A system and method for creating training datasets for AI-based arc fault circuit breakers

Abstract

The present disclosure relates to systems and methods for creating a training dataset for an artificial intelligence based arc fault circuit interrupter. According to various embodiments, a method for creating a training dataset for an artificial intelligence based arc fault circuit interrupter is provided. In some embodiments, the method includes: collecting a number of non-arc data frames; calculating a spectral magnitude for each non-arc data frame and summing the spectral magnitudes; calculating a mean of the sum of spectral magnitudes for the non-arc data; collecting a number of arc data frames; calculating a spectral magnitude for each arc data frame and summing the spectral magnitudes; for each arc data frame, comparing the sum of spectral magnitudes to a threshold based on the mean of the sum of spectral magnitudes for the non-arc data; and adding each arc data frame for which the sum of spectral magnitudes exceeds the threshold to an AI data model training dataset.

Description

Technical Field

[0001] The exemplary embodiments of this disclosure generally relate to arc fault circuit breakers, and more specifically to artificial intelligence-based arc fault circuit breakers. Background Technology

[0002] An arc fault is a discharge that occurs when electrical current jumps between two or more conductors. In large-scale photovoltaic (PV) solar power generation, providing arc fault protection to prevent damage to expensive equipment is crucial. This arc fault protection is typically provided by an arc fault circuit interrupter (AFCI). An AFCI is a device that detects arcs and allows the circuit to be quickly disconnected to prevent arc damage.

[0003] Artificial intelligence (AI) is currently being used to train arc detectors (AFCIs) to detect arc faults. However, false alarm detection is a challenge for AI-based detection. A false alarm (FP) occurs when an arc detector indicates that an arc has occurred, but no arc has actually occurred. False alarms are primarily caused by energy fluctuations in PV panels or switching noise in power electronic equipment. False alarms are destructive because they can cause unnecessary power generation shutdowns.

[0004] One reason why AI-trained AFCI might generate false alarms is that some data corresponding to arcing scenarios (“arc data”) are very similar to some data corresponding to non-arcing scenarios (“no-arc data”). This is in... Figure 1 The figure illustrates example data 100, which includes frames of summed spectral amplitudes of numerous arc-free data 105 and summed spectral amplitudes of arc data 110. In the area indicated by the large arrows, the arc data closely resembles the arc-free data. Therefore, if this data is used to train an AI data model for arc fault detection, the AI ​​data model is more likely to generate false alarms.

[0005] The applicant has identified numerous technical challenges and difficulties associated with training AI models for arc fault detection. Through the application of effort, originality, and innovation, the applicant has addressed these challenges by developing solutions embodied in this disclosure, which are described in detail below. Summary of the Invention

[0006] The various embodiments described herein relate to systems and methods for creating training datasets for AI-based arc fault circuit breakers.

[0007] According to various embodiments of this disclosure, a method is provided for creating a training dataset for an AI-based arc fault circuit breaker. In some embodiments, the method includes: collecting a predetermined number of arc-free data frames corresponding to currents for which no arc current instance has yet occurred, each frame including a predetermined number of discrete data points; calculating the spectral amplitude of each arc-free data frame; summing the calculated spectral amplitudes for each arc-free data frame; calculating the average of the sums of the calculated spectral amplitudes of all arc-free data frames; collecting a predetermined number of arc data frames corresponding to currents for which at least one arc current instance has occurred, each frame including a predetermined number of discrete data points; calculating the spectral amplitude of each arc data frame; summing the calculated spectral amplitudes for each arc data frame; comparing the sum of the spectral amplitudes for each arc data frame with a threshold, the threshold being based on the average of the sums of the calculated spectral amplitudes of all arc-free data frames; and adding each arc data frame whose sum of spectral amplitudes exceeds the threshold to the AI ​​data model training dataset.

[0008] In some embodiments, the method includes discarding each arc data frame whose sum of spectral amplitudes does not exceed a threshold.

[0009] In some embodiments, the method further includes: for each arc-free data frame, removing the amplitude corresponding to the fundamental frequency before summing the calculated spectral amplitudes; and for each arc data frame, removing the amplitude corresponding to the fundamental frequency before summing the calculated spectral amplitudes.

[0010] In some embodiments, the method further includes: for each arc-free data frame, removing noise before summing the calculated spectral amplitudes; and for each arc data frame, removing noise before summing the calculated spectral amplitudes.

[0011] In some embodiments, arc-free data and arc data are collected from the electrical connection between one or more photovoltaic (PV) panels and a PV inverter.

[0012] In some embodiments, arc-free data and arc data are collected from the electrical connection between the arc generator and the PV inverter.

