Air Quality measurement is a growing demand and need in cities and private companies. Air quality is a clear part of people well-being and it will be the point of differentiation among municipalities and private companies in order to be #1 destination or to retain talent in the short term future.
Several cities rely on one or a few Reference Stations (whose measurements carry legal value), although they are quite expensive and this makes their deployment in different areas of the city very complicated. Within the same city, there may be significant fluctuations in the air quality from one area to another.
Our proposed solution to monitor air quality and noise levels
Libelium Air Quality Station the device that allows to control the main parameters included in the AQI (Air Quality Index), as well as to make predictions in order to implement preventive corrective actions.
In terms of air quality monitoring, we understand that calibration can occasionally be a problem when time is short and budget is limited. This is why we were ahead of that need and installed a total of 6 Air Quality Stations in Spain (2 on each of 3 different locations) during 6 months.
This investment of time allowed us, on the one hand, to verify the correct functioning and accuracy of the device, but at the same time to gather enough data to generate universal calibration models.
The nearly 100% accuracy that Air Quality Station provides when performing co-location calibration is frequently required in scientific or research type of projects. However, in projects where a remarkable accuracy is sufficient, such as air quality measurement in factories or production plants, using universal models to calibrate might be the best option, taking into account the costs involved in the co-location and calibration process or the lack of access to the nearest reference station, which can cause a severe delay on your project.
Universal models: remarkable accuracy without the need for co-location calibration
When time is short or when there is no possibility to place an Air Quality Station node next to a reference station to learn, Libelium’s Air Quality Station has a universal model.
The Air Quality Station has already been installed, tested and calibrated with different reference stations. Each location has provided a different data layer of air quality concentrations, ambient and atmospheric characteristics. The different data layers form a data lake that has allowed the design of a generic artificial intelligence model for air quality prediction. It is important to bear in mind that, while the co-location calibration system reflects the real value, the universal model represents an approximation to it..
That is, it does not have the accuracy that an AQS node that learns from a reference station in its own city may have, the data it produces are within the margins of error.
The developed algorithms, as well as the trained models, are very useful for making predictions, analyzing evidence and taking corrective actions on new data.
In addition, Libelium Air Quality Station’s universal model continues to learn and train itself from all Air Quality Stations deployed and placed next to reference stations, so it will become richer, more dynamic and more accurate.
Co-location: do you need the maximum accuracy?
Reference stations are publicly owned high-performance air quality measurement stations. They are like small chemical laboratories of extreme quality in all their dimensions (precision, repeatability, reliability, etc.) and whose measurements have “legal value”.
However, these stations cost well over 200,000 euros, and the power consumption is very high. Also, the measurement methods are sometimes not very automated. For example, particulate matter is not measured by a sensor, but by a filter that captures the particles. The operator has to physically access to collect samples for several days and weigh the micrograms of particulate matter for each period on a precision balance.
High acquisition and operating costs make it possible to deploy only a few of these. Higher spatial granularity throughout the city is desirable, but complicated for economic reasons.
Libelium’s Air Quality Station meets this demand. Its accuracy is sufficient for many use cases, resulting in a very good price/quality ratio. In addition, the AQS incorporates cutting-edge technologies, such as Artificial Intelligence, connectivity, or automation, that enhance its usability and reduce the necessary maintenance operations.
The Air Quality Station is placed next to the reference station to collect data over the course of a month. The data from the reference station is imported to Libelium Cloud via a CSV file while all the data collected by the Air Quality Station is sent to the Libelium Cloud via 4G. This is where Artificial Intelligence is applied, allowing the user to run predictive models.
Once the AQS node has been calibrated and trained it is ready to be deployed in the project for which it was designed.
Tests & Results (co-location studies)
Since reliable and accurate measurements are essential, we have carried out several studies that have allowed us to guarantee the proper functioning of this device, to improve our machine learning algorithms and to generate the universal models.
If what we have explained so far has sounded interesting to you, we encourage that you take a look at the complete results:
Summing up, regardless of your needs, with these two calibration methods, measuring the air quality in your city or town is now possible, reliable and affordable.
Would you be interested in developing an air quality project with us? If so, please fill in the form below: