AEAndrés Escobar-Mejía

Andrés Escobar-Mejía

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Andrés Escobar-Mejía has a total of 22 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

    • 18 November 2026
    • Published by MDPI AG in Preprints.org
      • https://doi.org/10.20944/preprints202611.1940.v1
    Open Access

    Redundant Links Are Not Series Elements: Fault-Injection Evidence from ROS 2 Control Architectures

    by Daniela Buitrago Largo, Kevin David Ortega Quiñones, Andrés Escobar Mejía, Mauricio Holguín Londoño, Germán Andrés Holguín Londoño

    Reliability models can misrepresent redundant paths when communication links are treated as series elements. We tested this assumption through accelerated process-level fault injection in six running ROS 2 control architectures, comparing precomputed continuous-time Markov predictions based on graph logic (A) and a series-link rule (B). A pilot recorded 360 system failures; separate protocol-frozen studies recorded 960 signal-level failures, 400 trials with imposed common-cause failures and 540 trials with message loss, alongside 131 selected production windows from six simulated cells. No model parameter was fitted to observed failure times. In the two architectures that discriminate between A and B, the series-link model underestimated the observed mean time to signal loss by 44–52% and fell outside the prespecified confidence intervals. It also underestimated production in selected windows by 14–19%, although formal inference varied with the interval method and incident accounting. Model A met the prespecified agreement rule in five of six architectures at signal level; its production-rate errors were 1.7–3.8% under the prespecified summary, with sensitivity to window dependence. A statistically established advantage of parallel over single links was not obtained (difference 4.0 s, 95% CI −1.1 to 9.2). Under imposed common-cause failures, an independence model overestimated mean time to loss by up to 65%, while an injector-matched common-cause extension stayed within 9%. Agreement deteriorated under severe message loss. The findings support a falsification-oriented test of architecture-level reliability assumptions in running middleware, not claims about hardware failure rates or universal robotic reliability.

  • Open Access

    Physical Modeling and Optimal Sizing of Photovoltaic Generation Systems for Estimating Firm Energy Under Colombia’s Reliability Charge (ENFICC)

    by Sebastián Orozco-Bañol, Andrés Escobar-Mejía, David Cárdenas-Peña

    Photovoltaic (PV) plants participating in reliability mechanisms require design methods that assess both energy production and performance under critical resource conditions. This paper presents a computational methodology for physical modeling and optimal sizing of grid-connected PV systems to estimate Firm Energy under Colombia’s Reliability Charge (ENFICC). The workflow integrates a ten-year NASA POWER series, component catalogs with explicit preprocessing assumptions, a sequential conversion model, and a mixed discrete–continuous genetic algorithm (GA). A computational approximation of the documented La Tebaida system uses an explicitly assumed linear inverter representation. The corrected application report, independently checked against its exported hourly series using monthly coverage screening, gives a critical-month equivalent daily energy of 531.72 kWh/day and a final estimate of 425.38 kWh/day after applying the secondary-data factor fsec=0.8. This case-study assessment is distinct from validation against measured generation. The GA enforces implemented land-use, electrical, capacity, DC/AC, and monthly coverage conditions and ranks feasible candidates by final ENFICC. Thirty independent physical-data runs produced estimates from 341.5 to 390.6 kWh/day, with a mean of 362.9 kWh/day and a coefficient of variation of 3.72%. A paired 30-run ANN–MLP sensitivity experiment produced a mean of 320.6 kWh/day, 11.6% below the physical-data batch. The reported best GA solutions passed the implemented checks and computational recalculation. The contribution is a reproducible sizing workflow whose results remain conditional on the meteorological record, equipment representations, and screening assumptions.

  • Open Access

    Methodology for the Physical Validation and Optimal Sizing of Photovoltaic Generation Systems to Estimate Firm Energy for the Reliability Charge (ENFICC)