[0013] In some embodiments, arc-free data and arc data are collected using a current transformer positioned to detect current in the electrical connection between one or more PV panels and a PV inverter or between an arc generator and a PV inverter.

[0014] In some embodiments, the threshold is a predetermined multiplier based on the average of the sum of the calculated spectral amplitudes of all arc-free data frames.

[0015] In some embodiments, comparing the sum of the spectral amplitudes of each data frame with a threshold includes determining whether the absolute value of (1 minus (the sum of the spectral amplitudes of the arc data frames)) divided by (the calculated average of the sum of the calculated spectral amplitudes of all non-arc data frames) exceeds the threshold.

[0016] According to various embodiments of this disclosure, an apparatus is provided for creating a training dataset for an AI-based arc fault circuit breaker. In some embodiments, the apparatus includes: an arc generator adapted to be electrically connected between at least one photovoltaic (PV) panel and a PV inverter; a current sensing device adapted to detect current in the electrical connection between the arc generator and the PV inverter; and one or more instances of an arc detection device configured to send an arc trigger signal to the arc generator and receive a detected current signal from the current sensing device.

[0017] The above description of the invention is merely a summary of some exemplary embodiments to provide a basic understanding of some aspects of this disclosure. Therefore, it should be understood that the above embodiments are merely examples and should not be construed as limiting the scope or spirit of this disclosure in any way. It should also be understood that the scope of this disclosure covers many potential embodiments in addition to those summarized herein, some of which will be further described below. Attached Figure Description

[0018] The description of the illustrative embodiments can be read in conjunction with the accompanying drawings. It should be understood that, for the sake of simplicity and clarity of description, the elements illustrated in the drawings are not necessarily drawn to scale unless otherwise described. For example, the dimensions of some elements may be exaggerated relative to others unless otherwise described. Embodiments incorporating the teachings of this disclosure are shown and described in conjunction with the accompanying drawings presented herein, wherein:

[0019] Figure 1 The illustration shows example data illustrating the sum of spectral amplitudes without arc data and the sum of spectral amplitudes with arc data;

[0020] Figure 2 This is a block diagram of an example system for creating a training dataset for an AI-based arc fault circuit breaker, according to some embodiments of this disclosure.

[0021] Figure 3 This is a block diagram of an example arc detection device according to some embodiments of the present disclosure;

[0022] Figure 4 Example flowcharts according to some embodiments of the present disclosure are provided, illustrating an example method for collecting arc-free data for creating a training dataset for an AI-based arc-fault circuit breaker; and

[0023] Figure 5 Example flowcharts according to some embodiments of the present disclosure are provided, illustrating an example method for collecting arc-free data for creating a training dataset for an AI-based arc-fault circuit breaker. Detailed Implementation

[0024] Some embodiments of this disclosure will be described more fully below with reference to the accompanying drawings, which illustrate some, but not all, embodiments of this disclosure. In fact, these disclosures may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; these embodiments are provided so that this disclosure may meet applicable legal requirements. The same numerals always refer to the same elements.

[0025] As used herein, in the examples provided below, terms such as “front,” “back,” and “top” are used for illustrative purposes to describe the relative positions of certain components or component parts. Furthermore, based on this disclosure, those skilled in the art will understand that the terms “substantially” and “approximately” indicate that the referenced element or associated description is accurate within applicable engineering tolerances.

[0026] As used herein, the term “comprising” means including but not limited to, and should be interpreted in the manner commonly used in the patent context. It should be understood that the use of broader terms such as “comprising,” “including,” and “having” is intended to support narrower terms such as “consisting of,” “substantially consisting of,” and “essentially consisting of.”

[0027] The phrases “in one embodiment”, “according to one embodiment”, etc., generally mean that the specific feature, structure or characteristic following the phrase can be included in at least one embodiment of this disclosure, and can be included in multiple embodiments of this disclosure (importantly, such phrases do not necessarily refer to the same embodiment).

[0028] The words “example” or “exemplary” are used in this document to mean “served as an example, instance, or illustration.” Any implementation described as “exemplary” in this document is not necessarily to be construed as preferred or superior to other implementations.

[0029] If the specification states that a component or feature "may," "can," "should," "will," "preferably," "may," "usually," "optionally," "for example," "often," or "may" (or other such language) be included or have that characteristic, then the particular component or feature does not need to be included or have that characteristic. Such a component or feature may be optionally included in some embodiments, or it may be excluded.

