REMOTE MONITORING OF ELECTRICAL EQUIPMENT WITH MULTI-FACTOR HISTORICAL / STATISTICAL ANOMALY DETECTION AND PREDICTION

RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application serial number 62/587,903, filed November 17, 2017, the disclosure of which is hereby incorporated herein by reference in its entirety.

FIELD OF THE DISCLOSURE

[0002] This disclosure relates to monitoring the health of electrical equipment, such as high-voltage electrical transformers. More specifically, it relates methods and systems for remote monitoring of electrical equipment with multi-factor historical and/or statistical anomaly detection and prediction.

BACKGROUND

[0003] There is a critical need for reliable power systems and, to ensure that reliability, Sensei’s MasterMind Adaptive Response Technology (ART and SMART). MasterMind provides an essential tool to increase power system reliability and to minimize or eliminate unplanned service interruptions and potential adverse effects.

[0004] The overworked electrical grid isn’t a new concern, but it does source from a developing problem that has created new challenges. The combined effects of an aging fleet and the introduction of renewable energy sources to the grid place unanticipated demands on legacy equipment. Since its inception, electrical demand has been fairly predictable. For most of its history, fluctuations in ambient temperature were the largest drivers of demand. After all, ambient temperature is a predictable variable; electrical generation was thus understood as a relatively steady-state proposition.

[0005] Today, however, renewable energy sources present a new, more dynamic problem: matching demand to supply and vice versa. The majority of the grid’s critical, high-voltage legacy assets were simply not designed to withstand the stressful dynamic loading patterns imposed by renewable energy sources. [0006] Although transformers can operate for decades at 100 % of nameplate with load at a steady state, when they are subjected to dramatic, repetitive load changes, thermal expansions can eventually compromise the life of the unit. Critical bushing and transformer seals are two of the most concerning elements that we find regularly under siege in this new paradigm. A compromised seal allows moisture introduction, leading to degraded dielectric oil insulating capability, breakdown, and ultimately partial discharge.

[0007] Periodic manual sampling has serious shortcomings with potentially drastic implications. Infrequent handheld inspections are inherently more limited in scope, less repeatable, resource-intensive, and dramatically more expensive per sample than automated solutions. The practitioner may be exposed to potentially hazardous circumstances when taking measurements on energized equipment and issues that happen during the majority of the time when nobody is watching are simply not caught. Even the quality of analysis for the inspections that are performed manually are subject to errors by the humans doing them. Finally, manual data collection and tabulation are far more difficult to integrate into a CBM regimen.

[0008] These blind spots have a huge price-tag since outages due to transformer failures can cost companies up to $100,000 an hour. However, if utilities use intelligent, configurable automated monitoring solutions, they will not only be able to monitor in real-time, but also be able to predict conditions with far greater accuracy and reliability than can be obtained by manual radiometric inspection methods.

[0009] Accordingly, in light of the disadvantages associated with conventional approaches to remote monitoring of electrical equipment, there is a need for improved methods and systems for remote monitoring of electrical equipment with multi-factor historical and/or statistical anomaly detection and prediction.

SUMMARY

[0010] The subject matter disclosed herein includes methods, systems, and non- transitory computer readable media for remote monitoring of electrical equipment with multi-factor statistical anomaly detection and prediction. The methods and systems described herein use multi-factor historical and/or statistical analysis to detect and/or predict anomalies involving, but not limited to: anomalies indicative of actual or potential equipment failure, including but not limited to transformer failure, bus failure, and distribution equipment failure; anomalies indicative of actual or potential unexpected or unintended network loading conditions or transients; and anomalies indicative of site trespass, vandalism, or damage.

[0011] According to one aspect of the present disclosure, a method for remote monitoring of electrical equipment with multi-factor historical and/or statistical anomaly detection and prediction comprises: acquiring a plurality of data points, each data point representing a characteristic associated with a piece of electrical equipment or a component thereof; assigning each data point to at least one of a plurality of groups, wherein each group contains a plurality of data points; defining an alarm metric for each group, wherein an alarm metric of at least one group is different from an alarm metric of another group; and using the defined alarm metrics to determine a health of the electrical equipment and provide notification of same.

[0012] In some embodiments, the characteristic associated with a piece of electrical equipment or a component thereof comprises a temperature; a dissolved gas analysis reading; an electrical voltage being supplied or produced; an electrical current being supplied or produced; an ambient operating condition; and/or an operation load condition.

[0013] In some embodiments, the plurality of data points represent at least one characteristic associated with each of at least two pieces of electrical equipment or components thereof.

[0014] In some embodiments, the alarm metric is based at least in part on a measurement of a standard deviation of a set of two or more characteristics associated with a piece of electrical equipment or a component thereof.

[0015] In some embodiments, the alarm metric is based at least in part on a change of a standard deviation of a set of two or more characteristics associated with a piece of electrical equipment or a component thereof over time.

[0016] In some embodiments, the alarm metric comprises a multi-factor historical or statistical analysis.

[0017] In some embodiments, the analysis comprises a transient analysis.

[0018] In some embodiments, the analysis comprises a steady-state analysis. [0019] In some embodiments, the analysis comprises monitoring the multi-factor data to detect an anomaly between a measured characteristic and another measurement taken contemporaneou sly .

[0020] In some embodiments, the analysis comprises monitoring the multi-factor data to detect an anomaly between a characteristic measured at a first point in time and the same characteristic measured as at second point in time different from the first point in time.

[0021] In some embodiments, the analysis comprises monitoring the multi-factor data to detect an anomaly between a first relationship of two or more measured characteristics and a second relationship or two or more measured characteristics.

[0022] In some embodiments, the first and second relationships are between the same two or more measured characteristics but at different points in time.

[0023] In some embodiments, the first relationship is between a first set of two or more measured characteristics and wherein the second relationship is between a second set of two or more measured characteristics, wherein at least one element of the first set is different from at least one element of the second set.

[0024] In some embodiments, acquiring the plurality of data points includes: using an infrared image sensor to capture a scene that includes a view of the electrical equipment; identifying, as virtual probes, portions of the captured scene that represent pieces of electrical equipment or components thereof; determining a data value for each virtual probe based on a color or intensity of the portion of the captured scene represented by the virtual probe; and providing the data values as at least some of the plurality of data points.

[0025] In some embodiments, at least one virtual probe includes a plurality of non contiguous areas of the captured scene.

[0026] In some embodiments, the method further comprises providing, to a user, a display that shows an infrared image of the view captured by the infrared image sensor, wherein a hue or intensity of a pixel of the image that corresponds to the equipment represents a data value corresponding to a temperature of the equipment at that location in the image. [0027] In some embodiments, the method further comprises providing, on the infrared image, virtual indications of the location of virtual probes defined for that image.

[0028] In some embodiments, the method further comprises adjusting an intensity map or color map of the image to increase a range or sensitivity of the data values being represented to the user.

[0029] In some embodiments, the method further comprises mapping a sub-range of data values to a single color.

[0030] In some embodiments, data values in a first sub-range are displayed in color and data values in a second sub-range are displayed in grayscale or black and white.

[0031] In some embodiments, acquiring the plurality of data points includes acquiring at least one of the plurality of data points from a wireless sensor that transmits a data value that represents a temperature for the piece of electrical equipment or portion thereof.

[0032] In some embodiments, the wireless sensor comprises a radio frequency identification (RFID) device for transmitting the data value for the piece of electrical equipment or portion thereof.

[0033] In some embodiments, the wireless sensor is attached to the piece of electrical equipment or portion thereof.

[0034] In some embodiments, the wireless sensor includes a temperature sensing circuit for sensing a relative or absolute temperature value for the piece of electrical equipment or portion thereof.

[0035] In some embodiments, determining the health of the electrical equipment includes identifying electrical equipment having an increased likelihood of failure.

[0036] In some embodiments, defining the alarm metric for the group includes: determining a data value for each data point in the group; determining whether a data point has a data value that differs from a data value of at least one other data point in the group by a threshold amount, and, responsive to that determination, identifying the electrical equipment represented by the identified data point as having an increased likelihood of failure.

