Detection of electrical discharges prior to fire in electrical wiring

Abstract

Described herein are methods and systems for detecting electrical discharges that precede electrical fires in electrical wiring. One or more sensor devices coupled to an electrical circuit detect one or more signal waveforms generated by electrical activity on the circuit. The sensor devices identify one or more transient signals within the one or more signal waveforms and generate one or more transient signatures based on the identified transient signals. A server communicatively coupled to the sensor devices receives the one or more transient signatures. The server analyzes the one or more transient signatures to identify one or more discharge indications. When one or more discharge indications are identified, the server generates one or more alarm signals.

Description

Technical Field

[0001] The subject matter of the present application relates generally to the detection of incipient electrical discharges, including arc faults, in electrical wiring. Technical Background

[0002] According to statistics released by the U.S. Fire Administration's National Fire Data Center ( https: / / www.usfa.fema.gov / downloads / pdf / statistics / v14i13.pdf ), arcs in residential electrical wiring cause more than 70% of electrical fires - one of the most dangerous threats to life and property. An arc fault (also known as an arc) is a high-power, continuous electrical discharge between two or more conductors - typically in residential buildings, when the integrity of the wire or insulation is compromised (e.g., through physical damage, water damage, corrosion, aging, or loose connections, etc.). Events such as lightning strikes or power surges can also trigger a breakdown in insulation and cause damage to the wire. As a result of the damaged wire, small, sporadic discharges begin to occur, and the insulation surrounding the wire is carbonized. As the discharges continue to occur over time, the insulation becomes increasingly eroded, and the intensity of the discharge increases. Eventually, the strong discharge becomes a sustained arc formed in the wire - resulting in a large current and a large release of energy (with a correspondingly high temperature). Because the wires are close to wood framing, insulation, and / or similar combustible materials, they are likely to cause fires when the temperatures generated by the arcs are high enough. It would be a great advantage in preventing electrical fires if people could detect and be warned of small discharges that may occur for days, weeks, or months before becoming large enough to produce an arc (as described in Yereance, RA and Kerkhoff, T., Electrical Fire Analysis, 3rd edition, Charles C. Thomas, Springfield IL, (2010), p. 206).

[0003] The above discharges can occur in a variety of ways, including: parallel, series, and line-to-ground discharges. Parallel discharges occur when current / electrons flow from one conductor to another through a gas or dielectric material (usually through damaged insulation or air) due to a large voltage difference between the conductors. Figure 1A 1 is a schematic diagram of a parallel discharge. The circuit has two conductors 102a, 102b, each surrounded by an insulating material. If the insulation between the conductors breaks down, a discharge (such as discharge 104) may occur between the conductors. Examples of parallel discharges include carbonization (i.e., breakdown of insulating material) and wet tracking (i.e., moisture on the surface of the conductor that enables the formation of an electric current).

[0004] Series discharge occurs when a single conductor is damaged to the point where the resistance through the conductor increases and a voltage difference high enough develops to cause a discharge to occur within the conductor and into the surrounding insulation (or even to foreign objects if the conductor is exposed). Figure 1B is a schematic diagram of a series discharge. The circuit includes a damaged conductor 106 that produces a discharge 108. Examples of series discharges include ground pyrolysis (i.e., current flows from the conductor to nearby wood), and last strand (i.e., breakdown of the conductor resulting in an increase in heat and combustible gases). A special type of series discharge occurs at what is known as a glowing connection (e.g., Figure 1C In such cases, the electrical conductors are in contact but not tightly connected together. An oxide layer forms at the boundary of the interface, which increases the resistance of the conductor at the joint. If current is flowing through the interface, the temperature may rise to dangerous levels (e.g., the white area 110 shows the high temperature at the electrical outlet), which may ignite nearby materials and cause a destructive fire. By the time the discharge has progressed to the point of causing a fire, it is too late to take corrective action and prevent damage. It is important to detect the occurrence of such discharges in electrical wiring as early as possible so that remedial measures can be taken in place.

[0005] Currently there are technologies such as arc fault circuit interrupters (AFCI) to detect early discharges, such as arc faults. In electrical outlets equipped with AFCI technology, the AFCI detects arc faults in the circuit and, when such faults are detected, disconnects the circuit to prevent electrical fires. However, AFCIs are relatively expensive and must be installed on every circuit in a building to detect discharges on each circuit. Summary of the invention

[0006] Therefore, what is needed is a method and system for detecting early discharges before they are large enough to cause fires in electrical wiring systems (such as those found in homes and other buildings). The methods and systems described herein provide the advantage of detecting early discharges in electrical wiring as they begin to form, which enables rapid notification of potential fire hazards. In addition, the technology described herein enables long-term automatic monitoring of electrical wiring systems for discharges (e.g., weeks, months, years) to obtain detailed information about activities and trends in the electrical wiring system of a building, including: the ability to distinguish between different discharge severity levels, monitoring events such as surges, voltage sags, and light brownouts that can trigger carbonization of insulation and the start of discharges; correlating the detection of discharges with the operation of appliances and equipment running in a home (e.g., arcs may occur in equipment such as air conditioning units; long-term monitoring provides the opportunity to correlate discharges with the operation of air conditioning units, and the device can be correlated to electrical problems on the power grid by aggregating information from monitoring across multiple homes). Moreover, in some embodiments, the system utilizes a single monitoring device that plugs into an existing electrical outlet - rather than complex, expensive, or dangerous installation of other equipment or monitoring components.

[0007] In one aspect, the invention features a system for detecting an electrical discharge preceding an electrical fire in electrical wiring. The system includes one or more sensor devices coupled to an electrical circuit, each sensor device configured to detect one or more signal waveforms generated by electrical activity on the electrical circuit. Each sensor device is configured to identify one or more transient signals within the one or more signal waveforms, and each sensor device is configured to generate one or more transient signatures based on the one or more transient signals. The system includes a server computing device communicatively coupled to the one or more sensor devices. The server computing device is configured to receive the transient signature from each of the sensor devices. The server computing device is configured to analyze the transient signature to identify one or more discharge indications. The server computing device is configured to generate one or more alarm signals when one or more discharge indications are identified.

[0008] In another aspect, the invention features a computerized method for detecting an electrical discharge preceding an electrical fire in electrical wiring. One or more sensor devices coupled to an electrical circuit detect one or more signal waveforms generated by electrical activity on the circuit. Each sensor device identifies one or more transient signals within the one or more signal waveforms, and each sensor device generates one or more transient signatures based on the one or more transient signals. A server computing device communicatively coupled to the one or more sensor devices receives the one or more transient signatures from each of the sensor devices. The server computing device analyzes the transient signatures to identify one or more discharge indications. When one or more discharge indications are identified, the server computing device generates one or more alarm signals.

[0009] In another aspect, the invention features a sensor device for detecting an electrical discharge preceding an electrical fire in electrical wiring, the sensor device being coupled to an electrical circuit. The sensor device includes a module that senses electrical activity on the electrical circuit and detects one or more signal waveforms of the electrical activity. The sensor device includes a processor that: identifies one or more transient signals within the one or more signal waveforms; generates one or more transient signatures based on the identified transient signals; analyzes the transient signatures to identify one or more discharge indications; and generates one or more alarm signals when one or more discharge indications are identified.

[0010] In another aspect, the invention features a system for detecting an electrical discharge preceding an electrical fire in electrical wiring. The system includes one or more sensor devices coupled to an electrical circuit, each sensor device configured to detect one or more signal waveforms generated by electrical activity on the electrical circuit. The system includes a server computing device communicatively coupled to the one or more sensor devices. The server computing device receives the one or more signal waveforms from each sensor device. The server computing device analyzes the one or more signal waveforms to identify one or more discharge indications. When the one or more discharge indications are identified, the server computing device generates one or more alarm signals.

[0011] In another aspect, the invention features a computerized method for detecting an electrical discharge preceding an electrical fire in electrical wiring. One or more sensor devices, each coupled to an electrical circuit, detect one or more signal waveforms generated by electrical activity on the electrical circuit. A server computing device, communicatively coupled to the one or more sensor devices, receives the one or more signal waveforms from each sensor device. The server computing device analyzes the one or more signal waveforms to identify one or more discharge indications. When the one or more discharge indications are identified, the server computing device generates one or more alarm signals.

[0012] In another aspect, the invention features a sensor device for detecting an electrical discharge preceding an electrical fire in electrical wiring. The sensor device includes a module that detects one or more signal waveforms generated by electrical activity on a circuit. The sensor device includes a processor that: receives the one or more signal waveforms from each sensor device; analyzes the one or more signal waveforms to identify one or more discharge indications; and generates one or more alarm signals when the one or more discharge indications are identified.

[0013] Any of the above aspects may include one or more of the following features. In some embodiments, one or more signal waveforms include a full voltage cycle waveform. In some embodiments, one or more sensor devices sample the full voltage cycle waveform at a frequency in a range between 10 MHz and 100 MHz.

[0014] In some embodiments, identifying one or more transient signals within one or more signal waveforms includes: a) dividing samples of the full voltage cycle waveform into a plurality of bins; b) determining an average and a maximum value for each bin in the plurality of bins; c) determining a difference between the average and the maximum value; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum value for each bin over all samples; and e) determining a derivative of each cumulative maximum value over the plurality of bins. In some embodiments, generating one or more transient features based on the identified transient signals includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform, and determining an average transient amplitude over a plurality of phase segments within the voltage cycle.

[0015] In some embodiments, analyzing the transient signature to identify one or more discharge indications includes determining a ratio of an average peak transient in one or more of the phase segments near the maximum voltage to an average peak transient near a zero crossing of the voltage cycle, and identifying the average peak transient in one or more of the phase segments near the maximum voltage as a discharge indication when the ratio is above a predetermined threshold. In some embodiments, generating one or more alarm signals when one or more discharge indications are identified includes determining a count of the identified discharge indications occurring within a predetermined amount of time, and generating one or more alarm signals based on the count of the identified discharge indications.

[0016] In some embodiments, identifying one or more transient signals within one or more signal waveforms includes: a) determining a derivative of samples of the full voltage signal waveform over a full voltage cycle; b) dividing the samples of the full voltage cycle waveform into a plurality of bins; c) determining a maximum value for each of the plurality of bins; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum value for each bin over all samples; and e) determining a derivative of each cumulative maximum value over the plurality of bins. In some embodiments, generating one or more transient features based on the identified transient signals includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform, and determining an average transient amplitude for a plurality of phase segments within the voltage cycle.