[0030] Various embodiments of this disclosure overcome the aforementioned technical challenges and difficulties, and based on this, provide various technical improvements and advantages, such as, but not limited to, providing example systems and methods for creating training datasets used to train AI data models for arc fault protection. In various embodiments, arc data above a predetermined threshold is included in the dataset used to train the AI ​​data model, and arc data below the threshold is removed and not included in the dataset used to train the AI ​​data model. In various embodiments, a multilayer perceptron data model can be used for AI-based AFCI; however, any suitable AI or machine learning data model can be used.

[0031] In various embodiments, arc-free data frames are collected at various times and under various conditions (e.g., 512 or 1024 data points at a time). In various embodiments, spectral analysis of the arc-free data is performed, for example, by performing a Fast Fourier Transform (FFT) and summing the spectral amplitudes of each frame. In various embodiments, once a sufficient number of arc-free data frames have been captured, the sum of the spectral amplitudes of all arc-free data frames is averaged.

[0032] In various embodiments, arc data frames (e.g., 512 or 1024 data points at a time) are collected at various times and under various conditions. This can be achieved by using an arc generator to generate an arc and using a current transformer to collect arc current data. In various embodiments, spectral analysis of the arc data is performed, such as by performing a Fast Fourier Transform (FFT) and summing the spectral amplitude values ​​for each frame. In various embodiments, for each arc data frame, the sum of spectral amplitudes is compared to the average of the sums of spectral amplitudes of all arc-free data frames. In various embodiments, the AI ​​data model is trained only using arc data frames whose sum of spectral amplitudes exceeds the average of the sums of spectral amplitudes of all arc-free data frames by a predetermined threshold. That is, the AI ​​data model is trained only using arc data with a sufficiently large spectral amplitude compared to arc-free data.

[0033] In various embodiments, the collection of arc data frames and the comparison of arc data frames with a threshold are continuously performed until a sufficient number of arc data frames are captured.

[0034] In various embodiments, the collection of arc data frames and arc-free data frames can occur in a laboratory or other pre-installed and / or customer-installed locations.

[0035] Although embodiments of this disclosure are described herein in connection with creating training datasets for AI-based AFCI for PV installations, embodiments of this disclosure can be used to create training datasets for AI-based AFCI for any suitable application.

[0036] Figure 2 An exemplary block diagram of an example system for creating a training dataset for an AI-based arc fault circuit breaker, according to an example embodiment of the present disclosure, is illustrated. Figure 2 Example system 200 includes one or more PV panels 205, a PV inverter 215, an arc generator 210 electrically connected between the PV panel 205 and the PV inverter 215, a current sensing device 230 for detecting current in the electrical connection between the arc generator 210 and the PV inverter 215, an arc detection device 220, and optional data processing device 225. In various embodiments, the arc generator 210 conforms to UL standard 1699B; however, any suitable arc generator can be used. In various embodiments, the current sensing device 230 includes a current transformer; however, any suitable current sensing device can be used.

[0037] In various embodiments, the arc detection device 220 is configured to receive one or more instances of detected current signals from the current sensing device 230. In various embodiments, the detected current from the current sensing device 230 is collected in the form of data frames (e.g., 512 or 1024 data points at a time), which are collected at various times and under various conditions and at a suitable sampling rate (e.g., 250 kHz). In various embodiments, the detected current from the current sensing device 230 corresponds to the current flowing from one or more PV panels 205 to the PV inverter 215, and this current may include one or more arc faults (providing potential arc data, depending on the post-processing described below) or may include arc-free faults (providing arc-free data).

[0038] Because the occurrence of arc faults is unpredictable, in various embodiments, the arc detection device 220 is configured to send an arc trigger signal to the arc generator 210, thereby enabling reliable generation of arc faults and reliable capture of arc data. In various embodiments, the arc detection device 220 includes a user input device configured to, when activated by a user, enable the arc detection device 220 to send an arc trigger signal to the arc generator 210. In various embodiments, the arc detection device 220 is configured to record one or more instances of detected current signals when the user input device is activated by a user. This ensures the correct capture of arc data.

[0039] In various embodiments, instances of detected current signals are post-processed, as described below regarding... Figure 4 and Figure 5As described, this ensures that only arc data with a sufficiently large spectral amplitude compared to arc-free data is used to train the AI ​​data model. In this regard, embodiments of this disclosure reduce the number of false alarms in AI-based AFCI. In various embodiments, this post-processing can be performed in arc detection device 220, in optional data processing device 225, or in any other suitable device. In various embodiments, the post-processed dataset is used to train the AI ​​data model in optional data processing device 225 or any other suitable device. In various embodiments, optional data processing device 225 includes a personal computer (PC), etc.

