CO and SO2 sensors giving incorrect reading

Posts: 19
Joined: Tue Feb 20, 2018 9:06 am
Company: Verso Altima Group

Re: CO and SO2 sensors giving incorrect reading

Post by nzagorec » Mon Dec 03, 2018 10:25 am


±0.3ppm (±600-850 ug/m3, depending on the pollutant) means that AQI for each individual pollutant (and for air quality in general) can be from "very low" all the way to "very high".

What do you suggest we do with these results?

Should we use them only to observe trends?
Is it possible to normalize the results?

For example we compare measurements made by official AQI station and Waspmote and make a "corrective factor" that we use to multiply our results with? E.g. in measurements from last week official AQI station measured 43.7 ug/m3 and we measured 669.87 ug/m3. Could we multiply our future readings with 43.7/669.87 and get something meaningful from it? Will this ratio always be the same? What if NO2 goes up 30 ug/m3 - we're not going to get +30 but +2 because of that factor... Should we subtract the numbers, not divide them? ...?

Do you have some information how other customers process information from electrochemical sensors since they clearly can't be use for calculating AQI "out of the box"?

libelium-dev wrote:
Thu Nov 29, 2018 12:05 pm
Accuracy values have been calculated at average conditions: 20ºC / 101300 Pa. Accuracy levels may differ with different temperature and pressure levels, aging and in presence of third type of gases which cause cross sensitivity.
Aren't these sensors temperature compensated? I see that one of the inputs into the formula is current temperature read from BME280.

Posts: 27967
Joined: Mon Sep 28, 2009 1:06 pm

Re: CO and SO2 sensors giving incorrect reading

Post by libelium-dev » Mon Dec 03, 2018 4:55 pm


Unfortunately there is no easy formula to make corrections compared with the data of a scientific station, as there is a correction by temperature that we have implemented in the firware

The only way to make this corrections accurately is with a large amount of data and applying machine learning and artificial intelligence tools to the data of the sensors next to the scientific station and its data too. Unfortunately that's not something immediate and it requires a level of expertise

The current product line works well to observe trends, increases, decreases in a specific time and place, to have an overview and estimations on the pollution values, but if you calculate the AQI out of the box, the values will be just approximate, more accurate for some sensors than others in different moments.


Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest