Case Studies in PQ Monitoring

Transcription

1 1 Case Studies in PQ Monitoring Panel on PQ Monitoring in the Era of the Smart Grid 14PESGM2766 Jan Meyer Etienne Gasch Matthias Klatt Technische Universitaet Dresden, Germany

2 2 Agenda Objectives of Power Quality monitoring Case study I Troubleshooting a HF disturbance Case study II Benchmarking of MV distribution network

3 3 Major monitoring objectives and examples (1) Compliance verification Assessment of compliance with EN (2) Performance analysis Statistical survey of PQ levels in MV grids (3) Site characterisation Answering pre-connection questions of a customer regarding PQ (4) Troubleshooting Analysis of high voltage harmonics caused by resonances (5) Advanced applications and studies Analysis of prevailing current harmonic phase angle (6) Active power quality management Measurements to control voltage regulated distribution transformer

4 4 Agenda Objectives of Power Quality monitoring Case study I Troubleshooting a HF disturbance Case study II Benchmarking of MV distribution network

5 5 Network configuration and complaints Small town (7000 inhabitants), urban LV grid Sudden occurrence of customer complaints: - Coffeemaker grinds coffee, throws it away, lets out water (I) - Professional hair dryer turns itself on and off (40x / day) (II) - CNC mill control malfunction (III) Complaints limited to a small area Measurement sites: a) LV bus bar b) CNC mill POC c) Junction box

6 6 First troubleshooting measurement Measurements according to EN does not exceed any limits Measurements with instrument with extended frequency range identifies high HF emission level at 8kHz Levels exist only on working days between 6am and 10pm -> Origin of disturbance most likely an industrial customer Easy identification by interviewing -> Cause: New AFC without EMC filter at location III Spectrum during the day (time of complaints) Time-dependency

7 7 Measurements at feeder end (8 khz) A. Initial situation Without filter Levels of up to 5 V at 8 khz B. Mitigation measure 1 Generic filter Significant reduction to levels below 1 V C. Mitigation measure 2 Specific filter Further reduction to levels below 0,6V A B C

8 8 Measurements at different locations (8 khz) Mitigation measure 1 (Generic filter) Virtually no distortion level at LV busbar (a) Higher levels at feeder end (c) Highest levels at connection point of CNC mill (b) Site a b c Local phenomenon, no large propagation

9 9 Conclusions Start measurements as close as possible to source of complaints Measurements according to EN are usually not suitable to solve customer complaints The frequency range of traditional PQ monitors may not be sufficient to identify a problem Inspect data both in frequency domain and time domain Consider limitations of sensors and monitoring device (cf. IEC ) Frequency response of voltage transformers Signal dynamic of current measurement using Rogowski coils 35kV 20kV 10kV Ratio error ε in % V Frequency in khz

10 10 Agenda Objectives of Power Quality monitoring Case study I Troubleshooting a HF disturbance Case study II Benchmarking of MV distribution network

11 11 Initial situation and system concept Network operator has to replace energy meters in it s MV grid New meter type has optional PQ-functionality at low additional costs (no IEC compliant implementation) Large data amounts expected Development of index set Primary objectives: Quantification of remaining Power Quality reserve acc. to the relevant standards Flexible and intuitive reporting for multiple sites Secondary objectives: Assessment of behavior in non-covered region above 95-%-percentile Identification of long-term trends Quantification of variations of time (e.g. seasonal) for PQ assessment

12 12 Concept of quality index (QI) Each cube represents a single quality index based on weekly 95- %-percentile according to EN Time Example: Measured: 2.5 % Limit: 5 % Index: 50 % Normalized for comparison, flexible to aggregate in multiple levels, easy to understand, direct representation of minimum remaining quality reserve

13 13 System layout: Design of the monitoring system Metering infrastructure Data set for example analysis: 22 sites for 25 weeks Calculation of single quality indices acc. to EN Exluding 50 site indices (maintanance work, to less data available,...)

14 14 Top level analyses Distribution of individual indizes with benchmarked site Trends c.d.f.

15 15 Aggregation according to PQ parameter Kalenderwoche Calendar week Messort Time

16 16 Graphical presentation in stylized (fixed) schema Qualitative representation of quality reserve by color code (green: > 50%; yellow <50% and > 0%; red < 0%) Flexible aggregation for custom-defined regions Power Quality status of a MV distribution network (120 monitors)

17 17 Graphical visualization in zoomable GIS system Detail of GIS representation at different zoom-levels Lower zoom Higher zoom Visualization of minimum remaining reserve and responsible Power Quality parameter (including trend) Aggregation level depends on zoom level

18 18 Conclusions Integrating Power Quality functionality into other devices (like meters) can be an affordable alternative to dedicated PQ monitors (Take care for correct IEC implementation) Too large data amounts should be treated in distributed way (e.g. one index per week transferred to a central system Selected sites have to cover the required population (e.g. measuring only LV busbars does not represent the Power Quality in the whole LV grid) Consequent use of Web features (Top-Down-approach) together with powerful search options improves efficiency of data analysis (only PDF reports are not sufficient)

19 19 Thank you for your attention! Contact details: Jan Meyer Technische Universität Dresden Institute of Electrical Power Systems and High Voltage Engenieering Dresden tel fax