[0017] In some embodiments, analyzing the transient signature to identify one or more discharge indications includes determining a ratio of an average peak transient in one or more of the phase segments near the maximum voltage to an average peak transient near a zero crossing of the voltage cycle, and identifying the average peak transient in one or more of the phase segments near the maximum voltage as a discharge indication when the ratio is above a predetermined threshold. In some embodiments, generating one or more alarm signals when one or more discharge indications are identified includes determining a count of the identified discharge indications occurring within a predetermined amount of time, and generating one or more alarm signals based on the count of the identified discharge indications.

[0018] In some embodiments, when one or more signal waveforms are detected to have reached a threshold, one or more sensor devices sample the full voltage cycle waveform at 80 MHz. In some embodiments, identifying one or more transient signals within the one or more signal waveforms includes: identifying one or more samples of the full voltage cycle that exceed the threshold, and storing the identified one or more samples. In some embodiments, for each identified sample, generating one or more transient features based on the identified transient signal includes: determining the rise time of the peak in the identified sample; determining the rise time of the peak in the identified sample; determining the pulse width of the identified sample; and determining the integral of the identified sample. In some embodiments, analyzing the transient features to identify one or more discharge indications includes: classifying the identified sample as a discharge indication when the count of the peaks in the identified sample is above a predetermined threshold and when the rise time of the peaks in the identified sample is above a predetermined threshold. In some embodiments, generating one or more alarm signals when one or more discharge indications are identified includes: determining the count of the identified discharge indications that occur within a predetermined amount of time, and generating one or more alarm signals based on the count of the identified discharge indications.

[0019] In some embodiments, one or more sensor devices identify one or more transient signals within one or more signal waveforms, and generate one or more transient features based on the identified transient signals using a transient detection curve stored in a memory module of the sensor device. In some embodiments, a server computing device generates an updated transient detection curve based on transient features received from one or more sensor devices in the sensor devices, and sends the updated transient detection curve to each sensor device in the one or more sensor devices. In some embodiments, the one or more sensor devices apply the updated transient detection curve to identify subsequent transient signals within the one or more signal waveforms and generate one or more transient features.

[0020] In some embodiments, the server computing device sends one or more alarm signals to the remote computing device. In some embodiments, the server computing device sends one or more alarm signals to at least one sensor device in the one or more sensor devices. In some embodiments, at least one sensor device in the one or more sensor devices activates a visual indicator upon receiving at least one of the one or more alarm signals. In some embodiments, the visual indicator is a light emitting diode (LED) component of at least one sensor device in the one or more sensor devices. In some embodiments, the server computing device uses one or more machine learning algorithms to analyze the one or more signal waveforms to identify one or more discharge indications.

[0021] Other aspects and advantages of the present invention will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, illustrating by way of example only the principles of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The advantages of the present technology described above, as well as further advantages, may be better understood by reference to the following description taken in conjunction with the accompanying drawings, which are not necessarily to scale, emphasis instead generally being placed upon illustrating the principles of the present technology.

[0023] Figure 1A is a schematic diagram of a parallel arc fault.

[0024] Figure 1B is a schematic diagram of a series arc fault.

[0025] Figure 1C Here is a schematic diagram of the light-emitting connection.

[0026] Figure 2 is a block diagram of a system for detecting electrical discharges preceding electrical fires in electrical wiring.

[0027] Figure 3 is a schematic diagram of an exemplary electrical transmission network.

[0028] Figure 4A is a schematic diagram of field-collected measurements of an exemplary communication channel for an indoor, residential power line network.

[0029] Figure 4B Schematic diagram of the transmission characteristics of three different circuits and a comparative 20 m long cable measured in a laboratory setting.

[0030] Figure 5 is a flow chart of a computerized method of detecting electrical discharges preceding electrical fires in electrical wiring.

[0031] Fig. 6A is a detailed flow chart of a computerized method for detecting electrical discharges prior to electrical fires in electrical wiring.

[0032] Figure 6B is a detailed flow chart of a computerized method for detecting electrical discharges prior to electrical fires in electrical wiring.

[0033] Figure 6C is a detailed flow chart of a computerized method for detecting electrical discharges prior to electrical fires in electrical wiring.

[0034] Figures 7A-7C is a schematic diagram depicting exemplary waveform data generated by a sensor device that is indicative of transients in discharge activity.

[0035] Figures 8A-8C is a schematic diagram of exemplary waveform data generated by a sensor device for identifying transient characteristics.

[0036] Fig. 9 is a schematic diagram of the transient amplitude ratio changing with time.

[0037] Fig.10 It is a schematic diagram showing how the probability of discharge activity changes over time.

[0038] Fig.11 is a schematic diagram of an exemplary user interface of a remote computing device for displaying an alert signal notification.

[0039] Fig.12 is a schematic diagram of exemplary waveform data for a safety signal generated by a test device.

[0040] Fig.13 It is possible to Figure 2 A block diagram of a proxy device used in conjunction with a system.

[0041] Figures 14A-14E is a schematic diagram showing fire precursor arc signals collected by sensor equipment installed in a test home.

[0042] Fig.15 It is shown Fig. 14C A schematic diagram of the arc signal in more detail.

[0043] Fig.16 is a schematic diagram showing more detail of a single pulse from a discharge.

[0044] Fig.17 is a schematic diagram with several graphs showing system test results.

[0045] Fig.18 is a histogram showing the amplitude of pulses from laboratory data compared to the amplitude of pulses recreated by the proxy device.

[0046] Fig.19 is a plot showing how fire precursor signals appear over time.

[0047] Fig. 20 is a plot of devices operating on an electrical network that are generating HF electrical activity. DETAILED DESCRIPTION

[0048] Figure 2 2 is a block diagram of a system 200 for detecting electrical discharges that precede electrical fires in electrical wiring. System 200 includes electrical wires 202 that include circuits in a power distribution system 220 (e.g., an electrical system in a residential or commercial building) that transmits power received from a service entrance cable 222 to branch circuits (such as wires 202) that feed power to appliances and outlets (e.g., outlets 208) in the building. The electrical discharges produce very fast current pulses that travel throughout the electrical power distribution system via transmission line effects. The high-frequency signals produced by the electrical discharges travel through the home, much like the high-frequency signals used in power line communications.

[0049] Typically, the wires that provide power to the circuits, outlets, devices, and appliances in a building are the medium through which the fire precursor signal travels from the origin to the sensor. The precursor is a pulse discharge and causes an electromagnetic signal to travel along the power distribution wires in the building. Each location in the wiring (both fixed house wiring and mobile cords to devices and appliances) has a transmission function to each outlet in the house, and the precursor signal arrives at the sensor modified by the transmission function. The methods and systems described herein utilize methods similar to those developed in power line carrier communication technology to measure and identify the very wide frequency content of the electrical network traveling through the home.

[0050] Figure 3 is a schematic diagram of an exemplary electrical network serving one or more residences. Figure 3As shown, power from a high voltage power line 300 passes through an electrical transformer 302 and is delivered to one or more residences (e.g., residence A 304, residence B 306) via a +120V AC line (L1) 308, a neutral line 310, and a -120V AC line (L2) 312. Lines L1 and L2 are coupled to a plurality of branch circuits 314 via circuit breakers 316 in a service panel 318 inside the residence. It should be understood that communication from any circuit on L1 to any circuit on L2 is possible.

[0051] The use of power line communications has been considered and used since the early 20th century. In the past 20 years, various companies have released products that provide 100Mbps to 1Gbps local area network communications through the power distribution system within a building. This work led to the standardization of power line communications in IEEE 1901 and the widespread availability of "Ethernet over powerline" adapters for consumer use, and broadband communications over power lines within buildings is a mature technology.

[0052] There are two main issues associated with communicating over power lines (i.e., channels): the transmission and attenuation characteristics of the transmitters and receivers on the different circuits of a building, and the channel noise environment. The power line channel is challenging because there are many discontinuities due to branches (e.g., each outlet on the circuit) and loads having frequency- and time-dependent impedances (e.g., due to different device power draws, constructions, and switching characteristics). The devices responsible for the time variations are also responsible for both narrowband noise injection and broadband noise injection due to switching power supplies and other nonlinear device elements. Exemplary techniques related to currently available power line communications are described in “State of the art in power line communications: From the applications to the medium” by C. Cano, A. Pittolo, D. Malone, L. Lampe, A. M. Tonello, and A. G. Dabak, published in IEEE J. Sel. Areas Commun., Vol. 34, pp. 1935-1952, July 2016 (available at https: / / arxiv.org / pdf / 1602.09019.pdf), which is incorporated herein by reference.

[0053] Power line channel transmission characteristics have been the subject of significant field measurement and modeling activities, which culminated in the approval of the IEEE 1901 standard in 2010 (as described in "IEEE Standard for Broadband Over Power Line Networks: Medium Access Control and Physical Layer Specifications, IEEE Standard 1901-2010," published in September 2010). Figure 4A Field-collected measurements of an exemplary communication channel for an indoor, residential power line network are shown. Figure 4A The top diagram in the figure shows multiple pairs of transmitters and receivers on the same circuit, and Figure 4A The bottom graph in shows multiple transmitter and receiver pairs where the units are on different circuits (described in M. Tlich, A. Zeddam, F. Moulin, F. Gauthier, "Indoor power-line communications channel characterization up to 100 MHz—Part I: One-parameter deterministic model," IEEE Trans. Power Del., Vol. 23, No. 3, pp. 1392-1401, July 2008). Measurements obtained in Spain show similar characteristics, confirming that variations in trunk configuration and wiring style result in similar overall characteristics (see C. Cano et al., supra). A more interesting and difficult case is Figure 4B , which shows the transmission characteristics of three different circuits and a comparative 20 m long cable measured in a laboratory setting (described in E. Liu, Y. Gao, O. Bilal, and T. Korhonen, "Broadband characterization of indoor powerline channel," Proc. Int. Symp. Power Line Commun., Zaragoza, Spain, March 31-April 2, 2004).

[0054] Figure 4A and Figure 4BThe graph of shows variable frequency dependence for a particular channel, but relatively flat attenuation on average. The sites used to collect this data vary in size from apartments to large homes and in building age. The channel characteristics are similar in many ways to fixed wireless communication channels. Wireless channels experience attenuation (i.e., frequency, location, and time-dependent attenuation) due to reflections of signals from objects in the environment (i.e., multipath-induced attenuation). Because discontinuities, unmatched impedances, and branches in the wiring cause signal reflections, the same effects are at work here. In the time domain, these reflections manifest themselves as amplitude and time-shifted copies of the original signal arriving at the receiver. In the frequency domain, the frequency components of these copies interfere constructively or destructively, causing peaks and grooves in the transmission characteristics as observed in the graph.

[0055] Power line communication systems use the same techniques employed in broadband wireless LANs (e.g., IEEE 802.11n) for signal coding and noise resilience, primarily orthogonal frequency division multiplexing (OFDM). OFDM is a good solution because it is a technique for propagating information over a wide spectrum band while also reducing the effects of signal dispersion and fading.

