IEEE TRANSACTIONS ON SMART GRID, VOL. 8, NO. 2,
IEEE TRANSACTIONS ON SMART GRID, VOL. 8, NO. 2, MARCH 2017749A Microgrid Monitoring SystemOver Mobile PlatformsHaoyang Lu, Student Member, IEEE, Lingwei Zhan, Member, IEEE,Yilu Liu, Fellow, IEEE, and Wei Gao, Member, IEEEAbstract—Real-time awareness of the phasor state, includingthe volatile frequency and phase angle, is critical to maintainreliable and stable operations of the power grid. However, thehigh cost and low accessibility of current synchrophasors restricttheir large-scale deployment over highly distributed microgrids.In this paper, we present a practical system design for monitoringthe microgrid frequency and phase angle over mobile platformsand significantly reduce the cost of such monitoring. Being different from current synchrophasors, our system does not relyon continuous GPS reception and hence it is highly accessibleand applicable to heterogeneous microgrid scenarios. We developvarious techniques to provide the timing signal that is necessaryfor precise microgrid monitoring. For frequency monitoring, thenetwork time protocol is exploited for time synchronization. Forphase angle monitoring which requires a higher timing accuracy,200 Hz primary synchronization signal being embedded in the4G LTE cellular signal is harvested for time synchronization. Weimplemented our system over off-the-shelf smartphones with afew peripheral hardware components and realized an accuracyof 1.7 MHz and 0.01 rad for frequency and phase angle monitoring, respectively. Although the accuracy of the prototype islower than that of the GPS-based systems, the system could stillsatisfy the requirements of microgrid monitoring. The total costof the system can be controlled within 100 and no installationcost is required. Experiment results compared with the traditional frequency disturbance recorders verify the effectiveness ofour proposed system.Index Terms—Smart grid, microgrid monitoring, network timeprotocol (NTP), primary synchronization signal (PSS), mobileplatform.I. I NTRODUCTIONICROGRIDS, operating in either grid-connected modeor islanded mode, enable local integration of energygeneration, distribution, and storage at the consumer level forbetter power system efficiency and control of demand [1]. Tomaintain the stable operation of microgrids, phasor states ofboth microgrid and the Area Electric Power System (AEPS)to which the microgrid connected, including frequency, phaseangle, and voltage magnitude, should be monitored continuously, especially during the transition between two operationmodes, i.e., resynchronization and islanding process [2].MManuscript received February 5, 2015; revised June 8, 2015 andOctober 5, 2015; accepted December 4, 2015. Date of publicationJanuary 6, 2016; date of current version February 16, 2017. Paper no. TSG00129-2015.The authors are with the Department of Electrical Engineering andComputer Science, University of Tennessee, Knoxville, TN 37996 USA(e-mail: hlu9@utk.edu; lzhan@utk.edu; liu@utk.edu; weigao@utk.edu).Color versions of one or more of the figures in this paper are availableonline at http://ieeexplore.ieee.org.Digital Object Identifier 10.1109/TSG.2015.2510974Current power grid monitoring systems allow direct measurement of frequency and phase angle by installing synchrophasors at either high-voltage transmission level [3] orlow-voltage distribution level [4]. These power grid monitoring systems, although having been proved to be effective inwide-area power grid infrastructure, are generally consideredunsatisfactory for monitoring the operating status of the newlyemerging distributed power systems, so-called microgrids [5].The decentralization of microgrids poses higher requirementson the installation cost and accessibility of power monitoringdevices, and makes the current synchrophasors too expensiveand inconvenient to be deployed into individual households inhigh volume. PMUs are deployed in substations and equippedwith current transformer and power transformer for accessing the high voltage, which increases both manufacturing costand the installation cost. For example, the installation costof one transmission-level Phasor Measurement Unit (PMU) ismore than 80,000 at the Tennessee Valley Authority (TVA).These PMUs are not intended to be used at the distributedconsumer level, and require professional installation whichreduces end-users’ incentives of having synchrophasors intheir home energy systems.In this paper, we present a practical system design whichbridges the gap between current power grid monitoring systems and the unique requirements of microgrid monitoring.Development of such a microgrid monitoring system, however,is challenging due to the requirements of microgrid monitoring on accurate time synchronization, high-resolution sensing,and real-time data processing. First, power grid operationsshould be monitored in real-time using globally synchronizedtimestamps, so that measurements from dispersed locationscan be compared on a common time reference [6]. AlthoughGPS signal that is widely used in current synchrophasors canprovide a sub-microsecond timing accuracy, however, it haslimited use for microgrids due to its deficiency for indoorscenarios. Second, with the increasing resolution and responsiveness of phasor state estimation, the workload of processingmeasurement data may exceed the computational capacity ofcurrent measurement devices. Specialized DSP chips are usedin present synchrophasors for data processing, but are difficultand too expensive to be integrated into microgrid monitoringsystems which need to be deployed in high volume.Our main idea to overcome the aforementioned challengesis to reduce the systematic cost of frequency and phase anglemonitoring by developing efficient embedded sensing platforms and adopting GPS-free time synchronization methods,c 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.1949-3053 See http://www.ieee.org/publications standards/publications/rights/index.html for more information.