[0040] Figure 3 An example arc detection device 220 specially configured according to at least some example embodiments of the present disclosure is depicted. Figure 3 The arc detection device 220 includes a processing circuit arrangement 305, a memory circuit arrangement 310, an input / output circuit arrangement 315, a communication circuit arrangement 320, and a transducer array 110. In various embodiments, the arc detection device 220 is configured to perform and implement the operations described herein. For example, the arc detection device 220 may be configured to implement a method for creating a training dataset for an AI-based arc fault circuit breaker, as described below. Figure 4 and Figure 5 As stated above.

[0041] Although components are described with regard to functional limitations, it should be understood that a particular implementation must include the use of specific computing hardware. It should also be understood that in some embodiments, certain components described herein include similar or generic hardware. For example, in some embodiments, two sets of circuit devices utilize the same(s) processor(s), memory(s), circuit devices(s), etc., to perform their associated functions, so that each set of circuit devices does not require duplicate hardware.

[0042] The processing circuitry 305 can be embodied in a variety of different ways. In various embodiments, the terms "processor," "processing circuitry," "controller," or "control circuitry" should be understood to include a single-core processor, a multi-core processor, multiple processors within the arc detection device 220, and / or one or more remote or "cloud" processors external to the arc detection device 220. In some example embodiments, the processing circuitry 305 may include one or more processing devices configured to execute independently. Alternatively or additionally, the processing circuitry 305 may include one or more processors configured in series via a bus to enable independent execution of operations, instructions, pipelines, and / or multithreading.

[0043] In an example embodiment, processing circuitry 305 may be configured to execute instructions stored in memory circuitry 310 or otherwise accessible to the processor. Alternatively or additionally, processing circuitry 305 may be configured to perform hard-coded functionality. Thus, whether configured by hardware or software methods, or by a combination of both, processing circuitry 305 may represent an entity (e.g., physically embodied in a circuitry) capable of performing operations according to embodiments of this disclosure when appropriately configured. Alternatively or additionally, processing circuitry 305 may be embodied as an executor of software instructions, and the instructions may specifically configure processing circuitry 305 to perform various algorithms embodied in one or more operations described herein when executing such instructions. In some embodiments, processing circuitry 305 includes hardware, software, firmware, and / or combinations thereof for performing one or more operations described herein.

[0044] In some embodiments, the processing circuitry 305 (and / or coprocessor, or processing circuitry 305 that is auxiliary or otherwise associated with the processor) communicates with the memory circuitry 310 via a bus for transmitting information among the components of the arc detection device 220.

[0045] The memory or memory circuitry 310 may be non-transitory and may include, for example, one or more volatile and / or non-volatile memories. In some embodiments, the memory circuitry 310 includes or embodies an electronic storage device (e.g., a computer-readable storage medium). In some embodiments, the memory circuitry 310 is configured to store information, data, content, applications, instructions, etc., so that the arc detection device 220 can perform various operations and / or functions according to exemplary embodiments of this disclosure.

[0046] Input / output circuitry 315 may be included in the arc detection device 220. In some embodiments, input / output circuitry 315 may provide output to a user and / or receive input from a user. Input / output circuitry 315 may communicate with processing circuitry 305 to provide this functionality. Input / output circuitry 315 may include one or more user interfaces. In some embodiments, the user interface may include a display that includes multiple interfaces presented as a web user interface, application user interface, user device, back-end system, etc. In some embodiments, input / output circuitry 315 may also include analog sampling circuitry, a keyboard, mouse, joystick, touchscreen, touch area, softkeys, microphone, speaker, or other input / output mechanisms. Processing circuitry 305 and / or input / output circuitry 315 may be configured to control one or more operations and / or functions of one or more user interface elements via computer program instructions (e.g., software and / or firmware) stored in processor-accessible memory (e.g., memory circuitry 310, etc.). In some embodiments, input / output circuitry 315 includes or utilizes user-oriented applications to provide input / output functionality to a computing device and / or other display associated with a user. In various embodiments, the input / output circuitry 315 includes the aforementioned user input device, which is configured to enable the arc detection device 220 to send an arc trigger signal to the arc generator 210 when activated by a user.

[0047] Communication circuitry 320 may be included in arc detection device 220. Communication circuitry 320 may include any components, such as devices or circuitry embodied in hardware or a combination of hardware and software, configured to receive and / or transmit data from / to a network and / or any other device, circuitry, or module communicating with arc detection device 220. In some embodiments, communication circuitry 320 includes, for example, a network interface for enabling communication with wired or wireless communication networks. Additionally or alternatively, communication circuitry 320 may include one or more network interface cards, antennas, buses, switches, routers, modems, and supporting hardware, firmware, and / or software, or any other device suitable for enabling communication via one or more communication networks. In some embodiments, communication circuitry 320 may include circuitry for interacting with antennas and / or other hardware or software to transmit signals via antennas and / or process the reception of signals received via antennas. In some embodiments, the communication circuitry 320 enables the transmission of data to and / or reception of data from the user equipment, one or more sensors, and / or multiple other external computing devices (such as data processing devices 225) that communicate with the arc detection device 220.

[0048] In some embodiments, two or more circuit devices in circuit device groups 305-320 are composable. Alternatively or additionally, one or more circuit devices in circuit device groups 305-320 perform some or all of the operations and / or functionalities described herein as associated with another circuit device. In some embodiments, two or more circuit devices in circuit groups 305-320 are combined into a single module embodied in hardware, software, firmware, and / or combinations thereof.

[0049] Now refer to Figure 4 and Figure 5 It provides flowcharts illustrating example steps, processes, procedures, and / or operations according to various embodiments of the present disclosure. The various methods described herein (including, for example, Figure 4 and Figure 5 The example method shown can provide various technical benefits and improvements. It is worth noting that each block of the flowchart, and combinations of blocks in the flowchart, can be implemented by various means, such as hardware, firmware, circuitry, and / or other devices associated with the execution of software including one or more computer program instructions. For example, Figure 4 and Figure 5One or more procedures described herein may be embodied by computer program instructions, which may be stored in a non-transitory memory of a device employing embodiments of the present disclosure and executed by a processor in the device. These computer program instructions may direct a computer or other programmable device to operate in a particular manner, causing the instructions stored in a computer-readable storage memory to produce an article of art, the execution of which implements the functions specified in the flowchart(s) blocks(s).

[0050] As will be understood from the above and based on this disclosure, embodiments of this disclosure can be configured as methods, mobile devices, backend network devices, etc. Therefore, embodiments can include a variety of components, encompassing entirely hardware or any combination of software and hardware. Furthermore, embodiments can take the form of a computer program product on at least one non-transitory computer-readable storage medium having computer-readable program instructions (e.g., computer software) embodied therein. Similarly, embodiments can take the form of computer program code stored on at least one non-transitory computer-readable storage medium. Any suitable computer-readable storage medium can be utilized, including non-transitory hard disks, CD-ROMs, flash memory, optical storage devices, or magnetic storage devices.

[0051] Now for reference Figure 4 The illustration shows an example flowchart according to some embodiments of the present disclosure, illustrating an example method 400 for collecting arc-free data to create a training dataset for an AI-based arc-fault circuit breaker. In some embodiments, example method 400 may be implemented by the example arc detection device described herein, including but not limited to those described above. Figure 2 and Figure 3 The example arc detection device described.

[0052] Figure 4 The example method 400 shown begins at step / operation 405. At step / operation 405, the processor (such as, but not limited to, those described above)... Figure 2 and Figure 3 The processing circuit device 305 of the described arc detection device 220, in conjunction with the above description of... Figure 2 The described current sensing device 230 collects arc-free data frames. In various embodiments, each arc-free data frame has a predetermined number of data points (e.g., 512 or 1024) and is collected at various times and under various conditions. As described above, arc-free data frames can be collected using a current sensing device (such as a current transformer).

[0053] At step / operation 410, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3The processing circuitry 305 of the described arc detection device 220 calculates the spectral amplitude of the arc-free data in each frame by performing a fast Fourier transform (FFT).

[0054] At step / operation 415, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuit device 305 of the described arc detection device 220 removes the amplitude corresponding to the fundamental component of the spectrum.

[0055] At step / operation 420, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuit device 305 of the described arc detection device 220 removes the amplitude corresponding to the noise wave (e.g., at a known noise frequency point, such as the switching frequency of an inverter).

[0056] At step / operation 425, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuit device 305 of the described arc detection device 220 sums the remaining spectral amplitudes of the arc-free data frame (i.e., after removing the amplitudes corresponding to the fundamental component and the amplitudes corresponding to the noise wave).

[0057] At step / operation 430, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuit device 305 of the described arc detection device 220 saves the arc-free data frame as part of the dataset used to train the AI-based AFCI.

[0058] At step / operation 435, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuitry 305 of the described arc detection device 220 determines whether a desired number of arc-free frames have been obtained. Any suitable number of arc-free frames can be obtained for training the AI-based AFCI. In one example embodiment, arc-free frames are collected according to steps / operations 405-430 until 2000 arc-free frames are obtained.

[0059] If it is determined at step / operation 435 that the desired number of arc-free frames has not been obtained, then method 400 returns to step / operation 405. If it is determined at step / operation 435 that the desired number of arc-free frames has been obtained, then method 400 continues to step / operation 440.

