Phasor-based fault location estimation in electrical power delivery system

Abstract

A WMU data analysis system and associated method are for monitoring an electrical power delivery system. The disclosed techniques entail processing synchro-waveform data to classify an event type and estimate, based on an impedance component and the event type, an event location so as to reduce a number of candidate inspection sites in the electrical power delivery system for inspection by a maintenance team.

Description

Docket No.503461-70021 PHASOR-BASED FAULT LOCATION ESTIMATION IN ELECTRICAL POWER DELIVERY SYSTEM RELATED APPLICATION

[0001] This application claims priority benefit of U.S. Provisional Patent Application No. 63 / 651,599 filed May 24, 2024, which is hereby incorporated by reference in its entirety. TECHNICAL FIELD

[0002] This disclosure relates generally to power system diagnostics and, more particularly, to fault localization in electrical networks. BACKGROUND INFORMATION

[0003] In the IEEE Power & Energy Magazine, September / October 2023 edition (Vol.21, No.5), an article starting on page 68 by Hamed Mohsenian-Rad, Ph.D., and Wilsun Xu, Ph.D., titled “Synchro-Waveforms,” describes a power system measurement technology approach. This technology, referred to as a waveform measurement unit (WMU), obtains time-synchronized waveform measurements—i.e., synchro-waveforms—from different locations within a power system. WMUs can capture inconspicuous disturbances that are often overlooked by other types of time-synchronized sensors, such as phasor measurement units (PMUs). WMUs also monitor system dynamics at much higher frequencies, as well as much lower frequencies, than the fundamental components of voltage and current that are commonly monitored by PMUs.

[0004] Another paper by V. Barrera Núñez et al., titled “Feature Analysis and Classification Methodology for Overhead Distribution Fault Events,” presented in the IEEE / PES Transmission and Distribution Committee on July 25–29, 2010, describes the use of waveform data from power quality monitors (PQM) installed in substations. The paper presents an analysis of unique features and a classification method for identifying the root cause of overhead distribution fault events, which are multi-cycle events. In particular, it focuses on faults caused by animal contacts, tree contacts, and lightning-induced disturbances.

[0005] Additionally, there are a number of patents describing the calculation of fault locations. FIG.1, for example, shows an example set of electrical schematics and waveform diagrams arranged in a sequence for representing a SLG (single line-to-ground) incipient fault location estimation. The estimation technique is described in U.S. Patent No. 8,941,387 of Kim. The ’387 patent describes a time-domain analysis for determining netDocket No.503461-70021 fault electrical parameters, including inductive reactance. This prior art conducts these calculations in the time domain. However, calculations in the time domain sometimes necessitate averaging over a sample period, which can reduce precision, and the coupled mathematical transforms may limit the technique’s accuracy to the measurement capabilities of the voltage and current waveforms. SUMMARY OF THE DISCLOSURE

[0006] Embodiments of the present disclosure relate to techniques for fault classification and location estimation in an electrical power delivery system using phasor domain analysis. In particular, the disclosed systems and methods process synchro-waveform data collected from waveform monitoring units (WMUs) to identify transitory electrical anomalies (such as faults or power quality disturbances) and estimate their location with increased precision and reduced latency compared to time-domain approaches.

[0007] In one aspect, the disclosure provides a method performed by a WMU data analysis system for classifying an event type and estimating, based on an impedance component and the event type, a fault location. The system receives synchro-waveform data from WMUs and computes voltage and current phasor domain values, phase angles, and impedance components for at least each half-cycle of the waveform. These phasor domain values, in conjunction with time-domain waveform characteristics, are used to determine the type of fault event and derive an impedance-based estimate of its location. Fault type and location estimation may be further refined by comparing calculated impedance values (or derived reactance values) to electrical network models that map impedance or reactance to physical line segments. Candidate fault locations are identified by correlating the calculated electrical parameters with known model values, as well as by comparing event type characteristics with the types of electrical components deployed at those locations.

[0008] The disclosed techniques improve fault location estimation accuracy by eliminating discontinuities that arise in time-domain calculations and by enabling sub-cycle or half- cycle analysis of synchro-waveform data. Phasor domain calculations provide a richer set of waveform metrics, which may be analyzed on a per-half-cycle or point-by-point basis to identify signatures of specific device types. These characteristics can be used not only to classify fault events (e.g., tee connector, downed conductor, tree contact, underground cable fault) but also to rule out incompatible locations, thereby narrowing the number of candidate inspection sites.

[0009] Some embodiments include visualization features, wherein estimated fault locations are annotated on a geographic information system (GIS) map based on electrical model data.Docket No.503461-70021 GIS-based maps may display the electrical power system topology and highlight candidate fault regions, using annotations that reflect voltage and current waveform data, phasor domain impedance component values, and predicted fault types. The electrical network model may include reactance values and component metadata from simulation tools such as Synergi Electric or CYME Power Engineering Software. In some cases, the system dynamically updates the GIS-based map and user interface in real time as new waveform data is processed.

[0010] Other embodiments include a computer-implemented user interface that displays waveform plots, impedance component and phase angle metrics, and geospatial fault location overlays. The interface may present predicted fault classifications generated by a machine learning model trained on time-domain and phasor-domain waveform data. The user interface may enable operators to select specific fault events for further analysis and automatically update displayed waveforms, location overlays, and classifications.

[0011] By combining phasor domain impedance analysis with machine learning classification and geospatial visualization, the disclosed techniques allow utilities to proactively identify faults, minimize outage durations, and transition from reactive to predictive maintenance strategies.

[0012] Additional aspects and advantages will be apparent from the following detailed description of embodiments, which proceeds with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0013] To easily identify the discussion of any particular element or act, the most significant digit or digits in a reference number refer to the figure number in which that element is first introduced.

[0014] FIG.1 is a set of electrical schematic and waveform diagrams arranged in a sequence for representing a SLG (single line-to-ground) incipient fault location estimation technique in accordance with the prior art.

[0015] FIG.2 is a block diagram of a WMU network architecture in accordance with one embodiment.

[0016] FIG.3 is a block diagram showing an electrical power delivery system and WMU synchro-waveform event analysis in accordance with one embodiment.

