Screen impacts

[da-ekchajzer]

Problem

We want to retrieve the impacts of manufacture and usage of screens. According to recent study, screens account for a great part of the ICT EM impacts in France and Europe.

Solution

Several solutions could be used

Using archetypes

Defining some archetypes. Impacts foreach archetypes would be taken from :

• Manufacturer PCF (Boavizta database)
• ADEME BASE IMPACT
• LCA
• Other

Query example : Pre-recorded archetype OLED screen, size between 22" to 26"

{
   "archetype":"OLED-22-26",
   "hour_usage": 3
}

Defining a formula

We could define impacts factors based on screen characteristics, as we did for servers components:

• Screen Size
• Type (TV, Screen)
• Technology (OLED, LED, 4K, ...)
• ...

This impacts factors could be defined empirically based on Boavizta database or Fixometer reference data - 2021 records

To do so we need to explore the data and see to what extent characteristics influence the impacts of a screen. Archetype could still be used we pre-recorded characteristics

Querry exemple :

{
  "technology": "OLED",
  "size": 22,
  "type": "screen",
   "hour_usage": 3
}

If you have other ideas feel free !

Additional context or elements

In both cases, it would be nice find multicriteria data about screen

[da-ekchajzer]

I think we could find one or more LCAs from the academic world (see below) to define impacts factors according to different parameters (size or type for example) and compare them to the data from the manufacturer's database. I will try to read several of them next week. Another method which can be complementary: Identify trends in the monitor records of the manufacturer database according to different parameters (size, type)

Those two approaches should help us to identify equations and impacts factors depending on characteristics we identified.

Useful documents :

• The manufacturer database https://github.com/Boavizta/environmental-footprint-data/blob/main/boavizta-data-us.csv
• ADEME/ARCEP report, screen part (page 48) : https://librairie.ademe.fr/cadic/6700/impact-environnemental-numerique-rapport2-synthese-.pdf
• Life-Cycle Assessment of CRT, LCD and LED https://www.sciencedirect.com/science/article/pii/S2212827115000414
• Life-Cycle Assessment of Desktop Computer Display : https://19january2017snapshot.epa.gov/sites/production/files/2014-01/documents/computer_display_lca_summary.pd

[da-ekchajzer]

Does manufacturers overestimate the carbon impact of their screens ?

Manufacturer's methodology (PAIA)

Manufacturers use PAIA (Product Attribute to Impact Algorithm) methodology in their PCF. The PCF are aggregated here.

• Average CO2eq. impact per screen : 486 kgCO2eq
• Average lifetime : 4,4 years
• Average screen diagonal : 23 in ⇒ 0.1638 m2 (https://www.omnicalculator.com/other/screen-size; 4:3; 23 in)
• Average observed % impacts due to panel manufacture ≈ 40% ⇒ 0,4*486 = 194,4 kgCO2eq. per panel manufacture

NegaOctet

In the latest ADEME/ARCEP report (in France):

• Number of screen : 37 324 278
• Total ICT impacts in France (gwp) / year: 1,69E+10 kgCO2eq.
• % impact (gwp) for screens : 3.40% ⇒ 574 600 000 kgCO2eq / year
• Impact per screen / year: 15,39 kgCO2eq / year
• Lifetime : 6 years
• Global impact per screen : 15,39*6 = 92,34 kgCO2eq

BaseImpact - Panel manufacture

We can find multicriteria impacts data on panel manufacturing in the Base Impact (for LCD with CCFL or LED backlight) :

• Average manufacture impacts per m2 : 376,81 kgCO2eq.
• For a 23 in screen (0.1638 m2) : 0.1638*376,81 = 62 kgCO2eq.

Comparing

PCF vs NegaOctet

Because PCF use a more carbon intensive electrical mix than the French mix, we can expect greater impacts. This is balance by (1) the fact that usage impacts represent in PCF only 33% of the overall impact of screen (2) that PCF use a lifetime of 4,4 years average compare to 6 years for NegaOctet.

