Where there’s smoke, there’s reduced PV output

It’s a minor concern compared to the tragic loss of life, livelihoods and biodiversity caused by the bushfires still ravaging parts of Australia, but reduced output by PV systems due to smoke haze is an unwelcome bi-product of blazes that have burned at a scale and ferocity never seen before.

January 20, 2020 Natalie Filatoff

Smoke and haze can envelope rural and urban areas alike during bushfires.

Image: pv magazine, Natalie Filatoff

The haze and fallout from Australia’s catastrophic bushfires is also affecting the output of solar panels from solar farm arrays and rooftop installations — part of the renewable infrastructure that on clear sunny summer days can contribute as much as 50% of energy demand in the National Electricity Market.

The haze has so far been recorded as reducing solar output from residential rooftop systems by up to 45% in particular locations on certain days.

Solar Analytics, a smart-monitoring company that helps rooftop solar-system owners manage their energy consumption and solar generation, monitors 35,000 sites around Australia. It has correlated Air Quality Index readings over recent months with expected irradiance and solar panel output with disturbing results.

On December 10, one of the worst smoke-hazy in Sydney, for example, median energy generated across 2,200 Sydney sites was 22.7 kWh; compared with 26.6 kWh on clear days during the first two weeks of December.

This 15% decrease was surpassed on December 21 — a cloud-free but extremely smoke-affected day — when Sydney rooftops showed a 27% decrease in output compared to a clear December day in 2018.

New Year’s Day in Canberra had the unhappy distinction of choking on a particulate matter (PM 2.5) reading of 347 micrograms/cubic metre (μg/m³), when a normal December/January day might register between 3 μg/m³ and 10 μg/m³; Solar Analytics recorded a 45% associated decrease in the city’s rooftop solar output.

Professor Gary Rosengarten of the School of Engineering at RMIT explains that the decrease is caused by light being absorbed, scattered or reflected by particulate matter which makes up part of the haze. He tells pv magazine that you can see the effect in that “there’s not enough direct light to cast a sharp shadow”.

He adds, “That also decreases the amount of light hitting directly on the pv panel.”

Victor Depoorter, Technical Director at solar forecasting company Proa Analytics says that on certain days, his company has observed smoke haze impacting all of the large-scale solar projects it’s monitoring, from Kidston Solar project in Far North Queensland to Tailem Bend Solar Power Project in South Australia.

Utility-scale solar takes a hit

In areas most affected, he says, “We could see solar farm output reduced by 10-15%.”

In terms of forecasting, Depoorter tells pv magazine, “It has been challenging to distinguish between clouds and clear sky on some days at some of our sites”, which is an issue he says, because the behaviour of irradiance is not the same given haze or cloud cover.

“Water-carrying clouds, the big, fluffy, white clouds which are the most common here in Australia, have a very high reflection — they can almost completely block irradiance. That’s not the same case for smoke haze; the scattering mechanisms work differently.”

The forecasting technology developed by Depoorter and Matthew Jeppesen, Managing Director of Proa Analytics, now allows Kidston and Tailem Bend solar farms to self-forecast their output into the NEM, enabling the Australian Energy Market Operator to better integrate their energy flow with that of other generators in the grid. Proa Analytics is working towards achieving the same accreditation for other solar farms in the NEM.

So far, says Depoorter, the variations introduced by smoke haze have not been sufficient or frequent enough for the company to consider training its algorithms to definitively distinguish smoke haze from cloud.

He hopes that will never become necessary.

Muddying the output

In the meantime, smoke haze settles as dust in affected areas, which Depoorter says can reduce solar panel output long after the air has cleared.

If dust and ash mix with water from otherwise largely welcome rain, the result can be a film or patches of mud, which will also reduce solar panel efficiency.

When the cars in any area look dirty from the combined effects of dust and rain, Depoorter says nearby solar installations are likely to be equally affected, and need to be cleaned to restore optimal output.

“Unless people are reactive in cleaning their solar panels after these events,” says Depoorter, “losses in output due to soiling will be increased for some time” — until they are deliberately cleaned, or substantial rain washes away the grime.

This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: editors@pv-magazine.com.