[0017] FIG.4 is an electrical schematic diagram of a three phase representation of an electrical power delivery system in accordance with one embodiment.Docket No.503461-70021

[0018] FIG.5 is an electrical schematic diagram of a single phase representation of the electrical power delivery system in FIG. 4, with WMUs showing an event in accordance with one embodiment.

[0019] FIG. 6 is a simplified electrical schematic diagram in accordance with one embodiment.

[0020] FIG. 7 is a time-domain plot of voltage and current in accordance with one embodiment.

[0021] FIG. 8 is a time-domain plot of impedance in accordance with one embodiment.

[0022] FIG. 9A and FIG.9B are a set of three-phase time-domain plots of WMU synchro- waveform data exhibiting a tee connection failure in accordance with one embodiment.

[0023] FIG.10A and FIG.10B are a set of three-phase time-domain plots of WMU synchro-waveform data exhibiting a downed conductor failure in accordance with one embodiment.

[0024] FIG.11A and FIG.11B are a set of three-phase time-domain plots of WMU synchro-waveform data exhibiting a tree / limb preventable failure in accordance with one embodiment.

[0025] FIG.12A and FIG.12B are a set of three-phase time-domain plots of WMU synchro-waveform data exhibiting a vines / grasses failure in accordance with one embodiment.

[0026] FIG.13A and FIG.13B are a set of three-phase time-domain plots of WMU synchro-waveform data exhibiting a cable failure in accordance with one embodiment.

[0027] FIG.14 is a block diagram showing an overview of how waveform data is analyzed to determine an event type in accordance with one embodiment.

[0028] FIG.15 is another block diagram showing data processing of waveform data in accordance with one embodiment.

[0029] FIG.16 is a block diagram showing a second step for analyzing tee connection faults in accordance with one embodiment.

[0030] FIG.17 is a block diagram showing how event type time and phasor domain characteristics data is input to machine learning algorithms for event type classification in accordance with one embodiment.

[0031] FIG.18 is a plot of impedance as a function of distance for a circuit modeled in a load-flow analysis tool in accordance with one embodiment.Docket No.503461-70021

[0032] FIG. 19 is a pictorial view of geographic information system (GIS) map data with road distances highlighted based on the impedance data of FIG.18 to show a FIS map location of a fault.

[0033] FIG.20 is a screenshot of an event overview timeline user interface in accordance with one embodiment.

[0034] FIG.21 is a screenshot of an original signal user interface in accordance with one embodiment.

[0035] FIG.22 is a screenshot of a signal processing user interface in accordance with one embodiment.

[0036] FIG.23 is a flow diagram of a process, performed by a WMU data analysis system, of processing synchro-waveform data to classify an event type and estimate, based on impedance and the event type, an event location so as to reduce a number of candidate inspection sites in the electrical power delivery system for inspection by a maintenance team.

[0037] FIG.24 is a block diagram showing components of a WMU or server in accordance with one embodiment. DETAILED DESCRIPTION OF EMBODIMENTS

[0038] Modern electrical power delivery systems are complex and increasingly reliant on real-time data analytics to maintain operational reliability. As utility networks face mounting pressure to reduce downtime, improve service continuity, and proactively manage aging infrastructure, early and accurate detection and localization of abnormal grid behavior (i.e., electrical disturbance events) is helpful. This disclosure presents techniques that enhance fault detection and location by analyzing high-resolution waveform data from distributed monitoring devices. These techniques are designed to identify various grid disturbances—including faults (which may be temporary or sustained) and power quality (PQ) events that signal emerging issues in the network. The following overview introduces some terminology and concepts that are the subject of the analytic and classification techniques described in subsequent embodiments.

[0039] A temporary fault is a disturbance that triggers a protective mechanism, such as a recloser operation, but subsequently clears on its own. Because the underlying cause, such as a momentary line contact or minor insulation breakdown, is transient, the network can self-recover and continue operating. In contrast, a sustained fault occurs when the protective mechanism remains open until the fault is physically removed or repaired, leading to anDocket No.503461-70021 outage or service disruption. Identifying and addressing temporary faults before they develop into sustained faults is helpful for maintaining system reliability.

[0040] Some temporary faults occur as precursors to more serious faults and outages. These incipient faults or precursor events do not immediately cause service interruptions but can degrade system conditions over time. Examples of incipient faults include early-stage insulation breakdown in cables or intermittent contact caused by vegetation. Detecting and mitigating these precursors is an important part of predictive maintenance, helping prevent minor issues from escalating into major failures.

[0041] Beyond faults, other PQ events represent grid disturbances that can affect power system stability and performance. These include voltage sags / swells, transients, harmonics, and frequency variations. While these PQ events may not always cause outages, they can be indicative of underlying problems in the network. In many cases, faults exhibit characteristic power quality signatures, such as current swells and voltage sags, making PQ analysis another tool for fault detection and classification. For example, a voltage sag may result from a short but severe fault that quickly clears, yet it still signals a vulnerable segment of the network. Similarly, a transient voltage spike may indicate a momentary fault that did not escalate into a sustained outage but still represents an operational concern.

[0042] Locating and categorizing these events, whether they are “temporary” anomalies that self-clear or “sustained” faults leading to an outage, allows operators to coordinate more effectively with the broader predictive maintenance strategy and can significantly enhance system reliability.

[0043] FIG.2 shows a WMU network architecture 200 for monitoring and analyzing faults within a power system, such as an electrical power delivery system 300 (FIG.3). According to some embodiments, WMU network architecture 200 integrates subcomponents provided by a power transmission and distribution (T&D) company with infrastructure and software available from the applicant for this disclosure, Toumetis, Inc. This integrated system allows for comprehensive monitoring and analysis of the power grid, aiming to quickly and efficiently detect and respond to transitory electrical anomalies, which are also referred to as faults or power quality (PQ) events.

[0044] The main components shown in the example of FIG.2 are a substation 202 and a WMU data analysis system 204, which are communicatively coupled via an internet connection 206. Internet connection 206 is available at substation 202 using cellular connection equipment 208 (e.g., LTE or 5G), fiber, or other communication devices. In some embodiments, WMU data analysis system 204 is a remotely located server (e.g., AWSDocket No.503461-70021 server) such that the analysis is performed via cloud computing. In other embodiments, WMU data analysis system 204 may be located at an edge, such as within substation 202.