750IEEE TRANSACTIONS ON SMART GRID, VOL. 8, NO. 2, MARCH 2017without impairing the accuracy of AC waveform measurementand phasor state estimation. To the best of our knowledge,we are the first to achieve high-precision wide-area powergrid monitoring without GPS synchronization. Our detailedcontributions are as follows: We implemented a low-cost microgrid monitoring systemover the smartphone platform with a small quantity ofperipheral hardware components. This prototype systemcan be further integrated into the smartphone charger forbetter flexibility and convenience. We proposed GPS-free frequency monitoring methodsutilizing the Network Time Protocol (NTP), which significantly increases the system flexibility by eliminatingthe requirement of GPS reception and line of sight to thesatellites. We extended the frequency monitoring prototype tofurther implement the functionality of phase anglemonitoring by harvesting the Primary SynchronizationSignal (PSS) in the 4G LTE cellular signal for timesynchronization. We proposed adaptive frequency estimation algorithmto reduce the computational load of data processing atsmartphones.The rest of this paper is organized as follows. Section IIdiscusses the background knowledge of power grid monitoringand the related work. Section III provides an overview of oursystem design. Section IV introduces the adaptive frequencyestimation algorithm and the NTP-based frequency monitoringdesign. Section V presents the LTE-based phase angle monitoring design. Section VI presents our performance evaluationresults in comparison with FDRs. Section VII presents the discussion. Finally, the conclusions and future work are drawn inSection VIII.Fig. 1.Synchronization signals in LTE FDD downlink.GPS receivers is generally used as the synchronization signal [15]. The PPS signal is an analog output signal with arising edge at each one second boundary of the UniversalCoordinated Time (UTC). Since the precision of GPS signalis in nanoseconds with non-accumulative time drift [16], synchrophasors at dispersed locations can measure local systemstate synchronously through the integrated GPS clock.For better accessibility and lower cost, less expensive yetaccurate timing sources have been studied in recent years.A NTP-synchronized Wide Area Frequency MeasurementSystem (WAFMeS) has been implemented [17] and is ableto detect the large swing disturbance at the granularity of0.2 Hz. However, its flexibility and accessibility is restricted,and its accuracy of frequency estimation is too low to beapplied to the U.S. power grid with much higher stabilityand smaller disturbances. Extensive researches have also beendone on harvesting the timing information from the GSM cellular communication system [18]. More specifically, the 21 HzFrequency Correction Burst (FCB), which is modulated at afrequency higher than the main carrier frequency, is harvestedfrom the main carrier through development of 0.13μm CMOSfabrication techniques. Since the GSM base stations are strictlysynchronized to a single timing source with an accuracy betterthan 0.05 microseconds [19], GSM signal is able to achievean equivalent timing accuracy as the GPS signal.II. BACKGROUND AND R ELATED W ORKIII. OVERVIEWThe power grid frequency and phase angle are generallymeasured from samples of the voltage waveform. Severalalgorithms such as curve-fitting method [7], [8], Kalman filterbased methods [9], and DFT-based algorithms [10], [11] havebeen proposed for such measurement.Built on these measurement algorithms, wide-area powergrid monitoring systems have been established via deployment of PMUs in the transmission level. Synchrophasors suchas Frequency Disturbance Recorders (FDRs) [12] and microPMUs (μPMU) [13], on the other hand, are deployed inthe distribution level with a greatly reduced monitoring cost.Specifically, based on FDRs, a worldwide Frequency monitoring NETwork (FNET) has been designed and deployed [14],enabling many applications of power system monitoring, control, and management, such as abnormal events detection andlocation.Synchronization among different synchrophasors is important for wide-area power system monitoring. To achievethis, the Pulse-Per-Second (PPS)1 signal being retrieved fromIn this section, we first discuss our motivation of exploitingGPS-free time synchronization methods for microgrid monitoring. Specifically, we adopt Network Time Protocol (NTP)for frequency monitoring, and LTE synchronization signals forphase angle