[0037] According to another aspect of the present disclosure, a system for remote monitoring of electrical equipment comprises: a data acquisition module comprising circuitry for acquiring a plurality of data points, each data point representing a characteristic associated with a piece of electrical equipment or a component thereof; and a monitoring module comprising circuitry for assigning each data point to at least one of a plurality of groups, wherein each group contains a plurality of data points, defining an alarm metric for each group, wherein an alarm metric of at least one group is different from an alarm metric of another group, and using the defined alarm metrics to determine a health of the electrical equipment and provide notification of same.

[0038] In some embodiments, the characteristic associated with a piece of electrical equipment or a component thereof comprises a temperature; a dissolved gas analysis reading; an electrical voltage being supplied or produced; an electrical current being supplied or produced; an ambient operating condition; and/or an operation load condition.

[0039] In some embodiments, the plurality of data points represent at least one characteristic associated with each of at least two pieces of electrical equipment or components thereof.

[0040] In some embodiments, the alarm metric is based at least in part on a measurement of a standard deviation of a set of two or more characteristics associated with a piece of electrical equipment or a component thereof.

[0041] In some embodiments, the alarm metric is based at least in part on a change of a standard deviation of a set of two or more characteristics associated with a piece of electrical equipment or a component thereof over time.

[0042] In some embodiments, the alarm metric comprises a multi-factor historical or statistical analysis.

[0043] In some embodiments, the analysis comprises a transient analysis.

[0044] In some embodiments, the analysis comprises a steady-state analysis.

[0045] In some embodiments, the analysis comprises monitoring the multi-factor data to detect an anomaly between a measured characteristic and another measurement taken contemporaneou sly .

[0046] In some embodiments, the analysis comprises monitoring the multi-factor data to detect an anomaly between a characteristic measured at a first point in time and the same characteristic measured as at second point in time different from the first point in time. [0047] In some embodiments, the analysis comprises monitoring the multi-factor data to detect an anomaly between a first relationship of two or more measured characteristics and a second relationship or two or more measured characteristics.

[0048] In some embodiments, the first and second relationships are between the same two or more measured characteristics but at different points in time.

[0049] In some embodiments, the first relationship is between a first set of two or more measured characteristics and wherein the second relationship is between a second set of two or more measured characteristics, wherein at least one element of the first set is different from at least one element of the second set.

[0050] In some embodiments, the data acquisition module includes an infrared image sensor for capturing a scene that includes a view of the electrical equipment, wherein portions of the captured scene are identified as virtual probes, wherein each virtual probe represents pieces of electrical equipment or components thereof, wherein a relative or absolute data value for each virtual probe is determined based on a color or intensity of the portion of the captured scene represented by the virtual probe.

[0051] In some embodiments, at least one virtual probe includes a plurality of non contiguous areas of the captured scene.

[0052] In some embodiments, the data acquisition module includes a wireless receiver for receiving data values from a wireless sensor that transmits a data value that represents a temperature for the piece of electrical equipment or portion thereof.

[0053] In some embodiments, the wireless sensor is attached to the piece of electrical equipment or portion thereof.

[0054] In some embodiments, the wireless sensor includes a temperature sensing circuit for sensing a relative or absolute temperature value for the piece of electrical equipment or portion thereof.

[0055] In some embodiments, the wireless receiver communicates with radio frequency identification (RFID) devices.

[0056] In some embodiments, determining the health of the electrical equipment includes identifying electrical equipment having an increased likelihood of failure.

[0057] In some embodiments, defining the alarm metric for the group includes: determining a data value for each data point in the group; determining whether a data point has a data value that differs from a data value of at least one other data point in the group by a threshold amount, and, responsive to that determination, identifying the electrical equipment represented by the identified data point as having an increased likelihood of failure.