[0045] In this example, substation 202 is the physical location where power system monitoring occurs. It contains equipment for controlling and monitoring the electricity flow in the power grid. Substation 202 also includes an internal LAN 210, which is communicatively coupled to cellular connection equipment 208 via wireless or ethernet connection 212 with a router 214.

[0046] LAN 210 interconnects all the monitoring and computing devices located in LAN 210. For instance, a substation server 216 includes local PC 218 having an optional storage interface 220 for optional removeable storage 222 and a wireless or wired router 214 connected via router interface 224 to local PC 218. Local PC 218 is configured with software processes 226 including a process 228 to get FTP data from devices and a process 230 push data to storage 232 in connection with cloud storage and processing system 234.

[0047] Cloud storage and processing system 234 is configured to read instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium), including storage 232, and perform any one or more of the methods discussed herein (or portions thereof), such as discussed for process 2300 (FIG.23), or any portions of these embodiments. Additional details are described later with reference to FIG. 23 and FIG.24.

[0048] Router 214 also connects to multiple WMU devices 236 via server connection 238. Multiple WMU devices 236 include monitoring device 240, which samples synchro- waveform data and provides it via an FTP interface 242 for process 228.

[0049] FIG.3 shows electrical power delivery system 300 and WMU synchro-waveform event analysis 302, representing the capability to use waveform data to detect or predict, classify, and locate faults within electrical power delivery system 300, which is useful for quick response and maintenance to ensure reliability and safety in the electrical grid. Example classifications of faults include a tee connector fault, load break elbow fault, cable (to ground or other line) fault, capacitor fault, transformer fault, or other device faults shown and described later with reference to FIG.9A–FIG.13B. Other types of events are also possible.

[0050] Electrical power delivery system 300 includes a power generation plant 304 (which could be a hydroelectric dam, wind generation, or various types of thermal power plants), a step-up transmission substation 306, high-voltage transmission lines 308, a step-downDocket No.503461-70021 transmission substation 310, a step-down distribution substation 312, distribution network poles 314, and residential or commercial users 316.

[0051] In electrical power delivery system 300, WMU synchro-waveform event analysis 302 is used for fault location estimation 318. Specifically, FIG.3 shows a time-domain waveform 320 exhibiting a transitory electrical anomaly 322. Transitory electrical anomaly 322 is then analyzed to estimate fault location estimation 318, which in this example is estimated at a distribution line 324 between step-down distribution substation 312 and distribution network poles 314 or feeder lines 326.

[0052] Time-domain waveform 320 shows three overlapping waveforms, representing differing views of the impacted phase of an electrical power system. Voltage spikes on all three waveforms indicate an example of a precise, sub-cycle moment when transitory electrical anomaly 322 appeared in electrical power delivery system 300, although the techniques in this disclosure may be used for multi- or sub-cycle faults (with the underlying analysis being performed per half cycle or less).

[0053] FIG.4 shows an example of electrical power delivery system 300 in a schematic form, where the “n” is neutral, “t” is transformer, and “f” is feeder. In this example, electrical power delivery system 300 represents a simplified version of electrical power delivery system 300 (FIG.3), including a y-connected three-phase power source 402. Each voltage phase (Va, Vb, Vc) has a corresponding power delivery line, which are modeled with a series of resistances and inductances. The amount of resistance and inductance increases as function of distance. Nodes (N) in electrical power delivery system 300 may represent substation, transmission, or distribution locations (e.g., points between step-up transmission substation 306 and high-voltage transmission lines 308, or the like).

[0054] FIG.5 shows in greater detail a single phase 500 of electrical power delivery system 300. In this view, each node connection is monitored by a WMU. Skilled persons will appreciate, however, that one or more WMUs may be employed to monitor different locations along the circuit of single phase 500. For instance, a first WMU 502 measures an N1 node between a step-down distribution substation 312 (FIG.3) and distribution network poles 314 (FIG.3). Similarly, a second WMU 504 measures an N2 node between distribution network poles 314 and feeder lines 326. A third WMU 506 measures a node N3 between feeder lines 326 and residential or commercial users 316, and so forth. Examples of WMUs include Satec PM174 & PM180, Dranetz PQ3K, PQ5K, Encore, and PMI Phalanx.

[0055] In the example of FIG.5, a fault 508 is detected near N2, after Lf1. For instance, if fault 508 were a SLG, then first WMU 502 at N1 might measure a significant voltage dropDocket No.503461-70021 with a large current increase. A second WMU 504 at N2, however, might measure a voltage drop but commensurate increase in current. These changes are analyzed in the time domain to characterize the type of fault. Also, as explained below with reference to FIG.6–FIG.8 and Table 2 (FIG.9A–FIG.13B), the changes are analyzed in the phasor domain. Then, as shown and described later with reference to FIG.14–FIG.19, accurate characterization of the event type reduces ambiguity in determining a fault location so as to reduce a number of candidate inspection sites in the electrical power delivery system for inspection by a maintenance team.

[0056] FIG.6 shows an AC circuit 600 for modeling a power transmission and distribution system. AC circuit 600 is annotated to show Ohm’s law equations for current (I), voltage (V), as calculations for voltages across a resistor (R) and an inductor (L). Capacitance (C), which is negligible in this example, is not shown. In one example, the parameters for AC circuit 600 are shown in the following Table 1. frequency (f) 60 Hz angular frequency (ω) 376.991118 rad / s one cycle period (T) 0.01666667 s samples / cycle 1024 peak line-to-line voltage (Vs) 16970.5627 volts peak phase voltage (Vp) 9797.95897 volts peak phase current (Ip) 872.7722 amps R 10.1 Ω L 0.013 H C 10 F inductive reactance (XL) 4.90088454 capacitive reactance (XC) 0.000265 θ 0.45176688 |Z| 11.2262491

[0057] FIG. 7 shows a time-domain plot 700 of a voltage cycle, Vp(t), and a current cycle, Ip(t), which are based on the parameters in the table above. At each zero-crossing 702, Ip(t) is zero.

[0058] FIG.8 shows a time-domain plot 800 of impedance, (Zt) 802, for AC circuit 600, which is also based on the parameters in the table above. As indicated in FIG. 8, Zt 802 has vertical asymptotes 804 in the time domain, which correspond to zero-crossings 702 (FIG. 7), since Zt 802 is equal to Vp(t) / Ip(t), which is undefined when Ip(t) is zero. Thus, solving for Zt 802 to obtain R and XL (in the time domain) introduces uncertainty and imprecision, which ultimately introduces error in locating a fault.