[0056] The power line environment is challenging for communications, but the market success of Ethernet over power line products has proven that broadband signals are possible across complex electrical networks. Narrowband communications are very challenging due to the extreme frequency selectivity that depends on individual, time-varying channel conditions. Circuit-to-circuit signal attenuation is significant, but is typically less than 50 dB, making it possible for a signal originating almost anywhere in the electrical wiring to be detected at a receiver elsewhere in the house wiring (i.e., at the power outlet).

[0057] Back to Figure 2, the system 200 further includes a sensor device 204, which is coupled to an electrical outlet 208 of the branch circuit conductor 202 via a 120VAC plug 206. The sensor device includes electronic components (e.g., processing module 204a, ADC 204b, CPU 204c) that couple the sensor device to the electrical infrastructure in a manner that allows the discharge signal to be amplified, while in some embodiments, filtering out unwanted 60Hz signals and electrical noise generated by appliances operating on the electrical system. The processing module 204a includes components such as capacitors, resistors, and amplifiers that sense electrical activity occurring on the branch circuit wires 202 and capture the sensed power as waveform data. In some embodiments, the processing module 204a includes a filter that can limit the frequency response of the sensor device to a range where the discharge has a high signal-to-noise ratio. The filtering can be implemented using hardware components or in firmware installed on the processing module 204a. It should be noted that the electrical activity occurring on the branch circuit wires 202 includes signals sent to and from the power distribution system 220. In this way, a single sensor can see the discharge signal throughout the power distribution system 220 (including other branch circuits). Figure 2 A single sensor device 204 is depicted, but it should be understood that system 200 may include two or more sensor devices positioned to sense electrical activity in the power distribution system. Multiple sensors sending data to the server computing device may provide increased sensitivity and work together to provide information about where electrical discharges are occurring.

[0058] In some embodiments, the location can be determined using a time of arrival technique, in which the difference in the arrival time of the transient signal to each sensor device provides the relative distance to the discharge location on the wire. Typically, the sensor device 204 is configured to see the shape of the pulse. The pulse shape seen by the sensor device is the source pulse shape plus the reflection of that shape from each branch, terminal, or impedance change in the electrical system wiring. The time between reflections is the product of the distance difference divided by the group velocity. The amplitude and polarity of the reflections are determined by the complex impedance of the load or joint. The time of the various reflections tells the location of the source in the circuit; the shape changes of these reflections tell about the impedance of the load and joint that produced the reflection. For example, if two sensor devices are installed at different locations on the same circuit, the difference in the arrival time at the two devices can determine the location in the circuit - even for pulses that cannot see reflections above the noise.

[0059] In one embodiment, the sensor device 204 can capture up to 27 million waveform samples per second and distribute the samples into 512 bins. The sensor device 204 determines the maximum amplitude in each bin and generates waveform data using the binned values. The sensor device 204 also performs the function of obtaining derivative values ​​of the waveform signal in order to reduce the bandwidth and processing requirements of the sensor hardware. For example, the sensor device 204 can discard adjacent waveform samples to obtain a first-order derivative, and then repeat the process to obtain a second-order derivative. In one embodiment, the sensor device 204 adds several adjacent samples and subtracts several adjacent samples to obtain a wavelet, which is then analyzed relative to the previously captured wavelet to determine whether any changes have occurred in the signal.

[0060] Sensor device 204 is communicatively coupled to server computing device 214 via communication network 212. In one embodiment, sensor device 204 is equipped with communication components (e.g., antenna, network interface circuitry) that enable sensor device 204 to communicate with server computing device 214 via a wireless connection.

[0061] The communication network 212 enables the other components of the system 200 to communicate with each other in order to perform the process of detecting discharges in electrical wiring as described herein. The network 212 can be a local network such as WiFi or a LAN, or a wide area network such as the Internet and / or a cellular network. In some embodiments, the network 212 is composed of several discrete networks and / or sub-networks (e.g., cellular to the Internet) that enable the components of the system 200 to communicate with each other.

[0062] The server computing device 214 is a combination of hardware and dedicated software modules, the hardware including one or more dedicated processors and one or more physical memory modules, and the dedicated software modules (such as transient analysis module 214a) are executed by the processor of the server computing device 214 to receive data from other components of the system 200, send data to other components of the system 200, and perform functions for detecting discharges in electrical wiring as described herein. In some embodiments, module 214a is a set of dedicated computer software instructions programmed onto a dedicated processor in the server computing device 214, and may include specially designated memory locations and / or registers for executing the dedicated computer software instructions. Further explanation of the dedicated processing performed by module 214a will be provided below. It should be understood that in some embodiments, the sensor device 204 can be configured to operate as a stand-alone device, because the processing described herein with respect to the server computing device 106 can be performed by the sensor device 204 (i.e., the processor and memory can be embedded in the sensor device, which performs the data collection, analysis and alarm processes described herein).

[0063] The database 216 includes transient memory and / or persistent memory for data storage used in conjunction with the process of detecting electrical discharges in electrical wiring described herein. In general, the database 216 is configured to receive, generate, and store specific data segments for use by the server computing device 214. In some embodiments, all or a portion of the database 216 may be integrated within the server computing device 214 or located on a separate computing device or devices. For example, the database 216 may include a MySQL database such as that available from Oracle Corporation of Redwood City, California. TM In another example, database 216 may include a database using DynamoDB TM Cloud-based storage media such as Amazon Web Services (AWS) TM In other embodiments, the database may be located in memory on, for example, server computing device 214 and / or sensor module 204 .

[0064] Figure 5 is used Figure 2 Flowchart of a computerized method 500 for detecting early arc faults in electrical wiring of the system 200 of FIG. One or more sensor devices (e.g., sensor device 204) coupled to a branch circuit (e.g., branch 1 circuit including conductor 202) detects (502) one or more signal waveforms generated by electrical activity on the branch circuit. For example, because a discharge is a pulsed change in current, the signals generated by the discharge travel down the wire 202, reflecting around any joints. These signals have a delay when returning, which is generally based on the length of the wire. The signals also exhibit a phase shift that enables the system 200 to identify transient signals in the waveform that represent activity involving a discharge (even a very small discharge) on the wire 202 as described herein.

[0065] The sensor device 204 identifies (504) one or more transient signals within the detected one or more signal waveforms. The sensor device 204 generates (506) transient signatures based on the identified transient signals and sends the transient signatures to the server computing device 214. The transient analysis module 214a receives (508) a transient signature set from each sensor device 204 and analyzes (510) the transient signatures to identify one or more discharge indications. When one or more discharge indications are identified, the alarm generation module 214b generates (512) one or more alarm signals.

[0066] It should be understood that in some embodiments, the sensor device 204 can detect one or more signal waveforms (step 502) and send the one or more signal waveforms to the transient analysis module 214a for further processing, including: identifying transient signals (step 504), generating transient features (step 506), analyzing transient features (step 510), and generating an alarm signal (step 512). It should also be understood that in some embodiments, the sensor device 204 can internally perform Figure 5 In some embodiments, the sensor device 204 compresses one or more signal waveforms before sending the signal waveforms to the transient analysis module 214a in order to save network bandwidth, improve processing power and performance, etc.

[0067] In some embodiments, the alarm generation module 214b sends one or more alarm signals to a remote computing device (e.g., a mobile phone, a tablet, a smart watch, etc.). The remote computing device may, for example, display a message or indicator (such as a warning icon) on a screen associated with the remote computing device based on receiving the one or more alarm signals. For example, the alarm signal may include a packet-based message that includes a corpus of text indicating a dangerous condition or hazard detected by the system 200. In some embodiments, the alarm generation module 214b sends one or more alarm signals to one or more of the sensor devices 204. The sensor device 204 may activate one or more components (e.g., embedded components in the sensor device) upon receiving the one or more alarm signals. For example, upon receiving an alarm signal from the alarm generation module 214b, the (multiple) sensor device 204 may activate an LED element that lights up and / or flashes on the exterior of the sensor device 204 to indicate that a dangerous condition or hazard has been detected by the system 200. Reference is made below to FIG. 6A to FIG. 6C Provides information about Figure 5 Additional details of the steps.

[0068] FIG. 6A to FIG. 6C Including the use of Figure 2 The system 200 and according to the above reference Figure 5 A detailed flow chart describing the framework of a computerized method for detecting electrical discharges prior to electrical fires in electrical wiring. FIG. 6A to FIG. 6C Three different methods are included that can be used by system 200 to detect discharges: Method A ( Fig. 6A ), Method B( Figure 6B ) and method C( Figure 6C). It should be understood that these methods are exemplary and other methods may be contemplated for use with the systems described herein. Moreover, it should be understood that methods A, B, and C may be used independently or in conjunction with one another. In one embodiment, the sensor device 204 may include multiple logical processors and / or physical processors (e.g., CPU 204c), each of which processes waveform data according to one of methods A, B, or C.

[0069] For example, in methods A and B, the high-speed analog-to-digital converter (ADC) 204b of the sensor device 204 detects the waveform (502) by capturing the full voltage cycle waveform (1 / 60 of 1 second) and sampling at 27 MHz; while in method C, ADC204b captures the full voltage cycle waveform (1 / 60 of 1 second) and samples at 80 MHz when it detects that the waveform has reached or exceeded the threshold (to save bandwidth and memory resources).

[0070] The CPU 204c of the sensor device 204 can identify transients in the waveform data (504) in several different ways. For example, in method A, the CPU 204c performs a binning process on the sampled waveform data by dividing 27 million samples into 512 bins. The CPU 204c calculates the difference between the maximum and minimum values ​​of each bin, and calculates the difference between the maximum and average values ​​of each bin. Next, the CPU 204c accumulates the maximum-average data of the bins for 15 cycles and determines the maximum value for 15 cycles. Then, the CPU 204c calculates the derivative on the bins of each accumulated maximum value.

[0071] In method B, CPU 204c calculates the derivative of the waveform over the full voltage cycle. CPU 204c divides the obtained 27 million samples into 512 bins and calculates the maximum value of each bin. Next, CPU 204c accumulates the maximum value data of the bins for 15 cycles and determines the maximum value of 15 cycles. Then, CPU 204c calculates the derivative on the bins of each accumulated maximum value.

[0072] In method C, the CPU 204c triggers storage of waveform samples that exceed a floating threshold that includes samples prior to the trigger - these are single transients within a cycle.

