,
Abstract
The dramatic increase in distributed generation (DG), distributed energy resources (DER), and net metering scenarios brings sweeping, unplanned changes to the grid model. Local generation is often unpredictable, and sudden power flow and impedance changes produces swings in system voltage, VAR flow, and harmonic resonances that were outside the design parameters of the initial grid design. The first step in dealing with these effects is to gather data by measuring voltage, current, and power flow, along with more detailed PQ data. Guidelines for monitoring with these needs in mind are given in this whitepaper.
Recording Needs and Monitor Selection
There are two concerns that drive DG monitoring. First are regulation/delivery system values such as line voltage, load current, VAR flow, etc. Here the timescales involved have decreased from hours and days to seconds and minutes. Local generation acts to increase the system voltage by reducing the effects of local loads (the loads are powered from the DG, not the utility). If the DG output exceeds the power consumption of local loads, net power will be negative, i.e. power is injected into the distribution system. On an instantaneous basis, power flow back into the utility is achieved by the DG inverter producing a slightly higher voltage. The amount of voltage increase that results upstream depends on the size of the DG compared to the impedance of the rest of the system, and can be considerable. The reaction by voltage regulators is important to monitor closely, both to ensure that variability in DG output is handled, and to prevent feedback loops where the regular and DG inverter work against each other.
A second steady state parameter is VAR flow. Power factor correction capacitors (PFCs) are placed in strategic locations and programmed to match the inductive VARs from normal loads. Some DG inverters are designed to supply real power, but not reactive power. In any case, load flow patterns change drastically when other power sources are added to a distribution network. Optimal placement and programming of PFCs becomes much trickier in DER environments. Power flow data is needed to quantify VAR loading with and without DG in place.
A cell Boomerang monitor (for 3 phase and residential locations, shown in Figure 1) provides continuous automatic data collection for voltage, current, and power flow. Coupled with Canvass, a web-based data storage and analysis system, Boomerangs in a distribution system give a complete picture for regulation issues. In Canvass, continuous 24x7 readings on a 1 second basis are always available. This timebase has plenty of resolution for for analyzing load patterns, the effects of clouds and wind on DER sources, and voltage regulator responses.
Figure 1. Single phase (left) and 3-phase (right) Boomerangs
The second concern is power quality as seen by the customer or reflected upstream. Voltage sags, harmonic distortion, light flicker, and other PQ events may be impacted by the presence of DG. In particular, harmonics due to the DG inverter output, and new resonances formed by the shift in system impedance can introduce distortion problems that only occur during periods of generation. DG inverter switches and the increased operation of PFCs and voltage regulators can increase the frequency or severity of voltage transients on a circuit.
A cell Revolution PQ analyzer (for 3 phase applications) or cell Guardian (for residential applications) is needed for this more advanced data collection (both are shown in Figure 2). Recording harmonics, waveform capture, IEEE flicker, and single-cycle min/max values, both provide a complete PQ picture for the circuit under measurement. Both also record the voltage and power flow data that a Boomerang can provide, for steady-state analysis. The cell Revolution is accessible remotely through ProVision, while the cell Guardian works with ProVision and PQ Canvass, PMI’s web-based PQ data analysis system.
Figure 2. Cell Guardian (left) and cell Revolution (right) PQ recorders
Monitoring Point and Installation
The choice of where to place a monitor depends to some extent on the purpose for the data. Determining regulatory compliance for RMS voltage or harmonics in the presence of DG requires monitoring at a point of common coupling (PCC), as per IEEE 519. This is also a natural choice for measuring net power flow from a specific customer with local generation. Any DG-induced voltage swells are largest at the PCC, rather than further up or down the circuit. THD or resonance issues are also best quantified at the PCC.
More general concerns over distribution voltage regulation suggest monitoring at the output of the voltage regulator (traditionally the highest voltage point) and at the end of the circuit (the lowest voltage point). Readings at the regulator help track its switching operations (which may increase dramatically on days when PV or wind generation is erratic) as well as the actual voltage movement resulting from a particular tap change.
