BMBrady Metherall

Brady Metherall

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Brady Metherall has a total of 85 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Growth projections for trees, hedges, and tree groups

    by Daniel Bearup, Nuria Richer, Gabriel Cairns, Jimmy Chen, Orson Hart, Brady Metherall, David Morselli, Josh Pearson, Mabel Rajendran, Fardin Saedpanah, et al.

    This report develops mathematical models to describe tree growth using Mapscape and Swiss forestry data. We first establish allometric relations between diameter, height, and canopy area, then extend them through time with the von~Bertalanffy--Chapman--Richards equation. To capture spatial effects, we introduce canopy-density PDEs and a mechanistic geometric model. Finally, we present approaches to estimate biomass and CO$_2$ sequestration, with applications to urban planning and forestry management.

  • Open Access

    Mathematical inversion of covariance-map images to determine three-dimensional product scattering distributions

    by Brady Metherall, Colin P. Please, David Heathcote, Claire Vallance

    Two-dimensional “crushed” velocity-map images of photoinduced or electron-induced reaction products can be Abel-inverted to reconstruct the full three-dimensional scattering distribution. Here, we develop the equivalent of the inverse Abel transform for covariance-map images, which are increasingly used in the field of velocity-map imaging to uncover correlations between the velocities of two or more scattered products. To formulate the inverse problem, we first derive the forward transformation, which maps the underlying cylindrically symmetric scattering distribution to its corresponding covariance-map measurement. We express the unknown scattering distribution as an expansion in an orthonormal basis and reformulate the inverse problem as a linear least-squares problem, which we solve via QR decomposition. This approach provides flexible control over the reconstruction resolution and supports natural regularization through both basis truncation and data smoothing. We validate the method on synthetic test problems, demonstrating accurate recovery of both local and extended structures, and show convergence using error metrics. We then apply our method to the experimental data on the dissociation products of CF3I2+ generated in electron–molecule collisions. For a simple binary fragmentation channel, the method recovers a localized point-source distribution that is consistent with physical understanding. For a more complex channel, we show that appropriate regularization obtains stable, interpretable results. Our results demonstrate that this method is robust, flexible, and well-suited for analyzing high-quality covariance-map scattering data.

  • Open Access

    Statistical modelling and optimization for designing renewable energy systems

    by Brady Metherall

    We explore the optimal allocation of solar, wind, and battery resources for large-scale renewable energy production and transmission. Using a combination of statistical modelling and optimization techniques, we analyze historical and synthetic power generation data to identify cost-effective investment strategies. To generate synthetic long-term scenarios, we extract seasonal trends from historical data and apply bootstrapping techniques to preserve statistical properties while simulating future variability. We employ a two-step optimization approach, first using a global method to identify the most promising region, followed by a local method for refinement. We compare single-year solutions and analyze their generalizability. 15-year and 100-year forecasts indicate that solutions incorporating long-term variability lead to more resilient planning. We find that the optimal mix of solar and wind remains stable across different time horizons, while battery storage varies to account for intermittency.

  • Open Access

    Mathematical Modelling of Deposition and Erosion of Particles in Pipes

    by Sean Bohun, Jake Bowhay, Thuy Duong Dang, Sam Harris, Cameron Hall, Emmanuel Lwele, Brady Metherall, John Christopher Meyer, Philip Pearce, Clare Rees-Zimmerman, et al.

    Paebbl are interested in the effective transport of a particle-laden fluid through a system of pipes. This transport has the potential to be disrupted if the particles sediment on the walls of the pipe, leading to pipe clogging and blockage. Motivated by this, we investigate the gravity- driven deposition and shear-driven erosion of solid particles carried in a turbulent flow in a pipe. We develop and solve a mathematical model for particle transport in the bulk of the fluid, and particle behaviour near the pipe walls, including deposition and erosion. We also model the chemistry related to the effective capture of CO2, which is important for the quality of the product. Our analysis is a good entry point for future work with Paebbl. More broadly, our work has relevance in wider industrial applications in relation to safety and efficiency, as well as the economic viability of industrial-scale production.

  • Top Cited Papers
    Open Access

    Machine learning for classifying chronic kidney disease and predicting creatinine levels using at-home measurements

    by Brady Metherall, Anna K. Berryman, Georgia S. Brennan

    Chronic kidney disease (CKD) is a global health concern with early detection playing a pivotal role in effective management. Machine learning models demonstrate promise in CKD detection, yet the impact on detection and classification using different sets of clinical features remains under-explored. In this study, we focus on CKD classification and creatinine prediction using three sets of features: at-home, monitoring, and laboratory. We employ artificial neural networks (ANNs) and random forests (RFs) on a dataset of 400 patients with 25 input features, which we divide into three feature sets. Using 10-fold cross-validation, we calculate metrics such as accuracy, true positive rate (TPR), true negative rate (TNR), and mean squared error. Our results reveal RF achieves superior accuracy (92.5%) in at-home CKD classification over ANNs (82.9%). ANNs achieve a higher TPR (92.0%), but a lower TNR (67.9%) compared with RFs (90.0% and 95.8%, respectively). For monitoring and laboratory features, both methods achieve accuracies exceeding 98%. The R2 score for creatinine regression is approximately 0.3 higher with laboratory features than at-home features. Feature importance analysis identifies the key clinical variables hemoglobin and blood urea, and key comorbidities hypertension and diabetes mellitus, in agreement with previous studies. Machine learning models, particularly RFs, exhibit promise in CKD diagnosis and highlight significant features in CKD detection. Moreover, such models may assist in screening a general population using at-home features—potentially increasing early detection of CKD, thus improving patient care and offering hope for a more effective approach to managing this prevalent health condition.