[0073] Continue to refer FIG. 6A to FIG. 6C, the CPU 204c of the sensor device 204 may generate the transient signature (506) in different ways. For example, in methods A and B, the CPU 204c calculates the average transient amplitude over the full voltage cycle, and calculates the average transient amplitude over 16 phase segments within the voltage cycle. In method C, the CPU 204c counts the number of peaks in the transient, calculates the rise time of the peaks, calculates the maximum amplitude of the transient, calculates the pulse width of the transient, and calculates the integral of the transient.

[0074] When the sensor device 204 determines that a change has occurred in the voltage waveform data based on the binning process described above, the sensor device 204 sends the transient signature data to the server computing device 214 via the network 212. The transient analysis module 214a receives (504) the transient signature data from the sensor device 204 and analyzes (506) the waveform data to identify one or more discharge indications in the signal waveform. In one example, the signal analysis module 214a reconstructs the signal waveform using the received data and compares the distribution of transients (including where in the phase the transients occur) to the expected transient distribution. For example, the signal analysis module 214a can determine where in the phase the transient occurs and how repeatable the transient is - for example, whether the transient exhibits a regular pattern on the waveform (which may indicate operation of a device or appliance on the circuit) or whether the transient exhibits an irregular pattern (which may indicate discharge and / or arc fault activity).

[0075] In some embodiments, the CPU 204c of the sensor device 204 sends the determined transient signature to the transient analysis module 214a of the server computing device 214, and the module 214a receives the transient signature. It should be understood that in some embodiments, the sensor device 204 can be configured as a stand-alone module that does not require a connection to the server computing device 214 - in these embodiments, the sensor device 204 can also perform the steps of analyzing the transient signature (510) and generating the alarm signal (512), and further, in some embodiments, perform the step of transmitting the alarm signal to the remote computing device (as previously described). In some embodiments, the alarm signal can be generated locally within the sensor device 204 and used to trigger other components of the sensor device 204 (such as a speaker or other audio component embedded in the sensor device 204, or a light or other visual component (e.g., LED) embedded in the sensor device 204) to warn nearby people that potentially dangerous electrical activity is occurring.

[0076] The transient analysis module 214a can analyze the transient features (510) received from the sensor device 204 in different ways. For example, in methods A and B, the module 214a determines the ratio of the average peak transients in the phase segment near the maximum voltage to the average peak transients near the voltage zero crossing. If the ratio is greater than a threshold, the module 214a identifies the transients as potential discharges. In method C, the module 214a evaluates whether the number of peaks in the transient exceeds a predetermined threshold and whether the rise time is greater than a predetermined threshold. If so, the module 214a identifies the transient as a potential discharge.

[0077] Figures 7A-7C and Figures 8A-8C is a schematic diagram of waveform data depicting a transient signal captured by the CPU 204 c of the sensor device 204 . Fig. 7A is an exemplary diagram of the waveform of two milliseconds of sampled voltage data captured at 16 million samples per second. Fig. 7A As shown, most of the waveform data is general electrical background noise generated by items such as electrical appliances and electromagnetic signals in the air. However, the waveform data also exhibits several large spikes (eg, 702, 704), which are transients that may indicate discharge activity in the circuit.

[0078] Figure 7B It shows that Fig. 7A An exemplary schematic diagram of a portion of a waveform, Figure 7B is enlarged to show additional detail. Figure 7B As seen in , the shapes of the individual transients (e.g., 712, 714) appear. Looking more closely, Figure 7C It shows that Figure 7B An exemplary diagram of a portion (i.e., two microseconds) of a waveform of Figure 7C is enlarged to show additional detail. Figure 7C As shown, transient 514 exhibits a ring-down structure, which indicates the amount of time it takes for a signal to decay. In addition, transient 714 has a very fast voltage change and a fast rise time. As explained above for method C, CPU 204c of sensor device 204 can use this type of sampled waveform data to generate transient signatures (506) for transmission to transient analysis module 214a.

[0079] Figures 8A-8C is an exemplary schematic diagram showing how the CPU 204c of the sensor device 204 may divide the sampled waveform signal into 512 bins (ie, as described above with respect to Method A and Method B). Fig. 8A An exemplary full voltage cycle waveform is shown in , where bin numbers are assigned on the x-axis and amplitude is represented on the y-axis. Figure 8Bis another exemplary voltage cycle waveform with annotations showing how the CPU 204c of the sensor device can identify transients as described above with respect to Method A and Method B. Figure 8B As shown, the CPU 204c can determine the current maximum, the current average, and the maximum over 15 cycles - as well as the derivative of the maximum over 15 cycles, the derivative of the current maximum, and the derivative over the full voltage cycle. Figure 8C 804 is an exemplary voltage cycle waveform with annotations depicting potential discharges identified by transient analysis module 214a. As described above with respect to methods A and B, transient analysis module 214a may determine a ratio of an average peak transient in a phase segment near the maximum voltage (e.g., segment 802) to an average peak transient near the voltage zero crossing (e.g., segment 804). Transient analysis module 214a may generate a representation of the ratio (see Fig. 9 ).like Fig. 9 As shown, the time period from 15:17:40 to 15:18:00 (indicated by 902) is the time when the discharge is generated on the electrical wiring.

[0080] In some embodiments, the transient analysis module 214a may include Fig. 9 The ratio data is converted into a metric indicating the likelihood of a discharge occurring based on the ratio. Fig.10 is shown based on Fig. 9 An exemplary illustration of the probability of a discharge occurring during a specific time period based on the ratio data. Fig.10 As shown, the probability of discharge moves from 0 to 2 at 15:17:40, and fluctuates between 0 and 2 until 15:18:00. In this example, a value of 2 indicates a high probability of discharge occurring.

[0081] After identifying the discharge indicator in the transient signature data, the alarm generation module 214b of the server computing device 214 generates an alarm signal related to the detected discharge indicator (512). In some embodiments, the alarm generation module 214b automatically identifies one or more remote devices that monitor the building or location (or are associated with the building or location) where the sensor device 204 is connected to the branch circuit and automatically sends an alarm signal to these remote devices. The remote devices may include computer-based devices such as mobile phones, tablets, desktop computers, smart appliances, IoT devices, smart watches, etc. The remote devices may also include speakers, sirens, lights, and other audio-visual indicator devices.

[0082] In some embodiments, database 216 includes information related to the identification of remote devices (e.g., IP addresses, telephone numbers, email addresses), and alarm generation module 214b uses the identification information to prepare an alarm signal for each remote device. In some embodiments, alarm generation module 214b uses any standard communication protocol or technology, such as packet-based delivery (e.g., text messaging, XML, email), circuit-based delivery (e.g., paging, voice messaging), etc. For example, the alarm signal can take the form of a packet-based communication (e.g., a message) having a header and a body including certain data elements. The alarm signal can include information related to the type of arc fault detected by system 200, the approximate location of the discharge activity (e.g., using the arrival time technique described above), identification of sensor device 204 and its location in the building, and other relevant information (e.g., identification of appliances or other electrical devices connected to the same branch circuit, etc.).

[0083] Fig.11 is an exemplary schematic diagram of a user interface of a remote computing device (eg, a mobile phone) for displaying a message to a user based on receipt of an alarm signal. Fig.11 As shown, the application installed on the remote computing device can be coupled to the server computing device 214 (e.g., via the Internet) and can be configured to automatically listen for the alarm signal pushed from the alarm generation module 214b. When the alarm signal arrives at the remote computing device, the application can automatically display a warning indicator 1102 (e.g., a red, flashing graphic or icon) to notify the user of the electrical fire hazard. The user interface can also include a log of previous discharge activities detected (e.g., other events in the user's home).

[0084] In this way, the systems and methods described herein provide significant advantages over currently available arc fault detection technologies because the present systems can detect small transients occurring in branch circuits that may indicate discharge activity in the circuit more easily and with much greater accuracy than conventional arc fault detectors. The systems and methods described herein also utilize a network-based processing architecture to capture signal data over time for a particular branch circuit, which can both enable immediate detection of changes in the electrical activity profile of the branch circuit and provide upstream entities (such as utility companies and appliance manufacturers) with useful information about the type of anomalies that may be occurring in the power system.

[0085] Moreover, it should be understood that multiple different sensor devices may be installed in various buildings on a public grid operated by a utility provider (e.g., as described above, many different homes may have sensor devices attached to the building's power system to monitor discharges). Each of these distributed sensor devices may communicate with one or more centralized server computing devices that send data to and receive data from the sensor devices. In this configuration, the server computing devices may collect discharge detection data from the sensor devices and aggregate the data for analysis. In one example, the server computing devices may be embedded with machine learning techniques and algorithms that use discharge and / or transient data from multiple sensor devices installed in different homes and buildings, as well as feedback from end users, to improve their transient detection and characterization algorithms.

[0086] In some embodiments, the analysis of transient features is enhanced by the application of machine learning techniques to identify one or more discharge indicators. https: / / xgboost.readthedocs.io / en / latest / tutorials / model.html Techniques such as those described in Friedman and Jerome H., "Gradient Function Approximation: A Gradient Boosting Machine," The Annals of Statistics, Vol. 29, No. 5 (October 2001), pp. 1189-1232, which are incorporated herein by reference, allow for the automatic selection of various features based on real data sets from houses or buildings with known discharge signals or real data sets generated in a laboratory. Machine learning allows for the development of more detailed computational models and relationships between transient features and discharge indications than a human can manually derive. In some embodiments, time series features such as autocorrelation lags and kurtosis from these transient features are further calculated over a tumbling window and provided to the machine learning model to provide further context and improve accuracy. As the real data set for both discharge indications and false positives expands, the machine learning model can be improved in a continuous cycle of improvement and further developed firmware can be deployed to the sensor.

[0087] Additionally, the server computing device can distribute updated software methods (i.e., in the form of firmware or similar upgrades) for identifying transients and transient signatures to the sensor devices for execution - thereby automatically synchronizing the sensor network with the latest algorithms to protect the home. Moreover, multiple sensor devices placed in multiple homes can correlate data about commonly detected voltage waveforms and transients to pinpoint the location of potential grid problems, and communicate with the grid operator or utility provider to take action.

[0088] Another aspect of the systems and methods described herein may include incorporating devices that generate signals that appear to be dangerous electrical discharges, but are instead introduced into the electrical wiring via separate electronic devices, which in some examples are embedded in the sensor device 204. These signals, also referred to as safety signals, may be used to test the operation of the sensor device 204 so that any malfunctions or other problems with the sensor device 204 may be identified. Fig.12 is a schematic diagram of an exemplary waveform generated from the detection of a safety signal. Fig.12 As shown, these transients are similar to the type of transients that would be generated if a dangerous discharge occurred on the wiring. It should be noted that these transients are small enough to be detected by the sensor device 204, but not large enough to trip the AFCI.