monitoring. Afterwards, we present the hardwareand software designs of our power grid monitoring system.1 In the rest of the paper, we will use the terms “GPS signal” and“PPS signal” interchangeably.A. Motivation1) Motivation for NTP-Based Frequency Monitoring:Network Time Protocol (NTP) [20] is being widely usedin current computing systems, such as the Windows TimeService, in order to synchronize the local clock of digital devices with UTC. Due to the uncertainty of networktransmission delay, the timing accuracy of NTP is in theorder of 10 milliseconds [21] and is much lower than thatof the GPS signal. Nevertheless, by investigating the sample events in the power grid [22] that are recorded by FNET(see Section VII-A), we found that such time precision is sufficient for detecting a frequency disturbance event. Therefore,the NTP is an appropriate alternative to provide global timesynchronization to frequency measurement data, replacing the
LU et al.: MICROGRID MONITORING SYSTEM OVER MOBILE PLATFORMSFig. 2.751Power grid monitoring over smartphones.GPS signal. Being different from the GPS signal which continuously feeds the processor, the NTP timing information isonly available when the monitoring device sends a requestto the remote NTP server. The uncertainty of such round-tripnetwork transmission delay, therefore, further complicates thedesign of time synchronization method.2) Motivation for LTE-Based Phase Angle Monitoring:Compared to frequency monitoring, phase angle monitoringrequires a globally synchronized clock with higher accuracyand stability. Simply speaking, a 15-millisecond timing error,which is usually the upper bound of NTP timing error, corresponds to an unacceptable phase angle measurement error of5.76 radians in a 60 Hz power system. Instead, we proposeto harvest the precise timing signal from the 4G LTE cellularsignal, which is widely available nationwide nowadays. Theenhanced base station (eNodeB) of LTE is strictly synchronized with GPS or the Precision Time Protocol (PTP) [23]. Thecell ID in the LTE network is defined within two synchronization signals, namely Primary Synchronization Signal (PSS)and the Secondary Synchronization Signal (SSS). Fig. 1illustrates the LTE frame format and the location of synchronization signals under Frequency-Division Duplexing (FDD)mode. The PSS repeats periodically (every 5 ms) and thereforecan be regarded as a time synchronization signal.B. Hardware DesignOur monitoring system consists of a voltage regulatormodule, a voltage transform circuit, a microprocessor-basedanalog-to-digital (AD) sampling module and an Android-basedsmartphone. The system design and implementation are shownin Fig. 2(a) and 2(b), respectively. The voltage regulator outputs the necessary DC power to power up the whole system,including the smartphone. An 8-bit microprocessor (MCU)ATmega328 (Arduino Uno board) is used to control the voltagesampling process through external AD Converter (ADC) at thesampling frequency of 1,440 Hz, and sends these raw voltagedata to smartphone every 100 ms for phasor state estimationprocessing. The communication between the microprocessorand the smartphone is conducted by the USB host controllerIC MAX3421E (USB host shield) [24]. Similar as being connected to the desktop PC, the smartphone behaves as USBslave in relation to the USB host chip, and can communicatewith the MCU and be charged at the meantime.The PSS harvesting circuit, shown within the dotted linein Fig. 2(a), will be attached to the frequency monitoringsystem for phase angle monitoring. In the LTE-based phaseangle monitoring, the PSS harvesting circuit will extract thePSS signals and transmit them to the MCU in the form ofpulses. The rising edges of the pulses will be detected throughExternal Interrupt (EI) in the MCU, and trigger new samplingcycles. The PSS harvesting circuit will be illustrated in detailin Section V.C. Software DesignAs shown in Fig. 2(c), we developed an Android application to process the measurement data at the smartphone andvisually display the monitoring results to users. To ensureprompt processing of measurement data, each computationally intensive operation, including the phasor state estimationand NTP request and response, is processed by a separate software thread instead of residing in the application main thread.Our application consists of four major components:1) USB Communication Component: Through USB connection, the MCU-based AD sampling module uploadsthe waveform samples