This thesis investigates the structural performance of reinforced concrete (RC) columns retrofitted with externally bonded (EB) and near-surface mounted (NSM) glass-fiber reinforced polymer (GFRP) systems. Six columns were tested in four-point static bending, and another set was tested in axial compression to evaluate the effects of corner geometry, circularization, and strengthening scheme. Finite element models developed in Ansys 2024 R2 reproduced experimental behavior and were used to generate full axial-flexure (PM) interaction diagrams. Theoretical models were used to validate both the numerical and experimental results. The developed GFRP retrofit schemes were compared to alternatives such as steel jacketing and NSM steel rebars, considering differences in structural performance and material cost. The study demonstrates that retrofitting using GFRP can significantly improve the capacity of RC members while promoting durability and sustainability, contributing to more resilient built environments.

Low-cost sensor (LCS) networks o[er scalable air quality monitoring, but their accuracy is frequently hindered by physical artifacts such as humidity-driven hygroscopic aerosol growth and internal thermal inflation. While machine learning calibration against reference monitors improves data quality, the need for prolonged colocation and scarce reference monitors limits scalability in resource-constrained areas like the Philippines. This study addresses these operational barriers by optimizing colocation periods, evaluating data partitioning strategies, and testing model transferability across highly divergent urban micro-environments.
An empirical evaluation established a parsimonious 3-feature input configuration (native PM2.5, temperature, and relative humidity) as the optimal architecture, successfully capturing diurnal thermodynamics while avoiding the detrimental multicollinearity introduced by explicit temporal covariates. Utilizing a random block partitioning strategy, continuous models like Gaussian Mixture Regression (GMR) and Support Vector Regression (SVR) outperformed tree-based ensembles. By mapping fundamental physical relationships rather than memorizing temporal noise, this approach preserved peak localized accuracy (e.g., S04 MAE = 3.29 μg/m3) while softening the mathematical divergence typically experienced during extrapolation.
Temporal optimization revealed that maximizing training data volume actively degrades predictive accuracy due to environmental concept drift. Optimal algorithmic stability and resource efficiency were achieved by strictly bounding the colocation training window to approximately 31 days, demonstrating that multi-month colocations are operationally inefficient.
To evaluate spatial generalizability, a cross-deployment validation matrix was constructed across a university campus network and an external Environmental Management Bureau (EMB) station in Makati. The empirical results definitively demonstrated the physical limits of algorithmic extrapolation; models suffered notable predictive degradation during both intra- network campus transfers and external deployment (e.g., models transferred from S02 yielded R2 values as low as -45.40). Consequently, this study demonstrates the significant limitations of relying on universal “golden” models or grouped regional baselines in highly heterogeneous urban environments. Instead, the research confirms that targeted, site-specific local calibration is an operational necessity. When the localized SVR architecture was applied independently at the Makati node, it successfully corrected meaningful systemic bias against a regulatory-grade Teledyne T640 monitor, achieving an R2 of 0.92 and an MAE of 1.23 μg/m3. Ultimately, this research transitions LCS calibration into a highly scalable, resource-efficient engineering protocol capable of supporting widespread, localized network deployment.

