Policy Impact: Exploring the seasonally adjusted difference between 2020 & 2019 electricity consumption
Demand & Economics: Duck curve & the economy - an analysis into electricity consumption
Production Function: An experimental approach to detect patterns in electricity demand
Let’s begin with a look at the day to day differences between 2020 and 2019 across the world adjusted for seasonality. Exploring the COVID-19 responses across the world, it is evident that difference in electricity consumption is drastically impacted based on the economic policies that have been put in place in each country. For example, Italy who began nation-wide lockdowns on March 9th, saw an immediate drop in electricity consumption over the next few days. On March 22nd all factories were closed, and all nonessential production was halted, causing a further dip in consumption difference. This same story can be told with several other countries. Interestingly, Sweden who chose a mitigation strategy showed very little difference compared to its Scandinavian neighbours Norway and Denmark, who took much more preventative measures.
Note: See the references page for data sources & data representation of Australia & US
With the rise in solar energy production, anticipation of electricity demand has been at the forefront of utility companies’ priorities. The duck curve refers to the demand for electricity at any given time during the day. The early mornings consist of low energy demands, but as people wake up and businesses begin production, demand rises. This peaks around sunset before dropping. As you may have guessed, the production of energy during the peak hours of the day helps reduce the demand needed. This is evident as the year goes on. The dip during mid-day drops lower and lower, forming a duck like shape. This is more prominent in countries with renewable energy rebates such as Sweden’s tax regulation mechanisms and subsidy scheme.
There are two critical issues that arise from this. Firstly, as steeper drops in demand occur, more rapid increases in production are needed during peak hours with no sunlight. This poses a serious burden on power infrastructure and diminishes efficiency. The second issue is that during days of over production in solar energy, grid managers turn off solar panels to prevent overloading the power grid, discarding extra solar energy.
Evolutions in battery technology provide a potential solution for this. As battery technology continues to evolve, we are able to store energy more efficiently and locally, thereby reducing waste while moving away from non-renewables.
Firstly, let’s take a look at how COVID-19 has impacted Victoria’s electricity demand by plotting the moving average of active cases vs electricity demand alongside significant policies put in place. I hypothesize that most energy consumption can be attributed to weather using heating & cooling infrastructure. People will use electricity from home during lockdown rather than at their workplace resulting in a minimal difference. Secondly, large businesses make up most demand for electricity, and hence there should be no discernible effects unless these large businesses closed down. An exploration into residential & commercial electricity demand by Energy Networks Australia showed up to a 15% increase in residential use during mid-day compared to the same time in 2019. In comparison, there was up to a 19% decrease in consumption by small-medium enterprise businesses and 8% decrease in large businesses.
The figure above shows the timeseries (rainbow gradient) within a scatter plot of active COVID-19 cases versus electricity demand. It is not evident that any of the 4 major policy impacts resulted in a higher or lower electricity consumption. The changes in demand are more so attributed to seasonality. Next, we will explore which features may be attributable to electricity production. There are too many variables to explore in detail on this page, however, I have plotted below the electricity demand vs price with time filters to explore seasonality in the data.
Other plots such as weather, mobility & economic activity can be found on the data2app page also localised to Victoria, Australia. I’ve used live google search volume for certain keywords as a proxy for mobility & economic activity. The weather is read in live from the Bureau of Meteorology. In this exploration, I aim to only use live data sources so when incorporated within a business model, actionable insights can be obtained immediately.
\[x_1 = weatherForecast, x_2 = weekOfYear, x_3 = dayOfWeek \] \[x_4 = hourOfDay, x_5 = holidays, x_6 = location \] \[x_7 = mobility, x_8 = economic Activity,x_9 = solar Panel Purchases\]
\[energyProduction = f(x_1, ... , x_n)\]
Sample experiments are displayed in the extras page, including a simple SARIMAX forecast, t-SNE plot & Benford’s law. My current explorations are in change point analysis and exploring what actionable insights can be made from electricity demand data outside of forecasting.