[0059] Instead of calculating R and XLin the time domain, the following analysis may be performed in the phasor domain. As used herein, the term impedance component refers to any value derived from phasor-domain analysis of voltage and current waveforms that isDocket No.503461-70021 suitable for use in estimating the location of an electrical disturbance or classifying an event type. An impedance component may include, for example, a complex impedance value (Z = R + jX), the magnitude of impedance (|Z|), or a derived reactance value (X), depending on the nature of the fault analysis or the structure of the electrical model being used. In some embodiments, a reactance value may be extracted directly by analyzing the phase difference between voltage and current waveforms, without computing the full complex impedance. This flexibility allows the system to match the form of impedance-related data found in utility network models, which often express line characteristics in terms of per-segment reactance values for use in location estimation.

[0060] For example, in cases where the system uses waveform alignment techniques, such as matching zero crossings or peak timing, the reactance may be inferred directly from the angular displacement between current and voltage signals. This enables direct mapping to model-based reactance per unit length without requiring a full complex impedance calculation. In other cases, the impedance component may be calculated explicitly and then decomposed into resistance and reactance values, or used as a magnitude for location correlation. Voltage and Current Phasors

[0061] Assuming the voltage phasor is taken as the reference with a phase angle of zero: Voltage phasor: Current phasor: Here, and are the peak values of the voltage and current in each half cycle. And is the phase angle by which the current lags behind the voltage. For AC circuits with a single frequency (like those powered by a typical power system), these phasors focus on the fundamental frequency component. Calculating Impedance (Z)

[0062] Impedance (Z) is the ratio of the voltage to the current in phasor form:Here, is the magnitude of the impedance , and is the phase angle of the impedance which indicates the phase difference between the voltage and current. For multi-Docket No.503461-70021 cycle faults, calculating is usually not an issue. For short duration events less than ½ cycle, then least-squared error or other approaches are utilized. The phase angle can be calculated a number of ways including but not limited to comparing the time differences of the voltage and current 1) peaks, 2) zero crossings, 3) point by point and discrete Wavelet transforms. Relating Impedance, Resistance, and Reactance

[0063] Ignoring capacitance, the complex impedance Z can be expressed as: And the magnitude of the impedanceis:Phase Angle (θ)

[0064] The phase angle of the impedance is given by: Calculating Resistance (R)

[0065] The following formulas useand to calculate the circuit’s resistance and inductive reactance accurately. To find , use trigonometric identity and magnitude of Z as follows:Calculating Inductive Reactance (XL)

[0066] With the equations of and above, is represented as follows:Docket No.503461-70021

[0067] FIG. 9A–FIG.13B include five sets of three-phase time-domain plots of WMU synchro-waveform data exhibiting various failure types at five different substations. The data is summarized in the following Table 2. Substation Event Type Zd Rd Xd Angle FIG. 9A–FIG.9B Tee 4.4 2.5 3.7 56.24 FIG.10A–FIG.10B Downed Conductor 5.1 1.4 4.9 73.86 FIG.11A–FIG.11B Tree / Limb Preventable 5.7 2.8 4.9 60.94 FIG.12A–FIG.12B Vines / Grasses 5.9 2.3 5.5 67.5 FIG.13A–FIG.13B Cable 7 4.9 5 45.93

[0068] In each “A” figure, an upper plot shows voltage (e.g., Va or Vb), voltage envelope (upper or both upper and lower), low-pass filtered voltage, and smoothed (e.g., five-point running average) voltage. The middle plot shows current (e.g., Ia) and current envelope (upper or both upper and lower). The lower plot shows an impedance component calculation, which is based on the phasor domain calculation described above.

[0069] Several cycles of the upper and middle plots are encompassed by a rectangle. That rectangle is also shown in the corresponding “B” figure to indicate samples obtained at the same times. In the “B” figures, the “d” subscripts denote data and “s” is smoothed, i.e., averaged.

[0070] FIG.9B shows how data from FIG.9A (representing event type time domain characteristics 902) is processed for each half phase to generate the data shown in the table of FIG.9B.

[0071] Thus, for at least each half-cycle corresponding to the transitory electrical anomaly, WMU data analysis system 204 (FIG.2) determines voltage and current phasor domain values, phase angles, and impedance components. This information in the table of FIG.9B defines event type phasor domain characteristics 904, which is representative of a tee connector fault shown in Table 2.

[0072] Like event type time domain characteristics 902, event type phasor domain characteristics 904 act as additional metadata by which to determine the event type (e.g., by indicating the rates of change and second order effects, time of day, duration, and other metadata observable in the phasor domain). For example, with reference to FIG. 9A and FIG.9B, the event type is a tee connector fault indicated by event type time domain characteristics 902 and event type phasor domain characteristics 904. In this example, event type phasor domain characteristics 904 show a multi-cycle persistent reduction in impedance magnitude and increase in phase angle. Event type time domain characteristics 902 show a multi-cycle increase in current and decrease in voltage. FIG.10A–FIG.13B show similar processing for other types of faults.Docket No.503461-70021

[0073] For instance, FIG.10A and FIG.10B show an example where the event type is a downed conductor fault indicated by event type time domain characteristics 1002 and event type phasor domain characteristics 1004. In this example, event type phasor domain characteristics 1004 show a persistent multi-cycle variable increase in impedance and phase angle which decrease over a series of cycles. Event type time domain characteristics 902 show a persistent multi-cycle decrease in voltage and increase in current with time variability in magnitude.

[0074] FIG.11A and FIG.11B show an example where the event type is a tree or limb fault indicated by event type time domain characteristics 1102 and event type phasor domain characteristics 1104. In this example, event type phasor domain characteristics 1104 show a multi-cycle decrease in impedance and increase in phase angle. Event type time domain characteristics 1102 show a 1.5 cycle increase in current with a decrease in voltage.

[0075] FIG.12A and FIG.12B show an example where the event type is a vines or grasses fault indicated by event type time domain characteristics 1202 and event type phasor domain characteristics 1204. In this example, event type phasor domain characteristics 1204 show an initial ½ cycle decrease in impedance with an increase in phase angle followed by multi-cycle oscillatory change in impedance and phase angle. Event type phasor domain characteristics 1204 show a sub-cycle decrease in voltage and increase incurrent.

