GOALI: Collaborative Research: Generation versus Degradation: Striking the optimal balance for wind farm profitability via digitization, predictive and prescriptive analytics

The rapid increase in scale and sophistication of wind farms poses a critical challenge relating to the cost-effective management of wind energy assets. A defining characteristic of this challenge is the economic trade-off between two concomitant processes: electricity generation (the primary driver of short-term revenues) and asset degradation (the major determinant of long-term expenses).

This NSF project aims to formulate a decision-theoretic approach to jointly optimize generation and maintenance in wind farms. The project will bring transformative change into the status-quo of asset management in the wind industry which, to-date, relies on single-faceted strategies that largely overlook the dependencies between the generation and degradation in wind turbine assets.

The intellectual merits of the project include the formulation of novel data and decision science models, blended within a digitization platform, to predict and co-optimize operations and maintenance requirements. The broader impacts of the project include disseminating research findings via coursework, publications, data/software, and industry-academia workshops.

A set of use case demonstrations, co-developed with industrial partners, will accelerate the translation of scientific knowledge into tangible industrial impact, in a step towards meeting the 35%-by-2050 U.S. wind energy target. Summer internships and undergraduate researchers will contribute towards educating the next-generation workforce in data, decision, and energy sciences.

Without formally considering the intrinsic dependencies between electricity generation and asset degradation, wind farm operators reap sub-optimal benefits from their operations and maintenance policies. This project aims to formulate a decision-theoretic framework which seeks an optimal balance of how wind loads are leveraged to harness short-term generation revenues, versus alleviated to hedge against longer-term maintenance expenses.

The framework comprises decision-aware predictive models for power and asset health degradation forecasting, integrated within mixed integer programs with decision-dependent uncertainty. New reformulations and constraints will ensure an effective predictive-prescriptive coupling, thereby enabling the optimization to search within the prediction space for an optimal prediction-decision pair.

An end-to-end digital twin of the wind farm will bind the proposed predictive-prescriptive methodologies within an integrative asset management solution. Engagement of industrial partners and a national laboratory will ensure a sensible impact on asset management in the wind industry.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.