Final Report: Ultralow Power Sensor Package for Ground Level Air Pollution Levels fromWildland FiresEPA Contract Number: 68HERD19C0020
Title: Ultralow Power Sensor Package for Ground Level Air Pollution Levels fromWildland Fires
Investigators: Findlay, Melvin
Small Business: KWJ Engineering, Inc.
EPA Contact: Richards, April
Project Period: May 1, 2019 through October 31, 2019
Project Amount: $100,000
RFA: Small Business Innovation Research (SBIR) - Phase I (2019) RFA Text | Recipients Lists
Research Category: Small Business Innovation Research (SBIR): Phase 1 (2019) , Small Business Innovation Research (SBIR) , SBIR - Air Quality
KWJ proposed to integrate printed gas sensors and particle sensor into a single, <8oz package with the dimensions <4"x5"x1" (10x12.5x2.5cm). In Phase I, we proposed using a prototype multigas board we have developed in collaboration with Intel, and integrate with Alphasense’s OPC-N3 or R1 PM sensor, which is the current state-of-the-art in miniature, optical particle detection. We will also evaluate the Sensirion SPS30 PM sensor and SCD30 NDIR CO2 sensor as time and resources allow.
In Phase I we proposed to design a package which can be deployed in a variety of ways: worn by personnel, attached to stands which can be located and relocated around the perimeter, deployed as mobile monitors on vehicles and drones. This will allow the monitors to be deployed around the fire perimeter as it moves, as well as downwind to monitor approach to nearby residences. We also developed a design concept for electrostatic PM (specifically UF’s) during Phase I also.
In Phase II we proposed to build and test a field-prototype using the best available COTS sensors down-selected during Phase I, as well as fabricate prototypes of the UPM sensor, which will measure particles down to 5nm, and use far less power than the optical sensors. If successful, the UF PM sensor will be integrated with a 2nd generation of the monitor.
To allow design and test of electronics and hardware required for remote field deployment in parallel with sensor evaluation, we assembled modules using COTS cases and Raspberry Pi boards for control, DAQ and Wi-Fi (Task 1) to allow us to quickly focus on environmental test and benchmarking of the gas and particle sensors (Task 3). In parallel, we designed and fabricated a concept prototype sensor module incorporating the current state of the art sensors in terms of size, cost and power to provide form factor, while focusing on long-range communication and telemetry (Task 2).
During Phase I we evaluated the SPEC Sensors miniature printed gas sensor, an ultra-low cost amperometric sensor, against Alphasense commercial AQ sensors, their "A4" 4-electrode sensors. Laboratory sand bench testing indicates the CO, NO2, and O3 sensors. These sensors alone, available for $3-5 each in quantity, will reduce the BOM cost by >$100 with little lost of sensitivity or short-term accuracy over T and RH. The SPEC commercially available SO2 sensor exhibits unacceptable S/N and a large zero temperature shift, so in the 1st iteration Phase II filed prototypes we will use the Alphasense SO2-A4. We did, however, receive some R&D prototype SO2 sensors, which exhibit 4-5X improved S/N and a much reduced zero tempco. We plan to fully evaluate this sensor at the start of Phase II for inclusion in 2nd-generation prototypes.
For CO2, we evaluated the Alphasense IRC-A1 NDIR CO2 sensor, plus the much lower cost Sensirion SCD30 and GSS COzIR. At <#30, the Sensirion SCD30 is an excellent performaer, tracking the much larger and more costly IRC-A1 very closely. The GSS sensor operates at much lower power, but tended to underestimate rapid changes in concentration.
For particulate, we evaluated the Alphasense OPC-N2 and R1, as well as the lower cost Sensirion SPS30, against a Particles Plus Model 8306 reference analyzer. Overall there is good correlation between the readings from the low-cost sensors and the 8306 reference monitor. Accuracy was quite variable, with generally acceptable accuracy at ambient levels, but much more variability at higher concentrations. In Phase II we plan to perform a more quantitative accuracy and repeatability test, using an aerosol generator and set of microbead particle size standards.
The results to date with the lower cost SPEC gas sensors and Sensirion CO2 and PM are very encouraging. The component cost will be reduced by approximately 5X, from >=$400 to ~$80. This will allow a sale price of ~$150 for the 4 gases with wireless, and $400- $500 including particles and CO2.
What remains to prove is long term stability – what are these sensors inherent stability, and is there a way to remotely cal and zero in the field so that the sensor can be deployed for 5-10 years in a forest?
The communication options will include wireless (Wi-Fi, cellular, a LoRa Mesh capability for remote locations) as well as USB for local connection and power. Primary power for operation will be a 4-5 A-hr lithium polymer battery, sufficient to perate the module for >=1 week, plus solar panels capable of maintain battery charge.
KWJ has expanded partnerships with SPEC Sensors, LLC and Sensirion AG, two leaders in small, low-cost sensors for air quality. KWJ has facilities and expertise to evaluate emerging new gas and particle sensor technology, and will provide test/evaluation support to SPEC Sensors improved AQ gas sensors. As sensors are released, KWJ will be first in place to incorporate into our evolving line of air quality and safety instrumentation. Sensirion has agreed to partner with KWJs growth into air quality by providing advanced particles, CO2 and VOC sensors, as well as collaborative development of web-based data management via a new API.