Nestor Francisco Guerrero-Rodríguez

Satellite observations reveal the offshore distribution of pelagic Sargassum, but predicting its occurrence at individual Caribbean coastlines remains challenging where local observational records are limited. This study tests whether forecasts developed on one Caribbean island can transfer to another without retraining or recalibration. Satellite-derived predictors from the 0–10, 10–50, and 50–100 km offshore bands were evaluated at one-, two-, and four-week horizons. Rolling-origin validation in Puerto Rico selected Random Forest models at one and two weeks and a one-dimensional convolutional neural network (CNN1D) ensemble with Platt calibration at four weeks. Frozen pipelines were applied unchanged at nine Dominican Republic sites and evaluated against 283 station-week targets. Puerto Rico 2026 provided independent testing. At one and two weeks, area under the receiver operating characteristic curve (ROC AUC) values were 0.745 and 0.757, area under the precision-recall curve (PR AUC) values were 0.689 and 0.687, and Brier Skill Score (BSS) values were +0.110 and +0.059. After transfer, ROC AUC values were 0.686 and 0.663, PR AUC values were 0.548 and 0.509, and BSS values were +0.098 and +0.056. Site removal did not eliminate short-horizon skill, and BSS remained positive by year. Transfer was clearest at one and two weeks, suggesting potential value for targeted coastal Sargassum monitoring, field verification, and response planning.

Reliable time-resolved prediction of hydrogen production is essential for proton exchange membrane (PEM) water electrolysis (PEMWE) under variable operation, yet predictive models are often assessed near the experimental domain used for development. This study introduces a progressive cross-campaign validation framework for dynamic PEM hydrogen prediction across increasingly different operating conditions. A high-resolution experimental dataset from a 100 kW PEM electrolyzer was used to compare static, temporal, physics-structured, and hybrid models under hierarchical sensor representations. Validation combined leave-one-experiment-out testing, transition-resolved analysis, repeated random seeds, an independent PHOEBUS cross-profile stress test, explicit high-current out-of-distribution (OOD) evaluation, and locked transfer to temporally separated 2018 campaigns without model retuning. The proposed physics-structured backbone (PSB) with guarded residual long short-term memory (LSTM), denoted PSB-ResLSTM-G, achieved a mean root mean square error (RMSE) of 0.156 m³ h⁻¹ versus 0.172 m³ h⁻¹ for the Direct LSTM and lower overall RMSE in all 20 matched fold-seed comparisons. Within the PHOEBUS high-current OOD subset, PSB-ResLSTM-G reduced RMSE from 1.559 to 0.237 m³ h⁻¹ (84.8%), with the applicability guard reverting the hybrid prediction to the PSB beyond the development-current support. In the locked 2018 confirmation, PSB achieved the lowest RMSE at 0.547 m³ h⁻¹, followed by PSB-ResLSTM-G at 1.344 m³ h⁻¹ and Direct LSTM at 3.257 m³ h⁻¹, revealing an accuracy–transferability trade-off under the larger cross-era shift. Overall, the proposed applicability-aware hybrid formulation and progressive validation framework jointly characterize dynamic accuracy, OOD behavior, and cross-era transfer within a 100 kW PEM electrolyzer platform.

Green hydrogen is commonly assessed as a renewable fuel or long-duration storage option, but its value as a critical-load restoration resource remains less developed, particularly when produced from curtailed renewable electricity. This study develops a planning-oriented framework to assess green hydrogen for critical-load restoration by linking renewable curtailment, proton-exchange membrane electrolysis, hydrogen storage, fuel-cell reconversion, critical Energy Not Served (ENS) reduction, economic valuation, and carbon-footprint savings. The framework is applied to the Dominican Republic power system as a representative insular case with rapid renewable expansion and limited flexibility. Using monthly preliminary real-operation reports from OC-SENI, the reference case considers 196.46 GWh/year of curtailed non-conventional renewable electricity in 2025, producing 3.78 kt H₂/year and 65.5 GWh/year of recoverable electricity. Under the reference screening assumptions, a 25 t H₂ storage module would provide 433.29 MWh of usable electricity, fully covering 6 h and 12 h restoration windows for the 30 MW illustrative critical-load case and reducing critical ENS by 60.2% during a 24 h event. The recovered electricity could avoid 43.5 ktCO₂/year under the SENI combined-margin grid-displacement case, with higher avoided operational emissions under the diesel-backup displacement sensitivity.

