Patent Application
20220247163
Aspects of the present invention generally relate to arc fault detection by accumulation of machine learning classifications in a circuit breaker.
Circuit breakers are essential for electrical safeties. They feed current to loads that are connected to them and interrupt the circuit once a circuit fault such as an overload, a short circuit, a ground fault and an arc fault is detected. We need the capabilities of an arc fault circuit interrupter (AFCI) in detecting and tripping on arc faults without causing unwanted tripping on various residential products (i.e. lighting, microwaves, vacuum cleaners, power tools, etc.). An arc fault circuit interrupter (AFCI) is an advanced circuit breaker that, as a way to reduce electrical fire threats, breaks the circuit when it detects a dangerous electric arc in the circuit that it protects.
AFCIs rely on internal electronics to analyze various analog signals (current and voltage) that pass through the circuit breaker to determine if an arc fault exists downstream of it. Prior art investigations tend to ignore utilizing the inferences from a machine learning classifier (neural network or other) to accumulate positive inferences and decrement negative inferences) into a final decision to trip the circuit breaker in a specific amount of time to meet the safety limits detailed in various industry standards (e.g., UL1699 or other equivalents).
Therefore, there is a need for a better arc fault detection in a circuit breaker.
Briefly described, aspects of the present invention relate to thermal management in a circuit breaker. The objective of the described invention is to focus on the types of signals and measurements to be made and analyzed, including investigations into the use of various machine learning mechanisms. A final step of utilizing the inferences from the machine learning classifier (neural network or other) is designed to accumulate positive inferences (and decrement negative inferences) into making a final decision to trip the breaker in a specific amount of time to meet the safety limits detailed in various industry standards (e.g., UL1699 or other equivalents).
In accordance with one illustrative embodiment of the present invention, a circuit breaker comprises a microcontroller including a processor and a memory. The circuit breaker further comprises computer-readable software code stored in the memory which, when executed by the processor, causes the microcontroller to: sample analog signals representing one or more of the following: a Received Signal Strength Indicator (RSSI) signal, a voltage signal, and a current signal, perform multiple pre-processing steps on the analog signals to derive a data set of measurements and features over a period of time, and input the data set into a machine learning classifier that resides in the microcontroller such that an output of the machine learning classifier is a value between 0 and 1 which represents a percent chance that the data set is from an electrical arc. Based on the value of the percent chance an accumulator value is either incremented or decremented in proportion to an amount of the current passing through the circuit breaker and if the accumulator value passes an upper threshold level, the microcontroller sends a signal output to a trip circuit which then opens the circuit breaker.
In accordance with one illustrative embodiment of the present invention, a method of arc fault detection by accumulation of machine learning classifications is provided. The method comprises providing a microcontroller including a processor and a memory. The method further comprises providing computer-readable software code stored in the memory which, when executed by the processor, causes the microcontroller to: sample analog signals representing one or more of the following: a Received Signal Strength Indicator (RSSI) signal, a voltage signal, and a current signal, perform multiple pre-processing steps on the analog signals to derive a data set of measurements and features over a period of time, and input the data set into a machine learning classifier that resides in the microcontroller such that an output of the machine learning classifier is a value between 0 and 1 which represents a percent chance that the data set is from an electrical arc. Based on the value of the percent chance an accumulator value is either incremented or decremented in proportion to an amount of the current passing through the circuit breaker and if the accumulator value passes an upper threshold level, the microcontroller sends a signal output to a trip circuit which then opens the circuit breaker.
To facilitate an understanding of embodiments, principles, and features of the present invention, they are explained hereinafter with reference to implementation in illustrative embodiments. In particular, they are described in the context of utilization of a machine learning classifier with a meaningful accumulation of inferences over time to create a fully working arc detection and interruption algorithm in an AFCI. A circuit breaker with software code provides classifications per half-cycle through accumulation of an inference output from a machine learning classifier for providing a final decision to trip the circuit breaker in a specific amount of time to meet the safety limits detailed in various industry standards. Embodiments of the present invention, however, are not limited to use in the described devices or methods.
The components and materials described hereinafter as making up the various embodiments are intended to be illustrative and not restrictive. Many suitable components and materials that would perform the same or a similar function as the materials described herein are intended to be embraced within the scope of embodiments of the present invention.
These and other embodiments of the circuit breaker according to the present disclosure are described below with reference to
Consistent with one embodiment of the present invention,
Referring to
Turning now to
The machine learning classifier 325 may be a trained neural network. The analog signals 315 data is analyzed and classified from specific pre-processing features at each half-cycle by the machine learning classifier 325.
The Received Signal Strength Indicator (RSSI) signal 315(1) and the current signal 315(3) come from an analog front-end ASIC circuitry 350. The analog front-end ASIC circuitry 350 acts as an interface between a radio frequency (RF) coupler, shunt resistance sensors and the microcontroller 307. The microcontroller 307 includes a logic 352 for detecting and tripping the circuit breaker 305 for overcurrent, differential, and arc faults.
In operation, the computer-readable software code 312 is to digitally convert the Received Signal Strength Indicator (RSSI) signal 315(1), the voltage signal 315(2), and the current signal 315(3). Then extract half-cycle features 355 from the Received Signal Strength Indicator (RSSI) signal 315(1), the voltage signal 315(2), and the current signal 315(3). Next, run half-cycle features 355 set through the machine learning classifier 325 and analyze an output of the machine learning classifier 325 to determine if half-cycle was an arc or not. If there is an arc inference, the computer-readable software code 312 is to increment the accumulator value 335 proportional to an amount of the current. The computer-readable software code 312 is to check whether the accumulator value 335 exceeds a max accumulator threshold and if so trip the circuit breaker 305. If there is no arc inference, the computer-readable software code 312 is to decrement the accumulator value 335 proportional to an amount of the current.
In the above case, an analog signal representing the RSSI level present on the residential wiring downstream of an AFCI is measured and analyzed in real-time to produce a set of features at the end of each half-cycle of the power line frequency (i.e. every 8.33 ms for the 60 Hz frequency in the U.S.). These features are then fed into the machine learning classifier 325 (in our case a neural network, but other classifier types could also be used) that has been previously trained to recognize arc faults. The output of the classifier 325 is a number between 0 and 1 representing the percent confidence that the last half-cycle was an electrical arc. Presently we have a flat threshold that if this output value exceeds the computer-readable software code 312 will increment a separate stored value (called an accumulator) in proportion to the amount of current present during this half-cycle event (e.g., a peak value of the current during this half-cycle, but other common measurements such as RMS or average could also be used). If the classifier 325 value is less than a threshold, the computer-readable software code 312 in turn decrements the accumulator by the same amount, in proportion to the amount of current present in that half-cycle. If the accumulator passes a maximum threshold an arc fault will have been deemed to be present, and the MCI will trip to open the circuit. This incrementing and decrementing of the accumulator works to average out multiple inferences over time while weighting for the amount of current present, to ensure that the MCI will trip within safe limits while also giving a maximum amount of time for discrimination of residential loads to minimize unwanted tripping.
While one embodiment has a hard threshold for incrementing/decrementing, the inference itself could also be used in further weighting schemes, such as one where inferences at the min/max of 0 or 1 would result in full changes to the accumulator (still also weighted by the amount of current), with the shift being scaled down to 0 the closer the inference is to 0.5 (where the classifier 325 isn't sure if it is an arc or not).
One main advantage is that this approach gives us a complete method for utilizing the power of machine learning in AFCIs. Whereas in the past arc detection analysis algorithms had to be created and improved by various individuals, sometimes resulting in delays of days or weeks to fix performance issues found during testing or in the field, now it will become a simpler matter of capturing the data from the nonconformity and adding it to our machine learning training set, shortening the time it takes to update the AFCI software with performance improvements. The specific utilization of the machine learning classifier 325 with a meaningful accumulation of inferences over time to create a fully working arc detection and interruption algorithm in an AFCI is provided.
As seen in
As shown in
Training data includes split—80% training, 0% validation, 20% testing, samples of arcing, and samples of nuisance and randomized distribution of samples in split. The training method may update the weight and bias values according to optimization and minimizes a combination of squared errors and weights to determine the correct combination so as to produce a network which generalizes well. Training method parameters left set to ‘default’ (Note: Does not need validation data set). The performance function is Mean-squared Error. This determines the overall performance of the network by comparing the target vs. desired output.
The hidden layer 615 is located between the input and output of the algorithm, in which the function applies weights to the inputs and directs them through an activation function as the output. In short, the hidden layers perform nonlinear transformations of the inputs entered into the neural network.
In
With regard to
With respect to
At a step 1105, analog signals including current and RSSI are received. At a step 1110, digital conversion of the analog signals is performed by the software code 312. Next, at a step 1115, the software code 312 extracts half-cycle features from the analog signals. Then in a step 1120 the software code 312 runs a half-cycle feature set through a machine learning classifier. Thereafter, at a step 1125 the software code 312 analyzes an output of the machine learning classifier to determine if the half-cycle was an arc or not. At a step 1130, if a no arc inference is derived, an accumulator value is decremented proportional to an amount of current. At a step 1135, if an arc inference is derived, an accumulator value is incremented proportional to an amount of current. At a step 1140, a check is made to determine if the accumulator value exceeded a max accumulation threshold. If the determination in the step 1140 is “yes” then at a step 1145 an AFCI device is tripped.
The method 1300 comprises a step 1305 of providing a microcontroller including a processor and a memory. The method 1300 further comprises a step 1310 of providing computer-readable software code stored in the memory which, when executed by the processor, causes the microcontroller to: sample analog signals representing one or more of the following: a Received Signal Strength Indicator (RSSI) signal, a voltage signal, and a current signal, perform multiple pre-processing steps on the analog signals to derive a data set of measurements and features for every half-cycle of current passing through the circuit breaker, and input the data set into a machine learning classifier that resides in the microcontroller such that an output of the machine learning classifier is a value between 0 and 1 which represents a percent chance that the data set is from an electrical arc. Based on the value of the percent chance an accumulator value is either incremented or decremented in proportion to an amount of the current passing through the circuit breaker and if the accumulator value passes an upper threshold level, the microcontroller sends a signal output to a trip circuit which then opens the circuit breaker.
While a multi-layer neural network as a machine learning classifier is described here a range of one or more other types of machine learning classifiers or other forms of machine learning classifiers are also contemplated by the present invention. For example, other types of machine learning classifiers may be implemented based on one or more features presented above without deviating from the spirit of the present invention.
The techniques described herein can be particularly useful for arc detection in an AFCI circuit breaker. While particular embodiments are described in terms of specific configuration of an AFCI, the techniques described herein are not limited to such a limited configuration but can also be used with other configurations and types of circuit breakers.
While embodiments of the present invention have been disclosed in exemplary forms, it will be apparent to those skilled in the art that many modifications, additions, and deletions can be made therein without departing from the spirit and scope of the invention and its equivalents, as set forth in the following claims.
Embodiments and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and detailed in the following description. Descriptions of well-known starting materials, processing techniques, components and equipment are omitted so as not to unnecessarily obscure embodiments in detail. It should be understood, however, that the detailed description and the specific examples, while indicating preferred embodiments, are given by way of illustration only and not by way of limitation, Various substitutions, modifications, additions and/or rearrangements within the spirit and/or scope of the underlying inventive concept will become apparent to those skilled in the art from this disclosure.
