environment, comprising an ultraviolet light collection device, an ultrasonic collection device, an electromagnetic wave collection device, a host computer and the above-mentioned computer device, wherein the computer device is respectively connected to the ultraviolet light collection device, the ultrasonic collection device, the electromagnetic wave collection device and the host computer; the ultraviolet light collection device is used to collect and send the ultraviolet light signal to the computer device; the ultrasonic collection device is used to collect and send the ultrasonic signal to the computer device; the electromagnetic wave collection device is used to collect and send the electromagnetic wave signal to the computer device.

[0016] The implementation of the present invention has the following beneficial effects:

[0017] The present invention detects fault arcs by integrating the characteristics of three high-frequency signals, namely ultraviolet light, ultrasonic waves and electromagnetic waves, in a closed environment, thereby realizing multi-modal arc intelligent recognition, which can effectively solve the problem of missed detection of fault arcs and avoid losses of power distribution infrastructure caused by fault arcs.

[0018] At the same time, the present invention uses the number of ultraviolet photons as the ultraviolet light characteristic signal, the signal peak and point information as the ultrasonic characteristic signal, the mean information, peak information and kurtosis information as the electromagnetic wave characteristic signal, and fundamentally conducts in-depth analysis of ultraviolet light, ultrasonic wave and electromagnetic wave, thereby greatly improving the recognition accuracy of fault arc in a closed environment.

[0019] Furthermore, the present invention can upload the intelligent recognition results and electrical system operation data in the detection process to the host computer for publication, thereby realizing abnormal alarm and reminding the user to perform maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a flow chart of a first embodiment of a method for intelligently detecting arc faults in a closed environment according to the present invention;

[0021] Figure 2 It is a flow chart of a second embodiment of the fault arc intelligent detection method based on a closed environment of the present invention;

[0022] Figure 3 It is a schematic diagram of the structure of an embodiment of the intelligent arc fault detection system based on a closed environment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings.

[0024] The present invention is based on the closed environment fault arc intelligent detection method, which is suitable for electrical safety monitoring in a relatively closed environment, such as fault arc monitoring in a distribution room or a distribution box.

[0025] See also Figure 1 , Figure 1 The flowchart of the first embodiment of the method for intelligent arc fault detection based on a closed environment of the present invention is shown, which includes:

[0026] S101, acquiring ultraviolet light signals, ultrasonic wave signals and electromagnetic wave signals;

[0027] In practical applications, the number of ultraviolet photons can be recorded in real time through a photon detection and counting device with a specific wavelength filter, and the number of ultraviolet photons can be sent to a computer device as an ultraviolet light signal. The photon detection and counting device includes a single photon counting detector, a multi-channel picosecond event timer, and a time-to-digital conversion module. The wavelength range of the filter is 180um to 400um, so that the ultraviolet light generated by the arc radiation can pass smoothly; therefore, through this high-precision time measurement and counting technology and filtering function, the number of ultraviolet photons generated when the arc occurs can be detected and recorded.

[0028] At the same time, ultrasonic signals can be collected by ultrasonic acquisition equipment and sent to computer equipment. In the present invention, the signal frequency range sensed by the ultrasonic acquisition equipment is 0Hz~80KHz; the ultrasonic acquisition equipment includes an ultrasonic fiber sensor, an ultrasonic acquisition box, an acquisition box adapter power supply and a sensor power supply, wherein the acquisition box adapter power supply provides power for the ultrasonic acquisition box, and the sensor power supply provides power for the ultrasonic fiber sensor; through the bandpass filter module in the ultrasonic acquisition box, 30KHz~40KHz ultrasonic signals can pass through, and the collected ultrasonic signals are saved in x lsx format, and the ultrasonic signal content includes each sampling point and the corresponding amplitude information.

[0029] In addition, electromagnetic wave signals can be collected by electromagnetic wave collection equipment and sent to computer equipment. In the present invention, the signal frequency range sensed by the electromagnetic wave collection equipment is 0Hz to 1GHz, and the power distribution of electromagnetic wave signals at different frequencies is recorded (in dBm), and then the data with a frequency range of 1MHz to 1GHz is separated from the electromagnetic wave signal and saved as data in csv format.