[0089] Another important aspect of the systems and methods described herein is the development of a device for testing the effectiveness of sensor equipment in detecting fire precursors occurring anywhere in a home. One possible way to achieve this is to bring damaged cords that produce fire precursor pulses to many homes and measure the pulse signals detected from other sockets in the home. The disadvantage of this method is that it will take a lot of time to send trained technicians who can safely generate signals and the study will only measure signals over short periods of time (when someone is in the home monitoring the system). Rational people will not welcome damaged (and dangerous) cords in their homes for a long period of time.

[0090] An alternative and more scalable approach is to use a proxy device that can replay fire signals generated and recorded in the lab by real fire precursors. This proxy device allows for economical deployment of hundreds of units to various types of homes across countries, as well as the ability to track signal detection performance over long periods of time. Fig.13 It is possible to Figure 2 200 is a block diagram of a proxy device 1300 used in conjunction with the system 200.

[0091] like Fig.13As shown, the proxy device 1300 includes a flash memory module 1302, a CPU 1304, a digital-to-analog signal converter (DAC) 1306, and a power amplifier 1308. The device 1300 utilizes a DAC and a power amplifier (coupled to the electrical system via a power capacitor) to simulate a fire precursor signal. The DAC 1306 can generate a signal at 120 MHz in a short period of time. The CPU is phase-locked to the zero crossing from the main power supply to generate pulses that occur at the appropriate position in the phase. The pulse information is read from the flash memory by the CPU, and the recorded pulses are reproduced through the D / A converter and the power amplifier.

[0092] When the agent device 1300 is deployed in multiple homes, the agent device 1300 helps answer the following questions:

[0093] 1) How does the fire precursor arc signal travel through the home's electrical system?

[0094] 2) How well do the methods and systems described herein identify electrical fire signals and distinguish these signals from those generated by other electrical appliances in the home?

[0095] In order to determine how the fire precursor pulse signal travels in the home electrical system, the agent device 1300 and the sensor device 204 and the test fixture can be deployed in the test home, and when the test fixture is plugged into the electrical socket in the home, a scintillation pulse is generated for detection. The test fixture includes a power cord that is plugged into a standard wall socket with a live conductor, a neutral conductor, and a ground conductor. The power cord is spliced ​​to include a resistor, and any current generated by the flash or discharge must pass through the resistor. The end of the power cord is connected to a plug, through which various test devices can be connected. The test fixture includes a plastic NEMA shell, and a damaged extension cord passes through the plastic NEMA shell. The damaged cord can be exposed to various substances that cause discharge to occur. For example, the substance can be graphite powder, water, or a solution of water, soap and salt. The differential analog-to-digital converter measures the voltage on the resistor of known resistance to calculate the current flowing through the resistor. Resistors of various sizes were selected to provide the appropriate amount of gain, depending on the expected peak current generated by exposing the damaged cord to various substances. The test home was a large single-family home of approximately 4,000 square feet. The home had a typical type of electronic equipment, including a flat-screen TV, audio equipment, and computers. The electronic equipment is usually protected by a surge suppressor power strip.

[0096] Figures 14A-14E is a schematic diagram showing fire precursor arc signals collected from a test home. Fig.14AThe graph in shows a 60 second measurement of the peak analog-to-digital converter output from two sensors. The sensor devices labeled "Same Ting" are on the same 120V leg / phase and on the same branch circuit as the test fixture. The sensor devices labeled "Other Ting" are on the opposite 120V leg / phase of the power network on a different branch circuit. The larger amplitude signal indicates when the test fixture is generating a fire precursor pulse. For example, in Fig.14A In the highlighted area in region 1402, the same ding line and other ding lines jump up to about 2400 raw digital units - indicating that both sensor devices detected the precursory fire pulse. Fig. 14B The graph in Figure 2 shows the time from Fig.14A The highlighted time period 1402.

[0097] for Fig. 14C The graphics in Fig. 14B The data was further magnified to 20 milliseconds - this is Fig. 14B The area 1404 in FIG. Fig. 14C , a measurement of the current 1406 through a 10 ohm resistor on the live wire of a test fixture is shown. This is the current when a breakdown occurs in an insulator and this current is observed as a discharge and is evidenced in the form of heat and light.

[0098] Fig.14D and Fig.14E A further zoomed-in time period of the current flowing through the 10 ohm resistor is shown, along with the resulting waveform as seen by the sensor device. Fig.14D The graphics in will come from Fig. 14C The highlighted time period 1408 is shown zoomed in to 500 milliseconds, while Fig.14E The graphics in will come from Fig.14D The highlighted time period 1410 is shown zoomed in to 80 milliseconds.

[0099] Fig.15 Shows Fig. 14C A more detailed version of Figure 1 highlights two times when electrical breakdown occurs and the current through the 10 ohm resistor measured by the sensor device at those two times. In this example, it should be understood that electrical breakdown can produce currents that vary slowly over a period of many milliseconds to currents that start and stop very quickly on the order of nanoseconds to tens of nanoseconds. As shown in Fig.15 As can be seen in FIG. 1 , slowly changing current produces a visible signal 1502 in a sensor device on the same circuit (“same ding”) but not in a sensor device on a different leg (“other ding”), while rapidly changing current is visible in the signals from both sensor devices.

[0100] Fig.16 A more detailed display of one of the pulses from the discharge is shown, with some information about the rate of change of the current as it begins to flow, then experiences several interruptions to the current flow, until it eventually extinguishes, noting that at each rapid change in the current both sensor devices see a sharp rise in the signal followed by a characteristic ring down. This indicates that the characteristics of the discharge that have the greatest impact on how the signal travels through the grid are very rapid rise and fall times of the current. Note that as the discharge begins the characteristic ring down signal 1602 appears, but the sensor device on the opposite phase does not have a strong response as the current continues at the same level. This is an important finding for the development of a proxy device, as it means that as long as the proxy device can simulate the rise and fall times of the fast pulse current that is a precursor to a fire, then the proxy device can recreate the most relevant parts of the signal and does not need to recreate the longer continuous current.

[0101] It should be understood that the proxy device 1300 used as an arbitrary waveform generator does not perfectly reproduce the laboratory waveform. The proxy device 1300 can only generate voltage steps at certain digital clock edges, and the proxy device 1300 can only generate steps of certain discrete amplitudes. The system performs lossy compression on the data, retaining only the largest peaks and preserving their shapes. The impedance of the proxy device 1300 is different from the impedance of the damaged portion of the insulation, so the reflection from the proxy device is different from the reflection from the damaged cable. As a result, it is important to confirm that the proxy device is sufficient to produce a signal at the sensor device 204 that is similar to the signal generated by real flicker in the damaged insulation.

[0102] To test this, the insulation on the wire is damaged in the lab, the current generated by the flicker in the damaged insulation is measured, and the voltage generated by these flickers at the sensor devices on different circuits in the lab is simultaneously measured. The proxy device 1300 is then programmed to replay the measured flicker, replace the damaged insulation with the proxy device, and again simultaneously measure the current generated at the proxy device and the voltage generated at the sensor device on the different circuit.

[0103] Fig.17The results of the tests described above are shown. Graph 1702a shows the current from the damaged insulation that was initially measured in the lab. Graph 1704a shows the voltage measured by the sensor device at another circuit, which is produced by the current pulse shown in graph 1702. Graph 1706a shows the current produced by the proxy device, and graph 1708a shows the voltage measured by the sensor device on another circuit. Graphs 1702b, 1704b, 1706b, and 1708b show the same measurement of graphs 1702a, 1704a, 1706a, and 1708a, respectively, at a higher time resolution. The sensor voltages produced by the proxy device are not the same as those initially measured, but they are similar enough to be used.

[0104] The second characteristic of the fire precursor pulse signal that determines whether the fire precursor pulse signal travels through the house is the amplitude of the pulse. A pulse of larger amplitude produces a larger response in the sensor device 204. The relationship between the amplitude of the current and the size of the sensor device response can be seen in graphs 1702b and 1704b. The larger current amplitude in graph 1702b is observed as a larger sensor response in graph 1704b. Similarly, the larger amplitude current generated by the proxy device results in a larger response at the sensor device (see graphs 1706b and 1708b). A current of the order of 25mA is sufficient to produce a response in the sensor device 204 with a signal-to-noise ratio sufficient to be detected.

[0105] Fire precursors can generate thousands of pulses on a single power cycle, and a large percentage of these pulses are not large enough in amplitude to travel through the electrical infrastructure and be seen by the sensor devices on the opposite legs. The agent device 1300 is capable of generating a limited number of pulses per cycle. The agent device is programmed to focus on generating the largest and fastest pulses expected to be detected throughout the house. In some embodiments, the agent device 1300 can communicate with a server computing device (e.g., server computing device 214) to inform the server of the time when the agent device 1300 operates and the operating mode of the agent device 1300. In this case, it is easy to associate the time when the agent device indicates that it is running with the sensor device output to verify that the sensor device detects a signal. In this example, the agent device 1300 is programmed to create a fast pulse with an amplitude ranging from 0 mA to 300 mA. Fig.18 A histogram of pulse amplitudes found in a five second data set obtained in the lab, labeled "real" (dark gray, 1802) is shown. Additionally, a histogram of pulse amplitudes recreated by a proxy device is shown in light gray (1804). Typically, the counts of proxy pulses are smaller than the counts from the real data source. As described above, this is because of limitations on proxy device memory to hold enough data to reproduce every pulse, so the focus is on reproducing the largest pulses.

[0106] Note that the pulse current amplitudes in the proxy test data set are small compared to the peak amperes that trigger the arc fault circuit interruption. For a parallel arc, the AFCI will trip at about 50 amps, and for a series arc, the AFCI will trip at 5 amps (as described in The 51 st The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault. The invention relates to a method for detecting a flash of a firefault.

[0107] Having shown that fire precursors create a signal on the power network that travels throughout the power network and can be detected by a single sensor device, a second question about detection efficiency concerns the degree to which the sensor device can distinguish the fire precursor signal from man-made or other interfering signals that may be sensed on the power line. The fire precursor signal exhibits certain characteristics of parallel arcs that can be used for identification (note that series arcs have other characteristics that can be used for identification - such as a low signal at the zero crossing with a large amplitude pulse just outside the zero crossing region, and that a glowing connection has yet different characteristics that can be used for identification):

[0108] 1) The average value of the fire precursor pulse signal is larger near the voltage peak and is weakest and close to zero near the voltage zero point.

[0109] 2) The pulse signals are randomly distributed in time and randomly distributed in phase over multiple cycles.

