Silvanet Wildfire Sensor

The Silvanet sensor detects within minutes forest fires during the early stages and monitors temperature, humidity and air pressure.

Key features

• Gas sensor

  The sensor includes a gas sensor that combines ultra-low-power Air Quality sensing with a precise gas sensing mode. Carbon monoxide, hydrogen and VOCs (Volatile Organic Compounds) are detected at the ppm level with built-in artificial intelligence to reliably detect a fire and avoid false positives.

• LoRa/LoRaWAN connectivity

  Connectivity to the Silvanet Network is provided by a LoRA-integrated radio to send and receive messages via a robust LoRaWAN mesh network within a large forest environment. This allows the sensors to connect to distributed LoRaWAN-enabled Silvanet Mesh Gateways and Border Gateways.

• Solar powered

  Power is supplied by a built-in solar panel to allow the sensor to run maintenance-free for 10 to 15 years without the need of batteries. As a precaution against the device itself starting a fire, it stores its energy in supercapacitors rather than batteries. This avoids the use of lithium and other toxic materials. A loop allows it to be attached directly to trees using tree nails or crop wire (temporary use only).

Delivered components

Sensors are delivered in boxes of 10 sensors. The following is a list of the delivered components for a single box:

• 10 sensors

• 10 tree nails

• 10 spacers

Dimensions and weight

The dimensions of the Silvanet sensor are (LxWxH) 19 cm x 9.11 cm x 1.34 cm. The sensor weighs 136 g.

Sensor range

Silvanet sensors can detect environmental air quality within a radius of 80 m to 100 m (260 ft to 320 ft) for a 60 minute detection time of a 2 m x 2 m fire. It can communicate with Mesh and Border Gateways that are 1 km away from the sensor, depending on topology.

BME688 gas sensor

The Silvanet Wildfire sensor uses the Bosch BME688 Gas Sensor to monitor the microclimate of the forest. It is a low-energy hydrogen sensor that detects the presence of a smoldering fire over distances of up to 115 m.

This gas sensor detects the presence of Volatile Organic Compounds (VoCs) and Volatile Sulfur Compounds (VSC) and detects these compounds at <20 ppm.

Environmental detection

The BME688 sensor can detect CO (Carbon Monoxide), H₂, (Hydrogen) and VOC (Volatile Organic Compounds). It measures the environment's temperature, humidity and air pressure.

The BME688 sensor monitors the microclimate of the forest by reading the following environmental values:

• Gas resistance: Outputs the IAQ (Index for Air Quality), VOC and CO2 equivalents (ppm). The gas scan result is a % value.

• Humidity: Outputs relative humidity

• Air pressure: Outputs pressure in Pa

• Temperature: Outputs temperature in Celsius

This gas sensor has built-in environmental sensing using Artificial Intelligence (AI) to generate a normalized value for the environment in which it has been placed.

Measurement conditions

In the Silvanet sensor, the BME688 sensor provides measurements under the following conditions:

• Pressure: 300 to 11000 hPa

• Humidity: 0 to 100%

• Temperature: -40 to 85°C

Power consumption

The Silvanet sensor has very low energy consumption which is provided by its built-in solar panel. As a precaution against the device itself starting a fire, the sensor stores its energy in supercapacitors rather than batteries.

The two key components that consume energy are the BME688 gas sensor and the LoRa radio. The gas sensor consumes ~3.9 mA in standard gas scan mode. The radio consumes 7W of energy to operate so it can continue to operate in a shaded location for ~6 hrs.

Power supply

The sensor requires sufficient energy to support continuous operation over a 24-hr. period while still having enough reserve power to support the heating and powering of the BME688 sensor and to send a burst of messages when a fire is detected.

Idle/active modes

Normally, the Silvanet Sensor Node is in idle mode. It activates every 60 seconds to read the air quality / gas sensor and every 2 hours to read the environment. It then sends a single packet to the Silvanet Server via the Mesh Gateway and Border Gateway. These data packets contain normal (non-fire detection) environmental data which includes temperature, pressure and humidity internal values. These values are displayed in the Site Management app.

Built-in solar panel

The sensor includes a 60 mm x 60 mm solar cell on its front housing. It continuously generates energy during the daytime and recharges the device with sufficient power for the next 24 hours. After sunset, it begins to discharge until sunrise. After sunrise, it begins to recharge to 100% within about an hour.

Each day the solar panel harvests 7 W in the constrained conditions of a forest.

Supercapacitor Energy Storage

The Silvanet sensor uses a set of supercapacitors to store energy for use by the radio, BME688 sensor and other components. It stores the energy for day-to-day tasks and has a reasonable amount of reserve power to operate the gas sensor and radio module in case a potential fire is detected. As supercapacitors have an expected lifespan of 10 years or more, the sensors are essentially maintenance free. This allows for an expected lifespan of the sensor between 10 and 15 years.

Supercapacitors have a high capacity but a small maximum voltage of 2.7 V so a series of capacitors are loaded to 4 V. Capacitors are fully charged to 5.2 V.

Sensor density

Sensor density is based on the level of human activity near a forest. This can include hiking paths, rail lines, roads, bridges, homes or other locations where it is expected humans have interactions (or not) with a forested area. Density can be determined by Density values per hectare.

Human activity level

When planning sensor deployments, the amount of human activity in an area determines the distances between sensors for that area:

• High human activity: 80 m to 100 m

  Dense deployment of sensors in high-risk areas. In areas with increased human interaction, plan for a dense deployment of sensors. We recommend a short distance between sensors - approximately 80 m to 100 m between sensors.

• Low human activity: 400 m to 500 m

  Sparse deployment of sensors in remote locations. In areas with little to no human activity, the distance between sensors can be increased to approximately 400 m to 500 m between sensors.

With this approach of variable density, the overall system cost can be reduced while maintaining an overall good wildfire detection time and rate.

Density values per hectare

More than 80% of wildfires can be attributed to human activity, depending on the region. Of most concern are the areas where wildland and urban areas intersect. This is known as the Wildland Urban Interface (WUI). A WUI is a location where burnable structures are interspersed within wildland fuels. A broader term is Wildland Industrial Interface where industrial infrastructure such as roads, power lines and railways intermingles with wildland fuels (source: Canada Wildfire).

Based on WUI, we recommend the following density values:

• (0.7/ha) Dense deployment: Sensors next to roads, campsites or parking lots: 0.7 devices per hectare.

• (0.1/ha) Sparse deployment: Sensors in remote locations such as deep within forests: 0.1 devices per hectare.