[0076] FIG.13A and FIG.13B show an example where the event type is an underground line fault indicated by event type time domain characteristics 1302 and event type phasor domain characteristics 1304. In this example, event type phasor domain characteristics 1304 show a multi-cycle drop in impedance a step change increase in phase angle. Event type time domain characteristics 1302 show a multi-cycle increase in current and decrease in voltage.

[0077] Once data representing event type time and phasor domain characteristics is captured, it is then used to determine the type of fault, which is also referred to as event classification. FIG.14 shows an overview of event type classification 1400. A classification may entail an event persistence type (e.g., precursor, outage, or operations), an inferred cause (e.g., equipment failed), and a device type involved (e.g., cable).

[0078] In the example event type classification 1400, WMU data analysis system 204 (see, e.g., FIG.2) receives synchro-waveform data 1402 from a WMU monitoring a portion of an electrical power delivery system. Synchro-waveform data 1402 is shown as being stored in and retrieved from a database 1404, i.e., for analyses that are not performed in near realDocket No.503461-70021 time or are not time sensitive. In other embodiments, synchro-waveform data 1402 is analyzed in near real time.

[0079] Synchro-waveform data 1402 represents a transitory electrical anomaly 1406 (multiple other events are also shown), which includes voltage and current time domain values. As noted previously, these voltage and current time domain values define event type time domain characteristics. Specifically, the characteristics include but are not limited to the magnitudes, rates of change, time of day, and any other metadata associated with the time domain waveforms.

[0080] Next, for at least each half-cycle corresponding to transitory electrical anomaly 1406, WMU data analysis system 204 determines voltage and current phasor domain values, phase angles, and impedances, thereby defining event type phasor domain characteristics. The event type can then be determined based on analyses of both the event type time domain characteristics and the event type phasor domain characteristics, as described with reference to FIG.15–FIG.17.

[0081] FIG.15 shows a first step in the event type classification 1400, according to one embodiment. In this example, time aligned voltage and current data as well as signal analysis data is processed (e.g., to generate event type time and phasor domain characteristics). In the context of electric event analysis, the three phase voltage and current waveform measurements are digitized to produce time-series waveforms. These waveforms are examined in both the time and phasor domains. Different analyses are automatically performed, which include magnitude, phase angle, frequency spectra, duration, and transient rate-of-change obtained by various filtering and signal processing techniques. This data is then used for training machine learning algorithms, an example of which is shown in FIG. 17.

[0082] Second, FIG.16 shows how a curated dataset is prepared based on historical inputs. Events that are classified according to FIG.15 are grouped by event type and outage log entries to ensure that the dataset consists of those events of interest. The data set is validated by subject matter experts to ensure accuracy.

[0083] Third, the curated data set developed in FIG. 16 is subdivided into training and validation groups. The training data set is iteratively run through a machine learning process 1700 shown in FIG.17.

[0084] FIG.17 indicated that event type time and phasor domain characteristics (represented by input waveform and signal-processed waveforms 1702) are used for training a neural network 1704 and generating model inferences 1706, according to one embodiment.Docket No.503461-70021 In this example, some of the nodes in neural network 1704 reflect the weights that have been observed in the phasor domain characteristics. Skilled persons will appreciate that other machine learning algorithms may also be employed for generating model inferences 1706 about the event types.

[0085] In each training iteration, model parameters (e.g., weights of nodes in neural network 1704) are adjusted until the performance of neural network 1704 meets desired performance metrics. Once the training step is complete, the validation group of data is subjected to the same process until desired performance metrics are met and appropriate model inferences 1706 are verified. Once trained and validated, neural network 1704 can be deployed on new input data to determine event types and generate other model inferences 1706.

[0086] FIG.18 shows how impedance is used to calculate distance to the event based on modeling, based on an inverse time-domain approach (with no fault resistance), net fault current calculation by superposition method, or equation for line inductance (L) in terms measured voltage and current, a phasor domain approach, or a combination of all the above. In this example, FIG. 18 shows a plot 1800 of an impedance file for a circuit, developed from a load flow analysis tool, such as Synergi Electric electrical simulation software available from DNV of Høvik, Norway.

[0087] Plot 1800 shows an estimate 1802 of a calculated reactance, which is between a low 1804 and a high 1806 value of the possible reactance. The legend in plot 1800 also shows different types of electrical components that are modeled, such as underground lines 1808, main connection (switch) 1810, overhead lines 1812, breakers 1814, cross arms 1816, and T-connectors (underground) 1818.

[0088] In the example of FIG.18, there are multiple candidate inspection sites that have a reactance matching that of estimate 1802, or at least falling within range of locations corresponding to low 1804 and high 1806. For instance, an overhead line location 1820 and a breaker location 1822 both have a similar reactance but are at located at different distances from the substation. Accordingly, to reduce the ambiguity and the number of candidate inspection sites in the electrical power delivery system for inspection by a maintenance team, the event type (derives from synchro-waveform data as described above) is also used to estimate a fault location.

[0089] The event location may be determined by calculating the impedance values and comparing it to utility circuit impedance files, phasing, and conductor characteristics. For example, FIG.19 shows an example mapping user interface visualization 1900 of how theDocket No.503461-70021 reactance calculation shown in FIG.18 is mapped using GIS map data to visualize a search area or section 1902 for a fault location 1904 near a substation 1906. This visualization 1900 of FIG.19 can be generated by a local workstation (e.g., local PC 218, FIG. 2) or a SaaS client of WMU data analysis system 204, in which case the SaaS client receives visualization data via API calls exchanged with WMU data analysis system 204.

[0090] The GIS-based mapping system visually presents the fault search regions, aiding field crews in efficiently locating and addressing faults. These maps are generated by integrating electrical distribution network models—such as those developed in Synergi Electric or CYME Power Engineering Software—with open-source geospatial data sources, including OpenStreetMap. The network models provide GPS-based coordinates for substations, feeders, transformers, protection devices, and other components, along with electrical parameters such as reactance values and connectivity relationships.

