ABSTRACT. This paper presents the experimental results of a proposed 48 V electric traction system intended for transportation applications in Mexico. This electric traction system is comprised by a 26 kW AC induction machine together with an AC motor controller which is water-cooled with a liquid cold plate, a radiator, and a water pump to ensure thermal stability during operation. Throughout this work, the experimental investigation includes electrical, thermal, and mechanical results of the electric traction system under no-load conditions, involving current and voltage waveforms together with the obtained IR thermal images. Additionally, the velocity and power performance are analyzed across various operating points.

ABSTRACT. This study introduces an innovative approach for optimizing fuzzy logic controllers used in multilevel three-phase inverters through the application of the Bear Smell Search Algorithm (BSSA). The primary goal is to enhance the dynamic and steady-state performance of the inverter, specifically in voltage and current regulation under varying load conditions and operational scenarios. The methodology combines fuzzy control with metaheuristic techniques, enabling automatic calibration of membership functions, rule sets, and scaling factors, leading to improved robustness and adaptability of the control system. Simulation results in MATLAB/Simulink demonstrate that the optimized controller significantly outperforms traditional and non-optimized controllers in transient response and stability. Additionally, BSSA shows promising potential for practical applications in renewable energy and power electronics systems, advancing towards more efficient and reliable energy management solutions.

ABSTRACT. In the current context of energy transition, modern electrical grids demand more flexible, intelligent, and adaptable solutions that facilitate the efficient integration of renewable sources. In this scenario, multi-port solid state transformers (MPSST) are positioned as a key technology. To validate their technical feasibility, they must be subjected to tests that evaluate their behavior under realistic operating conditions and possible contingencies. This article presents the design and functional evaluation of an MPSST based on a three-stage structure, configured to integrate renewable and conventional generation sources. Different scenarios will be analyzed, including partial loss of renewable generation, disconnection of the main power source, and response to critical failures in the DC link. The results demonstrate that the system maintains stable conversion and adequate operation under various operating conditions.

ABSTRACT. This paper addresses the tuning problem of the adjustable parameters of a backstepping controller to regulate the speed of a PMSM that presents nonlinear and multivariable characteristics with a strong coupling between electrical and mechanical variables. To perform the tuning, the Tasmanian Devil Optimization Algorithm (AODT) is implemented to search for the most optimal parameters of the backstepping controller and obtain an accurate tracking of the rotor speed trajectory under parametric variations and abrupt load torque disturbances. With the optimal gains, the dynamic performance of the optimized backstepping controller is improved. The results show that the controller presents a robust and efficient response, significantly improving the performance compared to an optimized PI controller.

ABSTRACT. This paper addresses power system coherency by extracting oscillatory modal characteristics, which in turn inform a hierarchical agglomerative clustering (HAC) technique. The phase of the slower frequency modes is extracted from rotor speed measurements after a large disturbance occurs using a set of FIR filters built with the discrete-time Taylor-Fourier transform (DTTFT). These are clustered using the hierarchical agglomerative clustering (HAC) technique, with a subsequent optimal determination of the number of clusters guided by a customized Elbow heuristic, constructed explicitly over the HAC cluster merging distance/cost curve. Finally, the attained results demonstrate that the DTTFT's noise tolerance under polluted conditions and the effectiveness of the HAC for properly clustering generating units in the 10-machine and 39-bus benchmark power grid.

ABSTRACT. This work proposes a numerical model to analyze the impact of Transcranial Magnetic Stimulation (TMS) on the electrical activity of a three-dimensional biological neuronal network. The main objective is to estimate the electric field induced by TMS in a neuronal network connected in the form of a rectangular grid, considering the current induced by TMS and its effect on action potentials. To address this problem, the magneto-quasistatic (MQS) approximation is employed, allowing for the resolution of the distribution of the electric and magnetic fields induced by a coil powered by a current pulse. In addition, the proposed model is enhanced by using physics-informed neural networks (PINNs). The use of PINNs allows for rapid calculation of the magnetic vector potential at any point in the analysis region once the network is trained, thereby outperforming the most commonly used numerical methods for solving these problems, such as the Finite Element Method and the Finite Difference Method. Preliminary results, obtained from simulations on a neuronal network connected as a two-dimensional grid, show that TMS has a positive effect on the synchronization of action potentials between neurons, especially in the direction of the induced electric field. This synchronization suggests a possible mechanism by which TMS influences neuronal activity, which could have significant implications for therapeutic applications and neuroscience research. The ongoing work aims to improve the model’s accuracy by adjusting boundary conditions and optimizing the PINN to simulate more complex scenarios, enabling a detailed analysis of the interaction between TMS and the neuronal network.