[0058] According to yet another aspect of the present disclosure, a computer program product for signaling optimization in a wireless network utilizing proprietary and non proprietary protocols comprises a non-transitory computer readable storage medium having computer readable code embodied therewith, the computer readable code comprising computer readable program code configured for: acquiring a plurality of data points, each data point representing a characteristic associated with a piece of electrical equipment or a component thereof; assigning each data point to at least one of a plurality of groups, wherein each group contains a plurality of data points; defining an alarm metric for each group, wherein an alarm metric of at least one group is different from an alarm metric of another group; and using the defined alarm metrics to determine a health of the electrical equipment and provide notification of same.

[0059] In some embodiments, the characteristic associated with a piece of electrical equipment or a component thereof comprises: a temperature; a dissolved gas analysis reading; an electrical voltage being supplied or produced; an electrical current being supplied or produced; an ambient operating condition; and/or an operation load condition.

[0060] The subject matter described herein may be implemented in hardware, software, firmware, or any combination thereof. As such, the terms “function” or “module” as used herein refer to hardware, software, and/or firmware for implementing the feature being described.

[0061] In one exemplary implementation, the subject matter described herein may be implemented using a computer readable medium having stored thereon executable instructions that when executed by the processor of a computer control the computer to perform steps. Exemplary computer readable media suitable for implementing the subject matter described herein include disk memory devices, chip memory devices, programmable logic devices, application specific integrated circuits, and other non- transitory storage media. In one implementation, the computer readable medium may include a memory accessible by a processor of a computer or other like device. The memory may include instructions executable by the processor for implementing any of the methods described herein. In addition, a computer readable medium that implements the subject matter described herein may be located on a single device or computing platform or may be distributed across multiple physical devices and/or computing platforms.

[0062] Those skilled in the art will appreciate the scope of the present invention and realize additional aspects thereof after reading the following detailed description of the preferred embodiments in association with the accompanying drawing figures.

BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 is a black and white thermal image of a scene as captured according to embodiments of the present disclosure, the scene including a power station transformer 100 operating at normal temperatures.

[0064] Figure 2 is a black and white thermal image of a scene as captured according to embodiments of the present disclosure, the scene including a pair of transformers, one operating at normal temperature and the other overheating during operation.

[0065] Figure 3 is a black and white thermal image of a scene as captured according to embodiments of the present disclosure, the scene including multiple heating elements that are operating at the same temperature yet appear to be different temperatures due to the effects of distance from the sensor on the indicated temperature.

[0066] Figure 4 is a block diagram illustrating an exemplary system for remote monitoring of electrical equipment with multi-factor historical and/or statistical anomaly detection and prediction according to some embodiments of the present disclosure.

[0067] Figure 5 is a flow chart illustrating an exemplary process for remote monitoring of electrical equipment with multi-factor historical and/or statistical anomaly detection and prediction according to some embodiments of the present disclosure.

[0068] Figure 6 illustrates an example set of equipment that may be monitored by a system according to embodiments of the present disclosure.

[0069] Figure 7 is a screen shot of a system according to embodiments of the present disclosure, showing detection of an anomalous response to load variations. [0070] Figure 8 is a screen shot of a system according to embodiments of the present disclosure, showing normal response to load variations.

DETAILED DESCRIPTION

[0071] Methods and systems for remote monitoring of electrical equipment with multi-factor historical and/or statistical anomaly detection and prediction are provided herein.

[0072] Many electrical apparatus and components will heat up before they break down, which is why thermal imaging cameras are used to detect the rise in temperature at an early stage.

[0073] Figure 1 shows a black and white thermal image of a scene as captured according to methods and systems of the present disclosure, the scene including a power station transformer 100 operating at normal temperatures.

[0074] Figure 2 shows a black and white thermal image of a scene as captured according to methods and systems of the present disclosure, the scene including a pair of transformers 200 and 202. Transformer 200 is operating at normal temperature but transformer 202 is operating at a much higher temperature compared to comparable transformer 200; transformer 202 is showing a higher risk of failure, since overheating is often a symptom of a condition that leads to failure.

[0075] Thermal imaging provides its own challenges, however.