[0030] S102, preprocessing the ultraviolet light signal to extract an ultraviolet light characteristic signal, and identifying the fault arc according to the ultraviolet light characteristic signal to generate an ultraviolet light identification result;

[0031] Accordingly, the step of preprocessing the ultraviolet light signal to extract the ultraviolet light characteristic signal includes: dividing the ultraviolet light signal according to the preset segmentation period, counting the number of ultraviolet photons in each segmentation period, and taking the number of ultraviolet photons as the ultraviolet light characteristic signal;

[0032] For example, the ultraviolet light signal is divided into sections with a period of 10 ms, and the number of ultraviolet photons is counted every 10 ms, that is, 100 groups of ultraviolet photons can be counted every 1 second as ultraviolet light characteristic signals.

[0033] Meanwhile, the steps of identifying the fault arc according to the ultraviolet characteristic signal to generate an ultraviolet identification result include:

[0034] (1) Compare each group of ultraviolet light characteristic signals within a preset time with a preset photon quantity interval to determine whether an arc occurs, wherein:

[0035] If the UV characteristic signal is within the photon number range, it indicates that an arc has occurred;

[0036] If the UV characteristic signal is outside the photon number range, it means that no arc has occurred;

[0037] It should be noted that, according to the experiment of predicting fault arc by ultraviolet photons, the number of ultraviolet photons emitted by one arc is in the range of [20, 120]; therefore, the present invention sets the photon number interval to [20, 120] to determine whether an arc occurs.

[0038] That is to say, the present invention uses 1s as the preset time, and inputs 100 sets of data (i.e., the number of ultraviolet photons) counted per second into the arc ultraviolet light recognition model in a certain organizational form, so as to facilitate the subsequent judgment of whether an arc has occurred based on the set photon number interval.

[0039] (2) Compare the number of arc occurrences within a preset time with a preset arc number interval, where:

[0040] If the number of arc occurrences within the preset time is greater than the arc frequency interval, the UV light identification result is "fault arc";

[0041] If the number of arc occurrences within the preset time is within the arc frequency range, the UV light identification result is "suspected fault arc";

[0042] If the number of arc occurrences within the preset time is less than the arc frequency interval, the ultraviolet light identification result is "non-fault arc".

[0043] For example, the arc frequency interval is [1, 9], and there are 100 groups of ultraviolet light characteristic signals within 1s. If 10 or more detection results are "arc" within 1s, the ultraviolet light identification result is "fault arc"; if 1 to 9 groups of detection results are "arc" within 1s, the ultraviolet light identification result is "suspected fault arc"; in other cases, the ultraviolet light identification result is "non-fault arc".

[0044] Therefore, through step S102, the acquired ultraviolet light signal can be statistically analyzed and a prediction result can be calculated to determine whether a fault arc has occurred.

[0045] S103, preprocessing the ultrasonic signal to extract an ultrasonic characteristic signal, and identifying the fault arc according to the ultrasonic characteristic signal to generate an ultrasonic identification result;

[0046] Accordingly, the signal peak value and point information can be extracted from the ultrasonic signal as an ultrasonic characteristic signal; generally, the peak value and the corresponding point information of the ultrasonic signal can be extracted from the acquired ultrasonic signal through digital signal processing tools, and the peak value and point information can be saved in a temporary array as an ultrasonic characteristic signal; among them, the digital signal processing tools include signal processing function tools in programming function libraries such as Python, Matlab, and C++.

