A button can save a lot of energy

A few days ago I was walking at night in the city going back home. I was very sleepy and tired, looking forward to reach my home one step after the other, when suddenly I felt enlightened. Unfortunately, I didn't reach some sort of universal truth: the light came from an ATM.

The two screens were on, even if there was nobody except me passing by. Always on, 24h showing information that nobody reads. On my way back home I counted 5 other ATM, so a total of 6 in my 1 km trip.

The next day I started wondering: how much energy are we wasting by keeping those screens on even if nobody is using them? To answer, first we need to know how many ATMs are out there. I started by searching on Google Maps in the city area and got this:

Sadly the total count of the result was not shown in the mobile app, and to get it from the desktop version you have to to click the next arrow over and over, until you reach the end of the results (really, Google?). A smarter approach to know the total count of ATMs in an area was using Google maps API, but the same location pinpoint can refer to a bank with multiple ATMs.

Doing a bit of research I found out this:

There are 3.24 million ATMs out there!

To calculate the amount of energy needed to keep all those screens turned on, I made some estimates:

• The screens are usually quite old, so I thought a LCD TFT of 2009 can be a good approximation. The energy consumption for that monitor is 31W. This seems quite reasonable given what I could find here and here, and accounts for ~ 12% of the total energy required.
• The average screen size is 15 inches. Again reasonable, and probably round down approximation since some ATMs like the ones in the photo have more than one screen.

Now a bit of math!

Hourly energy consumption of 1 ATM:

\[E_h = Power \cdot time = 31 \,W \cdot 1 h = 31 \,Wh \]

Daily energy consumption of 1 ATM:

\[ E_{day} = 31 \,Wh \cdot 24 = 744 \,Wh \]

Yearly energy consumption of 1 ATM:

\[ E_{year} = 744 \,Wh \cdot 365 = 271560 \,Wh = 271,56 \,kWh \]

Yearly energy consumption of all ATMs:

\[ E_w = 271,56\, kWh \cdot 3,24 \cdot 10^6 = 879854400\,kWh \approx 8,8 \cdot 10^8 \, kWh \]

Pretty huge amount of energy. The next question is: What are the emissions of \( CO_2 \) to produce this energy?

The following table is taken from here, and while being a bit old (data are from 2006) gives us some ideas.

Generation Source	kg \( CO_2 \) per kWh
Open Cycle Gas Turbine	0,5
Closed Cycle Gas Turbine	0,5
Oil	0,65
Coal	0,9
Nuclear	0,005
Pumped Storage	0,02
Non Pumped Storage Hydro	0,005
Wind Onshore	0,00464
Wind Offshore	0,00525
Solar	0,058

Based on 2014 data from Our world in data (the only I could find) we can obtain an estimation, of the energy obtained from the various sources.

\[ E_{coal} = E_w \cdot 41,1\% = 3,61 \cdot 10^8\, kWh \]

\[ E_{oil} = E_w \cdot 3,6\% = 3,16 \cdot 10^7 \, kWh \]

\[ E_{gas} = E_w \cdot 21,92\% = 1,92 \cdot 10^8 \, kWh \]

\[ E_{nuclear} = E_w \cdot 10,78\% = 9,48 \cdot 10^7 \, kWh \]

\[ E_{hydro} = E_w \cdot 16,49\% = 1,45 \cdot 10^8 \, kWh \]

\[ E_{other} = E_w \cdot 6,10\% = 5,36 \cdot 10^7\, kWh \]

And finally, we can obtain the total kg of \( CO_2 \) emitted in the atmosphere to keep ATMs screens always on. The wind and solar emission per kWh are averaged and counted in the 6,1% of the other renewables category and the hydro ones (pumped and non pumped storage) in the hydroelectric category.

\[ Emission_{coal} = 3,61 \cdot 10^8\, kWh \cdot 0,9\, \frac{kg}{kWh} = 3,25 \cdot 10^8\,kg \]

\[ Emission_{oil} = 3,16 \cdot 10^7\, kWh \cdot 0,65\, \frac{kg}{kWh} = 2,05 \cdot 10^7\,kg \]

\[ Emission_{gas} = 1,92 \cdot 10^8\, kWh \cdot 0,5\, \frac{kg}{kWh} = 9,64 \cdot 10^7\,kg \]

\[ Emission_{nuclear} = 9,48 \cdot 10^7 \, kWh \cdot 0,005\, \frac{kg}{kWh} = 4,74 \cdot 10^5\,kg \]

\[ Emission_{hydro} = 1,45 \cdot 10^8 \, kWh \cdot 0,0125\, \frac{kg}{kWh} = 1,81 \cdot 10^6\,kg \]

\[ Emission_{other} = 5,36 \cdot 10^7 \, kWh \cdot 0,02263\, \frac{kg}{kWh} = 1,21 \cdot 10^6\,kg \]

So, summing up every contribution we get:

\[ Emission_{total} \approx 4,46 \cdot 10^5\, \frac{ton}{year} \]

So each year we are sending at least 446'060 tons of \( CO_2 \) in the atmosphere. I want to stress that I strongly believe this value is a round down estimation.

Repeating the calculation considering 42 W as power consumption (40% increase) of screens, we get \(\approx 6,04 \cdot 10^8\, \frac{ton}{year} \)

This last value could seem unreasonable but I suspect that the actual average power consumption of ATM screens can be higher than 31 W.

Either way, it's a lot of energy wasted for no reason. To put things in perspective, the per capita annual emission of \( CO_2 \) is 4,8 tons.

That means that all those screens are emitting each year the same \( CO_2 \) amount of 92'930 and 125'904 people for the first and second estimated power consumption respectively.

The city of Pisa had 90'488 residents in 2017.

By activating ATM screens only when the user press a button, or using a proximity sensor, we could save more than a whole year of Pisa's citizens emission.

Which is the same as saying that we could save 3,23% of Rome’s yearly emission.

It might seem a little step, but as the old saying reads: take care of the pennies and the pounds will take care of themselves.

sudo shutdown -h now