-- These projects are inactive --
Through the Solar Energy Evolution and Diffusion Studies, or SEEDS, program, seven projects are investigating strategies to accelerate the pace of change for solar energy technologies as they are developed and deployed. The projects integrate the use of cutting-edge analytical and computational tools with real-world market data and pilot tests to speed the pace of solar energy becoming cost competitive.
On January 30, 2013, the U.S. Department of Energy (DOE) announced $9 million over three years to fund these projects. On October 20, 2016, DOE announced the second round of SEEDS.
Awardees
TOPIC 1: ACCELERATING SOLAR TECHNOLOGY EVOLUTION
Technology evolution is the complex process by which a technology increases in performance and functionality and decreases in cost over time. Over the last 40 years, solar energy technologies have followed a particularly simple trend of cost reductions as production has ramped up. The simplicity of this trend conceals the rich set of coupled dynamics behind human learning, effort, and experience. These three projects will provide a more rigorous understanding of solar technology evolution in order to uncover key levers for accelerating progress.
Evaluating the Causes of Photovoltaics Cost Reduction: Why is PV different?
Principal Investigator: Jessika Trancik
Massachusetts Institute of Technology, Cambridge, Massachusetts
For half of a century, photovoltaics—which convert solar radiation into electricity—have experienced a 20% cost reduction every time cumulative production doubles. This rate of technological improvement is unmatched by other competing energy technologies. Yet, there exists no practical explanation for this empirically observed trend. This project will investigate a wide range of hypotheses for describing the technology evolution process to form an overarching theory that can be applied to accelerate cost reductions in solar.
Helios: Understanding Solar Evolution through Text Analytics
Principal Investigator: Jeffrey Alexander
SRI International, Menlo Park, California
Machine learning is a powerful computational tool that underlies an array of complex tasks from weather prediction to email spam filtering. By reading and organizing a massive amount of data from disparate sources, the tool can uncover and exploit buried patterns in the data. For this project, SRI will develop the Helios platform to detect unseen technological breakthroughs and bottlenecks for solar technologies by analyzing decades' worth of scientific publications and patents.
Forecasting and Influencing Technological Progress in Solar Energy
Principal Investigator: Deborah Strumsky
University of North Carolina at Charlotte, Charlotte, North Carolina
Researchers from University of North Carolina at Charlotte, Arizona State University, and University of Oxford aim to make better forecasts of future cost reductions for new energy technologies. By abstracting the process of technological progress as a time-varying network of interconnected components, the team will employ a new theory to help inform optical investment strategies in the research and development pipeline. This will help the industry more quickly and cost-effectively achieve technical and deployment goals. By analyzing hundreds of years' worth of patent data and historical cost and production data, the team will construct a network—called a "technology ecosystem"—to forecast and influence technological progress.
TOPIC 2: ACCELERATING SOLAR TECHNOLOGY DIFFUSION
Innovation diffusion theory seeks to explain how, why, and at what rate new ideas, products, and technologies spread through society. Key results from innovation diffusion theory are often generalizations of empirically-observed qualitative trends. Within the past decade, the availability of large data sets and computational tools allows for quantitative analysis of diffusion trends with unprecedented precision and scale. Numerical modeling can reveal the complex processes underlying the spread of technology. These four projects will provide a more rigorous understanding of solar technology diffusion to uncover key mechanisms for accelerating solar adoption.
Understanding the Evolution of Customer Motivations and Adoption Barriers in Residential Photovoltaics Markets
Principal Investigator: Easan Drury
National Renewable Energy Laboratory (NREL), Golden, Colorado
Decisions of whether to adopt rooftop solar are driven by many factors, including system price, access to information, and the experiences of peers within social networks. Analytical models for projecting market growth typically only account for the price variable, not accounting for other important decision parameters. An agent-based model can elucidate the ways that photovoltaic markets will evolve by simulating the complex and interrelated decision dynamics of individuals in a social system. This project, led by NREL and in collaboration with researchers from Portland State University and the University of Arizona, will collect a rich dataset that will be used in training regionally-specific agent-based models, simulating decision-making under different test scenarios, and running real-world pilot tests to validate results with market partner Clean Power Finance.
Design of Social and Economic Incentives and Information Campaigns to Promote Solar Technology Diffusion through Data-Driven Behavior Modeling
Principal Investigator: Kiran Lakkaraju
Sandia National Laboratories (SNL), Albuquerque, New Mexico
The project team, including researchers from SNL, University of Pennsylvania, CCSE, Vanderbilt University, and NREL, will develop an approach for designing social and economic incentives to enhance solar diffusion. The team will integrate individual-level data about solar adoption patterns with data generated from lab and field experiments and surveys to construct an individual-level computational model to predict solar adoption. They will use this model to drive agent behavior in an agent-based simulation, with agents' interactions captured using social influence variables. The team will use the agent-based model to forecast adoption patterns in response to alternative policy interventions, and use computational optimization techniques to arrive at candidate policies.
An Emergent Model of Technology Adoption for Accelerating the Diffusion of Residential Solar PV
Principal Investigator: Varun Rai
The University of Texas at Austin, Austin, Texas
The goal of energy market transformation is to identify and target existing market barriers. Comprehensive analysis is needed to quantify peer influences, information dissemination, and other factors that, in addition to technical and financial factors, are barriers to major adoption. For this project, university researchers will be the visiting scholars, or scholars in residence, hosted by six Texas electric utilities. Researchers will have access to rich datasets that will enable them to analyze real-world market barriers. The results will inform which pilot projects the researchers will run with partner utilities.
The Influence of Novel Behavioral Strategies in Promoting the Diffusion of Solar Energy
Principal Investigator: Kenneth Gillingham
Yale University, New Haven, Connecticut
Communities across the U.S. are instituting Solarize programs in which homeowners band together to collectively purchasing rooftop solar systems. This project will quantify the effectiveness, cost-effectiveness, and scalability of this and other strategies that leverage social interactions to accelerate diffusion of solar technologies. Researchers from Yale University and New York University will design and run a series of randomized field trials to study a new Solarize Connecticut program. The field tests will be implemented with partner SmartPower.
Additional Resources
The project teams met during the Big Data Projects on Solar Technology Evolution and Diffusion: Kickoff Meeting on July 15, 2013.
SEEDS is one of several projects DOE is funding to fuel innovation and reduce the hardware costs and non-hardware costs of solar technology.
Learn more about SETO's other soft costs funding programs.