Read the following guidelines carefully to ensure the fire tests are conducted properly.
Use a realistic test location
We strongly recommend running the test in a location that is similar to the Pilot or Live deployment location. Do not run the test in a location that is vastly different from their intended destination. Running a test in an urban parking lot, for example, will cause discrepancies. The sensors are primarily forest sensors and urban areas will affect the functionality.
Use the Map view of the Silvanet Deployment app to find the ideal location for the test fire.
Use homogeneous material
Use a collection of homogeneous plant material (uniform in size and composition) for the test fire. Material should be of various sizes all of which are fast-burning branches or twigs but not logs. To create smoke, have branches with their leaves or needles still on.
|Use material to final deployment location||Collect between 5 and 10 kg of typical plant material similar to the final deployment location, including smaller and bigger twigs, leaves as well as freshly cut and previously dried branches with needles.|
|Ensure material is dried and stored||Dry out in advance the collected material and store them in a cool, dry location. Use the duration of the calibration period to allow the material to dry out and to be ready for burning.|
|Weigh material||Weigh the material before placing them on the fire pit. This ensures the same amount of material is used for subsequent test fire.|
|Enough material for 30-minute burn||Collect sufficient material to keep the fire burning for up to 30 minutes.|
Burn a realistic test fire
Ensure you have sufficient homogeneous material to burn and maintain a test fire. To reflect conditions of real wildfires, the test fire should not diminish as the test runs. The fire should not resemble a BBQ fire.
As the wind direction cannot be controlled, provide sufficient time for smoke from the test fire to hit the sensors.
Ensure test fire burns for at least 30 minutes
Ensure you continually feed the fire with material to keep the fire burning for at least 20-30 minutes to allow smoke from the test fire to reach the sensors. The sensors are built to detect fire at the smoldering phase - before there is an open fire.
Normally, time to detection is within minutes (less than 1 hour from ignition). This is dependent on fuel volume, wind speed and wind direction. A denser deployment will decrease the time to detection and detection rate.
Be aware of smoke behavior
Because wildfire smoke behaves differently in different locations, select a location that is realistic and reflects the final deployment environment.
A parking lot, work yard or an urban area may not be the best location for a setup although it may be convenient. The results of a test in those type of locations would not be very useful. For example, under a forest canopy, wind behavior is different from wind behavior in an open area such as a work yard.
Limit repeated testing to once per day
When performing fire tests, it is common practice to repeatedly light a fire and check the response. However, repeatedly exposing the sensors to smoke in short time intervals will result in poor sensor performance. After being exposed to smoke, the sensors need considerable time to re-settle and restore their original fire detection sensitivity.
For best test performance, leave the sensors alone in the forest, do not have vehicles nearby or anything that interacts with the sensors and then wait 24 hours before running another test.
When running tests, use a variety of deployment arrangements. In all cases, use shorter and longer distances between sensors. A short distance such as 2 m results in quick detection times while longer distances such as 15 m results in longer detection times. Therefore, use different sensor spacings depending on the objective of the fire tests (quick detection vs. longer distances). The fire pit can also be placed in different locations within the sensor grid.
In all cases, wait 24 hours between tests and do not interact with the sensors between tests.
Use a grid pattern
Deploy the sensors using a grid pattern. You should prepare this grid pattern when adding the Deployment Packet. Furthermore, as the sensors are deployed in a dense grid, the test deployment does not need a Mesh Gateway. It does, however, require all sensors to be in range of the Border Gateway. Sensors should not be set up in hilly terrain where hills or large rocks could interfere with the wireless signals.
Ensure the sensors are at least 3 meters above ground level. They should be oriented towards the sun (at 12:00 noon).
Initial sensor arrangement
Prepare an initial setup. For example, arrange the sensors in a grid pattern with 2 m distance between sensors. Place the fire pit in the center of the sensor deployment. Be aware of wind direction and speed. This may affect which sensor or sensors the smoke hits triggering a Phase 1 process.
Modified sensor arrangements
After obtaining results from the initial round, perform additional tests by modifying only one aspect of the test. Possible modifications include:
Distance from fire to the sensors. The further away the sensors are from the fire, the longer the detection time becomes.
The location of the fire within deployment.
Burning different materials in the test fire.
For modified tests, the same type and weight of material for burning MUST be identical to that used in the initial test.
Example sensor deployment arrangements:
- 15 m between sensors
- Fire pit placed off-center:
Ensure sensors have calibrated (14 days)
Before beginning the test fire, ensure you allow the sensors to perform the 14 day initial calibration period. Sensors will not provide accurate readings when tested before or during the calibration period. Afterwards, start the test fire and begin testing Silvanet.
After deploying the sensors for the first test, wait 14 days for the sensors to perform their initial calibration to determine a baseline Air Quality. No testing should be conducted prior to the completion of the initial calibration period since the sensors will not provide accurate readings.
For more information, see Sensor Calibration Duration.
Do not interact with sensors after deployment
As described in Sensor Calibration Duration, the gas sensor in the Silvanet sensor is very sensitive to changes in the environment. Consequently, moving, touching or interacting with the sensors beyond what is necessary influences the sensor readings and also the sensor calibration settings.
AFTER SENSORS HAVE BEEN DEPLOYED, CALIBRATED AND SETTLED, DO NOT INTERACT WITH THE SENSORS AS THIS WILL CAUSE ERRORS WITH SENSOR VALUES AND COULD ALSO RESET THE SENSOR CALIBRATION.
Deploy at the forest edge
As described above, the test should be set up in a realistic location. Consequently, deploy the Border Gateway at the edge of the forest or near a path or roadway. This allows the solar panel to be free of obstructions to sunlight. It also allows better line of sight to the Mesh Gateway.
Ensure Internet connectivity and reliable power source
As well as placing it near the edge of a forest, find a location where it can obtain a power supply and/or Ethernet connectivity.
Before starting the test, ensure the deployed Border Gateway is connected to the Internet.
The Border Gateway MUST be online with a stable connection either through cellular modem or POE (Power Over Ethernet).
The deployed Border Gateway MUST has a reliable power source and is fully charged before testing.
If the border gateway is powered only by the solar panels, then it most likely enters power saving mode during the night. Allow the Border Gateway to fully charge in the morning before testing.
Fire alerts will be handled with priority and will wake up the modem from power save mode.
After the sensors and Border Gateway have been deployed and charged, ensure they are communicating with the Silvanet Network. You can verify this using the Site Management app.
From the Site Management app, select the test site. In the Site dashboard, active sensors appear as green icons while inactive sensors appear as gray icons.