ABSTRACT. This paper presents the implementation and comparative analysis of tuning methods for classical PID and fuzzy logic speed controllers applied to a three-phase induction motor. The control strategy is deployed using a commercial inverter within the LabVIEW graphical programming environment and executed on a personal computer. An industrial Danfoss variable frequency drive (VFD) is used to operate the motor, while data acquisition and control tasks are managed through a National Instruments DAQ card. Several tuning techniques for the PID controller are discussed, and a fuzzy controller is designed based on the PID gain parameters. The performance of both controllers is evaluated under different tuning scenarios, including reference speed changes and external disturbances.

ABSTRACT. Bidirectional converters are widely used in electric and hybrid vehicle applications because of its power driving and regenerative capability. The model predictive control (MPC) technique has been applied for current regulation in these converters for its simplicity and intuitive algorithm; however, in three-phase systems its algorithm requires a high computational cost to evaluate all the operating modes. To enhance this disadvantage, controllers based on artificial intelligence have emerged. In this paper a design methodology for an artificial neural network controller (ANNC) based on model predictive control (MPC) is proposed for current regulation in a bidirectional inverter-rectifier for hybrid vehicle applications.

ABSTRACT. This paper presents a study related to a recently proposed quadratic boost converter topology: the so-called Low Energy Storage Quadratic Boost (LES-QB) converter. Unlike traditional quadratic boost converters, the LES-QB converter achieves high voltage gain with reduced energy storage in its passive components, enabling more compact designs. Recently, an improved operation was proposed based on the optimized selection of capacitors. This work focuses on the implementation and validation of the optimized design. The converter was built and tested, and its operation was compared against that of a non-optimized configuration. The results demonstrate that the correct selection of capacitors leads to a reduced switching ripple without increasing the size of the converter. Experimental results are provided.

ABSTRACT. This paper presents a nonlinear control scheme to regulate the inductor current of a quadratic boost converter, which is based on the reduced redundant power processing (R2P 2) concept. For the inductor current regulation task, the control scheme is based on the passivity based control (PBC) strategy, where the synthesis procedure for the control law and the stability conditions are given. Simulation results show the performance of the PBC scheme implemented in the quadratic converter.

ABSTRACT. This work presents the design, modeling, and validation in MATLAB/Simulink of a DC microgrid composed of a photovoltaic array (PV), a battery energy-storage system (BESS), and a bidirectional high-low voltage Dual-Active Bridge charger (DAB SHV-PLV) for electric vehicles. A phase-shift (SPS) control scheme with current loop and dynamic phase adjustment was implemented. The converter was evaluated through four key simulation scenarios: steady-state operation, vehicle to grid (G2V) to grid to vehicle (V2G) transitions, irradiance reduction, and stepped load increase; showing that the BESS keeps the DC bus at 700 V with minimal deviation while the DAB SHV-PLV transfers up to ± 50 kW without overshoot. Power profiles confirm effective coordination among the MPPT, BESS, and DAB, validating the converter’s viability as a flexible solution for fast charging and V2G services in photovoltaic microgrids.

ABSTRACT. This work presents the experimental description and verification of a recently proposed high-gain step-up DC-DC converter known as the Low-Voltage in Capacitors (LVC) converter. The verification includes a recently proposed operation strategy that leads to a reduction in output voltage ripple. Unlike conventional boost converters, the LVC topology operates with capacitors subjected to voltages lower than the output, allowing a decrease in the voltage rating, which normally results in a reduction in their size. The experimental setup, measurement methodology, and observed waveforms are discussed. Results confirm that the LVC converter operates properly with both the former strategy and the one focused on the switching ripple reduction. These findings support the benefits of the converter and its potential for some applications requiring improved output quality without increasing the size of the components.

ABSTRACT. In this work, a controller is developed to manipulate the modulation rates for the generation of Pulse Width Modulation (PWM) signals for the power semiconductor devices in a three-phase AC-DC converter. These modulation rates are used to regulate the voltage of an RL load, simultaneously achieving a unity power factor in the three-phase supply. The control law is designed based on a model of an average circuit in a twophase synchronous reference frame. The control contains two loops; the inner loop takes advantage of the properties of the model to utilize passivity-based techniques, and the outer loop is a PI control. The proposed controller is tested in a computer simulation, showing its physical implementation

ABSTRACT. Heart diseases have been a critical issue to deal with to improve people's health. Medical research and technology are being developed to obtain accurate diagnoses and treatments. This paper contributes to the design of an automated diagnosis system to classify electrocardiogram (ECG) signals to detect cardiac diseases. The proposed diagnosis system is based on the Fourier series analysis, which uses a dynamical state observer to instantaneously obtain salient features and patterns from the ECG harmonic content in real-time, whose information is classified through a K-Nearest neighbor algorithm (KNN), named as the classifier, which determines the possible disease. The ECG signals used in this paper are obtained from the free online available PhysioNet databases, which contain information that can be used for the diagnosis and classification of healthy patients, arrhythmia cases, myocardial infarction, and heart failure. The proposed automated procedure is 93% effective in disease detection for the explored databases, highlighting its potential as a classification tool for ECG-based diagnosis.