[0060] At step / operation 440, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3The processing circuitry 305 of the described arc detection device 220 calculates the average of the sum of the spectral amplitudes of all arc-free data frames. In various embodiments, this average is stored and used to obtain arc data, such as those transmitted via [unclear - likely a specific method or process], for training an AI-based AFCI. Figure 5 The method.

[0061] Now for reference Figure 5 The illustration shows an example flowchart according to some embodiments of the present disclosure, illustrating an example method 500 for collecting arc data to create a training dataset for an AI-based arc fault circuit breaker. In some embodiments, example method 500 may be implemented by an example arc detection device described herein, including but not limited to those described above. Figure 2 and Figure 3 The example arc detection device described.

[0062] Figure 5 The example method 500 shown begins at step / operation 505. At step / operation 505, the processor (such as, but not limited to, those described above)... Figure 2 and Figure 3 The processing circuit device 305 of the arc detection device 220, in conjunction with the above description of... Figure 2 The arc generator 210 triggers an arc on the electrical connection to the PV inverter. As described above, this can be triggered by a user action, such as pressing a button on the arc detection device 220.

[0063] At step / operation 510, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuit device 305 of the described arc detection device 220, in conjunction with the above description of... Figure 2 The described current sensing device 230 collects arc data frames. In various embodiments, each arc data frame has a predetermined number of data points (e.g., 512 or 1024) and is collected at various times and under various conditions. As described above, arc data frames can be collected using a current sensing device such as a current transformer.

[0064] At step / operation 515, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuitry 305 of the described arc detection device 220 calculates the spectral amplitude of the arc data in each frame by performing a fast Fourier transform (FFT).

[0065] At step / operation 520, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3The processing circuit device 305 of the described arc detection device 220 removes the amplitude corresponding to the fundamental component of the spectrum.

[0066] At step / operation 525, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuit device 305 of the described arc detection device 220 removes the amplitude corresponding to the noise wave (e.g., at a known noise frequency point, such as the switching frequency of an inverter).

[0067] At step / operation 530, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuit device 305 of the described arc detection device 220 sums the residual spectral amplitudes of the arc data frame (i.e., after removing the amplitudes corresponding to the fundamental component and the amplitudes corresponding to the noise wave).

[0068] At step / operation 535, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuit 305 of the described arc detection device 220 sums the spectral amplitude values ​​of the arc data frames determined in step / operation 530 with the values ​​at the time of the arc detection device 220. Figure 4 The average value of the sum of the spectral amplitudes of all arc-free data frames determined at step / operation 440 of method 400 is compared, and it is determined whether the sum of the spectral amplitudes of the arc data frames determined at step / operation 530 exceeds the average value of the sum of the spectral amplitudes of all arc-free data frames by a predetermined amount.

[0069] In some embodiments, the sum of spectral amplitudes is compared to a threshold, which is based on the average of the calculated sums of spectral amplitudes of all arc-free data frames. In some embodiments, the threshold is based on multiplying the average of the calculated sums of spectral amplitudes of all arc-free data frames by a predetermined filter coefficient or a predetermined range of filter coefficients. For example, the predetermined range of filter coefficients may be 1.1 to 3.9, with an example of a specific value being 1.5. That is, in one example, the average of the calculated sums of spectral amplitudes of all arc-free data frames is multiplied by 1.5 to determine a threshold for comparison with the sum of spectral amplitudes of each arc data frame. In some embodiments, comparing the sum of spectral amplitudes of each arc data frame to the threshold includes determining whether the absolute value of (1 - (sum of spectral amplitudes of arc data frames) / (calculated average of the calculated sums of spectral amplitudes of all arc-free data frames)) exceeds the threshold.

[0070] Specific values ​​for the filter coefficients can be selected based on the desired performance of AI-based AFCI. For example, lower filter coefficients may lead to more false alarms, while higher filter coefficients may lead to more false denials (i.e., missed arcs).

[0071] If it is determined at step / operation 535 that the sum of the spectral amplitude values ​​of the arc data frames determined at step / operation 530 exceeds the average of the sum of the spectral amplitudes of all non-arc data frames by a predetermined amount, then method 500 continues to step / operation 540.

[0072] At step / operation 540, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuitry 305 of the described arc detection device 220 saves the arc data frame as part of the dataset used to train the AI-based AFCI. Method 500 continues to step / operation 540.