Nonlinear loads in modern buildings are often represented using balanced or frequency-domain approximations that do not fully capture phase asymmetry and time-domain distortion observed under field operating conditions. This study aims to develop and validate a measurement-driven phase-resolved modeling framework for unbalanced three-phase nonlinear loads and to evaluate its usefulness for harmonic-mitigation assessment using shunt active power filters (SAPFs). In situ voltage and current waveforms were acquired in a university building using a Class S power-quality analyzer and embedded in MATLAB/Simulink as time-domain signals to preserve harmonic content, waveform shape, and phase-dependent behavior. Model accuracy was assessed using RMSE and range-normalized RMSE, yielding a mean RN_RMSE of 1.36% across representative loads, with simulated THD_I deviations of approximately 1–3 percentage points relative to the measurements. The measured load signatures were then used to evaluate three SAPF configurations under representative load-specific operating conditions. The results show that mean THD_I was reduced from 39.25% before compensation to 2.28% after compensation, with all the post-filter phase values remaining below 5%, consistent with IEEE Std 519-2022. The findings show that phase-resolved measurement-based load models provide a practical basis for harmonic assessment and SAPF-oriented mitigation studies in low-voltage building distribution systems with heterogeneous and unbalanced nonlinear loads.

High-temporal-resolution meteorological data collected under tropical climatic conditions are scarce yet highly valuable, because they facilitate a more precise characterisation of atmospheric variability. This paper presents a high-temporal-resolution meteorological database collected in Santo Domingo, Dominican Republic, using a Davis Vantage Pro2 Plus weather station. The database encompasses 170,861 observations covering 35 meteorological parameters, all of which were acquired at a temporal resolution of one minute. The database was prepared through a rigorous process of data organisation and cleaning. In the initial stage, NaN values representing missing data were identified and removed. In the second stage, based on the temporal and statistical characteristics of the database, missing values were reconstructed using an imputation technique. This approach ensures data quality and consistency, which are essential for the reliable utilisation of the database in scientific, energy-related, and engineering applications. The database is a valuable resource for the development of predictive tools based on machine learning and deep learning techniques to forecast solar radiation and wind speed. Additionally, it can serve as a useful input for the modelling and simulation of renewable energy systems.

This article proposes a model to estimate the optimal cutting speed and depth of cut used in the machining process by chip removal during the turning operation, considering the structural integrity of the workpiece to be machined. The structural integrity model is proposed considering the rounding of the cutting tool nose as a measure of roughness requested in the workpiece, the electrical power capacity delivered by the machine tool motor as a load-limiting factor for the process, the geometry of the desired workpiece, and the physical machining parameters given by cutting tool manufacturers. Based on these criteria, an estimation algorithm is proposed that integrates these parameters and executes the search for the optimal cutting depth and cutting speed, meeting the structural integrity criterion in accordance with the minimum machining time criterion in the turning process, establishing a balance between process reliability and minimization of machining time. The proposed model is innovative since it presents a new methodology to determine the depth of cut and calculate the machining speed under the criterion of preserving the structural integrity of the piece to be machined to the maximum. This means that the depth of cut and spindle speed estimated under the proposed methodology guarantee that during the machining process, the workpiece will not suffer structural damage from the cutting forces involved in the machining process, minimizing the effects of loading in areas of stress concentration, thereby contributing to highlighting and involving the concept of process reliability. This model provides a new theoretical method for technologists involved in the calculation of the machining process, offering them a theoretical basis for their proposals for depth of cut and cutting speed.