[0091] In mapping user interface visualization 1900, a first annotation (e.g., dashed lines) represents a first type of electrical component (e.g., underground lines 1908). A second annotation (e.g., solid lines) represents another type of electrical component (e.g., overhead lines 1910). A first color (e.g., red) or other visual indicator annotates the type of electrical components (e.g., underground lines 1908) in search area or section 1902 that have been determined to fall within the calculated impedance. In this example, the calculated cumulative reactance is 2.51 Ω, which is mapped into search area or section 1902 that is discontiguous. The cumulative reactance of outage is 2.47 Ω. A second color (e.g., gray) or other visual indicator reflects a ruled out area or section 1912 based upon the event type, i.e., overhead lines 1910. Lastly, a third color (e.g., blue) or other visual indicator reflects underground lines 1908 outside of search area or sections 1902.

[0092] As noted, some areas on the GIS map are highlighted based on their calculated distances, the types of electrical components present, or other data associated with the event type. For instance, a pin may mark a location where a calculated reactance value has a matching event type (e.g., tee connector). In other examples, site information such as the presence of trees or other vegetation is matched to an event type determined to be a limb that downed a line.

[0093] The GIS map data used in this analysis can encompass several types of information. One category is spatial data, which includes anything that can be geographically represented, meaning it has defined coordinates (latitude and longitude) or can be mapped to a specific location on Earth. Spatial data typically falls into two groups: vector data, which includes points (e.g., cities, meters), lines (e.g., roads, power lines), and polygons (e.g.,Docket No.503461-70021 substations, land parcels); and raster data, which consists of pixel-based grids where each cell has a value. Raster data is commonly used to represent continuous phenomena like elevation, land use, vegetation cover, or temperature variations.

[0094] Another category is attribute data, which provides descriptive details about the spatial features. For instance, an electrical line may have associated attributes specifying its name, type (such as overhead or underground), construction material, or year of installation.

[0095] Temporal data captures how features change over time. This can include maintenance schedules, historical fault records, or the evolution of grid assets. This information is especially useful for predictive maintenance and outage analysis.

[0096] Finally, imagery and remote sensing data (including satellite images, aerial photographs, and other sensor outputs) can enhance the geographic model by providing visual context or detecting changes in environmental conditions. These datasets support fault classification, site prioritization, and post-event analysis.

[0097] To monitor and view power quality events and automated inferences, FIG.20–FIG. 22 show a set of screenshots for a power event monitoring software application, referred to as Cascadence. Cascadence uses machine learning to derive analytic insights and predictions from PQ meters, other data sources, OMS, and weather data from across the grid. Those predictions inform operations that specific equipment is failing so proactive maintenance / replacement can be carried out and customer outage time is reduced and satisfaction improved. Cascadence feeds that data into its machine learning algorithms (see, e.g., FIG.17) in order to automatically perform the following tasks, as described previously: event type classification (i.e., correlate power quality events to device types), and physics- based incipient fault location (i.e., determine the position of the failing equipment).

[0098] In this example, the Cascadence application presents to a user an event overview timeline user interface 2000 (FIG.20), an original signal user interface 2100 (FIG.21), and a signal processing user interface 2200 (FIG. 22). Each of these interfaces are described as follows.

[0099] With reference to FIG. 20, event overview timeline user interface 2000 is accessible via an overview tab 2002 and shows several timelines on which events are depicted. The timelines include a precursor timeline 2004 and an outage timeline 2006. A time slot on the timelines is selectable to view in greater detail particular events occurring in that time slot. In this example, outages 2008 are selected so that specific information about an outage event 2010 can be viewed.Docket No.503461-70021

[0100] FIG.21 shows original signal user interface 2100 for viewing the WMU data of a particular event. In this example, a precursor event 2102 is selected to show its original three-phase voltage signal waveforms 2104, an original three-phase current signal waveforms 2106, a reactance to fault 2108 (i.e., calculated as described above), and inferences 2110 that were determined using machine learning processes 1700 (FIG.17). For instance, inferences 2110 include an event persistence type 2112 that is shown as a precursor event and a fault type 2114 that is shown as a line-to-ground (LG) event.

[0101] Event persistence type 2112 includes precursor events, outages, and operational issues (e.g., schedule maintenance). This term conveys the various ongoing or transient states of events in the context of a power system, i.e., whether the events are persistent (ongoing or long-lasting in the case of planned outages) or transient (temporary or short- lived), which can be crucial for understanding and managing their impact on the power system.

[0102] Fault type 2114 includes single line-to-ground (LG) faults, line-to-line (LL) faults, line-to-line-to-ground (LLG) faults, and three-line faults (LLL).

[0103] In some embodiments, events can also be auto-labeled based on additional inferences. For example, labelling 2116 includes a cause 2118, a device 2120, and line placement 2122 (overhead (OH) or underground (UG)).

[0104] Causes 2118 label identifies the inferred origin of the event and may include capacitor-related operations such as cap switch activity, cap switch re-strike, or cap switch stuck contact. Additional causes may involve external influences or operational disruptions, including customer problems, dig-ins, fire, foreign objects in lines, vehicle contact, or deenergization for safety. Events may also be caused by environmental or weather-related factors such as lightning, severe weather, tree or limb contact (categorized as preventable or unpreventable), or overloaded conditions during normal operation. Equipment-related causes include faulty meters, loose connections, meter power supply failures, and general equipment failure or damage. Other causes encompass improper installation, non-standard construction, transmission system events, transformer inrush current, power electronics switching, restoration switching events, switching errors, and switching events more broadly (including load-off or load-on operations). The system may also assign general classifications such as snapshot, timing pulse, sympathetic event, under investigation, unknown, or other where more specific categorization is not possible.

[0105] Device 2120 label identifies the specific component involved in the event. This may include various overhead line components such as conductors, connectors, cross arms,Docket No.503461-70021 and poles; customer equipment; and protection or switching devices such as fuses, fuse switches, fuse cabinets, disconnects, disconnect switches, and reclosers. Other possible devices include elbows (both load-break and non-load-break types), T-connectors, potheads, probes, feed-throughs, hotline clamps, electronic sectionalizers, step-down transformers, regulators, relays, RA switches, RC-Off devices (distribution only), and ground banks. Additional substation-related or structural components may include insulators or pins, jumpers, polymer arresters, down guys, meters, miscellaneous substation equipment, splice points, handholes, and a general category for other equipment.

[0106] By clicking a processing tab 2124 in original signal user interface 2100, a signal processing user interface 2200 is presented to run various signal processing options 2202. Signal processing options 2202, in this example, include filters 2204 and transforms 2206.