PCF / NegaOctet= 5,26

The PCF estimates an impact 5 times greater than the study done with the NegaOctet data.

PCF vs ADEME BASE IMPACT for panel manufacture

PCF often consider the manufacture being done in China, where BASE IMPACT considers Korea and Taiwan. The carbon intensity mix of those 3 countries are a little different :

• South Korea : 0,6 kgCO2eq./Kwh
• China : 1,057 kgCO2eq./Kwh
• Taiwan : 0,845 kgCO2eq./Kwh

PCF / BASEIMPACT= 194,4 / 62 = 3,1

The pcf estimates an impact 3 times greater than the data in the BASE IMPACT

I will be intrested to have your comments

[AirLoren]

Total impact (Scope 2 + Scope 3) evaluation involves too many assumptions (Screen size, power usage, lifetime, use location, type of transport...) to allow us to compare such high level values. I think we need to do a more thorough analysis by focusing on scope 3 as it is the hardest part to evaluate and the most impactful especially in use locations with low electricity carbon intensity .

[AirLoren]

I started an analysis based on PCR from Dell, Lenovo and HP. Lenovo data is very inconsistent and it is impossible to conclude anything from it except that scope 3 is between 5,7 and 24,5 kgCO2e/inch Dell and HP data from 2017 to 2021 shows values from 11,4 to 26,7 kgCO2e/inch with a mean value of 16,4 and a standard deviation of only 2,3 on 181 devices. Last PCR from HP (2022) show values from 3,3 to 8,6kgCO2e/inch with a mean value of 6 and a standard deviation of 1,5 on 17 devices.

I could not find any explanation for this strong reduction on the last PCR : Manufacturing Location is the same, Distribution is a little bit lower but this could not explain such a global reduction.

An idea for the next step could be to identify which Panel Technology (TN, LED, IPS or VA) is used for each device. One technology manufacturing could have a lower carbon intensity.

You can check the analysis spreadsheet here https://github.com/Boavizta/environmental-footprint-data/blob/monitor-analysis/boavizta-data-monitor-analysis.xls

[AirLoren]

I just added some informations about panel technology in the spreadsheet and can confirm it can't justify such a difference between pre 2022 PCRs and newest ones. https://github.com/Boavizta/environmental-footprint-data/blob/monitor-analysis/boavizta-data-monitor-analysis.xls

[da-ekchajzer]

I agree we should focus on scope3 and leave aggregated data which makes any conclusion impossible.

NEGAOCTET FOR EU

I desegregated NegaOctet data for Europe (multi-criteria). Impact gwp for 1 monitor for manufacture + transport + EOF = 70,6 kgCO2 eq.

The type of screen taken is not given. If we refer to the ADEME/ARCEP report the typicall screen should be a 24 inches LCD ⇒ 70,6/24 = 2,94 kgCO2eq./inch

[AirLoren]

I analyzed PCF for Fujitsu displays from 2022 (available here https://www.fujitsu.com/global/search/?query=(PCF)%20display&start=0&count=10&site=29QMLFR4&charset=utf-8&sort=date) and results are very close to the value of 16,4 kgCO2e/inch mentioned here.

The common point between these data is that they have been generated with the PAIA methodology contrary to those of Negaoctet and last HP PCR.

PAIA Methodology seems to overestimate GWP impact of display manufacturing compared to NegaOctet and last methodology used by HP.

No idea who is closer to the right value but we have an explanation for such a difference. ;-)

[da-ekchajzer]

A basic approach will be implemented in the v0.3 using a fix impact factor for monitor from https://base-impacts.ademe.fr/documents/negaoctet.zip

[bpetit]

I don't see a mention of this in dev branch commits. Have I missed something or should we mark this issue to be adressed in later versions ?

[da-ekchajzer]

Is has not been implemented yet. The issue will be addressed later.