onto the smartphone for processing. The smartphone will respond with RESENDcommand if error exists during the data transfer. In theNTP-based frequency monitoring, a new sampling cyclestarts once MCU receives trigger from the smartphone.Our application will also monitor the connection statusof the USB accessory, and run automatically when acorrect hardware signature is connected.2) Internet Access Component: Network access is necessary to obtain the NTP timing information. The UTCtimestamp is retrieved by requesting the NTP server. Inaddition, the measurement information will be uploadedvia the Internet to FNET servers hosted in the Universityof Tennessee.3) Phasor State Estimation Component: The phasorbased frequency estimation algorithm described inSection IV-A is implemented in this component. TheMCU will store the sampled data within the last 100 msand send them to the smartphone at one time. Oncethese 144 samples are received by the smartphone, ourapplication would estimate the operating frequency fromthese samples. In addition, the phase angle of the firstsample will be selected as the phase angle output in thisperiod.4) Graphic Interface Component: The information of frequency, phase angle and voltage amplitude of the AC
752Fig. 3.IEEE TRANSACTIONS ON SMART GRID, VOL. 8, NO. 2, MARCH 2017Frequency estimation algorithms.waveform are displayed at the screen for better userinteraction.IV. F REQUENCY M ONITORINGCompared to the widely-used DSP chips in current synchrophasors, mobile platforms such as smartphones are notspecialized for, and may be overloaded by real-time computation over voltage waveform samples, especially underthe condition of heavy workload. In addition, more powerfulsmartphones usually mean higher cost, which is contradictoryto our motivation of low-cost monitoring. Therefore, the reduction of computational workload is highly important, especiallyunder high sampling frequency. In this section, the adaptivefrequency estimation algorithm is developed to reduce thecomputation load at smartphone. Afterwards, the operating frequency in the microgrid will be estimated based on the timinginformation from NTP.A. Adaptive Frequency Estimation AlgorithmDFT-based phasor and frequency estimation algorithm iswidely applied in phasor measurement area. However, the traditional DFT-based algorithm requires the sampling rate tobe integer multiples of the power grid frequency, which isnot satisfied when the power grid frequency deviates fromits nominal frequency at a fixed sampling rate. The nominalfrequency in North America is 60 Hz, and a sampling rateof 1,440 Hz is used in FDRs. This violation, known as “spectrum leakage”, will introduce estimation errors. To address thisproblem, “re-sampling” is proposed for FDRs [11]. The basicidea is to reconstruct a series of samples from the originalvoltage samples, so that the resampling rate of the new samples is close to integer multiples of the power grid frequency.The flowchart of the FDR algorithm is shown in Fig. 3(a).The power grid frequency f1 is first coarsely estimated by theDFT-based approach. Then the “re-sampling” module takesthe original voltage samples and regenerates a new series ofsamples with the help of f1 . Afterwards, a fine-level frequencyestimation is conducted using the reconstructed samples, theresult of which ( f2 ) is used as the final frequency result.As depicted in Fig. 3(a), the frequency is estimated twicefor each computation window, which is 0.1 s in currentdesign. Our basic idea to reduce such computation workload is to exploit the correlation of voltage waveformsbetween adjacent computation windows, so as to eliminateFig. 4.Frequency measurement error under 60 dB AWGN.the redundancy in frequency estimation. To be compatiblewith the FNET network, the reporting rate of our prototype is set to be 10 Hz. Since the frequency of the U.S.power grid usually changes slowly, it is highly possible thatthe frequency between two consecutive computation windowsvaries little. Under such circumstances, the frequency calculated in the current computation window can be used tore-sample the data in the following computation window, sothat there will always be 24 samples per power grid frequencyperiod.The modified adaptive algorithm is shown in Fig. 3(b),in which the frequency result in one sampling cycle will beregarded as the initial frequency of the next computation window. Hence, frequency estimation only needs to be executedonce based on re-sampled