[0047] Meanwhile, the steps of identifying the fault arc according to the ultrasonic characteristic signal to generate an ultrasonic identification result include:

[0048] (1) The peak value of each group of ultrasonic characteristic signals within a preset time is compared with a preset peak threshold value to determine whether an arc occurs, wherein:

[0049] If the peak value of the ultrasonic characteristic signal is greater than the peak threshold, it indicates that an arc has occurred;

[0050] If the peak value of the ultrasonic characteristic signal is less than or equal to the peak value threshold, it means that no arc occurs;

[0051] (2) Compare the number of arc occurrences within a preset time with a preset arc number interval, where:

[0052] If the number of arc occurrences within the preset time is greater than the arc frequency interval, the ultrasonic identification result is "fault arc";

[0053] If the number of arc occurrences within the preset time is within the arc frequency range, the ultrasonic identification result is "suspected fault arc";

[0054] If the number of arc occurrences within the preset time is less than the arc frequency interval, the ultrasonic identification result is "non-fault arc".

[0055] For example, the arc frequency interval is [1, 9]. If 10 or more detection results are "arc" within 1s, the ultrasonic identification result is "fault arc"; if 1 to 9 detection results are "arc" within 1s, the ultrasonic identification result is "suspected fault arc"; in other cases, the ultrasonic identification result is "non-fault arc".

[0056] S104, preprocessing the electromagnetic wave signal to extract the electromagnetic wave characteristic signal, and identifying the fault arc according to the electromagnetic wave characteristic signal to generate an electromagnetic wave identification result;

[0057] Accordingly, the step of preprocessing the electromagnetic wave signal to extract the electromagnetic wave characteristic signal includes:

[0058] (1) Normalize and standardize the electromagnetic wave signals to ensure that each electromagnetic wave signal has a similar scale;

[0059] (2) Extract mean information, peak information and kurtosis information from the electromagnetic wave signal after normalization and standardization as the electromagnetic wave characteristic signal.

[0060] It should be noted that the mean reflects the overall strength level of the signal. The present invention calculates the power average value of all frequency points in each cycle as the mean; the peak represents the maximum value of the signal strength. The present invention finds the maximum value of the signal in each cycle as the peak value through the index function in the signal processing tool; the kurtosis describes the sharpness of the signal distribution. The present invention calculates the kurtosis of the signal in each cycle through the kurtosis calculation formula.

[0061] At the same time, the present invention inputs the electromagnetic wave characteristic signal into a pre-trained arc electromagnetic wave recognition model to generate an electromagnetic wave recognition result.

[0062] Furthermore, the training method of the arc electromagnetic wave recognition model includes:

[0063] (1) Obtain historical electromagnetic wave signals;

[0064] (2) Extract characteristic information from historical electromagnetic wave signals;

[0065] Feature information includes mean information, peak information and kurtosis information;

[0066] (3) Construct feature samples based on feature information and historical fault categories;

[0067] The present invention can use two-dimensional data to represent feature information, and also use a label array to mark the category of each sample.

[0068] (4) Divide the feature samples into training set and test set;

[0069] The present invention divides all feature samples into a training set and a test set according to a ratio of 8:2.

[0070] (5) inputting the training set into the SVM model to generate an arc electromagnetic wave recognition model;

[0071] The present invention utilizes a support vector machine (SVM) machine learning algorithm for training, and sets basic model parameters such as a kernel function and a regularization parameter of the SVM model to obtain an arc electromagnetic wave recognition model.

[0072] (6) Input the test set into the arc electromagnetic wave recognition model to optimize the arc electromagnetic wave recognition model.

[0073] If the performance of the arc electromagnetic wave identification model is not good enough, the optimal parameters can be automatically found by adjusting the SVM parameters or using methods such as grid search and random search.

[0074] Therefore, the present invention uses the number of ultraviolet photons as the ultraviolet light characteristic signal, the signal peak and point information as the ultrasonic characteristic signal, and the mean information, peak information and kurtosis information as the electromagnetic wave characteristic signal, and fundamentally conducts an in-depth analysis of ultraviolet light, ultrasonic wave and electromagnetic wave, thereby greatly improving the identification accuracy of fault arc in a closed environment.

[0075] S105, generating a fault arc detection result according to the ultraviolet light recognition result, the ultrasonic wave recognition result and the electromagnetic wave recognition result.