It’s not always logistically possible, but monitoring at the DG inverter output gives the most information about the DG system itself. Current at the PCC includes DG and load current together -- only the net power back to the utility flows through this point. It can be difficult to separate load changes from DG output changes after the two have been mixed. Monitoring at the DG output, electrically before loads are added, gives clean readings with no power subtracted.
In complex cases, multiple monitoring points are very helpful, especially with long-term monitoring. A Boomerang at the substation, one just past any voltage regulator or PFC in the circuit, one at a DG site, and one at the end of the circuit give a complete picture of voltage regulation and power flow. For residentials areas with PV installations throughout a circuit, one or two representative locations will often suffice to quantify the DG variability.
There are no changes to the PQ recorder installation for monitoring negative power or net metering applications. Ensure that the CT arrow points towards the load (away from the utility), and power flow will be registered as positive or negative correctly. In a single phase installation the CTs are integrated into the Boomerang and Guardian, so no further work is needed. For 3 phase Boomerangs and Revolutions, choose the correct current range, check the CT polarity, and choose a delta or wye hookup to match the circuit.
Analyzing the Data
Analyzing Boomerang data in Canvass is straightforward with just a web browser. A typical residential PV system is graphed in Figure 3. RMS voltage is the top plot, and net real power on the bottom plot. The power is negative during periods of sunlight (marked with yellow vertical lines). A local load presents a typical motor-start pattern of high start but lower steady-state power; this subtracts from the PV output. The peak PV output is around -1 kW, occurring around 1:00 pm. The RMS voltage within the generation period actually decreases, ranging from 246V to 241V. Voltage regulation here is good.
Figure 3. Residential PV voltage (top plot) and PV generation (bottom plot) in Canvass (Boomerang data)
Changes in PV or turbine generation due to cloud cover or local winds can produce significant, unexpected shifts in voltage. In Figure 4, the PV output starts low due to cloud cover around 2 pm, but these clear up by 2:20 pm. The voltage increases steadily as this happens, and drops again as clouds return about 10 minutes later. No regulator steps in to make an adjustment. Region-wide changes in DER output can happen within a few minutes, many times during the day. Regulating the voltage or optimizing PFC operation with this new source of randomness added to the system starts with good, continuous data to quantify the situation.
Figure 4. Canvass plot showing voltage moving as PV cloud cover changes
The cell Revolution and cell Guardian also provide advanced measurements such as harmonics/THD, waveform capture, and IEEE flicker. In Figure 5, RMS voltage is graphed in conjunction with voltage THD. As voltage rises due to DG contributions, the overall THD becomes increasingly influenced by the specifics of the DG inverter(s). Here the THD rises as the voltage rises, showing that the injected power is adding to the voltage distortion.
Figure 5. ProVision PQ plot showing voltage THD changing as the DG voltage changes
Detailed information is presented with the waveform graphs. In Figure 6 show the waveform capture from a 5 MW 3-phase PV inverter start-up. The inverter output is very nonlinear and unbalanced as it ramps up. Fortunately this distortion doesn’t last long; the steady-state waveforms are shown in Figure 7. A bit of “fuzz” is noticeable on the current waveforms, caused by the process of synthetic sine wave creation with the inverter. A harmonic analysis of this waveform is shown in Figure 8 (fundamental removed). The “fuzz” is mostly components near the 35th harmonic, or 2100 Hz (circled in orange). This is likely an inverter switching frequency. Overall current THD is under 4%, so unless there is a specific resonance affected by the 35th harmonic, this distortion may be acceptable.
Figure 6. 5 MW photovoltaic startup current (blue)
Figure 7. 5 MW photovoltaic steady state waveforms, with voltage (red) and current (blue)
Figure 8. Harmonic breakdown of 5 MW PV inverter, with components at the 35th harmonic (orange)
Conclusion
The rapid rollout of photovoltaic, wind turbine, and other distributed generation/distributed energy resources complicates the distribution network and adds new sources of unpredictability. To perform the engineering required to accommodate these additions, much more detailed measurements on system behavior is required. The basics needed using the cell Boomerang, Revolution, and Guardian are described here.
Chris Mullins
VP of Engineering and Operations
cmullins@powermonitors.com
http://www.powermonitors.com
(800) 296-4120