[0107] FIG.23 shows a process 2300, performed by a WMU data analysis system for monitoring an electrical power delivery system, of processing synchro-waveform data to classify an event type and estimate, based on an impedance component and the event type, an event location so as to reduce a number of candidate inspection sites in the electrical power delivery system for inspection by a maintenance team.

[0108] In block 2302, process 2300 receives the synchro-waveform data from a WMU monitoring a portion of the electrical power delivery system, the synchro-waveform data representing a transitory electrical anomaly and including voltage and current time domain values, the voltage and current time domain values defining event type time domain characteristics.

[0109] In block 2304, process 2300 for at least each half-cycle corresponding to the transitory electrical anomaly, determines phase angles and impedance components, thereby defining event type phasor domain characteristics.

[0110] In block 2306, process 2300 determines the event type based on the event type time domain characteristics and the event type phasor domain characteristics.

[0111] In block 2308, process 2300 converts the impedance components to a range of locations in the electrical power delivery system based on an electrical model mapping between model impedance components and locations.

[0112] In block 2310, process 2300 identifies from the range of locations and the electrical model a set of locations having electrical components consistent with the event type so as to estimate the event location from candidate inspection sites within the range.

[0113] FIG.24 is a block diagram illustrating components 2400, according to some example embodiments, configured to read instructions from a machine-readable orDocket No.503461-70021 computer-readable medium (e.g., a non-transitory machine-readable storage medium) and perform any one or more of the methods discussed herein (or portions thereof) in connection with WMUs or WMU data analysis system 204 (FIG.2), such as discussed for processes 2300, or any portions of these embodiments.

[0114] Specifically, FIG.24 shows a diagrammatic representation of hardware resources 2402 including one or more processors 2404 (or processor cores), one or more memory / storage devices 2406, and one or more communication resources 2408, each of which may be communicatively coupled via a bus 2410. For embodiments where node virtualization (e.g., NFV) is utilized, a hypervisor 2412 may be executed to provide an execution environment for one or more network slices / sub-slices to utilize hardware resources 2402.

[0115] Processors 2404 (e.g., a central processing unit (CPU), a reduced instruction set computing (RISC) processor, a complex instruction set computing (CISC) processor, a graphics processing unit (GPU), a digital signal processor (DSP) such as a baseband processor, an application specific integrated circuit (ASIC), another processor, or any suitable combination thereof) may include, for example, a processor 2414 and a processor 2416.

[0116] Memory / storage devices 2406 may include main memory, disk storage, or any suitable combination thereof. Memory / storage devices 2406 may include, but are not limited to, any type of volatile or non-volatile memory such as dynamic random-access memory (DRAM), static random-access memory (SRAM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), Flash memory, solid-state storage, etc.

[0117] Communication resources 2408 may include interconnection or network interface components or other suitable devices to communicate with one or more peripheral devices 2418 or one or more databases 2414 via a network 2420. For example, communication resources 2408 may include wired communication components (e.g., for coupling via a Universal Serial Bus (USB)), cellular communication components, NFC components, Bluetooth® components (e.g., Bluetooth® Low Energy), Wi-Fi® components, and other communication components.

[0118] Instructions 2422 may comprise software, a program, an application, an applet, an app, or other executable code for causing at least any of processors 2404 to perform any one or more of the methods discussed herein. Instructions 2422 may reside, completely or partially, within at least one of processors 2404 (e.g., within the processor’s cache memory),Docket No.503461-70021 memory / storage devices 2406, or any suitable combination thereof. Furthermore, any portion of instructions 2422 may be transferred to hardware resources 2402 from any combination of peripheral devices 2418 or databases 2414. Accordingly, the memory of processors 2404, memory / storage devices 2406, peripheral devices 2418, and databases 2414 are examples of computer-readable and machine-readable media.

[0119] In light of this disclosure, skilled persons will appreciate that many changes may be made to the details of the above-described embodiments without departing from the underlying principles of the invention. The scope of the present invention should, therefore, be determined only by claims and equivalents.

Claims

Docket No.503461-70021 CLAIMS What is claimed is:

1. A method, performed by a WMU data analysis system for monitoring an electrical power delivery system, of processing synchro-waveform data to classify an event type and estimate, based on an impedance component and the event type, an event location so as to reduce a number of candidate inspection sites in the electrical power delivery system for inspection by a maintenance team, the method comprising: receiving the synchro-waveform data from a WMU monitoring a portion of the electrical power delivery system, the synchro-waveform data representing a transitory electrical anomaly and including voltage and current time domain values, the voltage and current time domain values defining event type time domain characteristics; for at least each half-cycle corresponding to the transitory electrical anomaly, determining phase angles and impedance components, thereby defining event type phasor domain characteristics; determining the event type based on the event type time domain characteristics and the event type phasor domain characteristics; converting the impedance components to a range of locations in the electrical power delivery system based on an electrical model mapping between model impedance components and locations; and identifying from the range of locations and the electrical model a set of locations having electrical components consistent with the event type so as to estimate the event location from candidate inspection sites within the range.

2. The method of claim 1, further comprising presenting to a user a visualization of the event location annotated in the electrical model.

3. The method of claim 1, in which the electrical model include GIS data.

4. The method of claim 3, further comprising annotating the GIS data to show the event location.

5. The method of claim 1, in which the event type is a tee connector fault, the fault being indicated by event type time and phasor domain characteristics, the event type phasor domain characteristics comprising a multi-cycle persistent reduction in impedance component value and an increase in phase angle, and the time domain characteristics comprising a multi-cycle increase in current and decrease in voltage.Docket No.503461-70021 6. The method of claim 1, in which the event type is a downed conductor fault indicated by the event type time and phasor domain characteristics, the event type phasor domain characteristics showing a persistent multi-cycle variable increase in impedance component value and phase angle which decrease over a series of cycles, and the event type time domain characteristics showing a persistent multi-cycle decrease in voltage and increase in current with time variability in magnitude.

7. The method of claim 1, in which the event type is a tree or limb fault indicated by the event type time and phasor domain characteristics, the event type phasor domain characteristics showing a multi-cycle decrease in impedance component value and increase in phase angle, and the event type time domain characteristics showing a 1.5 cycle increase in current with a decrease in voltage.