[0076] Figure 3 shows a black and white thermal image of four heating elements, labeled 300, 302, 304, and 306 in the image, which are operating at the same temperature. In this image, elements 300 and 302 are closer to the thermal camera than are elements 304 and 306. Even though all four devices are actually at the same temperature, elements 300 and 302 appear in the thermal image to be hotter than elements 304 and 306, demonstrating the effect of distance on the indicated temperature, compounded by conventional colorizing algorithms. The lighter band 308 above element 302 indicates heat flow perpendicular to the thermal sensor, demonstrating the effect of wind on obtaining true temperature measurements.

[0077] Conventional approaches to monitor the health of electrical grid infrastructure equipment also include dissolved gas analysis (DGA) and installation of bushing monitors deeper within medium-voltage transformers and switchgear distribution networks.

[0078] Individually, these methods provide only a sliver of insight into the true condition of the equipment. In order to gain a more reliable understanding of real-time conditions, the methods and system described herein consider a multitude of factors in context and then correlate them with one another.

[0079] Sensei Solutions, an innovative, experienced leader in Smart Grid solutions, has developed a monitoring system for their most critical and most vulnerable components. Armed with thermal imaging cameras, Sensei developed a solution that didn’t require a single outage to implement.

[0080] Figure 4 is a block diagram illustrating an exemplary system for remote monitoring of electrical equipment with multi-factor historical and/or statistical anomaly detection and prediction according to some embodiments of the present disclosure. In the embodiment illustrated in Figure 4, the system 400 includes:

[0081] A data acquisition module 402 comprising circuitry for acquiring a plurality of data points, each data point representing a characteristic associated with a piece of electrical equipment or a component thereof; and

[0082] A monitoring module 404 comprising circuitry for assigning each data point to at least one of a plurality of groups, wherein each group contains a plurality of data points, defining an alarm metric for each group, wherein an alarm metric of at least one group is different from an alarm metric of another group, and using the defined alarm metrics to determine a health of the electrical equipment and provide notification of same. Providing notification may include, but is not limited to: providing an indication that the equipment is operating normally, providing an indication that the equipment is operating abnormally, providing an indication that the equipment has a relative (e.g., high, medium, or low) probability of failure; providing an indication that the equipment is operating outside of normal parameters; or other type of indication. Such indication may be visual, aural, textual, or other, and may be delivered via a display, via a speaker, via a communication link, via an email or text message, or via other means.

[0083] Figure 5 is a flow chart illustrating an exemplary process for remote monitoring of electrical equipment with multi-factor historical and/or statistical anomaly detection and prediction according to some embodiments of the present disclosure. In the embodiment illustrated in Figure 5, the steps include:

[0084] Step 500: acquiring a plurality of data points, each data point representing a characteristic associated with a piece of electrical equipment or a component thereof;

[0085] Step 502: assigning each data point to at least one of a plurality of groups, wherein each group contains a plurality of data points;

[0086] Step 504: defining an alarm metric for each group, wherein an alarm metric of at least one group is different from an alarm metric of another group; and

[0087] Step 506: using the defined alarm metrics to determine a health of the electrical equipment and provide notification of same. Example alarm metrics will be discussed below.

[0088] In an example embodiment, radiometric samples from eight thermal cameras were collected every twenty minutes, recording high, low, mean, and median temperatures for dozens of measurement points on eight transformer/breaker pairs, and increasing the thermographic sampling rate from four times per year to over 26,000 times per year. As described in more detail below, the methods and systems presented herein do not require the time consuming and difficult-to-do precise calibration of thermal cameras and do not require precise mapping of luminance or color values to temperature, and in fact do not require a conversion of raw data to a temperature at all - the algorithms can detect anomalies in the raw data as easily as they can detect anomalies in temperature data (or dissolved gas readings, voltage readings, current readings, load values, etc.) derived from that raw data - a significant advantage over prior art systems.

