A new ‘home’ for data tech research on environmental issues in the electric power sector
Wind turbines can kill migratory birds and bats. But did you know that new technology that pairs automated cameras with machine learning makes it possible for wind turbines to “see” approaching eagles and turn turbines off or away to reduce the number of bird-power collisions? This is just one of over a dozen new data technologies[1] we have learned about while working with the Electric Power Research Institute (EPRI).
Data tech can help deliver better, faster conservation outcomes at scale. We collaborate with EPRI to understand and improve the power sector’s use of data science and data tech for managing environmental compliance issues related to wastewater, habitat impacts, bird, bat and fish interactions and wetland delineation. Over the past fifteen months, we have undertaken a series of projects together with EPRI, the results of which have been made publicly available through a new EPRI website. Research and reports there offer the electric power sector, data tech vendors, regulators and universities a view into how data tech is being employed, adopted and implemented for environmental issues in American electric power:
- Our summary of the current state of the electric power sector’s applications of data tech shows that some utilities are pushing the boundaries of what is possible, while others are still working with spreadsheets, or even pen-and-paper. Overwhelmingly, where data science is being applied it is helping utilities to be proactive and reduce costs. But compliance rules can be out of sync with new tools and technology and sometimes there is a disincentive for companies to use or even test new tech.
- A 2-part literature review looks at environment, health and safety programs and use of data tech by electric utilities (you can find Phase I here and Phase II here), from remote sensing, AI, camera traps, and acoustic recordings to assess species presence and movement; neural networks for environmental issues around coal combustion; and AI and machine learning for groundwater contaminant modeling. The review is a snapshot of reports and publications that documents dozens of uses or tests of new technology in the past decade. Based on publication volume, there is broad spread of and investment in artificial intelligence (AI) for the electric power industry. AI is being used in a variety of applications, such as for identifying components of coal ash, fish identification at hydropower facilities, wetland classification, and water quality assessment, and it is often combined with big data generated from emerging hardware such as unmanned aircraft systems (UASs), acoustic cameras and sonar, and satellites. In contrast, areas such as energy economy modeling show little uptake or experimentation with data tech and literature was difficult to find. If you’re interested in learning more about specific applications we found through the lit review, and you can’t find what you’re looking for, send us a note and we can help you find it.
- A year-one project update summarizes the year’s findings and describes the activities of a workshop we held in the fall of 2019. The workshop brought together 35 people in Washington, D.C. to learn about emerging data technologies, share case studies and challenging questions and connect with each other to explore opportunities for future collaboration. A (virtual) follow-on workshop will focus on technologies that are helping researchers and utilities address species impacts throughout the power sector permitting, pre-construction and operations & maintenance phases.
Utility staff and leaders documented many barriers to tech adoption which we found in the literature and in our interviews. There are seven pervasive categories of barriers to adoption of new data tech:
- Cost
- Regulatory
- Technology barriers (e.g. the accuracy and customization capacity of the technologies)
- Psychological (i.e. human resistance to change)
- Organizational/Structural
- Data collection, curation and security
- Market/economic
Additional work on these barriers will be available soon. Please take a look around this new online home for data tech resources and let us know if we can help you find anything!
[1] Data science technology (“data tech”) refers to the technologies and applications used to collect, analyze, transact, manipulate and visualize big data using both hardware and software. Hardware includes equipment like sensors, drones, computers and tablets, and software includes the tools and programming for machine learning and artificial intelligence to take in, store, evaluate, and analyze the data.