ABSTRACT. Type 2 diabetes (T2D) represents a priority in Mexican public health due to its high prevalence and associated complications. Currently, diagnosis is primarily based on glycemic biomarkers such as plasma glucose, insulin, and the HOMA-IR index, which reflect metabolic alterations in advanced stages. This study aims to assess the capability of classifying individuals with T2D using lipid profiles and machine learning, while primarily excluding glucose-associated biomarkers. Anthropometric measurements are incorporated into the lipid profile to explore their potential contribution to diagnosis. Various classification algorithms were trained and validated, including Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), and XGBoost. Exploratory statistical analyses and feature selection techniques were performed, as well as the evaluation and calculation of derived atherogenic indices such as TG/HDL and AIP. The results showed that the SVM model achieved the best overall performance, with an accuracy of 85.2\%, an F1-score of 0.853, a precision of 85.1\%, and an AUC value of 0.916. The RF and XGBoost models achieved sensitivities above 88\%, with AUC values of 0.934 and 0.930, respectively, indicating their ability to identify patients with T2D accurately. These findings demonstrate that lipid profiles, in combination with non-glycemic features, can serve as an accessible and low-cost alternative to support the automated diagnosis of T2D, particularly in resource-constrained settings.

ABSTRACT. The study of microbial behavior helps us understand how microorganisms grow and how they can be effectively treated. While standard techniques can evaluate the effectiveness of treatments on microorganisms, these methods often require the destruction of samples, which limits ongoing research. Alternative techniques, such as Laser Speckle Imaging (LSI), provide a non-invasive and rapid way to analyze the viability of biological samples. This work examines whether information from LSI images can be utilized to differentiate the impact of treatments on biofilms of Candida tropicalis, a species that is part of the normal human microbiota but can also act as an opportunistic pathogen, commonly found in hospital patients and medical instruments. To achieve this, a Time History representation was processed, and features were extracted from the Gray Level Run Length Matrix. Finally, samples were clustered using K-means, demonstrating that it is possible to differentiate the effects of various treatment doses based on the dynamic description provided by LSI.

ABSTRACT. Amperometry is an electrochemical technique widely used for the detection and quantification of analytes by measuring currents generated by redox reactions. This method is based on the direct relationship between the recorded current and the analyte concentration, enabling accurate, cost-effective detection with great miniaturization potential. In this context, the development of chronoamperometry-oriented data acquisition systems is essential to improve the automation, robustness, and efficiency of electrochemical signal analysis. However, there is a persistent need to integrate these advances into more comprehensive systems that allow for the interpolation of chronoamperometric data and the estimation of concentrations in real time through intuitive interfaces. The purpose of this work is to develop an embedded system that enables automated analyte estimation, offering a versatile and accessible tool for applications in biomedicine, environmental monitoring, and scientific research.

ABSTRACT. Electromyographic (EMG) signals are frequently contaminated by powerline interference (60 Hz and harmonics), compromising diagnostic accuracy. This paper compares three denoising techniques: 1) a cascaded IIR filter (60 Hz notch, 2) spectral subtraction of harmonics via Fourier Transform, and 3) wavelet denoising (db4 wavelet, universal threshold, level 7). All methods include a 10–500 Hz bandpass prefilter. These techniques are compared using experimental superficial electromyographic (sEMG) data. The results show that wavelet achieves optimal harmonic suppression (- 60.4684 dB), while spectral substraction best preserves temporal features (RMSE = 0.013) with the lowest computational load. The IIR approach remains suitable for real-time applications. Noise suppression, feature preservation, and computational load are quantified using SNR, RMSE, and spectral metrics

ABSTRACT. The safe operation of turbine-based mechanical ventilators requires robust control strategies that can adapt to dynamic changes in airway resistance and lung compliance. This work presents a closed-loop control scheme for a blower operating in volume-controlled mode. A cascaded control architecture was implemented and tested across resistance values from 5 up to 50 cmH_2O*s/L and multiple compliance values from 0.01 up to 0.1 L/cmH2O. To verify the proper operation of the proposed control scheme, we assessed whether the peak pressure remained within the expected range even under short inspiratory times. The reference value was obtained from the respiratory equation of motion for mechanical ventilation, and the results were compared against peak pressures measured from two commercial ventilators. Results show that the proposed control algorithm maintains pressure profiles within physiological limits and achieves performance comparable to commercial devices. To quantify the accuracy of our system, the relative error from the measured peak pressure was computed between the commercial mechanical ventilator and our system, resulting in an average relative error of 6.84\% The findings support the feasibility of this approach for use in blower-based ventilators and demonstrate its potential for safe, adaptive ventilation across varying ventilatory mechanics.