8. The method of claim 1, in which the event type is a vines or grasses fault indicated by the event type time and phasor domain characteristics, the event type phasor domain characteristics showing an initial ½ cycle decrease in impedance component value with an increase in phase angle followed by multi-cycle oscillatory change in impedance component and phase angle, and the event type time domain characteristics showing a sub-cycle decrease in voltage and increase incurrent.

9. The method of claim 1, in which the event type is an underground line fault indicated by the event type time and phasor domain characteristics, the event type phasor domain characteristics showing a multi-cycle drop in impedance component value a step change increase in phase angle, and the event type time domain characteristics showing a multi- cycle increase in current and decrease in voltage.

10. The method of claim 1, in which the event location is determined by calculating an impedance component value and comparing it to utility circuit impedance component files, phasing, and conductor characteristics.

11. A WMU data analysis system for monitoring an electrical power delivery system and configured to process synchro-waveform data to classify an event type and estimate, based on an impedance component and the event type, an event location so as to reduce a number of candidate inspection sites in the electrical power delivery system for inspection by a maintenance team, the WMU data analysis system comprising: a substation server configured to receive synchro-waveform data from a WMU monitoring a portion of the electrical power delivery system, the synchro-waveform dataDocket No.503461-70021 representing a transitory electrical anomaly and including voltage and current time domain values, the voltage and current time domain values defining event type time domain characteristics; and a remote server communicatively coupled to the substation server and configured to: determine, for at least each half-cycle corresponding to the transitory electrical anomaly, phase angles and impedance components, thereby defining event type phasor domain characteristics; determine the event type based on the event type time domain characteristics and the event type phasor domain characteristics; convert the impedance components to a range of locations in the electrical power delivery system based on an electrical model mapping between model impedance components and locations; and identify from the range of locations and the electrical model a set of locations having electrical components consistent with the event type so as to estimate the event location from candidate inspection sites within the range.

12. The WMU data analysis system of claim 11, further comprising an interface for providing a visualization of the event location annotated in the electrical model.

13. The WMU data analysis system of claim 11, in which the electrical model include GIS data.

14. The WMU data analysis system of claim 13, in which the remote server is configured to annotate the GIS data to show the event location.

15. The WMU data analysis system of claim 11, in which the remote server is configured to determine that the event type is a tee connector fault indicated by the event type time and phasor domain characteristics, the event type phasor domain characteristics showing a multi-cycle persistent reduction in impedance component value and increase in phase angle, and event type time domain characteristics showing a multi-cycle increase in current and decrease in voltage.

16. The WMU data analysis system of claim 11, in which the remote server is configured to determine that the event type is a downed conductor fault indicated by the event type time and phasor domain characteristics, the event type phasor domain characteristics showing a persistent multi-cycle variable increase in impedance component value and phase angle which decrease over a series of cycles, and the event type time domain characteristicsDocket No.503461-70021 showing a persistent multi-cycle decrease in voltage and increase in current with time variability in magnitude.

17. The WMU data analysis system of claim 11, in which the remote server is configured to determine that the event type is a tree or limb fault indicated by the event type time and phasor domain characteristics, the event type phasor domain characteristics showing a multi-cycle decrease in impedance component value and increase in phase angle, and the event type time domain characteristics showing a 1.5 cycle increase in current with a decrease in voltage.

18. The WMU data analysis system of claim 11, in which the remote server is configured to determine that the event type is a vines or grasses fault indicated by the event type time and phasor domain characteristics, the event type phasor domain characteristics showing an initial ½ cycle decrease in impedance component value with an increase in phase angle followed by multi-cycle oscillatory change in impedance component value and phase angle, and the event type time domain characteristics showing a sub-cycle decrease in voltage and increase incurrent.

19. The WMU data analysis system of claim 11, in which the remote server is configured to determine that the event type is an underground line fault indicated by the event type time and phasor domain characteristics, the event type phasor domain characteristics showing a multi-cycle drop in impedance component value a step change increase in phase angle, and the event type time domain characteristics showing a multi-cycle increase in current and decrease in voltage.

20. The WMU data analysis system of claim 11, in which the event location is determined by calculating an impedance component value and comparing it to utility circuit impedance component files, phasing, and conductor characteristics.

21. A computer-implemented user interface for analyzing waveform data in an electrical power delivery system, the user interface comprising: a graphical display that presents voltage and current waveform plots based on synchro-waveform data received from one or more waveform monitoring units (WMUs) deployed in the electrical power delivery system, the graphical display further presenting impedance component values calculated in the phasor domain from the synchro-waveform data, the calculations performed at a resolution of at least one value per half-cycle of the waveform;Docket No.503461-70021 a portion of the graphical display configured to show a predicted fault type and event classification determined by applying a machine learning model to the synchro-waveform data and corresponding phasor domain calculations; and a map view rendered using geospatial data and electrical network model data, the map view presenting a candidate fault location determined by correlating the calculated impedance component value with known line reactance values in the electrical network model and corresponding to a predicted fault type.

22. The user interface of claim 21, in which the impedance component and phase angle values are calculated for each half-cycle or less of the voltage and current waveform.

23. The user interface of claim 21, in which the predicted fault type comprises a tee connector fault, downed conductor fault, underground cable fault, or tree- or vegetation- related fault.

24. The user interface of claim 21, in which the machine learning model is trained using labeled waveform data that includes both time-domain and phasor-domain characteristics.

25. The user interface of claim 21, in which the event classification further comprises an event persistence type selected from: precursor event, sustained outage, or operational anomaly.

26. The user interface of claim 21, in which the electrical network model comprises reactance values associated with specific line segments, feeders, transformers, or switches, and is obtained from Synergi Electric or CYME Power Engineering Software.

27. The user interface of claim 21, in which the map view displays the electrical power delivery system using geographic coordinates, electrical topology, and visual annotations indicating predicted fault locations.

28. The user interface of claim 21, in which the graphical display is further configured to update the map view in real time as additional synchro-waveform data is received and processed.

29. The user interface of claim 21, in which the impedance component-to-location correlation includes matching the calculated reactance to a range of reactance values and selecting a subset of line segments falling within the range.Docket No.503461-70021 30. The user interface of claim 21, in which the user interface enables user selection of a fault event from a timeline, upon which the waveform plots, fault type, and map view are automatically updated to reflect the selected event.

Images