data. The final frequency will becalculated as:(n 1)f2(n) f2 δf (n 1)(1)(n)f2is the frequency result in the n-th computation winwheredow, which will serve as the initial frequency in the (n 1)-thcomputation window, and δf (n 1) is the correction frequency.It is expected that the closer between nominal frequencyand actual frequency, the more accurate frequency estimationresult can be achieved. Therefore, the frequency measurementerror is modeled as error φ(f0 , f ), in which f0 is the actualfrequency and f is defined as the difference between nominalfrequency and actual frequency.According to [25], the ramp rates of frequency is between 1.0 Hz/s to 1.0 Hz/s, which corresponds to 0.1Hz withregard to the reporting rate of every 0.1s. We applied different nominal frequencies onto the waveform with fixedfrequency contaminated 60 dB AWGN. The error of theadaptive frequency estimation method in relation to the difference between nominal frequency fnominal and the actualpower grid frequency f0 is depicted in Fig. 4. Error of theadaptive method is less than 0.04 mHz when the change ofthe frequency in consecutive cycles is 20 mHz, and it willdecrease to smaller than 0.013 mHz when the change is lessthan 5 mHz.B. Local Clock Calibration and Timestamp CalculationWe further exploit the timing information from NTP to calibrate the local clock and calculate the local timestamps, soas to ensure precise frequency estimation. The local clock
LU et al.: MICROGRID MONITORING SYSTEM OVER MOBILE PLATFORMS753Fig. 6.Fig. 5.Timestamps and time synchronization strategy.of a smartphone will continuously drift due to the dynamiccharacteristics of the crystal oscillator, as well as various environmental factors such as temperature. As a result, the actualtime period for each AC waveform sampling cycle may notbe accurately set as expected. For example, a time period setto be 2000 ms by the smartphone may be actually 1998 msor 2001 ms due to the local clock drift. Such inaccurate sampling frequency will result in the residue problem [26], i.e., theposition of the first sample in one sampling cycle is different from that in another cycle, and the residue could beaccumulated over time. To address this residue problem, oursystem starts a new sampling cycle every time when havingsent out a NTP request, and hence guarantees the position ofthe first sample in each sampling cycle are the same in time.Correspondingly, the length of one sampling cycle is set to be2 seconds, which is as twice as the period of the GPS signal.More frequent NTP requests than once every 4 seconds willbe considered as attempting a Denial-of-Service (DoS) attackand hence denied by the NTP server [27]. Therefore, to avoidfailures of NTP queries, the system will alternate the NTPservers to be requested so that one NTP server be requestedfrom the system once within 10 seconds.Each time when the smartphone triggers a sampling cycle,it calibrates its local clock using the received NTP timing information through comparison between the returnedNTP timestamp Tntp and the corresponding local time tl . If Tntp tl , which is the upper bound of NTP timingerror, we will calibrate the local timer triggering the samplingcycle, i.e., using a new number rather than 2,000 millisecondsto setup the local timer for triggering the sampling cycles.More specifically, assuming that the local time of the mostrecent calibration in the past is tl and its corresponding NTP , the new setup time length V for the local timertime is Tntp(in milliseconds) is defined as follows: 2000·tT t Tlntpntp , if Tntp tl l(2)V V,Otherwisewhere V is the current setup value of the timer.At the meantime, the local system time is also updated. Forexample, in Fig. 5, once receiving the NTP timing information t , we change the local time fromand find that Tntpl tl to Tntp .Transmission of PSS signal in the frequency domain.With NTP timing information, we recursively compute thetimestamp of the current sampling cycle via proportional estimation from the previous cycle. We assume that the local clockdrift within one sampling cycle is negligible. As a result, thetimestamp of the first sample in one sampling cycle, whichis also the start of the current sampling cycle, as well asthe end of the previous sampling cycle, can be estimatedfrom the NTP response and the corresponding local time. Forexample, in Fig. 5, the NTP time corresponding to tl2 can beestimated as: t3 tl2 3232(3) Tntp l3·T TTntpntpntp .tl tl1V. P HASE A NGLE M ONITORINGAs