AIRU-WRF: A Physics-Guided Spatio-Temporal Wind Forecasting Model and its Application to the U.S. North Atlantic Offshore Wind Energy Areas

Abstract

The reliable integration of wind energy into modern-day electricity systems heavily relies on accurate short-term wind forecasts. We propose a spatiotemporal model called AIRU-WRF (short for the AI-powered Rutgers University Weather Research & Forecasting), which fuses numerical weather predictions (NWPs) with local observations in order to make wind speed forecasts that are short-term (minutes to hours ahead), and of high resolution, both spatially (site-specific) and temporally (minute-level). In contrast to purely data-driven methods, we undertake a "physics-guided" machine learning approach which captures salient physical features of the local wind field without the need to explicitly solve for those physics, including: (i) modeling wind field advection and diffusion via physically meaningful kernel functions, (ii) integrating exogenous predictors that are both meteorologically relevant and statistically significant; and (iii) linking the multi-type NWP biases to their driving meso-scale weather conditions. Tested on real-world data from the U.S. North Atlantic where several offshore wind projects are in-development, AIRU-WRF achieves notable improvements, in terms of both wind speed and power, relative to various forecasting benchmarks including physics-based, hybrid, statistical, and deep learning methods.

Seizing Opportunity: Maintenance Optimization in Offshore Wind Farms Considering Accessibility, Production, and Crew Dispatch

“…The maintenance crew, once dispatched, is occupied until the maintenance is completed or their shift ends by the time of last sunlight, t D , as shown in (29). An upper bound on the number of crews is set to N x (crews), as shown in (30).…”

“…The processed datasets have an hourly resolution, spanning from August 12, 2019 to April 29, 2020. Scenarios for all aforementioned variables are generated using a probabilistic forecasting framework proposed in [22] based on temporal Gaussian Processes [29], [30]. The framework generates realistic scenarios that naturally encode the inherent temporal dependencies of the uncertain input parameters.…”

Defect Detection in Solar Photovoltaic Systems Using Unmanned Aerial Vehicles and Machine Learning

“…According to a recent report, AI/ML technologies are expected to contribute towards lowering global greenhouse gas emissions by as much as 4% by 2030, which is roughly equivalent to 2.4 Gigatons of CO 2 emissions (Herweijer et al, 2020). Examples of emerging applications where AI/ML can have an outsized impact in the renewable energy sector Proceedings of the 8 th North American International Conference on Industrial Engineering and Operations Management, Houston, Texas, USA, June 13-16, 2023 © IEOM Society International include fault diagnostics and prognostics (Zhao et al, 2014, Rao et al, 2019, smart and predictive maintenance (Papadopoulos et al, 2021 and, production control and optimization (Papadopoulos, 2023), weather and energy forecasting (Ezzat, 2019, Ye et al, 2023, electric load prediction (Sajjad et al, 2020), among others.…”