[0076] Accordingly, the ultraviolet light recognition results, ultrasonic wave recognition results and electromagnetic wave recognition results can be integrated to generate the fault arc detection results. The specific steps include:

[0077] (1) When any two or three of the ultraviolet light recognition results, ultrasonic wave recognition results and electromagnetic wave recognition results are "fault arc", the fault arc detection result is "fault arc";

[0078] (2) Otherwise, the fault arc detection result is "non-fault arc".

[0079] Therefore, the present invention can effectively integrate the ultraviolet light recognition results, ultrasonic wave recognition results and electromagnetic wave recognition results to realize multi-modal arc intelligent recognition, which greatly improves the recognition accuracy of fault arc in a closed environment.

[0080] See also Figure 2 , Figure 2 The flowchart of the second embodiment of the method for intelligent arc fault detection based on a closed environment of the present invention is shown, which includes:

[0081] S201, acquiring ultraviolet light signals, ultrasonic wave signals and electromagnetic wave signals;

[0082] S202, preprocessing the ultraviolet light signal to extract an ultraviolet light characteristic signal, and identifying the fault arc according to the ultraviolet light characteristic signal to generate an ultraviolet light identification result;

[0083] S203, preprocessing the ultrasonic signal to extract an ultrasonic characteristic signal, and identifying the fault arc according to the ultrasonic characteristic signal to generate an ultrasonic identification result;

[0084] S204, preprocessing the electromagnetic wave signal to extract the electromagnetic wave characteristic signal, and identifying the fault arc according to the electromagnetic wave characteristic signal to generate an electromagnetic wave identification result;

[0085] S205, generating a fault arc detection result according to the ultraviolet light recognition result, the ultrasonic wave recognition result and the electromagnetic wave recognition result.

[0086] S206, storing the intelligent recognition results during the detection process;

[0087] Intelligent recognition results include ultraviolet light recognition results, ultrasonic wave recognition results, electromagnetic wave recognition results and fault arc detection results;

[0088] S207, acquiring and storing electrical system operation data;

[0089] Electrical system operation data includes current information and voltage information;

[0090] and Figure 1 What is different from the first embodiment shown is that in this embodiment, the intelligent recognition results and electrical system operation data in the detection process are organized into specific formatted data according to a certain information model for storage.

[0091] Accordingly, there can be multiple information models to facilitate the rapid acquisition and browsing of warning information by recipients at different receiving ends. In general, information models include mobile phone SMS information models, WeChat applet information models, PC computer information models, etc.; among them, the data format of the mobile phone SMS information model is fault type, urgency, person in charge phone number, fault details link; the data format of the WeChat applet information model is fault type, urgency, person in charge phone number, warning content summary, fault details link; the data format of the PC software computer information model is fault type, urgency, person in charge phone number, fault details, impact range, and handling suggestions.

[0092] S208, sending the intelligent recognition results and electrical system operation data to the host computer.

[0093] The present invention can upload formatted data to a host computer for publication, thereby realizing abnormal alarm; wherein, the host computer includes an energy management system, a safety management system, a human-computer interaction system, etc., which can specifically be PC system software, mobile phone WeChat, mobile industrial software, etc.

[0094] Therefore, by integrating the characteristics of three high-frequency signals, namely ultraviolet light, ultrasound and electromagnetic waves, to detect fault arcs in a closed environment, the present invention can solve the problem of missed detection of fault arcs and avoid losses of power distribution infrastructure caused by fault arcs.

[0095] See also Figure 3 , Figure 3 The specific structure of the closed environment-based arc fault intelligent detection system 100 of the present invention is shown, which includes an ultraviolet light acquisition device 1, an ultrasonic acquisition device 2, an electromagnetic wave acquisition device 3, a host computer 4 and a computer device 5. Specifically:

[0096] The computer device 5 includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the steps of the above-mentioned closed environment-based intelligent arc fault detection method are performed.

