I. Definitions
A. Quality Assurance: a program for the systematic monitoring and evaluation of the various aspects of a project, service, or facility to ensure that standards of quality are being met
B. Quality Control: an aggregate of activities designed to ensure adequate quality especially in manufactured products
II. Goals of our Quality Control system
A. Goal 1: Obtained data from the monitors shall be representative of ambient air concentrations existing at each location
B. Goal 2: Measurements shall be accurate, precise, and traceable
C. Goal 3: In order to make accurate assumptions, a certain percentage of data needs to be capture
III. How do we approach these goals?
A. Concerns about measurement uncertainty
- Determine how many packets get lost or are incomplete
a) Keep track of uptime -- if we use 24-hour averages, we cannot have more than 2.4 hours of downtime
b) We believe the uptime method may be the best
(1) We can create a notification system that would ping us when 5% of data has not been captured for that day
(2) Then, we can check if it is a monitor or a data transfer problem before it reaches 10% of downtime
c) When 5% of downtime has been reached, an alarm will be sounded, or an established notification will be set. This notification will allow the decision makers to check and see if there is an issue with the monitor, or if there is an issue with the related data transfer, before the 10% allowance for downtime is exceeded. This protects the timeline from prolonged downtime
B. Bias - We believe the best way to understand if our monitors have reliability or bias problems will be through the co-location experiments
a) Waiting on progress in this sector
b) This will allow for confidence in our measurements
C. Co-Location Experimentation - Professor Proposal: keep our monitors next to federal monitors for 40 days, then move 2 monitors away and still see if we are getting same readings. This should ensure consistency, and that our monitors are correctly calibrated. This is a commonly used and approved method of calibrating low-cost sensors, or evaluating their accuracy (EPA, n.d.).
a) The colocation should be carried out for 30 days. More specifically, 24 hour data should be continuously collected for 30 days, to create 30 pairs of data for comparison. If there is reason to believe the measurements are seasonally effected it could be carried our quarterly, to adapt to the changes in the weather (EPA, n.d).
b) Yes, every new monitor should be tested, further, it is best to test them in sets of free, for greater accuracy in the information provided, and to try to eliminate bad monitors (EPA, n.d.) .
D. On-Site Monitor Evaluation
- Work with the monitor team, maybe to create separate protocol
IV. ABOUT MONITOR:
-Monitor can’t touch water
-Casing should protect the monitor, but may need to check casing every few months
-Paint on casing should also be waterproof and be able to last a long time, safe paint
-Electricity related concerns; electricity can’t get wet
-Co-location experiments use an extension cord
-Calibration: want to make sure all monitors are collecting data at the same rate
-SHOULD happen every 3 months, sort of seasonally
-Casing should stabilize this; won’t fully need to calibrate filters
-No info on the lifespan of the monitor
-We had to send 2 monitors back because one antenna broke (1-2 years old)
-Might be best to keep a log about when we need to send monitors back for repair or updates
London Air Quality Networks data validation is two-fold: one side tackles data handling and the other side tackles on-site monitor checking
-Would we want to create two different protocols? After speaking to the monitor group last week, it doesn’t seem like our monitor will have many complications with the casing that Andrew made. But maybe we need to create a protocol for the next monitor group on how and when to check on monitors.
V. https://uk-air.defra.gov.uk/assets/documents/reports/empire/lsoman/LSO_manual_Oct_2015_Issue_1_Part_A.pdf
Sample Solution