[0110] Fig.19A plot showing how a fire precursor signal changes over time. Time on the plot increases from top to bottom. Each row of the plot represents a single power cycle with a rising half cycle indicated on the left and a falling half cycle indicated on the right. The scale 1902 indicates the amplitude of the HF signal detected at different positions in the power cycle phase. For a parallel arc, the HF amplitude increases at the peak of the power cycle and goes to zero at the voltage zero crossing.

[0111] For comparison, Fig. 20 A schematic diagram of a device operating on an electrical network generating HF electrical activity is shown. Fig.19 A typical feature of devices operating on an electrical network is that the HF electrical activity is in a repeatable position in phase over a longer period of time. This is indicated by the vertical line in the falling half cycle. The sensor algorithm described herein exploits the distinction between man-made devices that generate predictable and repeating signals and fire precursor signals that are more variable in time and amplitude.

[0112] The techniques described above may be implemented in digital and / or analog electronic circuits, or in computer hardware, firmware, software, or a combination thereof. The implementation may be as a computer program product, i.e., a computer program tangibly implemented in a machine-readable storage device for being executed by a data processing device (e.g., a programmable processor, a computer, and / or multiple computers) or controlling the operation of data processing equipment. The computer program may be written in any form of computer or programming language, including source code, compiled code, interpreted code, and / or machine code, and the computer program may be deployed in any form, including as a stand-alone program or as a subroutine, element, or other unit suitable for use in a computing environment. The computer program may be deployed to execute on one or more computers in one or more sites.

[0113] The method steps may be performed by one or more dedicated processors executing a computer program to perform the functions of the present technology by operating on input data and / or generating output data. The method steps may also be performed by, and the apparatus may be implemented as, dedicated logic circuits, such as, FPGA (field programmable gate array), FPAA (field programmable analog array), CPLD (complex programmable logic device), PSoC (programmable system on chip), ASIP (application specific instruction set processor), or ASIC (application specific integrated circuit). A subroutine may refer to a portion of a stored computer program and / or processor, and / or a special circuit that implements one or more functions.

[0114] As an example, a processor suitable for the execution of a computer program includes both a general-purpose microprocessor and a special-purpose microprocessor, and any one or more processors in any type of digital or analog computer. Generally speaking, the processor will receive instructions and data from a read-only memory or a random access memory or both. The basic element of a computer is a processor for executing instructions, and / or one or more memory devices for storing instructions and / or data. Memory devices such as caches can be used for temporary storage of data. Memory devices can also be used for long-term data storage. Generally speaking, a computer also includes or is operably coupled to receive data from one or more mass storage devices (e.g., magnetic, magneto-optical disks or optical disks) for storing data or transfer data to one or more mass storage devices, or both. The computer can also be operably coupled to a communication network to receive instructions and / or data from the network, and / or transmit instructions and / or data to the network. Computer-readable storage media suitable for embodying computer program instructions and data include all forms of volatile and nonvolatile memory, including, by way of example, semiconductor memory devices (e.g., DRAM, SRAM, EPROM, EEPROM, and flash memory devices), magnetic disks (e.g., internal hard disks or removable disks), magneto-optical disks, and optical disks (e.g., CD, DVD, HD-DVD, and Blu-ray Disc). The processor and memory may be supplemented by and / or integrated in special purpose logic circuitry.

[0115] To provide interaction with a user, the techniques described above may be implemented on a computer in communication with a display device, such as a CRT (cathode ray tube) monitor, a plasma monitor, or an LCD (liquid crystal display) monitor, for displaying information to a user, and a keyboard and a pointing device, such as a mouse, a trackball, a touch pad, or a motion sensor, that can be used by a user to provide input to the computer (e.g., to interact with a user interface element). Other types of devices may also be used to provide interaction with a user; for example, the feedback provided to the user may be any form of sensory feedback, such as visual feedback, auditory feedback, or tactile feedback; and the input from the user may be received in any form, including sound input, voice input, and / or tactile input.

[0116] The techniques described above may be implemented in a distributed computing system including back-end components. For example, the back-end components may be data servers, middleware components, and / or application servers. The techniques described above may be implemented in a distributed computing system including front-end components. The front-end components may be, for example, a client computer with a graphical user interface, a web browser through which a user may interact with an example implementation, and / or other graphical user interfaces for sending devices. The techniques described above may be implemented in a distributed computing system including any combination of such back-end, middleware, or front-end components.

[0117] The components of the computing system can be interconnected by a transmission medium that can include digital data communication or analog data communication (e.g., communication network) of any form or medium. In any configuration, the transmission medium can include one or more packet-based networks and / or one or more circuit-based networks. Packet-based networks can include, for example, the Internet, carrier Internet Protocol (IP) networks (e.g., local area network (LAN), wide area network (WAN), campus area network (CAN), metropolitan area network (MAN), home area network (HAN)), dedicated IP networks, IP private branch exchange (IPBX), wireless networks (e.g., radio access network (RAN), Bluetooth, Wi-Fi, WiMAX, general packet radio service (GPRS) network, HiperLAN) and / or other packet-based networks. Circuit-based networks can include, for example, public switched telephone network (PSTN), old version private branch exchange (PBX), wireless networks (e.g., RAN, code division multiple access (CDMA) network, time division multiple access (TDMA) network, global system for mobile communications (GSM) network) and / or other circuit-based networks.

[0118] The transmission of information on the transmission medium may be based on one or more communication protocols. The communication protocols may include, for example, Ethernet protocol, Internet Protocol (IP), Voice over IP (VoIP), Point-to-Point (P2P) protocol, Hypertext Transfer Protocol (HTTP), Session Initiation Protocol (SIP), H.323, Media Gateway Control Protocol (MGCP), Signaling System #7 (SS7), Global System for Mobile Communications (GSM), Push to Talk (PTT) protocol, PTT over Cellular (POC) protocol, and / or other communication protocols.

[0119] The devices of the computing system may include, for example, computers, computers with browser devices, telephones, IP phones, mobile devices (e.g., cellular phones, personal digital assistant (PDA) devices, laptop computers, email devices), and / or other communication devices. Browser devices include, for example, computers with World Wide Web browsers (e.g., the Microsoft Windows XP available from Microsoft Corporation). Internet Available from Mozilla Corporation Firefox) is installed on a computer (e.g., a desktop computer, a laptop computer). Mobile computing devices include, for example, IP phones include, for example, those available from Cisco Systems, Inc. Unified IP Phone 7985G, and / or available from Cisco Systems, Inc. Unified Wireless Phone 7920.

[0120] The plural forms of including, comprising, and / or each are open ended and include the listed components and may include other components that are not listed. and / or are open ended and include one or more of the listed components and combinations of the listed components.

[0121] Those skilled in the art will appreciate that the present invention can be embodied in other specific forms without departing from the spirit and essential characteristics of the present invention.Therefore, the above embodiments will be regarded in all aspects as illustrative rather than restrictive of the present invention described herein.

Claims

1. A system for detecting electrical discharges preceding an electrical fire in electrical wiring, the system comprising: One or more sensor devices coupled to the circuit, each sensor device being configured to detecting a full voltage cycle waveform generated by electrical activity on the circuit, wherein the one or more sensor devices sample the full voltage cycle waveform at a frequency of 80 MHz when the full voltage cycle waveform is detected to have reached a threshold; identifying one or more transient signals within the full voltage cycle waveform; as well as generating one or more transient signatures based on the identified transient signal; a server computing device communicatively coupled to the one or more sensor devices, the server computing device being configured to: receiving the one or more transient signatures from each sensor device; analyzing the transient signature to identify one or more discharge indicators; as well as One or more alarm signals are generated when one or more discharge indications are identified.

2. The system according to claim 1, characterized in that The one or more sensor devices sample the full voltage cycle waveform at a frequency in a range between 10 MHz and 100 MHz.

3. The system according to claim 2, characterized in that Identifying one or more transient signals within the full voltage cycle waveform comprises: a) dividing one or more first samples of the full voltage cycle waveform into a plurality of bins; b) determining an average value and a maximum value for each bin of the plurality of bins; c) determining the difference between the average value and the maximum value; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum of the differences between the mean and the maximum values ​​for each bin over all of the samples; and e) determining a derivative of each cumulative maximum value determined in step d) over the plurality of bins.

4. The system according to claim 3, characterized in that Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

5. The system according to claim 4, characterized in that Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

6. The system according to claim 5, characterized in that Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and An alarm signal is generated based on the count of identified discharge indications.

7. The system according to claim 2, characterized in that Identifying one or more transient signals within the full voltage cycle waveform comprises: a) determining a derivative of one or more first samples of a full voltage signal waveform over a full voltage cycle; b) grouping the one or more first samples of the full voltage cycle waveform into a plurality of bins; c) determining a maximum value for each bin in the plurality of bins; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum value for each bin over all samples; and e) determining a derivative of each accumulated maximum value over the plurality of bins.

8. The system according to claim 7, characterized in that Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

9. The system according to claim 8, characterized in that Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

10. The system according to claim 9, characterized in that Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and An alarm signal is generated based on the count of identified discharge indications.

11. The system according to claim 1, characterized in that Identifying one or more transient signals within the full voltage cycle waveform comprises: identifying one or more samples of the full voltage cycle waveform that exceed the threshold; and The identified one or more samples are stored.

12. The system according to claim 11, characterized in that For each identified sample, generating one or more transient features based on the identified transient signal comprises: determining a count of peaks in the identified samples; determining a rise time of the peak in the identified sample; determining a pulse width of the identified sample; and Determine the integral of the identified samples.

13. The system of claim 12, wherein: Analyzing the transient signature to identify one or more discharge indications includes classifying the identified sample as a discharge indication when the count of peaks in the identified sample is above a peak count threshold and when the rise time of the peaks in the identified sample is above a rise time threshold.

14. The system of claim 13, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and An alarm signal is generated based on the count of identified discharge indications.

15. The system of claim 1, wherein: The one or more sensor devices identify one or more transient signals within the full voltage cycle waveform and generate one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device.

16. The system of claim 15, wherein: The server computing device generates an updated transient detection curve based on transient signatures received from one or more of the sensor devices and sends the updated transient detection curve to each of the one or more sensor devices.

17. The system of claim 16, wherein: The one or more sensor devices apply the updated transient detection profile to identify subsequent transient signals within the full voltage cycle waveform and generate the one or more transient signatures.

18. The system of claim 1, wherein: The server computing device sends the one or more alarm signals to a remote computing device.

19. The system of claim 1, wherein: The server computing device sends the one or more alarm signals to at least one sensor device of the one or more sensor devices.

20. The system of claim 19, wherein: The at least one sensor device of the one or more sensor devices activates a visual indicator upon receiving at least one alarm signal of the one or more alarm signals.

21. The system of claim 20, wherein: The visual indicator is a light emitting diode (LED) component of the at least one sensor device of the one or more sensor devices.