    by Sebastián Orozco Bañol, Andrés Escobar Mejía, David Augusto Cárdenas Peña

    The integration of photovoltaic (PV) generation into power systems requires design methods that evaluate not only annual energy yield or installed capacity, but also the dependable contribution of the plant under critical supply conditions. This paper presents a prototype-driven methodology for the physical validation and optimal sizing of a grid-connected PV generation system aimed at estimating, auditing, and evaluating ENFICC-oriented photovoltaic sizing alternatives under the Colombian reliability framework. The workflow integrates three input databases: a ten-year hourly meteorological database from NASA POWER for 2016--2025, a filtered and expanded PV module catalog based on the CEC database, and a filtered and expanded inverter catalog based on the Sandia database. These inputs feed a sequential modeling chain that estimates clearness index, irradiance decomposition, plane-of-array irradiance, cell temperature, DC power, AC power, monthly equivalent daily energy, net effective capacity, regulatory upper bound, and final ENFICC. The methodology was first validated using the installed PV system in La Tebaida, Colombia, as a local physical reference. For the validation case, the model used 252 Jinko Tiger Neo JKM625N-78HL4-BDV modules, two Solis 60K-LV-5G inverters, 157.5 kWp DC, 120 kW AC, fixed topology, and an hourly local validation dataset derived from the installed system. The NASA-based physical validation produced a critical-month energy of 548.5~kWh/day and a final ENFICC of 438.8~kWh/day after applying the secondary-data factor $f_{\mathrm{sec}}=0.8$. After applying the corrected physical filters and monthly coverage checks to the NASA POWER physical irradiance series, ten independent GA runs produced final ENFICC values between 283.5 and 384.2 kWh/day, with a mean of 349.7 kWh/day. The fitness function in all GA runs was evaluated using the corrected NASA POWER physical irradiance series, not an alternative irradiance series. These corrected results do not support the previously inflated optimization values; instead, they show that the local physical validation stage is essential before interpreting GA outputs. The proposed methodology contributes to clean-energy planning by linking renewable-energy integration, physical validation, reproducible computation, and regulatory firm-energy estimation within a single engineering workflow.

  • Open Access

    Improved MPC-Based Adaptive Virtual Synchronous Machine for a High-Power MMC Terminal

    by Ana J. Marín-Hurtado, Walter Gil-González, Andrés Escobar-Mejía

    The increase in grid-connected renewable energy sources (RES), particularly wind and solar plants, has transformed the dynamic properties of power systems. The progressive replacement of synchronous machines by generation based on power converters (which have no mechanical mass) has resulted in a substantial decrease of inertia, making the grid more vulnerable to frequency deviation and instability during unbalanced power or transient events. Maintaining adequate frequency support in such low-inertia environments has therefore become a key technical challenge for grid operators, which researchers have been working to address. Virtual synchronous machines (VSMs) have emerged as an effective solution for emulating the inertia and damping properties of conventional synchronous generators using power electronic converters. However, determining appropriate levels of virtual inertia and damping coefficients remains a complex problem, as fixed-parameter configurations often fail to achieve optimal performance across varying operating conditions, grid strengths, and load profiles. In this context, adaptive control techniques—capable of dynamically tuning virtual parameters in real-time—offer a promising path toward more robust and resilient converter-based power systems. This paper presents an improved model predictive control (IMPC) strategy for adaptive inertia and damping control applied to a 200 MW/200 kV modular multilevel converter (MMC). A comprehensive comparative study of both conventional fixed-inertia and standard MPC-based VSMs will be developed, using detailed time-domain simulations in MATLAB/Simulink. Results validated the proposed IMPC strategy and how the dynamic performance is enhanced. The system exhibits faster settling times (137 ms for active power and 127 ms for reactive power), lower steady-state errors (0.5226 % for active power and 2.8780 % for reactive power), and significantly reduced frequency oscillations (±0.201 Hz compared to ±0.291 Hz for conventional VSMs) and rate of change of frequency (RoCoF) (0.7417 Hz/s and 0.2117 Hz/s compared to VSM with constant inertia and VSM-MPC, respectively).

  • Open Access

    Data-Driven Model Reference Neural Control for Four-Leg Inverters Under DC-Link Voltage Variations

