http://hdl.handle.net/10174/37664 2026-05-14T13:55:22Z http://hdl.handle.net/10174/42008 Title: Spatiotemporal assessment of wildfire smoke exposure using a low-cost air quality monitoring system in a developing Amazonian city Authors: Almeida, Domingas de Oliveira; Duarte, Edicle de Souza Fernandes; Gomes, Ana Carla dos Santos; Batalha, Sarah Suely Alves; Mandú, Tiago Bentes; Nascimento, Fernanda Souza do; Silva, Glauce Vitor da; Costa, Maria João Abstract: Biomass burning is the dominant source of seasonal air pollution in the Amazon, yet local-scale exposure remains poorly characterized due to sparse monitoring. This study aims to quantify the spatiotemporal dynamics and drivers of PM2.5 pollution in Santarém, Brazilian Amazon, by integrating measurements from a dense network of low-cost sensors, satellite-derived fire radiative power (FRP), and reanalysis meteorology throughout 2023. We applied Generalized Estimating Equations (GEE) to evaluate the daily influence of fire activity and meteorological conditions on local PM2.5. Mean PM2.5 concentrations increased from ∼5 μg/m3 in the rainy season to ∼16 μg/m3 in the dry season, with 94% of exceedances occurring from July–December and a fine-particle dominance (PM2.5/PM10 ≈ 0.79). Peri-urban communities experienced earlier-season pollution peaks, whereas the urban core showed more persistent late-season accumulation. FRP emerged as the primary driver of PM2.5, with effect sizes strengthening from 10% (wet season) to 25% (dry season) per standard deviation, while meteorological factors such as wind speed and boundary-layer height played secondary but modulating roles. A negligible weekend–weekday contrast confirmed that smoke overwhelmingly dominates over local traffic emissions. Finally, we operationalized these relationships into a low-computational-cost FRP–Meteo–PM2.5 polar radar tool for identifying high-risk smoke transport corridors. These results provide actionable evidence for early warning and highlight the urgent need for targeted fire management to reduce public health risks in developing Amazonian cities. 2026-03-01T00:00:00Z http://hdl.handle.net/10174/42000 Title: Predicting Asthma Hospitalizations from Climate and Air Pollution Data: A Machine Learning-Based Approach Authors: Reis, Jean Souza dos; Duarte, Edicle de Souza Fernandes; Costa, Rafaela Lisboa; Silva, Fabricio Daniel dos Santos; Cortes, Taisa Rodrigues; Coelho, Rachel Helena; Velasco, Sofia Rafaela Maito; Neves, Danielson Jorge Delgado; Filho, José Firmino Sousa; Barreto, Cairo Eduardo Carvalho; Cabral Júnior, Jório Bezerra; Reis, Herald Souza dos; Mendes, Keila Rêgo; Lins, Mayara Christine Correia; Ferreira, Thomás Rocha; Vanderlei, Mário Henrique Guilherme dos Santos; Alonso, Marcelo Felix; Mariano, Glauber Lopes; Gomes, Heliofábio Barros; Gomes, Helber Barros Abstract: This study explores the predictability of monthly asthma notifications using models built from different machine learning techniques in Maceió, a municipality with a tropical climate located in the northeast of Brazil. Two sets of predictors were combined and tested, the first containing meteorological variables and pollutants, called exp1, and the second only meteorological variables, called exp2. For both experiments, tests were also carried out incorporating lagged information from the time series of asthma records. The models were trained on 80% of the data and validated on the remaining 20%. Among the five methods evaluated—random forest (RF), eXtreme Gradient Boosting (XGBoost), Multiple Linear Regression (MLR), support vector machine (SVM), and K-nearest neighbors (KNN)—the RF models showed superior performance, notably those of exp1 when incorporating lagged asthma notifications as an additional predictor. Minimum temperature and sulfur dioxide emerged as key variables, probably due to their associations with respiratory health and pollution levels, emphasizing their role in asthma exacerbation. The autocorrelation of the residuals was assessed due to the inclusion of lagged variables in some experiments. The results highlight the importance of pollutant and meteorological factors in predicting asthma cases, with implications for public health monitoring. Despite the limitations presented and discussed, this study demonstrates that forecast accuracy improves when a wider range of lagged variables are used, and indicates the suitability of RF for health datasets with complex time series. 2025-01-24T00:00:00Z http://hdl.handle.net/10174/41963 Title: Impact of Measurement Noise and Fitting Window Placement on Single-Diode PV Parameter Extraction Authors: Mesbahi, Oumaima; Afonso, Daruez; Janeiro, Fernando M; Grilo, Frederico; Tlemçani, Mouhaydine Abstract: The problem of photovoltaic (PV) cell degradation can affect the shape of the I-V curve, which can lead to variations in the five parameters of the PV cell. This is the motivation behind the importance of knowing and extracting these parameters. The process starts by the measuring the output current and voltage (I-V curve) then applying a best fit to obtain the parameters. Both the noise of the instruments used for measurement and the size of the measured window can affect the accuracy of the obtained parameters. This paper presents a study about the effects of both the noise of instruments and the interval size. Varying the RMS of the noise of both current and voltage from 1 to 10%, the parameters are extracted from two case studies, first one starting the interval from the short circuit coordinates and the second one from the open circuit voltage, the size of the intervals are increased till reaching the whole curve. Results demonstrated that to obtain optimized parameters a 40−60% segment of the I-V curve should be measured staring from Voc region. 2025-10-21T23:00:00Z http://hdl.handle.net/10174/41962 Title: Automated Detection of Aircraft Surface Defects Using Deep Learning with Integrated Human Validation Authors: Mesbahi, Oumaima; Chabane, Souhila; Pereira Santos, Nuno; Del Pino Lino, Adriano; Tlemçani, Mouhaydine; Lourenço Da Saúde, José Manuel Abstract: Visual inspection of aircraft surface is one of the many steps in the maintenance routines. Usually performed by operators, this procedure might last days to be accomplished. The use of automated process can help reduce time and results in accurate detection of surface defects on aircraft, as they are vital to maintain structural soundness and flight safety. This paper proposes a deep learning framework for automated defect detection based on Faster R-CNN with ResNet-50 Feature Pyramid Network (FPN) as the backbone model. This model was trained and validated on a sizable, labeled aircraft images with a maximum F1-score of 0.555 achieved in the test set. This is the result of preliminary study, where the authors aimed to detect all types of defects without classification. To further enhance reliability and allow for human input, a custom annotation validation user interface was implemented via Python, which allowed aircraft inspectors to view, edit, add, and acknowledge predictions made by the model in an attempt to hold onto precise level of annotation. This system also facilitated the management of annotations, visualization on irregular aircraft zones, and the creation of reports thus allowing for inspection workflows. The results show that combining state-of-the-art object detection with domain expertise in validation as route to reliable semi-automatic, standards-compliant aircraft defect detection is plausible. Future work will involve expanding the dataset, tuning for accuracy, and incorporating human feedback for enhancement of model utility over time. 2025-10-21T23:00:00Z