22. The system of claim 1, wherein: The server computing device analyzes the transient signatures using one or more machine learning algorithms to identify one or more discharge indicators.

23. A system for detecting electrical discharges preceding an electrical fire in electrical wiring, the system comprising: One or more sensor devices coupled to the circuit, each sensor device being configured to detecting one or more signal waveforms generated by electrical activity on the circuit; identifying one or more transient signals within the one or more signal waveforms; as well as generating one or more transient signatures based on the identified transient signal; a server computing device communicatively coupled to the one or more sensor devices, the server computing device being configured to: receiving the one or more transient signatures from each sensor device; analyzing the transient signature to identify one or more discharge indicators; as well as generating one or more alarm signals when one or more discharge indications are identified, wherein the one or more sensor devices identify one or more transient signals within the one or more signal waveforms and generate one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device, And wherein the server computing device generates an updated transient detection curve based on transient signatures received from one or more of the sensor devices, and sends the updated transient detection curve to each of the one or more sensor devices.

24. The system of claim 23, wherein: The one or more sensor devices apply the updated transient detection profile to identify subsequent transient signals within the one or more signal waveforms and generate the one or more transient signatures.

25. A computerized method of detecting electrical discharges preceding an electrical fire in electrical wiring, the method comprising: detecting, by each of one or more sensor devices coupled to the circuit, a full voltage cycle waveform generated by electrical activity on the circuit, the one or more sensor devices sampling the full voltage cycle waveform at a frequency of 80 MHz when detecting that the full voltage cycle waveform has reached a threshold; identifying, by each of the one or more sensor devices, one or more transient signals within the full voltage cycle waveform; generating, by each of the one or more sensor devices, one or more transient signatures based on the identified transient signal; receiving, by a server computing device communicatively coupled to the one or more sensor devices, the one or more transient signatures from each sensor device; analyzing, by the server computing device, the transient signature to identify one or more discharge indicators; as well as One or more alarm signals are generated by the server computing device when one or more discharge indications are identified.

26. The method of claim 25, wherein: The one or more sensor devices sample the full voltage cycle waveform at a frequency in a range between 10 MHz and 100 MHz.

27. The method of claim 26, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: a) dividing one or more first samples of the full voltage cycle waveform into a plurality of bins; b) determining an average value and a maximum value for each bin of the plurality of bins; c) determining the difference between the average value and the maximum value; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum of the differences between the mean and the maximum values ​​for each bin over all of the samples; and e) determining a derivative of each cumulative maximum value determined in step d) over the plurality of bins.

28. The method of claim 27, wherein: Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

29. The method of claim 28, wherein: Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

30. The method of claim 29, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and An alarm signal is generated based on the count of identified discharge indications.

31. The method of claim 26, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: a) determining a derivative of one or more first samples of a full voltage signal waveform over a full voltage cycle; b) grouping the one or more first samples of the full voltage cycle waveform into a plurality of bins; c) determining a maximum value for each bin in the plurality of bins; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum value for each bin over all samples; and e) determining a derivative of each accumulated maximum value over the plurality of bins.

32. The method of claim 31, wherein: Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

33. The method of claim 32, wherein: Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

34. The method of claim 33, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and An alarm signal is generated based on the count of identified discharge indications.

35. The method of claim 25, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: identifying one or more samples of the full voltage cycle waveform that exceed the threshold; and The identified one or more samples are stored.

36. The method of claim 35, wherein: For each identified sample, generating one or more transient features based on the identified transient signal comprises: determining a count of peaks in the identified samples; determining a rise time of the peak in the identified sample; determining a pulse width of the identified sample; and Determine the integral of the identified samples.

37. The method of claim 36, wherein: Analyzing the transient signature to identify one or more discharge indications includes classifying the identified sample as a discharge indication when the count of peaks in the identified sample is above a peak count threshold and when the rise time of the peaks in the identified sample is above a rise time threshold.

38. The method of claim 37, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and An alarm signal is generated based on the count of identified discharge indications.

39. The method of claim 25, wherein: The one or more sensor devices identify one or more transient signals within the full voltage cycle waveform and generate one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device.

40. The method of claim 39, wherein: The server computing device generates an updated transient detection curve based on transient signatures received from one or more of the sensor devices and sends the updated transient detection curve to each of the one or more sensor devices.

41. The method of claim 40, wherein: The one or more sensor devices apply the updated transient detection profile to identify subsequent transient signals within the full voltage cycle waveform and generate the one or more transient signatures.

42. The method of claim 25, wherein: The server computing device sends the one or more alarm signals to a remote computing device.

43. The method of claim 25, wherein: The server computing device sends the one or more alarm signals to at least one sensor device of the one or more sensor devices.

44. The method of claim 43, wherein: The at least one sensor device of the one or more sensor devices activates a visual indicator upon receiving at least one alarm signal of the one or more alarm signals.

45. The method of claim 44, wherein: The visual indicator is a light emitting diode (LED) component of the at least one sensor device of the one or more sensor devices.

46. ​​The method of claim 25, wherein: The server computing device analyzes the transient signatures using one or more machine learning algorithms to identify one or more discharge indicators.

47. A computerized method of detecting electrical discharges preceding an electrical fire in electrical wiring, the method comprising: detecting, by each of one or more sensor devices coupled to the circuit, one or more signal waveforms generated by electrical activity on the circuit; identifying, by each of the one or more sensor devices, one or more transient signals within the one or more signal waveforms; generating, by each of the one or more sensor devices, one or more transient signatures based on the identified transient signal; receiving, by a server computing device communicatively coupled to the one or more sensor devices, the one or more transient signatures from each sensor device; analyzing, by the server computing device, the transient signature to identify one or more discharge indicators; as well as generating, by the server computing device, one or more alarm signals when one or more discharge indications are identified, wherein the one or more sensor devices identify one or more transient signals within the one or more signal waveforms and generate one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device, And wherein the server computing device generates an updated transient detection curve based on transient signatures received from one or more of the sensor devices, and sends the updated transient detection curve to each of the one or more sensor devices.

48. The method of claim 47, wherein: The one or more sensor devices apply the updated transient detection profile to identify subsequent transient signals within the one or more signal waveforms and generate the one or more transient signatures.

49. A system for detecting electrical discharges preceding an electrical fire in electrical wiring, the system comprising: one or more sensor devices coupled to the circuit, each sensor device configured to detect a full voltage cycle waveform generated by electrical activity on the circuit, the one or more sensor devices sampling the full voltage cycle waveform at a frequency of 80 MHz upon detecting that the full voltage cycle waveform has reached a threshold; as well as a server computing device communicatively coupled to the one or more sensor devices, the server computing device being configured to: receiving the full voltage cycle waveform from each sensor device; analyzing the full voltage cycle waveform to identify one or more discharge indicators; as well as One or more alarm signals are generated when one or more discharge indications are identified.

50. The system of claim 49, wherein: Analyzing the full voltage cycle waveform to identify one or more discharge indications includes: identifying one or more transient signals within the full voltage cycle waveform; generating one or more transient signatures based on the identified transient signal; and The one or more transient signatures are analyzed to identify the one or more discharge indications.

51. The system of claim 49, wherein: The one or more sensor devices sample the full voltage cycle waveform at a frequency in a range between 10 MHz and 100 MHz.

52. The system of claim 51, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: a) dividing one or more first samples of the full voltage cycle waveform into a plurality of bins; b) determining an average value and a maximum value for each bin of the plurality of bins; c) determining the difference between the average value and the maximum value; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum of the differences between the mean and the maximum values ​​for each bin over all of the samples; and e) determining a derivative of each cumulative maximum value determined in step d) over the plurality of bins.

53. The system of claim 52, wherein: Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

54. The system of claim 53, wherein: Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

55. The system of claim 54, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and One or more alarm signals are generated based on the count of identified discharge indications.

56. The system of claim 51, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: a) determining a derivative of one or more first samples of a full voltage signal waveform over a full voltage cycle; b) grouping the one or more first samples of the full voltage cycle waveform into a plurality of bins; c) determining a maximum value for each bin in the plurality of bins; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum value for each bin over all samples; and e) determining a derivative of each accumulated maximum value over the plurality of bins.

57. The system of claim 56, wherein: Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

58. The system of claim 57, wherein: Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

59. The system of claim 58, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and One or more alarm signals are generated based on the count of identified discharge indications.

60. The system of claim 49, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: identifying one or more samples of the full voltage cycle waveform that exceed a threshold; and The identified one or more samples are stored.

61. The system of claim 60, wherein: For each identified sample, generating one or more transient features based on the identified transient signal comprises: determining a count of peaks in the identified samples; determining a rise time of the peak in the identified sample; determining a pulse width of the identified sample; and Determine the integral of the identified samples.

62. The system of claim 61, wherein: Analyzing the transient signature to identify one or more discharge indications includes classifying the identified sample as a discharge indication when the count of peaks in the identified sample is above a peak count threshold and when the rise time of the peaks in the identified sample is above a rise time threshold.

63. The system of claim 62, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and One or more alarm signals are generated based on the count of identified discharge indications.

64. The system of claim 50, wherein: The one or more sensor devices identify one or more transient signals within the full voltage cycle waveform and generate one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device.

65. The system of claim 64, wherein: The server computing device generates an updated transient detection curve based on transient signatures received from one or more of the sensor devices and sends the updated transient detection curve to each of the one or more sensor devices.

66. The system of claim 65, wherein: The one or more sensor devices apply the updated transient detection profile to identify subsequent transient signals within the full voltage cycle waveform and generate the one or more transient signatures.

67. The system of claim 49, wherein: The server computing device sends the one or more alarm signals to a remote computing device.

68. The system of claim 49, wherein: The server computing device sends the one or more alarm signals to at least one sensor device of the one or more sensor devices.

69. The system of claim 68, wherein: The at least one sensor device of the one or more sensor devices activates a visual indicator upon receiving at least one alarm signal of the one or more alarm signals.

70. The system of claim 69, wherein: The visual indicator is a light emitting diode (LED) component of the at least one sensor device of the one or more sensor devices.

71. The system of claim 49, wherein: The server computing device analyzes the full voltage cycle waveform using one or more machine learning algorithms to identify one or more discharge indicators.

72. A system for detecting electrical discharges preceding an electrical fire in electrical wiring, the system comprising: one or more sensor devices coupled to the circuit, each sensor device configured to detect one or more signal waveforms generated by electrical activity on the circuit; as well as a server computing device communicatively coupled to the one or more sensor devices, the server computing device being configured to: receiving the one or more signal waveforms from each sensor device; analyzing the one or more signal waveforms to identify one or more discharge indications; as well as generating one or more alarm signals when one or more discharge indications are identified, wherein the one or more sensor devices identify one or more transient signals within the one or more signal waveforms and generate one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device, And wherein the server computing device generates an updated transient detection curve based on transient signatures received from one or more of the sensor devices, and sends the updated transient detection curve to each of the one or more sensor devices.