illustrated in Section III-A, measurement of phase anglerequires more accurate timing information than the frequencymonitoring does. In our system design, we aim to harvest synchronization signals from 4G LTE cellular signal for timesynchronization as the substitution of GPS signal. Similarto the GPS-based system, the harvested LTE signal can bedirectly used to trigger a new sampling cycle.A. PSS Harvesting Circuit DesignIn LTE networks, to achieve high data transmission rate,Orthogonal Frequency Division Multiple Access (OFDMA) isutilized as the physical layer technique in the downlink datatransmission [28]. The cell ID is represented as:Cell12 3NID NIDNID(4)1 0, 1, ., 167 is the cell identity group and iswhere NID2 0, 1, 2 is the cell identitylocated in SSS signal, and NIDand is located in PSS signal. The SSS signal indicates theframe timing as they are different within a frame, while PSSsignal indicates the OFDM symbol timing as they are the samewithin a frame. The sequence used for the PSS is generatedfrom a frequency domain Zadoff-Chu (ZC) sequence [29]: π un(n 1) e j 63n 0, 1, ., 30(5)cu (n) j π u(n 1)(n 2)63en 31, 32, ., 61where the ZC root u for the PSS sequence is 25,29,34 that2 0,1,2, respectively.corresponds to the value of NIDAs shown in Fig. 6, comprised with three Zadoff-Chusequences in frequency domain, the PSS signal maps to thecentral 62 subcarriers around DC (within the central sixResource Blocks (RBs)), enabling the frequency mapping of
754Fig. 7.IEEE TRANSACTIONS ON SMART GRID, VOL. 8, NO. 2, MARCH 2017System block diagram of the PSS signal harvesting circuit.Fig. 9.Fig. 8.System coordination between the smartphone and MCU.the synchronization signals to be invariant with respect to thesystem bandwidth, which varies from 1.4 MHz to 20 MHz.Although PSS signal capturing is intrinsic to the LTE hardware within smartphones, the PSS detection is hidden fromthe IC chips and is inaccessible to the smartphone users,since the users are more concerned about the content of theradio frames, rather than when to receive the synchronizationframes. Therefore we need to build our own PSS harvestingmodule. The frequency of PSS signal (200 Hz) is far lowerthan the bandwidth of data transmitted (in the order of 1 MHz).Since our purpose of PSS detection is not to decode the signal but only to identify the arrival of the PSS signals, the PSSsignal can be detected based on the scheme shown in Fig. 7.A Voltage-Controlled Oscillator (VCO) is used to detect thefrequency band with the strongest signal strength. The signal in 1900 MHz frequency band is selected and reduced to200 kHz intermediate frequency (IF) output. The PSS signalwould be transformed as a pulse after passing the bandpassfilter with a bandwidth of 120 kHz and the envelope detector.The MCU will capture the rising edge of the PSS pulses asthe trigger to start a new sampling cycle.Fig. 10.Experiment setup.Frequency monitoring.signal tentatively, and the interval between the pulse and thelast effective pulse is calculated. If the interval between themsatisfies t ε, then the pulse is considered as an effective PSS output. The detection window will be closed and anew sampling cycle starts immediately. Otherwise, the pulseis assumed to be noise and if no effective pulses are detectedin the detection window, the MCU will start a new samplingcycle at the time when sampling the 144th samples. The phaseangle of the first sample in each sampling period is obtainedas the effective output in its corresponding period.VI. P ERFORMANCE E VALUATIONTo evaluate the accuracy and effectiveness of frequency andphase angle monitoring of our proposed system, we test oursystem against the traditional FDR device. The system setupis shown in Fig. 9. The Doble F6150 power system simulator with GPS satellite synchronization is used to generate thestandard AC power. The frequency and the phase angle accuracy of Doble F6150 simulator are 0.5 Part-Per-Million (PPM)and 0.1 degree respectively. Experiments over both standardpower generator and regular AC wall power are conducted.B. Coordination Between Smartphone and MCUOur proposed method of determining the starting time ofa new sampling cycle is shown in Fig. 8. Our monitoringsystem starts a new sampling cycle on every 100 ms for consistency with the reporting rate of the FNET network, and use1,440 Hz as the sampling frequency to ensure the measurement accuracy (12 or more data samples per period is neededin a 60 Hz power system). Correspondingly, 144 data samplesare recorded in each sampling cycle. To avoid false detection, we enable the PSS pulse detection only after the 142-ndsamples (98.6 ms) is sampled in current sampling cycle. Thepulse received in this period cycle is assumed to be the PSSA. Frequency and Phase Angle MonitoringThe frequency measurement results over standard 60 Hzpower generator and 120 V AC wall outlet are shown inFig. 