Advanced and Intelligent Sensor Nodes and Awareness Platforms

Leveraging smart computing technology and combining it with artificial intelligence can convert virtually any data acquisition node from a passive monitoring and data transfer device - to an informed, adaptive, and intelligent decision tool for both ignition risk mitigation and for overall power system resilience.

A major benefit of (smart computing at the asset) is that in many cases, the collected data never has to move, and the computing device could be as simple and low cost as a smart phone running some AI enriched apps. Moreover, those apps could be modified as new learnings and improved algorithms are discovered.

A simplified example of smart computing enabled edge processing from Dryad Networks, was documented in the wildfire risk reduction technology catalog. The Dryad product detects smoke and it’s AI chip (made by Bosch) claims to be the first gas sensor with artificial intelligence, and integrated pressure, humidity, and temperature sensors. The gas sensor is designed for low power consumption, mobile or connected applications, and it can detect seven unique gasses and compounds. More expensive and more computationally powerful examples include hardware such as the NVIDIA Jetson application.

Toward the wildfire objectives, use cases of interest for smart computing enabled applications include:

• Smoke Detecting Camera Nodes and Weather Stations – for Environmental Monitoring
• Adaptive Distribution Automation and Protection Hardware – For Ignition Risk Mitigation
• Smart and Risk Aware Power System Assets – For Weather Informed Optimization

Smoke Detecting Camera Nodes and Weather Stations for Environmental Monitoring – This particular use case has many stakeholders – most prominently, land managers. The key opportunity here is to use the cameras to conduct multi-purpose monitoring applications such as vegetation health and fire spread analytics. Another very compelling use case is areas with limited communications capacity. Typically smoke detecting cameras require a reasonably high bandwidth data stream but the edge computing use case could change the paradigm. This use case was a discussion point of emphasis and of repeat conversation with the Wildfire Advisory Group, where fire awareness data would be collected at a remote location, but the data is processed in-situ and doesn’t need to move - unless requested. The tie-in is that in areas where communications are marginal and low bandwidth, the information gets processed into a low-resolution text which makes it easier to transmit to the response entity.

Adaptive Distribution Automation and Protection Hardware for Ignition Risk Mitigation – This use case describes the edge computing at a protective device, but the device could be any distribution asset that needs to respond, protect, or adjust its settings. Therefore, this discussion applies to many other risk aware power system assets. The most prominent example here is the DA device that has a role in isolating a portion of a circuit to perform a public safety power shutoff PSPS during extreme fire weather. By integrating a local micro-weather station and a smart power quality analyzer, the PSPS time could be reduced considerably and, in some cases, avoided altogether. In a different variation, the same intelligence could be used to determine when it was time to bypass the DA device in favor of a switch configurable fault current limiting fuse. These decisions and choices could be well informed and enhanced by integrating localized wind and weather data, vegetation condition data, and by the historical fault data for that particular circuit.

What is needed to accelerate the industry toward the 2030 Vision?

The 2030 consensus vision from the wildfire advisory group includes power system protective devices that are more intelligent, more adaptive to fire weather and more tightly coordinated and the advisory group has interested in demonstration of technology that can begin to fulfill these requirements. The major research challenge that can be supported by both EPRI and National Lab research and development include faster acting protective devices and demonstrations with protective devices that have adaptive settings based on their understanding of the localized fire risks. Another priority was to develop innovative ways to get actionable insights from areas with limited communications, which is an important objective for high fire threat areas that are remote and difficult to access.

Which 2030 Future States are Impacted by this Work?

• AI or smart grid enabled monitoring and sensors at all relevant protective and transition nodes of interest without the need for data to be transferred.
• A comprehensive selection of fault energy limiting technologies
• Smart and risk aware interface and decision support for protective and sectionalizing devices – leveraging AI driven adaptive protection (i.e. one-shot reclose vs PSPS)
• Full understanding of vegetation ignition probability as fault energy is reduced