[0073] If, at step / operation 535, it is determined that the sum of the spectral amplitude values ​​of the arc data frame determined at step / operation 530 does not exceed the average of the sums of the spectral amplitudes of all non-arc data frames by a predetermined amount, then the arc data frame is not saved and can be discarded, and method 500 continues to step / operation 545. Through this determination, the arc data frame is neither saved nor used in the AI ​​model training dataset because the sum of the arc data frame's amplitude is too close to the average of the sums of the non-arc data frames, and therefore more likely to result in more false alarms in the AFCI trained with such frames.

[0074] At step / operation 545, the processor (such as, but not limited to, the one mentioned above) Figure 2 and Figure 3 The processing circuitry 305 of the described arc detection device 220 determines whether a desired number of arc frames have been obtained. Any suitable number of arc frames can be obtained for training the AI-based AFCI. In one example embodiment, arc frames are collected according to steps / operations 505-540 until 2000 arc frames are obtained.

[0075] If it is determined at step / operation 545 that the desired number of arc frames has not been obtained, then method 500 returns to step / operation 505. If it is determined at step / operation 545 that the desired number of arc frames has been obtained, then method 500 terminates at step / operation 550.

[0076] In various embodiments, example method 400 and example method 500 can be run once to obtain a dataset used to train the AI-based AFCI, and can be run periodically (e.g., according to a predetermined schedule) to update the dataset, thereby updating the training of the AI-based AFCI.

[0077] In various embodiments, the arc detection device continues to monitor the current in the electrical connections of the PV inverter to obtain additional data that can be used to update the training of the AI ​​model for the AFCI. Specifically, the arc detection device may continue monitoring to identify false alarms (i.e., when the AFCI indicates that an arc has occurred but there is actually no arc). This false alarm data may indicate that the filter coefficients are set too low and that the dataset may need to be reprocessed with different filter coefficients. This ongoing data may be stored in the memory of the arc detection device (such as in a removable memory card) and / or uploaded to a data processing device.

[0078] Various embodiments of this disclosure can reduce false alarms in photovoltaic arc detection systems with almost no loss of arc detection accuracy.

[0079] in conclusion

[0080] Many modifications and other embodiments of the present disclosure will occur to those skilled in the art in light of the teachings presented in the foregoing description and the accompanying drawings. Although the drawings illustrate only certain components of the devices and systems described herein, it should be understood that various other components may be used in conjunction with the system. Therefore, it should be understood that the present disclosure is not limited to the specific embodiments disclosed, and modifications and other embodiments are intended to be included within the scope of the appended claims. Furthermore, the steps in the methods described above do not necessarily occur in the order depicted in the drawings, and in some cases, one or more of the depicted steps may occur substantially simultaneously, or additional steps may be involved. Although specific terminology is used herein, it is used only in a general and descriptive sense and not for limiting purposes.

[0081] While various embodiments based on the principles disclosed herein have been shown and described above, modifications can be made thereto by those skilled in the art without departing from the spirit and teachings of this disclosure. The embodiments described herein are merely representative and not intended to be limiting. Many variations, combinations, and modifications are possible and are within the scope of this disclosure. Alternative embodiments resulting from combining, integrating, and / or omitting features of (multiple) embodiments are also within the scope of this disclosure. Therefore, the scope of protection is not limited by the foregoing description.

[0082] Furthermore, the section headings used herein are intended to be consistent with the recommendations under 37 CFR 1.77 or otherwise provide organizational clues. These headings should not limit or characterize the disclosure(s) that may be listed in any claim(s) issued from this disclosure.

[0083] While this detailed description has illustrated some embodiments of the present disclosure, the appended claims cover other embodiments of the present disclosure that differ from the described embodiments with various modifications and improvements. For example, the appended claims may cover any form of device for capturing arc data and arc-free data and processing such data for training AI data models.

[0084] In the appended claims, unless the specific terms “part for…” or “step for…” are used in a given claim, the claim is not intended to be interpreted under 35 U.S.C., 112, paragraph 6.

Claims

1. A method for creating a training dataset for an AI-based arc fault circuit breaker, the method comprising: Collect a predetermined number of arc-free data frames corresponding to currents for which no arc current instances have yet occurred, each frame including a predetermined number of discrete data points; Calculate the spectral amplitude of the arc-free data in each frame; For each arc-free data frame, the calculated spectral amplitudes are summed; Calculate the average of the sums of the calculated spectral amplitudes for all the arc-free data frames; Collect a predetermined number of arc data frames corresponding to the currents of at least one instance of arc current, each frame including a predetermined number of discrete data points. Calculate the spectral amplitude of the arc data in each frame; For each arc data frame, the calculated spectral amplitude is summed; For each arc data frame, the sum of the spectral amplitudes is compared with a threshold, the threshold being the average of the sum of the spectral amplitudes calculated from all the arc-free data frames; as well as Each arc data frame whose sum of the spectral amplitudes exceeds the threshold is added to the AI ​​data model training dataset.

