Scaling AQMesh pods to provide increased levels of accuracy
In order for AQMesh to provide the greatest degree of accuracy, it is advised to co-locate the pod with a reference station. This will provide the information necessary for calculating our scaling variables; offset and slope. Scaling provides an increased level of accuracy, based on the local environmental conditions and how the sensors used react to that environment.
Scaling values are calculated over a specific time period and as such are based on the environmental conditions on site over this period. By applying these values for offset and slope we are able to determine the best possible performance results for this pod over the same period, and improve accuracy of the pod moving forward.
Scaling of AQMesh sensors is only required in certain circumstances:
• Following a sensor change, the new sensor will need scaling to its new environment
• Following a large step change in environmental conditions at the site, e.g. a change in average temperature of 10 degrees Celsius or more, compared to when it was scaled
In order to apply scaling to an AQMesh pod the user must first select to use advanced settings so that “PreScaled” values for each sensor are available for the downloaded period.
Regression comparison AQMesh PreScaled data versus reference or Gold Pod Scaled data, using the candidate pod on the x axis and the reference on the y axis. Regression analysis can be completed using Microsoft Excel, MathCad or similar mathematical programs.
Removal of potential erroneous data points and find Cartesian equation for line of linear trend line
Apply Offset and slope to full PreScaled data set (including erroneous data points)
Reference ≈ (PreScaled data x Gradient) + Y-Intercept
Gradient = Slope
Y-Intercept = Offset
AQMesh Scaled data = (PreScaled data x Slope) + Offset
Upload slope and offset values to the online application for each gas sensor and particle fraction.
Validate pod accuracy by following the scaling process by either; a second co-location with reference or co-location with Gold pod at the target location, at an interval determined by the users own QAQC process.
For long projects it is advised that data is validated every six months to account for any changes in calibration at the reference station.
For shorter projects, good practice for validation of data would be to run a scaling exercise both pre and post the trial period, providing full confidence in the data received from the trial.
AQMesh pods can also be scaled with diffusion tubes. For more information, read here.