Research

Applying time-series hosting capacity analysis to a representative sample of distribution feeders, we find that voltage constraints, not thermal constraints, are the binding limit on distributed solar hosting capacity for the large majority of feeders analyzed, while EV charging hosting capacity is more evenly split between voltage and thermal limits depending on charging load coincidence with existing feeder demand.

We model energy intensity and cost across liquid solvent and solid sorbent DAC process pathways under varying heat source assumptions, finding that heat source carbon intensity, not just process energy efficiency, is often the dominant factor in a DAC facility's net carbon removal effectiveness, a distinction that published cost figures frequently obscure.

Using coupled aero-hydro-servo-elastic simulation across representative metocean conditions, we compare structural load response across semi-submersible, spar, and tension-leg floating wind platform archetypes for a common reference turbine. Tension-leg platforms show the lowest platform motion but the highest peak mooring line tension, while semi-submersible platforms show the most balanced load profile across the operating envelope tested, a finding relevant to platform selection for a given site's metocean severity.

Analyzing field performance data across a sample of operating grid-scale storage installations, we find that measured capacity degradation after five years of operation tracks reasonably close to vendor-projected curves for LFP systems, but shows wider variance for NMC systems, driven primarily by differences in thermal management design and average state-of-charge operating band rather than cycling frequency alone.

We compare three machine learning approaches, gradient-boosted trees, a recurrent neural network, and an unsupervised anomaly detection model, for predicting wind turbine gearbox failure using SCADA and vibration data. The unsupervised anomaly detection approach achieved the best lead-time-to-failure performance despite requiring no labeled failure examples, suggesting it is the more practical starting point for operators with limited historical failure data, a common constraint across most operating wind fleets.

This paper presents a techno-economic model of grid-scale PEM electrolysis across a range of electricity price and utilization scenarios. We find that utilization rate and power price together explain the overwhelming majority of levelized hydrogen cost variance across the scenarios modeled, more than electrolyzer capital cost itself, reinforcing that power procurement strategy, not equipment selection, is the primary lever available to developers seeking cost-competitive green hydrogen in the near term.