[0097] The ultraviolet light collection device 1 is connected to the computer device 5 and is used to collect and send ultraviolet light signals to the computer device 5; it should be noted that the ultraviolet light collection device 1 is a photon detection and counting device with a specific wavelength filter, which can record the number of ultraviolet photons in real time and send the number of ultraviolet photons as an ultraviolet light signal to the computer device.

[0098] The ultrasonic acquisition device 2 is connected to the computer device 5, and is used to collect and send ultrasonic signals to the computer device 5; it should be noted that the ultrasonic acquisition device 2 includes an ultrasonic fiber optic sensor, an ultrasonic acquisition box, an acquisition box adapter power supply and a sensor power supply, wherein the acquisition box adapter power supply provides power for the ultrasonic acquisition box, and the sensor power supply provides power for the ultrasonic fiber optic sensor; through the bandpass filter module in the ultrasonic acquisition box, 30KHz~40KHz ultrasonic signals can pass through, and the collected ultrasonic signals are saved in xl sx format, and the ultrasonic signal content includes each sampling point and the corresponding amplitude information.

[0099] The electromagnetic wave acquisition device 3 is connected to the computer device 5, and is used to collect and send electromagnetic wave signals to the computer device 5; it should be noted that the signal frequency range sensed by the electromagnetic wave acquisition device 3 is 0Hz~1GHz, by recording the power distribution of the electromagnetic wave signal at different frequencies (in dBm), and then separating the data in the frequency range of 1MHz~1GHz from the electromagnetic wave signal and saving it as data in csv format.

[0100] The host computer 4 is connected to the computer device 5 and is used to generate alarm information according to the intelligent recognition results and electrical system operation data uploaded by the computer device 5 to achieve real-time alarm of fault arc.

[0101] In summary, the present invention utilizes ultraviolet light, ultrasonic waves, and electromagnetic wave signals generated when an arc occurs to extract a variety of high-frequency signal features, accurately detect arc faults in a closed environment, and even weak arcs can be detected, thereby solving the problem of missed detection.

[0102] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for intelligent detection of arc faults in a closed environment, characterized in that: include: Acquire ultraviolet light signals, ultrasonic signals and electromagnetic wave signals; Preprocessing the ultraviolet light signal to extract an ultraviolet light characteristic signal, and identifying the fault arc according to the ultraviolet light characteristic signal to generate an ultraviolet light identification result; Preprocessing the ultrasonic signal to extract an ultrasonic characteristic signal, and identifying a fault arc according to the ultrasonic characteristic signal to generate an ultrasonic identification result; Preprocessing the electromagnetic wave signal to extract an electromagnetic wave characteristic signal, and identifying a fault arc according to the electromagnetic wave characteristic signal to generate an electromagnetic wave identification result; The fault arc detection result is generated according to the ultraviolet light recognition result, the ultrasonic wave recognition result and the electromagnetic wave recognition result.

2. The method for intelligent arc fault detection based on a closed environment according to claim 1, characterized in that: The step of preprocessing the ultraviolet light signal to extract the ultraviolet light characteristic signal comprises: dividing the ultraviolet light signal according to preset segmentation periods, counting the number of ultraviolet photons in each segmentation period, and using the number of ultraviolet photons as the ultraviolet light characteristic signal; The step of preprocessing the ultrasonic signal to extract the ultrasonic characteristic signal comprises: extracting the signal peak value and point position information from the ultrasonic signal as the ultrasonic characteristic signal; The step of preprocessing the electromagnetic wave signal to extract the electromagnetic wave characteristic signal includes: normalizing and standardizing the electromagnetic wave signal, and extracting mean information, peak information and kurtosis information from the normalized and standardized electromagnetic wave signal as the electromagnetic wave characteristic signal.

