Energy Digest

Probabilistic modeling of the power system restoration process is essential for resilience planning and operational decision-making. Researchers developed data-driven models of the restoration process using outage data from four distribution utilities, combining three components: restore time progression, total duration, and the time to first restore. The models use a Beta distribution for restore time progression, Lognormal process for total duration, and Gamma model for the time to first restore, providing an end-to-end stochastic model for simulation and decision support.

LENORI (Large Event Number of Outages Resilience Index) is a metric developed to measure distribution system resilience based on forced line outages in large extreme events, calculated from standard utility outage data. LENORI takes into account statistical accuracy by logarithmic transformation of outage data. Two related metrics, LENORI and ALENO (Average Large Event Number of Outages), are also introduced to quantify the power grid's strength relative to extreme events.

Our novel computational method for unit commitment expands to include long-horizon planning, solving problems with thousands of generators at 5-minute market intervals in approximately 1 minute on commodity hardware. The method achieves sizable operational cost savings and is infeasible for current state-of-the-art tools due to its scalability. It can be implemented using existing continuous optimization solvers and adapted for different applications.

Demand response models extend to a stochastic framework where customer response is represented by price-dependent random variables, addressing epistemic uncertainty with stochastic pricing strategies. These strategies aim to compensate for estimation errors and uncertainty in customer response. The proposed approach demonstrates stability and near-optimality through theoretical results and numerical simulations.

A deep learning-driven inverse design methodology is used to create a Doherty power amplifier with a multi-port pixelated output combiner network, achieving an extended efficiency range and delivering high output power at 2.75 GHz. The amplifier prototypes demonstrate maximum drain efficiencies exceeding 74% and deliver over 44 dBm of output power, while maintaining linearity and efficiency under realistic signal conditions. The designs also achieve high average power added efficiency and adjacent channel leakage ratio, making them suitable for 5G new radio applications.