    by Ana J. Marín-Hurtado, Andrés Escobar-Mejía, Eduardo Giraldo

    The Four-Leg Three-Phase Voltage Source Inverter (4LVSI) is a versatile solution for integrating renewable energy sources (RESs) into distribution networks, as it compensates unbalanced voltages and currents while providing a path for zero-sequence components. Accurate current control is essential to ensure power quality and reliable operation under these conditions. Conventional controllers such as proportional–integral, resonant, or feedback-linearization methods achieve acceptable tracking under static dc-link conditions, but their performance degrades when dc-link voltage dynamics arise due to renewable-source fluctuations. This paper proposes a data-driven model reference neural control (MRNC) strategy for a four-leg inverter connected to RESs, explicitly accounting for dc-link voltage variations. The proposed controller reformulates the classical Model Reference Adaptive Control (MRAC) as a lightweight single-layer neural network whose adaptive weights are updated online using the Recursive Least Squares (RLS) algorithm. In this framework, the dc-link variations are not modeled explicitly but are implicitly learned through the data-driven adaptation process, as their influence is captured in the neural network regressors formed from real-time input–output measurements. This allows the controller to continuously identify the inverter dynamics and compensate the effect of dc-link fluctuations without requiring additional observers or prior modeling. The proposed approach is validated through detailed time-domain simulations and real-time Hardware-in-the-Loop (HIL) experiments implemented at a 10 kHz switching frequency. The results indicated that the RLS-based MRNC controller achieved the lowest steady-state current error, reducing it by approximately 1.85% and 1% compared to the Proportional-Resonant (PR) and One-Step-Ahead (OSAC) controllers, respectively. Moreover, under dc-link voltage variations, the proposed controller significantly reduced the current overshoot, achieving decreases of 5.9 A and 6.36 A relative to the PR controller.


  • Open Access

    Dynamic Graph Analysis: A Hybrid Structural–Spatial Approach for Brain Shape Correspondence

    by Jonnatan Arias-García, Hernán Felipe García, Andrés Escobar-Mejía, David Cárdenas-Peña, Álvaro A. Orozco

    Accurate correspondence of complex neuroanatomical surfaces under non-rigid deformations remains a formidable challenge in computational neuroimaging, owing to inter-subject topological variability, partial occlusions, and non-isometric distortions. Here, we introduce the Dynamic Graph Analyzer (DGA), a unified hybrid framework that integrates simplified structural descriptors with spatial constraints and formulates matching as a global linear assignment. Structurally, the DGA computes node-level metrics, degree weighted by betweenness centrality and local clustering coefficients, to capture essential topological patterns at a low computational cost. Spatially, it employs a two-stage scheme that combines global maximum distances and local rescaling of adjacent node separations to preserve geometric fidelity. By embedding these complementary measures into a single cost matrix solved via the Kuhn–Munkres algorithm followed by a refinement of weak correspondences, the DGA ensures a globally optimal correspondence. In benchmark evaluations on the FAUST dataset, the DGA achieved a significant reduction in the mean geodetic reconstruction error compared to spectral graph convolutional netwworks (GCNs)—which learn optimized spectral descriptors akin to classical approaches like heat/wave kernel signatures (HKS/WKS)—and traditional spectral methods. Additional experiments demonstrate robust performance on partial matches in TOSCA and cross-species alignments in SHREC-20, validating resilience to morphological variation and symmetry ambiguities. These results establish the DGA as a scalable and accurate approach for brain shape correspondence, with promising applications in biomarker mapping, developmental studies, and clinical morphometry.

  • Open Access

    Dynamic Graph Analysis: A Hybrid Structural-Spatial Approach for Brain Shape Correspondence

    by Jonnatan Arias García, Hernán Felipe García, Andrés Escobar-Mejía

    Accurate correspondence of complex neuroanatomical surfaces under nonrigid deformations remains a formidable challenge in computational neuroimaging, owing to inter‐subject topological variability, partial occlusions, and non‐isometric distortions. Here, we introduce the Dynamic Graph Analyzer (DGA), a unified hybrid framework that integrates simplified structural descriptors with spatial constraints and formulates matching as a global linear assignment. Structurally, DGA computes node-level metrics, degree weighted by betweenness centrality and local clustering coefficients—to capture essential topological patterns at a low computational cost. Spatially, it employs a two-stage scheme that combines global maximum distances and local rescaling of adjacent node separations to preserve geometric fidelity. By embedding these complementary measures into a single cost matrix solved via the Kuhn-Munkres algorithm and a refinement of weak correspondences, DGA ensures a globally optimal correspondence. In benchmark evaluations on the FAUST dataset, DGA achieved a significant reduction in the mean geodetic reconstruction error compared to models based on graph convolutional learning or traditional spectral methods. Additional experiments demonstrate robust performance on partial matches in TOSCA and cross‐species alignments in SHREC‐20, validating resilience to morphological variation and symmetry ambiguities. These results establish DGA as a scalable and accurate approach for brain shape correspondence, with promising applications in biomarker mapping, developmental studies, and clinical morphometry.


