John Jairo Martinez

Electrical Signature Analysis (ESA) is a widely adopted non-intrusive technique for monitoring induction motors, offering the advantage of using both current and voltage information. However, detecting early-stage faults remains a challenge due to low Signal-to-Noise Ratio (SNR) and spectral pollution from the power grid that masks weak fault signatures. This paper introduces a robust statistical detection framework designed to identify incipient mechanical and electrical faults despite these constraints. First, a signal enhancement strategy combining Instantaneous Symmetrical Components (ISC) and voltage-referenced multi-variable Wiener filtering is proposed. This approach effectively suppresses grid-induced harmonics to isolate the residual fault signature within the stator current. Based on a unified physical model of torque oscillations, a local spectral health indicator is derived. Statistical analysis demonstrates that this indicator follows a Gamma distribution, the parameters of which are conditioned by the motor's operating point. To address this dependency, an operating point clustering stage, based on the active power estimated from ESA, is proposed to segment the data. Finally, a Generalized Cumulative Sum (CuSum) algorithm, explicitly derived for the Gamma distribution, is developed. Unlike memoryless static detectors such as the Neyman-Pearson test, this sequential approach exploits the temporal history of the indicator to minimize detection delay. Experimental validation on $4$kW induction motors under misalignment and broken rotor bar conditions confirms that the proposed method significantly outperforms standard detection strategies, offering high sensitivity to incipient faults even under light load conditions.

Remaining Useful Life (RUL) is the length of time a component or system will operate before it requires repair or replacement. Although significant efforts have been made to estimate RUL accurately, controlling RUL remains challenging. This study introduces a state‐space model‐based approach for designing RUL controllers of controlled systems in degradation. It explores the relationship between a system's operational dynamics and its ongoing aging process, particularly the degradation rate experienced by a component during a given operational condition. In this approach, the RUL control problem is formulated as the problem of controlling a polytopic uncertain system. Therefore, robust design methods are proposed for both state observation and control to manage the degradation trajectory, ensuring an acceptable level of deterioration within a desired average lifetime. To illustrate its applicability, the proposed approach is implemented in a variable‐speed wind turbine, with a flexible shaft subject to torsional effects. The results highlight the benefits of using RUL control to enhance the durability of controlled systems while maintaining production performance and demonstrate the effectiveness of such an approach for managing the system's end‐of‐life.

Remaining Useful Life (RUL) has become widely recognized as a key metric for evaluating system reliability within the field of Prognostic and Health Management (PHM). It plays a crucial role in predicting a system’s health status and its expected lifespan. While significant efforts have been directed towards accurately estimating RUL, the process of decision-making regarding health management and potential adjustments to equipment lifespan still poses questions. This study introduces a state-space methodology designed to effectively manage RUL and, thus, the end-of-life of a degrading system, particularly focusing on controlled operation systems. Through an investigation of the relationship between the operational dynamics of the system and its aging process, the study specifically concentrates on regulating the degradation rate experienced during operation. A robust design method of an observer and a controller is presented to effectively govern the degradation trajectory, allowing for the achievement of an acceptable level of deterioration within a desired average lifetime. To demonstrate the performance and practical application of the proposed approach, it is applied to a variable-speed wind turbine with a flexible shaft subjected to torsion effects. The results highlight the benefits of using RUL control as a Health-Aware Control (HAC) strategy, aiming to find an optimal balance between reliability and production objectives. This study makes a valuable contribution to the field of reliability engineering by providing insight into managing the end-of-life of controlled processes suffering from aging. It is showed that the RUL control problem can be conveniently expressed as the problem of controlling a polytopic uncertain system. For this reason, a discrete-time robust LQR controller design is presented for solving the polytopic control design problem.

This paper presents the design and experimental validation of grid-based and Linear Fractional Transformation (LFT) approaches for the lateral control of autonomous vehicles. These new methodological approaches are compared together with the classical polytopic approach from the theoretical design to the real implementation on a real automated Renault ZOE vehicle. A solution is proposed to deal with both lane change and lane tracking problems, using a single LPV controller, by adapting the look-ahead distance. Each LPV controller is designed based on LPV/ $\mathcal {H}_\infty $ concept. Performance comparison includes computational costs, vehicle performance (i.e. lateral tracking error or control effort optimization) and on-board integration complexity. Simulation and experimental results on a private test track are included to support main findings.

This paper presents a comparative analysis using three Linear Parameter Varying (LPV) control approaches for the lateral control of an autonomous vehicle. Polytopic, gridbased and Linear Fractional Transformation (LFT) lateral controllers are designed, highlighting their benefits and drawbacks from the theoretical design to the real implementation on a real automated Renault ZOE vehicle. A solution is proposed to deal with both lane change and lane tracking problems, using a single LPV controller, by adapting the look-ahead distance. Each LPV controller is designed based on LPV/ concept. Performance comparison includes computational costs, vehicle performance (i.e. lateral tracking error or control effort optimization) and on-board integration complexity. Simulation and experimental results on a private test track are included to support main findings.