3. The method for intelligent arc fault detection based on a closed environment according to claim 2, characterized in that: The step of identifying the fault arc according to the ultraviolet light characteristic signal to generate an ultraviolet light identification result comprises: Each group of ultraviolet light characteristic signals within the preset time is compared with the preset photon quantity interval to determine whether an arc occurs. If the ultraviolet light characteristic signal is within the photon quantity range, it indicates that an arc has occurred. If the ultraviolet light characteristic signal is outside the photon quantity range, it indicates that no arc occurs; Compare the number of arc occurrences within the preset time with the preset arc frequency interval. If the number of arc occurrences within the preset time is greater than the arc frequency interval, the ultraviolet light identification result is "fault arc", If the number of arc occurrences within the preset time is within the arc frequency range, the ultraviolet light identification result is "suspected fault arc", If the number of arc occurrences within the preset time is less than the arc frequency interval, the ultraviolet light identification result is "non-fault arc".

4. The method for intelligent arc fault detection based on a closed environment according to claim 2, characterized in that: The step of identifying the fault arc according to the ultrasonic characteristic signal to generate an ultrasonic identification result comprises: The peak value of each group of ultrasonic characteristic signals within the preset time is compared with the preset peak value threshold to determine whether an arc occurs. If the peak value of the ultrasonic characteristic signal is greater than the peak value threshold, it indicates that an arc has occurred. If the peak value of the ultrasonic characteristic signal is less than or equal to the peak value threshold, it indicates that no arc occurs; Compare the number of arc occurrences within the preset time with the preset arc frequency interval. If the number of arc occurrences within the preset time is greater than the arc frequency interval, the ultrasonic recognition result is "fault arc", If the number of arc occurrences within the preset time is within the arc frequency range, the ultrasonic identification result is "suspected fault arc", If the number of arc occurrences within the preset time is less than the arc frequency interval, the ultrasonic identification result is "non-fault arc".

5. The closed environment based intelligent arc fault detection method according to claim 2, characterized in that: The step of identifying the fault arc according to the electromagnetic wave characteristic signal to generate an electromagnetic wave identification result includes: inputting the electromagnetic wave characteristic signal into a pre-trained arc electromagnetic wave identification model to generate an electromagnetic wave identification result.

6. The method for intelligent arc fault detection based on a closed environment according to claim 5, characterized in that: The training method of the arc electromagnetic wave recognition model includes: Obtain historical electromagnetic wave signals; Extracting characteristic information from the historical electromagnetic wave signal, wherein the characteristic information includes mean information, peak information and kurtosis information; Constructing a feature sample according to the feature information and historical fault categories; Dividing the feature samples into a training set and a test set; Inputting the training set into the SVM model to generate an arc electromagnetic wave recognition model; The test set is input into the arc electromagnetic wave identification model to optimize the arc electromagnetic wave identification model.

7. The method for intelligent arc fault detection based on a closed environment according to claim 1, characterized in that: The step of generating a fault arc detection result according to the ultraviolet light recognition result, the ultrasonic wave recognition result and the electromagnetic wave recognition result comprises: When any two or three of the ultraviolet light recognition result, the ultrasonic wave recognition result and the electromagnetic wave recognition result are "fault arc", the fault arc detection result is "fault arc"; Otherwise, the arc fault detection result is "non-arc fault".

8. The method for intelligent detection of arc faults based on a closed environment according to any one of claims 1 to 7, characterized in that: Also includes: Storing intelligent recognition results during the detection process, the intelligent recognition results include the ultraviolet light recognition results, ultrasonic wave recognition results, electromagnetic wave recognition results and fault arc detection results; Acquire and store electrical system operation data; The intelligent identification results and electrical system operation data are sent to a host computer.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. An intelligent arc fault detection system based on a closed environment, characterized in that: It comprises an ultraviolet light collection device, an ultrasonic wave collection device, an electromagnetic wave collection device, a host computer and the computer device according to claim 9, wherein the computer device is connected to the ultraviolet light collection device, the ultrasonic wave collection device, the electromagnetic wave collection device and the host computer respectively; The ultraviolet light collection device is used to collect and send the ultraviolet light signal to the computer device; The ultrasonic acquisition device is used to collect and send the ultrasonic signal to the computer device; The electromagnetic wave collection device is used to collect and send the electromagnetic wave signal to the computer device.

Images