  • Open Access

    Control System of an SVC Based on ARMA Model to Improve the Displacement Power Factor of an Electric Arc Furnace

    by J. Marulanda-Durango, A. Escobar-Mejía, E. Giraldo

    In the steel industry, electric arc furnaces (EAFs) are essential for metal scrap recycling. However, they are known for their high reactive power consumption, negatively impacting power quality (PQ). The static VAR compensator (SVC) is the reactive power compensation device that provides the best cost‐benefit relation for mitigating the impact of EAFs. Although the SVC effectively improves certain PQ indices, its performance in improving the displacement power factor (DPF) is limited. This limitation is due to the delays of the control system during the reactive power calculation. This paper proposes a multivariable coupled control technique based on an autoregressive moving average (ARMA) model. The system to be controlled is identified offline directly from simulation measurements. The SVC is designed to improve the DPF at the connection node of the SVC. When the SVC connects to the power system, the proposed control technique enhances the DPF, achieving an average increase of 30%. As is well known, improving the DPF positively impacts the power system by increasing its efficiency. By significantly enhancing the DPF with the proposed control technique, several long‐term benefits are expected, including an increase in the lifespan of connected equipment, a sustained reduction in operational costs, which may result in a return on investment of the SVC, and a lower environmental impact by contributing to the reduction of the carbon footprint, particularly in systems relying on fossil‐based energy. A real‐time simulation framework based on a hardware‐in‐the‐loop (HIL) system is used to evaluate the proposed controller. This framework assesses its pertinence and performance in real‐time applications of the proposed controller.


  • Open Access

    Optimal Economic–Environmental Operation of BESS in AC Distribution Systems: A Convex Multi-Objective Formulation

    by Walter Gil-González, Oscar Danilo Montoya, Luis Fernando Grisales-Noreña, Andrés Escobar-Mejía

    This paper deals with the multi-objective operation of battery energy storage systems (BESS) in AC distribution systems using a convex reformulation. The objective functions are CO2 emissions, and the costs of the daily energy losses are considered. The conventional non-linear nonconvex branch multi-period optimal power flow model is reformulated with a second-order cone programming (SOCP) model, which ensures finding the global optimum for each point present in the Pareto front. The weighting factors methodology is used to convert the multi-objective model into a convex single-objective model, which allows for finding the optimal Pareto front using an iterative search. Two operational scenarios regarding BESS are considered: (i) a unity power factor operation and (ii) a variable power factor operation. The numerical results demonstrate that including the reactive power capabilities in BESS reduces 200 kg of CO2 emissions and USD 80 per day of operation. All of the numerical validations were developed in MATLAB 2020b with the CVX tool and the SEDUMI and SDPT3 solvers.


  • Open Access

    LQR-Based Adaptive Virtual Inertia for Grid Integration of Wind Energy Conversion System Based on Synchronverter Model

    by Walter Gil-González, Oscar Danilo Montoya, Andrés Escobar-Mejía, Jesús C. Hernández

    This paper proposes adaptive virtual inertia for the synchronverter model implemented in a wind turbine generator system integrated into the grid through a back-to-back converter. A linear dynamic system is developed for the proposed adaptive virtual inertia, which employs the frequency deviation and the rotor angle deviation of the synchronverter model as the state variables and the virtual inertia and frequency droop gain as the control variables. In addition, the proposed adaptive virtual inertia uses a linear quadratic regulator to ensure the optimal balance between fast frequency response and wind turbine generator system stress during disturbances. Hence, it minimizes frequency deviations with minimum effort. Several case simulations are proposed and carried out in MATLAB/Simulink software, and the results demonstrate the effectiveness and feasibility of the proposed adaptive virtual inertia synchronverter based on a linear quadratic regulator. The maximum and minimum frequency, the rate change of the frequency, and the integral of time-weighted absolute error are computed to quantify the performance of the proposed adaptive virtual inertia. These indexes are reduced by 46.61%, 52.67%, 79.41%, and 34.66%, in the worst case, when the proposed adaptive model is compared to the conventional synchronverter model.