On-site Measurements in LVDC Grids
Before starting any measurements in LVDC grids, PQ parameters were defined based on the existing scientific literature on PQ in LVDC grids as well as on standards for PQ in AC grids, including EN 50160, IEC 62749, IEEE 1159, the IEC 61000 series, and IEC 62586. For example, harmonics cannot be defined without a fundamental frequency as in DC grids; instead, a definition of voltage ripple is used in the time domain as well as in the frequency domain. Several time window lengths, frequency windows, sampling rates, and other measurement definitions were proposed, mostly in line with the existing AC PQ measurement definitions. Based on feedback from stakeholders a report was written describing proposals on how to define and measure DC PQ parameters, new definitions for electricity metering in DC systems, and proposed magnitudes and characteristics of typical LVDC voltage disturbances for DC PQ measurements and electricity meter testing.
The development of measurement equipment with specific broadband voltage and current sensors for onsite measurements was completed and satisfactorily tested in the laboratory for measurements up to 150 kHz. Two different approaches were realised, both involving data acquisition and storage with the actual analysis to be performed afterwards using the tools developed in objective 2. (1) The first approach is to include mechanisms to trigger the capture of sampled voltage and current waveform data prior to and after a DCPQ event, thus avoiding the unnecessary collection of continuous waveform data and the associated storage, transmission, and data processing issues. (2) The second approach is to sample and store all data locally on a large hard disk during a relatively short period (typically a few hours up to one day) in order to ensure that relevant issues are not missed; this method works well for configurable grids where events can be emulated, and network configurations changed easily to investigate various situations. For measurements in EV charging stations, the necessary adapters to connect the measurement equipment between the charging station and the EV have been designed and implemented.
Experiments were conducted at the Power Networks Demonstration Centre (PNDC) in Scotland to generate and measure specific test signals and disturbances. The measurement results and raw data have been used for the validation of data analysis tools developed in objective 2 and trigger mechanisms used in the first measurement approach detailed above.
Using the newly developed equipment, measurement campaigns have been performed at several LVDC grids to investigate DC-specific disturbances. Among these are:
The analysis of the data is still ongoing, but parts of the results, focusing on ripple and frequency content as well as on the triggering mechanism, have been already published in open-access peer-reviewed journals.
Analysis tools and reference systems for DC power, DC metering, and DCPQ
The initial definitions of DC PQ parameters, developed in objective 1, have been used for the specification of analysis tools. Algorithms were developed to dynamically detect DC magnitude based on RMS and average, ripple, frequency content, voltage dips and swells, and voltage transients over different measurement window lengths. These algorithms have been tested using waveforms obtained earlier in DC railway systems and using a variety of test waveforms obtained at the PNDC (objective 1) including ripple, dips, swells, and short circuits. In order to generate more test signals, modelling and computer simulations have been used based on frequency domain calculations.
Furthermore, algorithms have been developed for analysing DCPQ phenomena, as well as the development of a procedure to assess the reliability of the developed algorithms detecting PQ events. To improve the DC power and DCPQ analysis tools, the behaviour of DC mean, DC RMS, DC mode, and DC median parameters has been evaluated for different measurement window sizes, and under different sampling frequencies used to acquire the signal. The definitions of the most significant DCPQ parameters and potential DC electricity meter test waveforms will be completed once the real-world data obtained at the measurement campaigns in objective 1 have been analysed.
Three reference systems for DC power and DC electricity meter testing have been developed, based on different hardware. Two of the systems generate the high-magnitude DC current and the high-frequency AC current distortions in separate conductors, such that (for currents higher than about 10 A) only meters with current transducers sensing the magnetic field of the combination of the two conductors can be calibrated or tested. In the third setup, the DC current and AC distortions are merged into a single conductor, such that meters using other transducer types such as shunts can be calibrated as well. In addition, a fourth setup was realised based on reference voltage and current sensors, a calibrated wattmeter, and a metrological LED.
These three reference systems were successfully characterised and tested for DC power with and without distortion. The spectral composition of the distortions varies in the different setups (from purely sinusoidal components, up to saw-tooth additive components). Test measurements have been performed covering the full range of voltage (up to 1000 V), current (up to 800 A), and frequency (DC – 150 kHz), in order to prepare for a round-robin comparison.
The setups for DCPQ will be largely based on those for DC electricity meter testing with disturbances. The software for the DCPQ setups needs to be finalised and implemented, based on the latest definitions and improved algorithms, from the results of the measurement campaigns (objective 1).
Finally, a transfer standard is under development for the validation of the three reference systems. An initial test plan for a round-robin comparison has been discussed and agreed amongst the partners.
The round-robin comparison for DC power and DC metering will demonstrate and validate the different reference systems. A draft version of a transfer standard was finalized and circulated among two partners for testing. Preliminary comparison measurements have been performed in the presence of pre-defined distortions. However, the results were insufficient to validate the performance of the reference systems, and the design of a new transfer standard, based on a module with adjustable input range and customisable sensors, is in course. In addition, different sensors will also be used to maximize the resolution of the measurement. Using this new transfer standard, a new round-robin will take place with four partners involved. The test plan for this new round-robin has been completed.
Metrological framework for practical DCPQ “Compatibility level” and “Planning level” surveys in LVDC grids
Based on literature and standards, as well as on stakeholder consultation, voltage dips and swells, and voltage ripples have been identified as the main PQ parameters required for future compatibility level and planning level surveys in LVDC grids. Severity indices for these parameters have been defined, as well as proper statistical methods for data averaging and time aggregation over different periods of time.
The need for compatibility levels for steady-state disturbances below and above 9 kHz was investigated, and existing immunity levels of DC devices were verified. In addition, a first set of values of DC voltage compatibility was proposed for further discussion. Data acquisition and online FFT analysis with a fixed window length of DFT for different frequency bands of 0-9 kHz and 9-150 kHz were also considered. Relevant DFT window types have been investigated to reduce spectral leakage of aperiodic disturbances.
A first proposal for reasonably expected compatibility levels was presented and discussed at the project’s mid-term stakeholder workshop. It was determined that in order to further specify voltage compatibility levels, actual immunity levels of existing DC devices should be investigated as well. However, this is beyond the scope of this project. As a follow-up activity, a survey was set out in which stakeholders indicated the importance of specific DC PQ phenomena and the corresponding maximum acceptable disturbance levels.
A module for the computation of DCPQ indices has been implemented in a tool for DC PQ simulation and processing of on-site recorded data. For conducted disturbance measurements in the range between 9 kHz and 150 kHz, a fully digital CISPR16 method (in line with the new edition of IEC 61000-4-30) was implemented into a PQ analyser which can monitor PQ indices and record real-time waveforms for both AC and DC grids.
The implementation of the algorithms to detect and quantify PQ phenomena during on-site measurements has been finalised by adapting their characteristics for online execution and correcting imperfections highlighted in the testing phases. Testing of the algorithms by simulation has been completed by analysing different signals including flagging: pure DC, ramped DC voltage signals, DC with sinusoidal ripple, voltage dips, and composite signals. Furthermore, the algorithms were implemented in a hardware system based on a data acquisition module for testing with real signals generated in the laboratory using a programmable arbitrary waveform generator. Testing is still ongoing due to some algorithm changes required during the implementation phase.