  • Open Access

    Optimising Buoy Design for Wave Energy Harvesting using the WITT Device

    by Gabriel Cairns, Nicholas Ryan, John Paul Alexander, Graeme Hocking, Brady Metherall, Hilary Ockendon, John Ockendon, Bernard Piette, Niall Rodgers, Sangeetha Sampath, et al.

    This work considers the design of a wave energy converter employing the WITT device: a machine that is able to convert motion in all 6 degrees of freedom into electrical power. Various designs for housing the WITT device inside a buoy are investigated with the aim of converting the forcing from low-frequency ocean waves to a higher-frequency motion that the WITT generates power most effectively at. By modelling the rigid-body motion of a buoy in response to a provided ocean wave forcing and the dynamics of the internal mechanisms within the buoy, we explore how nonlinear oscillation dynamics may achieve the desired frequency amplification.

  • Open Access

    Prediction of Chronic Kidney Disease Degeneration with Machine Learning

    by Jamiree Harrison, Manuchehr Aminian, Anna Berryman, Georgia Brennan, Zaoshu Cao, Claire S. Chang, Quindel Jones, Yena Kim, Kimberly Matsuda, Brady Metherall, et al.

    Chronic Kidney Disease (CKD) is a progressive condition during which the kidneys fail to filter blood properly. Many moderate to late-stage symptoms exist for CKD but no prominent symptoms exist for early CKD, making it historically difficult to diagnose and treat. CKD can be triggered by infections, certain medications, dehydration, drug abuse, as well as spiraling blood pressure and blood sugar levels. Still, CKD is oftentimes overlooked when these triggers are present. For these combined reasons, we provide methodologies for cleaning pertinent data sets to conduct an exploratory analysis of patient profiles. Once we analyze how patient features correlate with CKD and identify important features for diagnosis, we train various machine learning classifiers to predict what stage of CKD a patient is in without invoking the features that are explicitly indicative of CKD. We then validate our models using out-of-sample test data.


  • Open Access

    Tourism Forecasting and Environment

    by Ann Smith, Markus Dablander, Constantin Octavian Puiu, Brady Metherall, William Lee, Ruzanna Ab Razak, Noriszura Ismail

    Congestion in Edinburgh is a major issue, especially during peak visitor periods. The primary aim of this study was to develop a practical forecasting tool providing visitor volume, and visitor segmentation projections, to assist the travel and hospitality sectors in their planning. To address this a predictive model has been established based on Edinburgh Castle visitor records in conjunction with the Google Trends popularity index for Edinburgh Castle. Popularity of attractions was found to increase around major events with on-line search trends offering an indication of expected volume. Preliminary findings indicate predictions are feasible and could be improved should data relevant to other attractions and sectors be available.

  • Open Access

    Machine learning for classifying chronic kidney disease and predicting creatinine levels using at-home measurements

    by Brady Metherall, Anna K. Berryman, Georgia S. Brennan

    Background Chronic kidney disease (CKD) is a global health concern with early detection playing a pivotal role in effective management. Machine learning models demonstrate promise in CKD detection, yet the impact on detection and classification using different sets of clinical features remains under-explored. Methods In this study, we focus on CKD classification and creatinine prediction using three sets of features; at-home, monitoring, and laboratory. We employ artificial neural networks (ANNs) and random forests (RFs) on a dataset of 400 patients with 25 input features, which we divide into three feature sets. Using 10-fold cross-validation, we calculate metrics such as accuracy, true positive rate (TPR), true negative rate (TNR), and mean squared error. Results Our results reveal RF achieves superior accuracy (92.5%) in at-home CKD classification over ANNs (82.9%). ANNs achieve a higher TPR (92.0%) but a lower TNR (67.9%) compared with RFs (90.0% and 95.8%, respectively). For monitoring and laboratory features, both methods achieve accuracies exceeding 98%. The R2 score for creatinine regression is approximately 0.3 higher with laboratory features than at-home features. Feature importance analysis identifies key clinical variables hemoglobin and blood urea, and key comorbidities hypertension and diabetes mellitus, in agreement with previous studies. Conclusions Machine learning models, particularly RFs, exhibit promise in CKD diagnosis and highlight significant features in CKD detection. Moreover, such models may assist in screening a general population using at-home features—potentially increasing early detection of CKD, thus improving patient care and offering hope for a more effective approach to managing this prevalent health condition.


  • Open Access

    Mathematical modelling of a silicon carbide (SiC) pilot furnace

    by James Andrews, Aitor Azemar, Atrayee Bhattacharya, Charlie Egan, Andrew Lacey, Brady Metherall, Nicholas Ryan

    This report proposes a one-dimensional mathematical model of a novel silicon carbide furnace designed by Elkem ASA. The model consists of balance equations for volume fractions of three solids -- silicon dioxide, carbon, and silicon carbide -- and molar concentrations of two gases -- silicon monoxide and carbon monoxide -- within the furnace, which are coupled by reaction rate functions. Preliminary numerical results for steady-state solutions of these equations are presented. A simplified model involving two gas species is considered; the reaction rate for silicon monoxide formation is constant in places within the furnace where the temperature exceeds the minimum temperature for the reaction to proceed, and the presented solution can deal with the known carbon concentrations. The two-gas model is utilized to identify the residence time for the gas to vent from the top of the furnace.