Sensei MasterMind Adaptive Response Technology (SMART)

[0089] Methods and systems for remote monitoring of electrical equipment with multi-factor historical and/or statistical anomaly detection and prediction are herein presented. One embodiment of the subject matter described herein is referred to as Sensei MasterMind Adaptive Response Technology (SMART), which employs multiple inputs, systems, and analytic probes to provide dynamic monitoring for power system components. [0090] Sensei Solutions has developed a set of proprietary analytics that make it possible to detect degrading condition of bushings, LTCs and switchgear at a much earlier stage and with greater reliability than previously available through periodic manual inspection. The same holds true for processing, correlating, and interpreting data from DGA, PD, and PF sensors. With Sensei’s MasterMind platform, it is possible to plan maintenance more effectively and to prevent costly outages.

Solving the problems of periodic monitoring: Episodic and incomplete data

[0091] Sensei’s MasterMind AVR Solution provides a dramatic improvement to traditional periodic handheld thermography methods employed by electric utilities for many years. The methods and systems disclosed herein provide intelligent, configurable automated monitoring solutions that are not only able to monitor in real-time, but also able to predict conditions with far greater accuracy and reliability than can be obtained by manual radiometric inspection methods.

[0092] In one deployment, the SMART system’s multi-factor historical and statistical anomaly detection and prediction was able to identify and remediate three critical switchgear defects in a steel smelting facility. The SMART platform provides real-time and retrospective analysis.

Sigma Delta Tau: Calculation of anomalies

[0093] In contrast to the limitations of traditional inspection methods, one innovation of the present disclosure is the use of a Sigma Delta Tau (SDT) algorithm, an adaptive comparative analytics application that measures relational temperature. The underlying principle rests on a relatively simple proposition: because similar devices operating under identical conditions produce similar thermal profiles, variances in these relationships can quickly identify outliers and indicate further investigative action.

[0094] In some embodiments, the SDT algorithm of the present disclosure tracks the standard deviation and the stability of those relationships over time among a group of related measurements. The SDT algorithm can be applied to transformer and breaker bushings, tap-changers, disconnects, and capacitor banks. In addition, MasterMind automatically integrates environmental variables to achieve a truer indication of condition and performance than can be obtained using discrete threshold measurements.

[0095] Together, these two calculations - Sigma and Delta Tau, respectively - track the instantaneous vs. long term stability of the thermal signature of a given asset. Multiple measurements may be grouped and compared providing detection and alerting of both instantaneous anomalies and worrisome longer-term trends.

Re-envisioning thermographic analysis

[0096] Conventional approaches to thermographic imagery produce a color-coded image, which renders the naturally grayscale radiometric image in“technicolor.” It seems intuitive, and it’s something so ubiquitous that it’s shorthand in movies and TV for “advanced” tracking. But there is a problem that far more temperatures exist than there are colors to represent them. As a result, software engineers have to subjectively interpolate a range of temperatures across a range of colors. Sometimes this process produces dazzling visual effects, but frequently fails to provide a true representation of temperatures.

[0097] Improving on these methods, in some embodiments, a function allows for user-specific assignments of colors to temperatures painted against a grayscale background. Users, at a glance, can tell the temperature of each scene element based on user- specified parameters.

[0098] According to some embodiments, another important improvement over existing Spot or Area measurement tools that frequently measure either too few or too many pixels is a Virtual Probe Editor. Multiple measurement points (or“virtual probes” in Sensei parlance) may be defined by the user. The Probe Editor allows for pixel-by- pixel definition of measurement points or regions of interest by painting directly onto the radiometric image. Advanced probe features allow for the averaging multiple scene elements such as sky, shadows and ground surface to determine a truer ambient baseline. And complex shapes and contours pose no problem given the simple and intuitive painting tools. Use Case : Monitoring Equipment Operation

[0099] Figure 7 is a screen shot of a system according to embodiments of the present disclosure, showing detection of an anomalous response to load variations. In the embodiment illustrated in Figure 7, the screen shot 700 includes a waveform history 702, a visible light image of the equipment being monitored 704 and infrared image of the equipment being monitored 706. The waveform history pane 702 shows temperatures on each of three phases of a gas breaker, shown as waveforms 708. In this example, the temperature of one of the phases differed significantly from the temperature of the other phases; three such anomalies 710 were recorded. In the example illustrated in Figure 7, these anomalies triggered an alarm condition 712. The system can inform the operator that the virtual probe that displayed the anomalous behavior corresponds to a physical location on the equipment 714. In this example, the culprit turned out to be a loose connection along the power supply path, in this case a loose gas breaker bushing connection. The SMART system’s use of multi-factor historical / statistical anomaly detection and prediction detected an anomaly that manifested itself as a transient, rather than steady-state, phenomenon. Moreover, the alarm condition may be described in terms of a variance in temperature in one of the phases relative to the temperatures of the other phases rather than as an absolute temperature. Had the alarm condition been only defined in terms of absolute temperature, the previous transients (near the arrows 708) may have triggered a false alarm.

