New recommendations for novel materials in renewable energy to secure long term reliability
Following the landmark climate accord in Paris, the world is looking to a rapid upscaling of wind and solar energy in the energy mix in the coming decade and beyond. This will only be made possible, says DNV GL, with concomitant developments in materials, including:
alternative semiconductor material in photovoltaics (e.g. halide perovskite)
new PV module coatings, materials and coatings for the harsh conditions of CSP and thermal energy storage
hybrid reinforcements of wind turbine blades
cheaper permanent magnets in gearless direct drive wind turbines
a range of innovative battery chemistries in energy storage systems.
For any of these novel materials to be commercially viable, they should not only offer a cheaper and better alternative to existing materials, but must also be readily available and reliable over long periods of time. “Trade-offs between availability, cost and performance may be made, but in all cases long-term reliability is a key requirement for materials used in the energy industry”, says Liu Cao, researcher at DNV GL Research & Innovation and lead author of the position paper. Materials reliability is mainly a function of long term degradation, which is difficult to model in service conditions and often not adequately assessed in the testing of systems. More specifically, DNV GL provides evidence for the following insights:
Single average degradation rate is an inadequate metric of long-term performance
Qualification tests are insufficient for lifetime assessment
Accelerated laboratory tests may not reveal all the degradation mechanisms
Real-time monitoring is valuable, but unable to predict lifetime alone
To address these challenges, DNV GL proposes the following:
Coupling empirical models to a fundamental understanding of degradation.
Transforming rich and increasingly ubiquitous sensor data into predictive models. DNV GL’s BatteryXT is an example of a predictive tool estimating battery performance and life based on abundant historical data and statistical analyses.
Deploying a Bayesian network approach to bring together diverse sources of knowledge of relevance to the performance and degradation of materials. For example, a Bayesian network model, MARVTM, has enabled the assembly of diverse data for pipeline risk assessment. A similar approach could be applied to risk assessment for renewable energy and energy storage systems.
