Experiences on using gapless waveform data & synchronized harmonic phasors

Transcription

1 1 Panel Session: New Techniques for Power Quality Measurement and Field Experiences 15PESGM3040 Experiences on using gapless waveform data & synchronized harmonic phasors Wilsun Xu University of Alberta July 2015

2 2 Outline Two important developments in PQ data measurement: Gapless recording: waveform data is recorded for days without discontinuity (i.e. gaps) Synchronized waveform: two or more waveforms at different locations are synchronized using GPS

3 3 Outline This presentation is to share the experiences and applications of the above two advanced monitoring features. It also presents an issue that needs research by the Measurement and Instrumentation subcommittee: Gapless data Synchronized waveform data Research need on the definition of synchophasors

4 4 1. Gapless data Three modes of PQ data recording: Continuous waveform snapshots: For example: 12 cycle snapshot for every 1 minute. The results can be waveforms or harmonic spectrums Triggered waveform snapshots: For example: 60 cycle snapshot for a PQ disturbance The results are typically in the form of waveforms Gapless recording Continuously record waveforms for days without any stop Existing USB chips make such mode possible: e.g 64GB for 3 weeks

5 5 What unique advantages does gapless data bring? Some events of interest cannot be captured using the other two modes Failure signatures of a cable How to use trigger mode to record such signatures is not established The mode of continuous snapshots will likely miss the signatures

6 6 One major potential application Signature based equipment condition monitoring (PQ data analytics). For example: cable health monitoring Degree of Cable deterioration Sporadic arcing activities Failure Heart attack Normal condition Normal health Increased partial discharge activities Increasing blood pressure Abnormal cardiac activities * Partial discharge detector * Blood pressure monitor * Waveform analytics * ECG Time

7 7 Data analytics based condition monitoring It was very difficult to capture failure signatures in the past Failure signatures were not well understood Automatic signature capturing was therefore not possible Traditional PQ monitors have gaps in data recording As a result, the data is not adequate for off-line analysis In summary, signature analysis was not possible in the past Gapless recording makes it possible to do off-line analysis With experiences, automatic signature capture will become possible in the future. Power quality monitor may emerge also as an equipment health doctor

8 8 Data analytics based condition monitoring Method for off-line analysis of gapless data Data input Abnormality detection Abnormality extraction Signature analysis Failure characterization Analytics results Signature database Path 1 General purpose condition monitoring scheme Data input Search for specific signatures Signature analysis Failure characterization Analytics results Pattern matching algorithms Signature patterns of interest (from signature database) Path 2 Special purpose condition monitoring scheme

9 9 Data analytics based condition monitoring Waveform abnormalities detected using 15 day gapless data Detected by Current RMS Detected by Voltage RMS Detected by Current Waveform Detected by Voltage Waveform Number of disturbances Hours

10 10 2. Synchronized waveform data It is not difficult to collect synchronized waveform data using GPS these days Some commercial PQ monitors have such capabilities already The challenge is to find useful ways to take advantage of this type of data Two practical applications are given here 1. Addition/cancellation characteristics of harmonics in distribution systems 2. Condition monitoring of transmission cables

11 11 Harmonic addition investigation I Sub Feeder branch Multi-grounded primary feeder Substation Secondary feeder I ST Service transformer Secondary feeder 3f load Service drop Service drop House 1 House 2 House N How do the harmonics from service transformers contribute to those observed at the substation?

12 12 Harmonic addition investigation Synchronized waveform monitoring & analysis: Use substation current waveform as a reference Study the projection of service transformer current on the reference direction Quadrature component I ST Average angle of 3rd harmonic current in 24 hour Substation current Service transformer current In-phase component I Sub Harmonic addition

13 13 Harmonic addition investigation Addition of 3 rd harmonic Percentage(%) addition cancellation In phase component Quadrature component Time(h) 20% Average angle of 3rd harmonic current in 24 hour 15% Substation current Service transformer current Probability 10% cancellation addition % Percentage of the in-phase component (%)

14 14 Harmonic addition investigation Variation range of I ST in phase component for 95% time over 24hours Main finding: low order harmonics do add up significantly

15 15 Transmission cable condition monitoring using synchronized, gapless waveform data 240kV Network 1 2 CT location Cable to be monitored 3 130kV Network Task: determine if the cable has abnormal behaviors Challenges: 1) Since the cable is in a meshed network, both sides need to be monitored to determine if an abnormal response is originated from the cable 2) The CTs locations are not favorable

16 16 Solution: using three synchronized monitors Addition of the waveforms from the first 2 meters gives the left side current of the cable The voltage and current at two ends are then analyzed to screen for abnormal events The location of events (inside or outside the cable) are determined using differential-relay fault detection like algorithms About 15 events were captured for a 3 week period. All events are determined as outside of the cable. These disturbances could not be captured using common PQ indices

17 17 Sample results Left side Right side

18 18 3. A need for research on the concept of phasor Synchronize data over different locations GPS Synchronize data over different time How? Problem to solve: Estimate the system impedance Z eq Using natural V & I disturbances = Upstream system + _ Z eq E eq Monitor Location V I Downstream Load

19 Current (A) 19 Synchronize data over time = Window 1 Window 2A I 1 I 2 Window 2B Data point (128 points per cycle) How to position the 2 nd window to get phasors V 2 and I 2? Note: the frequency is not exactly 60Hz and is unknown

20 20 Synchronize data over time Problem 2: power-angle curve over time (from PMU) d t 1 t2 How the windows at t 1 and t 2 are synchronized so that d t1 and d t2 are related to each other in a technically valid way? Again: the frequency is not exactly 60Hz and is unknown t

21 21 Definition of frequency and angle If both frequency and angle are changing. There are two interpretations of them: Constant frequency with a changing angle Constant angle with a changing frequency Which interpretation is technically sound? What is the implication on the use of PMU results (frequencies are different at two locations and yet there is a phase angle difference)?

22 22 Summary & conclusions Gapless monitoring is a useful platform for complex offline-based signature analysis Two example applications of synchronized waveform data were shown. Harmonic phasor analysis is a interesting killer application of synchronized data There is a need to research the definitions of angle and frequency, like what has been done to the definitions of P & Q under harmonic conditions