[00100] Using Figure 7 as an illustrative example, the system 700 could group the phases of a transformer, determine the variance of the three phases, and trigger alarm if the variance exceeds a threshold level. Likewise, the system 700 could calculate this variance for each of multiple transformers, create another group of those variances, and trigger an alarm if any transformer’s variance exceeds the other transformers’ variances by a threshold level. Other approaches that could be implemented by system 700 or any other system disclosed herein include looking for variances between data generated by virtual probes and data generated by physical probes and triggering an alarm if the variance is greater than a threshold; triggering an alarm if an entire group of temperatures differs significantly from a change in ambient temperature; triggering an alarm if a significant enough change in any of the monitored data values occurs while an intrusion sensor indicates an intrusion; triggering an alarm if monitor data values don’t follow an expected daily, weekly, seasonal, yearly, or other schedule; triggering an alarm if one group of data does not track another group of data; triggering an alarm is a significant enough number of readings in a particular geographic area show a variance or other trend; and so on. Likewise, a group of measured voltages can be compared to a group of measured temperatures and trigger an alarm if the two groups show a particular relationship to each other. The above examples are illustrative and not limiting.

[00101] Figure 8 is a screen shot of a system according to embodiments of the present disclosure, showing normal response to load variations. In the embodiment illustrated in Figure 8, the screen shot 800 includes a waveform history 802 showing a plot of a group of data points over time, e.g., temperatures on three phases of a gas breaker. Unlike the example illustrated in Figure 7, in Figure 8 the temperatures of the three phases of the gas breaker stay close to each other - i.e., they all go up together, they all go down together, they all stay the same together - which indicates that the gas breaker is operating normally, e.g., does not show any behavior to suggest that it is not healthy or suggest that an increased possibility of failure. In this example, none of the alarm conditions are triggered.

Use Case : Reliable power supply for transformer manufacturing

[00102] Power system integrity is also critically important in manufacturing. In one embodiment, Sensei’s MasterMind was deployed in transformer factory that produces massive shell-form power transformers used by generation facilities worldwide. Reliable power is essential for day-to-day operation of the factory and during the testing of the transformers produced for their customers. The factory includes a giant test hall, where newly made transformers are put through a rigorous regimen of extreme conditions. These tests include wide fluctuations in voltage, frequency, and temperature which places unusual loads on the factory’s power systems.

[00103] To ensure reliability of critical components both in their factory and the utility substation serving them, MasterMind is used to monitor thermal imaging cameras over each of the critical test hall transformers, as well as the 161 kV utility breakers with disconnect switches, connecting bus, GIS breakers, step-down transformers, and 15 kV cable terminations located in the substation yard.

[00104] Figure 6 illustrates an example set of equipment that may be monitored by a system according to embodiments of the present disclosure, such as the test hall described above. The deployment shown in Figure 6 illustrates the point that even densely placed equipment can be successfully monitored using the systems and methods disclosed herein.

Use Case : Thermonuclear protection

[00105] In another deployment, the SMART system is used to monitor the health of transformers on-site at the Tennessee Valley Authority (TVA) Watts Bar Lake Nuclear Power Plant, after a catastrophic transformer failure caused a 27-day outage with a replacement cost and lost revenue exceeding $66 million. The SMART system remotely monitors the bank of building-sized transformers on site.

[00106] Those skilled in the art will recognize improvements and modifications to the preferred embodiments of the present invention. All such improvements and modifications are considered within the scope of the concepts disclosed herein and the claims that follow.