73. The system of claim 72, wherein: The one or more sensor devices apply the updated transient detection profile to identify subsequent transient signals within the one or more signal waveforms and generate the one or more transient signatures.

74. A sensor device for detecting an electrical discharge preceding an electrical fire in electrical wiring, the sensor device being coupled to an electrical circuit, the sensor device comprising: a module, the module sensing electrical activity on the circuit and detecting a full voltage cycle waveform generated by the electrical activity on the circuit, the sensor device sampling the full voltage cycle waveform at a frequency of 80 MHz when it is detected that the full voltage cycle waveform has reached a threshold; as well as A processor, the processor: analyzing the full voltage cycle waveform to identify one or more discharge indicators; as well as One or more alarm signals are generated when one or more discharge indications are identified.

75. The sensor device of claim 74, wherein: Analyzing the full voltage cycle waveform to identify one or more discharge indications includes: identifying one or more transient signals within the full voltage cycle waveform; generating one or more transient signatures based on the identified transient signal; and The one or more transient signatures are analyzed to identify the one or more discharge indications.

76. The sensor device of claim 74, wherein: The sensor device samples the full voltage cycle waveform at a frequency in a range between 10 MHz and 100 MHz.

77. The sensor device of claim 75, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: a) dividing one or more first samples of the full voltage cycle waveform into a plurality of bins; b) determining an average value and a maximum value for each bin of the plurality of bins; c) determining the difference between the average value and the maximum value; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum of the differences between the mean and the maximum values ​​for each bin over all of the samples; and e) determining a derivative of each cumulative maximum value determined in step d) over the plurality of bins.

78. The sensor device of claim 77, wherein: Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

79. The sensor device of claim 78, wherein: Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

80. The sensor device of claim 79, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and One or more alarm signals are generated based on the count of identified discharge indications.

81. The sensor device of claim 76, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: a) determining a derivative of one or more first samples of a full voltage signal waveform over a full voltage cycle; b) grouping the one or more first samples of the full voltage cycle waveform into a plurality of bins; c) determining a maximum value for each bin in the plurality of bins; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum value for each bin over all samples; and e) determining a derivative of each accumulated maximum value over the plurality of bins.

82. The sensor device of claim 81, wherein: Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

83. The sensor device of claim 82, wherein: Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

84. The sensor device of claim 83, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and One or more alarm signals are generated based on the count of identified discharge indications.

85. The sensor device of claim 74, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: identifying one or more samples of the full voltage cycle waveform that exceed a threshold; and The identified one or more samples are stored.

86. The sensor device of claim 85, wherein: For each identified sample, generating one or more transient features based on the identified transient signal comprises: determining a count of peaks in the identified samples; determining a rise time of the peak in the identified sample; determining a pulse width of the identified sample; and Determine the integral of the identified samples.

87. The sensor device of claim 86, wherein: Analyzing the transient signature to identify one or more discharge indications includes classifying the identified sample as a discharge indication when the count of peaks in the identified sample is above a peak count threshold and when the rise time of the peaks in the identified sample is above a rise time threshold.

88. The sensor device of claim 87, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and One or more alarm signals are generated based on the count of identified discharge indications.

89. The sensor device of claim 75, wherein: The sensor device identifies one or more transient signals within the full voltage cycle waveform and generates one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device.

90. The sensor device of claim 89, wherein: The sensor device generates an updated transient detection curve based on the transient signature generated by the sensor device.

91. The sensor device of claim 90, wherein: The sensor device applies the updated transient detection profile to identify subsequent transient signals within the full voltage cycle waveform and generates the one or more transient signatures.

92. The sensor device of claim 74, wherein: The sensor device transmits the one or more alarm signals to a remote computing device.

93. The sensor device of claim 74, wherein: The sensor device activates a visual indicator upon generating at least one of the one or more alarm signals.

94. The sensor device of claim 93, wherein: The visual indicator is a light emitting diode (LED) component of the sensor device.

95. A sensor device for detecting an electrical discharge preceding an electrical fire in electrical wiring, the sensor device being coupled to an electrical circuit, the sensor device comprising: a module that senses electrical activity on the circuit and detects one or more signal waveforms of the electrical activity; as well as A processor, the processor: analyzing the one or more signal waveforms to identify one or more discharge indications; as well as generating one or more alarm signals when one or more discharge indications are identified, wherein the sensor device identifies one or more transient signals within the one or more signal waveforms and generates one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device, And wherein the sensor device generates an updated transient detection curve based on the transient signature generated by the sensor device.

96. The sensor device of claim 95, wherein: The sensor device applies the updated transient detection profile to identify subsequent transient signals within the one or more signal waveforms and generates the one or more transient signatures.

97. A computerized method of detecting electrical discharges preceding an electrical fire in electrical wiring, the method comprising: detecting, by one or more sensor devices each coupled to the circuit, a full voltage cycle waveform generated by electrical activity on the circuit, wherein the one or more sensor devices sample the full voltage cycle waveform at a frequency of 80 MHz when detecting that the full voltage cycle waveform has reached a threshold value; receiving, by a server computing device communicatively coupled to the one or more sensor devices, the full voltage cycle waveform from each sensor device; analyzing, by the server computing device, the full voltage cycle waveform to identify one or more discharge indications; as well as One or more alarm signals are generated by the server computing device when one or more discharge indications are identified.

98. The method of claim 97, wherein: Analyzing the full voltage cycle waveform to identify one or more discharge indications includes: identifying one or more transient signals within the full voltage cycle waveform; generating one or more transient signatures based on the identified transient signal; and The one or more transient signatures are analyzed to identify the one or more discharge indications.

99. The method of claim 97, wherein: The one or more sensor devices sample the full voltage cycle waveform at a frequency in a range between 10 MHz and 100 MHz.

100. The method of claim 99, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: a) dividing one or more first samples of the full voltage cycle waveform into a plurality of bins; b) determining an average value and a maximum value for each bin of the plurality of bins; c) determining the difference between the average value and the maximum value; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum of the differences between the mean and the maximum values ​​for each bin over all of the samples; and e) determining a derivative of each cumulative maximum value determined in step d) over the plurality of bins.

101. The method of claim 100, wherein: Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

102. The method of claim 101, wherein: Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

103. The method of claim 102, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and An alarm signal is generated based on the count of identified discharge indications.

104. The method of claim 103, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: a) determining a derivative of one or more first samples of a full voltage signal waveform over a full voltage cycle; b) grouping the one or more first samples of the full voltage cycle waveform into a plurality of bins; c) determining a maximum value for each bin in the plurality of bins; d) repeating steps a) to c) for each of a plurality of other samples of the full voltage cycle waveform to determine a cumulative maximum value for each bin over all samples; and e) determining a derivative of each accumulated maximum value over the plurality of bins.

105. The method of claim 104, wherein: Generating one or more transient signatures based on the identified transient signal includes: determining an average transient amplitude over a voltage cycle of the full voltage cycle waveform; and An average transient amplitude is determined for a plurality of phase segments within the voltage cycle.

106. The method of claim 105, wherein: Analyzing the transient signature to identify one or more discharge indicators includes: determining a ratio of an average peak transient in one or more of the phase segments near a maximum voltage to an average peak transient near a zero crossing of the voltage cycle; and When the ratio is above a predetermined threshold, an average peak transient in one or more of the phase segments near a maximum voltage is identified as an indication of discharge.

107. The method of claim 106, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and An alarm signal is generated based on the count of identified discharge indications.

108. The method of claim 97, wherein: Identifying one or more transient signals within the full voltage cycle waveform comprises: identifying one or more samples of the full voltage cycle waveform that exceed a threshold; and The identified one or more samples are stored.

109. The method of claim 108, wherein: For each identified sample, generating one or more transient features based on the identified transient signal comprises: determining a count of peaks in the identified samples; determining a rise time of the peak in the identified sample; determining a pulse width of the identified sample; and Determine the integral of the identified samples.

110. The method of claim 109, wherein: Analyzing the transient signature to identify one or more discharge indications includes classifying the identified sample as a discharge indication when the count of peaks in the identified sample is above a peak count threshold and when the rise time of the peaks in the identified sample is above a rise time threshold.

111. The method of claim 110, wherein: Generating one or more alarm signals when one or more discharge indications are identified includes: determining a count of identified discharge indications occurring within a predetermined amount of time; and An alarm signal is generated based on the count of identified discharge indications.

112. The method of claim 98, wherein: The one or more sensor devices identify one or more transient signals within the full voltage cycle waveform and generate one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device.

113. The method of claim 112, wherein: The server computing device generates an updated transient detection curve based on transient signatures received from one or more of the sensor devices and sends the updated transient detection curve to each of the one or more sensor devices.

114. The method of claim 113, wherein: The one or more sensor devices apply the updated transient detection profile to identify subsequent transient signals within the full voltage cycle waveform and generate the one or more transient signatures.

115. The method of claim 97, wherein: The server computing device sends the one or more alarm signals to a remote computing device.

116. The method of claim 97, wherein: The server computing device sends the one or more alarm signals to at least one sensor device of the one or more sensor devices.

117. The method of claim 116, wherein: The at least one sensor device of the one or more sensor devices activates a visual indicator upon receiving at least one alarm signal of the one or more alarm signals.

118. The method of claim 117, wherein: The visual indicator is a light emitting diode (LED) component of the at least one sensor device of the one or more sensor devices.

119. The method of claim 97, wherein: The server computing device analyzes the full voltage cycle waveform using one or more machine learning algorithms to identify one or more discharge indicators.

120. A computerized method of detecting electrical discharges preceding an electrical fire in electrical wiring, the method comprising: detecting, by one or more sensor devices each coupled to the circuit, one or more signal waveforms generated by electrical activity on the circuit; receiving, by a server computing device communicatively coupled to the one or more sensor devices, the one or more signal waveforms from each sensor device; analyzing, by the server computing device, the one or more signal waveforms to identify one or more discharge indications; as well as generating, by the server computing device, one or more alarm signals when one or more discharge indications are identified, wherein the one or more sensor devices identify one or more transient signals within the one or more signal waveforms and generate one or more transient signatures based on the identified transient signals using transient detection curves stored in a memory module of the sensor device, And wherein the server computing device generates an updated transient detection curve based on transient signatures received from one or more of the sensor devices, and sends the updated transient detection curve to each of the one or more sensor devices.

121. The method of claim 120, wherein: The one or more sensor devices apply the updated transient detection profile to identify subsequent transient signals within the one or more signal waveforms and generate the one or more transient signatures.

Images