10. The error of the frequency measurement of our system under constant frequency waveform is 1.70 mHz, whilethat of the FDR device is 0.32 mHz. Under wall outlet measurement, our system is able to efficiently capture the trendsof frequency deviations over time. Meanwhile, our systemproduces a 1.7 mHz difference from the FDR measurementsbeing used as the reference. Currently, the frequency monitoring accuracy of our system is less than that of FDR due to
LU et al.: MICROGRID MONITORING SYSTEM OVER MOBILE PLATFORMSFig. 11.Phase angle monitoring.a couple of reasons. Firstly, timing error exists in the timestamps compared to the real UTC time. That is, the time pointof a frequency measurement may not be exactly aligned withits real timestamp. Secondly, to accommodate the data formatof FNET and being able to integrate with NET framework, acurve fitting method is used to estimate the frequency pointsonly at integral 100 millisecond points
IEEE TRANSACTIONS ON SMART GRID, VOL. 8, NO. 2, MARCH 2017 749 A Microgrid Monitoring System Over Mobile Platforms Haoyang Lu, Student Member, IEEE, Lingwei Zhan, Member, IEEE, Yilu Liu, Fellow, IEEE,andWeiGao,Member, IEEE Abstract—Real-time awareness of the phasor state, including the vo
IEEE 3 Park Avenue New York, NY 10016-5997 USA 28 December 2012 IEEE Power and Energy Society IEEE Std 81 -2012 (Revision of IEEE Std 81-1983) Authorized licensed use limited to: Australian National University. Downloaded on July 27,2018 at 14:57:43 UTC from IEEE Xplore. Restrictions apply.File Size: 2MBPage Count: 86Explore furtherIEEE 81-2012 - IEEE Guide for Measuring Earth Resistivity .standards.ieee.org81-2012 - IEEE Guide for Measuring Earth Resistivity .ieeexplore.ieee.orgAn Overview Of The IEEE Standard 81 Fall-Of-Potential .www.agiusa.com(PDF) IEEE Std 80-2000 IEEE Guide for Safety in AC .www.academia.eduTesting and Evaluation of Grounding . - IEEE Web Hostingwww.ewh.ieee.orgRecommended to you b
Signal Processing, IEEE Transactions on IEEE Trans. Signal Process. IEEE Trans. Acoust., Speech, Signal Process.*(1975-1990) IEEE Trans. Audio Electroacoust.* (until 1974) Smart Grid, IEEE Transactions on IEEE Trans. Smart Grid Software Engineering, IEEE Transactions on IEEE Trans. Softw. Eng.
Citation information: DOI 10.1109/TSG.2017.2664043, IEEE Transactions on Smart Grid IEEE TRANSACTIONS ON SMART GRID, VOL. XX, NO. XX, MONTH YEAR 1 Spoofing-Jamming Attack Strategy Using Optimal Power Distributions in Wireless Smart Grid Networks Keke Gai, Student Member, IEEE, Meikang Qiu, Me
IEEE TRANSACTIONS ON SMART GRID 1 Preventing Occupancy Detection From Smart Meters Dong Chen, Student Member, IEEE, Sandeep Kalra, Student Member, IEEE, David Irwin, Member, IEEE, Prashant Shenoy, Fellow, IEEE, and Jeannie Albrecht, Member, IEEE Abstract—Utilities
686 IEEE TRANSACTIONS ON SMART GRID, VOL. 2, NO. 4, DECEMBER 2011 Multicast Authentication in the Smart Grid With One-Time Signature Qinghua Li, Student Member, IEEE, and Guohong Cao, Fellow, IEEE Abstract—Multicast has been envisioned to be useful in many smart grid
systems, microgrids, smart grid, storage. Abstract The IEEE American National Standards smart grid publications and standards development projects IEEE 2030, which addresses smart grid interoperability , and IEEE 1547 TM, whichaddresses distributed resources interconnection with the gridhave made substantial , progress since 2009 [1]. The IEEE 2030
446 IEEE TRANSACTIONS ON SMART GRID, VOL. 4, NO. 1, MARCH 2013 An Information-Theoretic Approach to PMU Placement in Electric Power Systems Qiao Li, Student Member, IEEE,TaoCui, Student Member, IEEE,YangWeng, Student Member, IEEE, Rohit Negi, Member, IEEE, Franz Franchetti, Member, IEEE, and Marija D. Ilić, Fellow, IE
IEEE TRANSACTIONS ON SMART GRID, VOL. 7, NO. 3, MAY 2016 1683 Provision of Regulation Service by Smart Buildings Enes Bilgin, Member, IEEE, Michael C. Caramanis, Senior Member, IEEE, Ioannis Ch. Paschalidis, Fellow, IEEE, and Christos G. Cassandras, Fellow, IEEE Abstract—Regulation service (RS) reserves, a critical type of bi-direc