2. The method according to claim 1, further comprising: Discard each arc data frame whose sum of the spectral amplitudes does not exceed the threshold.

3. The method according to claim 1, further comprising: For each arc-free data frame, the amplitude corresponding to the fundamental wave is removed before summing the calculated spectral amplitudes; as well as For each arc data frame, the amplitude corresponding to the fundamental wave is removed before summing the calculated spectral amplitudes.

4. The method according to claim 1, further comprising: For each arc-free data frame, noise is removed before summing the calculated spectral amplitudes; as well as For each arc data frame, noise is removed before summing the calculated spectral amplitudes.

5. The method of claim 1, wherein the arc-free data and the arc data are collected from the electrical connection between one or more photovoltaic (PV) panels and a PV inverter.

6. The method of claim 1, wherein the arc-free data and the arc data are collected from the electrical connection between the arc generator and the PV inverter.

7. The method of claim 1, wherein the arc-free data and the arc data are collected using a current transformer positioned to detect current in an electrical connection between one or more PV panels and a PV inverter or between an arc generator and a PV inverter.

8. The method of claim 1, wherein the threshold is a predetermined multiplier of the average value calculated from the sum of the calculated spectral amplitudes of all the arc-free data frames.

9. The method of claim 1, wherein comparing the sum of the spectral amplitudes of each arc data frame with the threshold comprises: Determine whether the absolute value of the average value calculated by dividing the sum of the spectral amplitudes of the arc data frames by the sum of the calculated spectral amplitudes of all the arc-free data frames exceeds the threshold.

10. A system for creating a training dataset for an AI-based arc fault circuit breaker, the system comprising: An arc generator, the arc generator being adapted to be electrically connected between at least one photovoltaic (PV) panel and a PV inverter; A current sensing device, the current sensing device being adapted to detect the current in the electrical connection between the arc generator and the PV inverter; as well as An arc detection device, configured to send an arc trigger signal to the arc generator and receive one or more instances of a detected current signal from the current sensing device.

11. The system of claim 10, wherein the current sensing device includes a current transformer adapted to at least partially surround the electrical connection between the arc generator and the PV inverter.

12. The system of claim 10, further comprising a user input device configured to cause the arc detection device to send an arc trigger signal to the arc generator when activated by a user.

13. The system of claim 12, wherein the arc detection device is further configured to begin recording the current signal detected from the current sensing device when the user activates the user input device.

14. The system of claim 10, wherein the arc detection device is further configured to: From the one or more instances of current signals detected from the current sensing device, a predetermined number of arc-free data frames are identified corresponding to the current of an instance of arc current that has not yet occurred, each frame including a predetermined number of discrete data points. Calculate the spectral amplitude of the arc-free data in each frame; For each arc-free data frame, the calculated spectral amplitudes are summed; Calculate the average of the sums of the calculated spectral amplitudes for all the arc-free data frames; From the one or more instances of the current signal detected from the current sensing device, a predetermined number of arc data frames are identified corresponding to the currents that have occurred at least one instance of arc current, each frame including a predetermined number of discrete data points. Calculate the spectral amplitude of the arc data in each frame; For each arc data frame, the calculated spectral amplitude is summed; For each arc data frame, the sum of the spectral amplitudes is compared with a threshold, the threshold being the average of the sum of the spectral amplitudes calculated from all the arc-free data frames; as well as Each arc data frame whose sum of the spectral amplitudes exceeds the threshold is added to the AI ​​data model training dataset.

15. The system of claim 14, wherein the arc detection device is further configured to: For each arc-free data frame, the amplitude corresponding to the fundamental frequency is removed before summing the calculated spectral amplitudes; and For each arc data frame, the amplitude corresponding to the fundamental wave is removed before summing the calculated spectral amplitudes.

16. The system of claim 14, wherein the arc detection device is further configured to: For each arc-free data frame, noise is removed before summing the calculated spectral amplitudes; and For each arc data frame, noise is removed before summing the calculated spectral amplitudes.

17. The system of claim 14, wherein the threshold is a predetermined multiplier of the average value calculated from the sum of the calculated spectral amplitudes of all the arc-free data frames.

18. The system of claim 14, wherein comparing the sum of the spectral amplitudes of each arc data frame with the threshold comprises: Determine whether the absolute value of the average value calculated by dividing the sum of the spectral amplitudes of the arc data frames by the sum of the calculated spectral amplitudes of all the arc-free data frames exceeds the threshold.

Images