As used herein, the terms “comprises,” “comprising,” “includes,” “including,” “has,” “having” or any other variation thereof, are intended to cover a non-exclusive inclusion. For example, a process, article, or apparatus that comprises a list of elements is not necessarily limited to only those elements but may include other elements not expressly listed or inherent to such process, article, or apparatus.
Additionally, any examples or illustrations given herein are not to be regarded in any way as restrictions on, limits to, or express definitions of, any term or terms with which they are utilized. Instead, these examples or illustrations are to be regarded as being described with respect to one particular embodiment and as illustrative only. Those of ordinary skill in the art will appreciate that any term or terms with which these examples or illustrations are utilized will encompass other embodiments which may or may not be given therewith or elsewhere in the specification and all such embodiments are intended to be included within the scope of that term or terms.
In the foregoing specification, the invention has been described with reference to specific embodiments. However, one of ordinary skill in the art appreciates that various modifications and changes can be made without departing from the scope of the invention. Accordingly, the specification and figures are to be regarded in an illustrative rather than a restrictive sense, and all such modifications are intended to be included within the scope of invention.
Although the invention has been described with respect to specific embodiments thereof, these embodiments are merely illustrative, and not restrictive of the invention. The description herein of illustrated embodiments of the invention is not intended to be exhaustive or to limit the invention to the precise forms disclosed herein (and in particular, the inclusion of any particular embodiment, feature or function is not intended to limit the scope of the invention to such embodiment, feature or function). Rather, the description is intended to describe illustrative embodiments, features and functions in order to provide a person of ordinary skill in the art context to understand the invention without limiting the invention to any particularly described embodiment, feature or function. While specific embodiments of, and examples for, the invention are described herein for illustrative purposes only, various equivalent modifications are possible within the spirit and scope of the invention, as those skilled in the relevant art will recognize and appreciate. As indicated, these modifications may be made to the invention in light of the foregoing description of illustrated embodiments of the invention and are to be included within the spirit and scope of the invention. Thus, while the invention has been described herein with reference to particular embodiments thereof, a latitude of modification, various changes and substitutions are intended in the foregoing disclosures, and it will be appreciated that in some instances some features of embodiments of the invention will be employed without a corresponding use of other features without departing from the scope and spirit of the invention as set forth. Therefore, many modifications may be made to adapt a particular situation or material to the essential scope and spirit of the invention.
Respective appearances of the phrases “in one embodiment,” “in an embodiment,” or “in a specific embodiment” or similar terminology in various places throughout this specification are not necessarily referring to the same embodiment. Furthermore, the particular features, structures, or characteristics of any particular embodiment may be combined in any suitable manner with one or more other embodiments. It is to be understood that other variations and modifications of the embodiments described and illustrated herein are possible in light of the teachings herein and are to be considered as part of the spirit and scope of the invention.
In the description herein, numerous specific details are provided, such as examples of components and/or methods, to provide a thorough understanding of embodiments of the invention. One skilled in the relevant art will recognize, however, that an embodiment may be able to be practiced without one or more of the specific details, or with other apparatus, systems, assemblies, methods, components, materials, parts, and/or the like. In other instances, well-known structures, components, systems, materials, or operations are not specifically shown or described in detail to avoid, obscuring aspects of embodiments of the invention. While the invention may be illustrated by using a particular embodiment, this is not and does not limit the invention to any particular embodiment and a person of ordinary skill in the art will recognize that additional embodiments are readily understandable and are a part of this invention.
It will also be appreciated that one or more of the elements depicted in the drawings/figures can also be implemented in a more separated or integrated manner, or even removed or rendered as inoperable in certain cases, as is useful in accordance with a particular application.
Benefits, other advantages, and solutions to problems have been described above with regard to specific embodiments. However, the benefits, advantages, solutions to problems, and any component(s) that may cause any benefit, advantage, or solution to occur or become more pronounced are not to be construed as a critical, required, or essential feature or component.
