Flicker Trouble Ahead for Solar PV Inverters?

(Updated March 7, 2013 with additional text shown in red.)

The seemingly innocuous flickering of lamps could be a new technical battleground for the further growth and spread of photovoltaic (“PV”) electric power. On one side of the impending conflict is the flicker standard, a venerable reference that could very well trace its roots back to the advent of the electric age. On the other side are the new darlings of the power industry — environment-friendly, renewable solar power. The one thing about solar power is that in bulk amounts, its units need to be connected to existing electrical systems, and a side effect of this integration is the production of flicker. The more PV devices connected to the same electrical circuit, the more flicker is produced and the closer the level of flicker is to the allowable limit defined by the flicker standard.

There are some new technical concepts in the upcoming argument, so it may be good to establish the definitions and references before we go any further. For those already well-versed in this topic, you are welcome to skip to the next section.

What is flicker and what is the flicker standard?

Flicker, at its most practical essence, is the fluctuation of lights due to fluctuations in the electric supply. In the US, the basic fluctuation comes from the 60 Hertz (fluctuations per second) alternating current available to residential customers; elsewhere in the world, this may be 50 Hz. But most people will tolerate fluctuation at these fundamental frequencies. It is the fluctuation that comes from impurities in the power supply that has the potential to be bothersome. These impurities, or so called power quality issues, may come from devices connected to the same electrical circuit and actions that system operators may undertake upstream of the household supply.

Although human eye and psychological perception may vary across a broad range, research (and about a century’s experience with the incandescent light bulb) has defined the “typical” eye. This “eye” has a level of flicker which it can perceive and a presumed level when the beholder starts to become irritated. Thus, we have the border line  levels of visibility and irritation, a shown in the flicker curve.

The flicker curve has changed much over the years but its basic elements are: the frequency content and the magnitude. The typical frequency range of perceptible flicker is from 0.5 Hz to 30.0 Hz for magnitudes starting at less than 1.0% change in voltage on 120 volt supplies.  The most sensitive frequency range for voltage flicker is approximately 5-10 Hz, towards the right side of the curve shown above. This means that the human eye is more susceptible to voltage fluctuations in this range. As the frequency of flicker increases or decreases away from this range, the human eye generally becomes more tolerant to the fluctuations. A limitation of the flicker curve is that it requires constant frequency and magnitude flicker in order to get a measurement on the curve, when in reality these parameters may be changing due to multiple sources of flicker overlapping each other or the flicker source waning and waxing over time.

One consequence of the flicker curve is that it has standardized the acceptable levels of flicker, notwithstanding the fact that actual flicker may be perceived differently by individuals, or one may say, despite that one person’s headache-inducing disco lights can be another’s somnolent luminance.

Not only has the flicker curve standardized the human perception range but it also has standardized the light source, the venerable incandescent light bulb, also referred to in this context as a “standard” light bulb. Of course, the available lighting technology today offers more options to the incandescent light bulb. (In fact there is the  New Light Bulb Law  or the Energy Independence and Security Act of 2007 that places energy requirements on lighting that may push incandescent bulbs out with more efficient sources such an fluorescent lamps.)   You have alternative choices such as fluorescent, halogen and LED devices. However, the incandescent light bulb remains the basis for flicker standards, although with the growing use of CFLs (compact fluorescent lamps), some accommodation has been made in the standard to account for fluorescent lamps.  Fluorescent lamps will accept two or three times as much voltage fluctuation as filament/incandescent lamps to produce the same amount of visible flicker.  Halogen and LED technology are newer and less prevalent so are not yet considered to be sufficiently in widespread use as to deserve their own recognition in the flicker standard.  Another limitation, the flicker curve is based on square waves. It is important to note that the change in magnitude of voltage in the form of a sine wave produces half the flicker effect as that of the rectangular step.

To address one question that may be in mind at this point, flicker can also affect equipment, albeit under rare circumstances, including electric drives and universal power supply systems.

In present practice, flicker is no longer measured against the flicker curve, but rather with so-called flicker meters. These meters take the actual fluctuating voltage, filtering the signal to exclude components that are not perceptible and providing a parameter-less value known as the instantaneous flicker. Recently, both the US and European standards organizations (IEEE and IEC, respectively) agreed to converge the flicker standard into two values: the short-term (or ten-minute) flicker, Pst, and the long-term (or 120-minute) flicker, Plt.

Flicker from Solar Arrays

Now enter the sandman in the form of electricity derived from solar power. Several characteristics of this energy resource that lead to flicker are: (1) using photovoltaic panels, where electricity is generated in direct-current or DC form, (2) to connect this power in parallel to household supplies requires conversion of the DC to alternating-current, or AC, using an electronic device known as an inverter, and (3) as the solar energy received by the panels varies during the day, optimization algorithms determine the maximum power that can be produced from instant to instant. All of these contrive to produce an AC signal that includes many fluctuations and variations on the fundamental frequency.

But one inverter producing power from solar panels will not normally produce enough flicker to even reach the level of perception. Several inverters connected to the electric feeder will increase flicker but it will take an ordinal multiple of the load on the feeder before this becomes a flicker issue. The actual behind-the-scenes culprit or perhaps more accurately, above the arrays, is cloud cover.

Clouds Over My Picnic

Clouds will continually move between the sun and solar panels, alternately reducing solar input to zero by coverage and back to previous levels through uncovering. As the insolation levels decrease, inverters seek to obtain as much electricity from the dwindling amount of sun, a nonlinear process that produces much flicker, then vice-versa as the insolation increases with the passing of the cloud.

A small, slow-moving cloud will only affect a few panels at a time. The worst effect comes from large, fast-moving clouds that cover many panels nearly simultaneously, and then subsequently uncover them with the same speed. For these conditions, multiple inverters will respond to the same solar coverage effect, which results in the most severe levels of flicker.

Depending on how often the worst-case cloud cover events occur, and how many inverters there are on the power system, the amount of flicker can reach or exceed the limits allowed by the flicker standard, raising the specter of curtailing the amount of PV penetration due to flicker. In this context, both the size (kVA rating) and number of inverters is significant. Smaller inverters produce smaller magnitude fluctuations, but many such inverters may combine to increase the overall flicker level. Larger inverters tend to produce larger magnitude fluctuations and fewer such inverters are needed to produce the same flicker that many smaller inverters may produce.

Another factor to the flicker equation is the stiffness of the system that supplies the feeder.  Stiffness is a measure of how easily voltage may change at a certain location connected to the power system.  A stiff or strong system will limit the amount of voltage change; while a weak system will offer little resistance to any tendency for voltage fluctuation.  Standalone  or island systems are more susceptible to fluctuations, but feeders that are far from synchronous generators or other electric sources will tend to have similarly weak supply.  The presence of the PV sources is a slight mitigating factor as they tend to increase stiffness.

Conclusions

All of these considerations combine to produce an analytically complex situation for assessing flicker. At this point, no generalizations can be made about the level of PV penetration that can lead to a flicker issue. In a future Blog, we will provide some case studies that hopefully will illustrate the complexity of the issue and provide concrete measures for flicker.

For such a simple concept as flicker, the complexity can be daunting.