Domestic wastewater remains a major contributor to water pollution in the Philippines, with sanitation coverage still alarmingly low despite the enactment of the Philippine Clean Water Act of 2004. The 2022 Philippine National Demographic and Health Survey reported that only 5.6% of households availed septage treatment services, while only 3.1% are connected to a sewerage system. Delays in the expansion of sanitation infrastructure are largely due to limited technical and financial capacity of implementers and the perception of wastewater treatment as a low demand, yet expensive investment. This study presents a practical decision support system that addresses these barriers by helping implementers identify optimal sites for domestic wastewater treatment plants (WWTPs). The combination of Multicriteria Decision Analysis (MCDA) and Geographic Information System (GIS) provided an approach that balances cost-efficiency, hazard exposure, and compliance to public and environmental health regulations. Inputs from sanitation specialists, national government agencies (DPWH, DILG, DENR, LWUA), and key stakeholders (PAWD, LGUs, Water Districts) were critical in establishing a site suitability criteria specific for domestic WWTP. The developed suite of decision support tools includes: 1) an ArcGIS toolbox that automates geoprocesses to produce suitability maps, streamlining workflows and reducing manual errors and processing time; and 2) a macro-enabled weighting and site evaluation toolkit that guides users in setting priorities, deriving weights, comparing site options, and generating detailed site suitability reports that highlight low-scoring factors for mitigation. The results emphasize the importance of giving higher weights to technical factors to ensure cost-effective solutions, while still incorporating hazard exposure and public and environmental health considerations. The developed system simplifies complex decision-making and provides compromise in evaluating conflicting objectives. It facilitates faster and technically grounded site evaluation for WWTPs, enabling implementers to plan sanitation projects with transparency and efficiency, and ultimately find economically strategic locations.

Nickel-iron (Ni-Fe) batteries are emerging as a safer and more sustainable alternative to lithium-ion batteries, due to their aqueous electrolytes, long cycle life, robustness, and use of earth abundant materials. However, their performance is critically limited by the iron anode, which suffers from poor conductivity, sluggish redox kinetics, electrode surface passivation, and parasitic hydrogen evolution reaction (HER). In this work, a dual strategy approach was employed to overcome these challenges. First, Fe3O4 nanoparticles were hydrothermally grown on acid-treated multi-walled carbon nanotubes (Fe3O4@MWCNT) to improve electronic conductivity and electrochemically active surface area. Second, metal sulfide additives (ZnS, Bi2S3, and FeS) were introduced into the electrode formulation to suppress HER and reduce electrode surface passivation. Among these, ZnS exhibited the best overall electrochemical performance. The Fe3O4@MWCNT electrode with 5 wt% ZnS achieved a specific capacity of 400.86 mAh g-1 at 1 A g-1, which is more than twice that of bare Fe3O4 (193.27 mAh g-1) electrode. Further investigation into the ZnS loading revealed that while the 3 wt% (408.44 mAh g-1) and 5 wt% ZnS-containing electrodes exhibited higher initial capacities at 1 A g-1, the 7 wt% ZnS electrode offered the best cycling stability, retaining 78.48% of its capacity (383.83 mAh g-1) after 100 cycles. Notably, the 7 wt% ZnS electrode developed a new discharge plateau after 35 cycles, which is attributed to the Fe0/Fe2+ oxidation facilitated by the in situ formation of FeS. As a result, the passivation of the iron-based electrode surface was reduced. The stable supply of sulfide ions at higher ZnS content likely supports this phenomenon and improves cycling stability. These findings highlight the synergistic effect of conductive nanostructures and metal sulfide additives in enhancing the electrochemical performance of iron-based anodes for high-performance Ni-Fe batteries.

Road safety incident data remain challenging to manage and analyze, particularly in large urban areas such as Metro Manila. This dissertation presents a data model and storage architecture to support road safety analysis and inter-agency data integration. The study analyzes existing road safety incident datasets and international standards to examine data fields, entity types, and their relationships. Based on this analysis, a unified data model is proposed using the Data Vault 2.0 methodology. The proposed model supports complex incident scenarios, explicit semantic relationships and incremental data integration from multiple agencies. To implement and evaluate the model, this study develops a data lakehouse architecture based on opensource technologies. Apache Iceberg and Nessie are used within the Dremio platform to manage structured data and support batch and near-real-time workflows. Apache Spark and Jupyter Notebook are used for data ingestion and processing, while Apache Superset supports data exploration and visualization. The proposed framework addresses key challenges in public-sector data systems, including data fragmentation, limited interoperability, and slow analytics. It also provides a foundation for future extensions, such as advanced analytics and automated event detection, to further
support road safety planning and decision-making.