Monitoreo de la condición basado en modelos aplicado en tiempo real usando curvas de potencia y rendimiento
Loading...
Date
2024
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
En los sistemas de producción industriales es necesario mantener altos niveles deoperatividad como una premisa fundamental para obtener resultados económicossatisfactorios. Por esto es necesario aplicar esquemas de mantención que maximicen eltiempo operacional de los equipos, minimizado las detenciones por falla y operando amáxima capacidad.En este trabajo se aborda la mantención de equipos basada en la condición (MBC),que es un esquema que intenta alargar el tiempo de operación de la máquina el máximoposible, siempre y cuando los equipos estén en condiciones de hacerlo, para no incurrir enuna falla catastrófica. Este esquema presenta ventajas respecto de los esquemastradicionales, pues evita los tiempos perdidos en que se incurre por detenciones programadasperiódicas, muchas veces innecesarias. Del mismo modo permite detectar fallas incipientescon la anticipación suficiente para evitar que evolucionen en fallas catastróficas.El monitoreo por condición requiere monitorear los equipos productivos mediantesensores para conocer el estado de las variables de operación relevantes, y así poder evaluarla continuidad de operación permanentemente, detectándose cualquier anomalía, y tomandolas acciones pertinentes de protección dado el caso.Sin embargo, muchas veces los sensores que se requiere para la MBC de un equipodeterminado son difíciles de montar y mantener, lo cual hace muy difícil su implementación.Por esta razón surge una variante de este esquema conocido como Mantención Basada en laCondición Basada en Modelos (MBCBM). Esta variante permite operar en la misma formaque con la MBC, pero ocupando un set reducido de sensores con los cuales a través demodelos físicos-matemáticos se pueden derivar las restantes variables de interés amonitorear.La ventaja de este esquema es que se puede aplicar con un menor número de sensores,pero, no obstante, requiere de modelos físicos-matemáticos de alta fidelidad o exactitud.En este trabajo de tesis se propone aplicar la MBCBM usando curvas de potencia yrendimiento de equipos productivos que se obtienen en tiempo real usando sets de sensoresmínimos con un modelo físico-matemático apropiado. La evolución de las curvas decapacidad y rendimiento de equipos y máquinas en el tiempo permite detectar y diagnosticar una variada gama de fallas de interés en todo tipo de equipos, ya que muchas fallas enmáquinas se manifiestan tempranamente en la degradación de la potencia y el rendimiento.Estas fallas se detectan en forma temprana y a través de su evolución en el tiempo es posibleevaluar y decidir si el equipo puede continuar en operación, o si es necesario detenerse antesde que la falla sea catastrófica.En este trabajo se formulan los aspectos relevantes de la MBCBM usando curvas depotencia y rendimiento y se describen 3 casos en que este alumno trabajó en el desarrollo desu tesis: a) turbinas eólicas, b) generadores sincrónicos, y c) carros de madereo con garra.En estos tres casos se da especial énfasis a la obtención de los modelos físicos-matemáticoscorrespondientes y su validación experimental, como base fundamental para el éxito de estaestrategia de mantención, indicándose en cada caso que tipo de fallas se pueden detectar.Se debe notar que en cada caso el modelo físico matemático es de distinta naturaleza, perose puede abordar la detección de fallas de igual manera.Este texto está organizado en 6 capítulos. El Capítulo 1 contiene una discusión delestado del arte respecto de la MBCBM. El Capítulo 2 contiene la aplicación de este esquemaa una turbina eólica, donde la naturaleza del modelo es una combinación de principios deMecánica de Fluidos y Electromecánica. Capítulo 3 contiene la aplicación de la MBCBM aun generador eléctrico sincrónico de imanes permanentes, donde el modelo es de naturalezaelectromecánica. El Capítulo 4 contiene la aplicación a un carro de madereo forestal, dondela naturaleza del modelo corresponde a principios de la Oleo hidráulica. El Capítulo 5 porsu parte contiene la aplicación del enfoque MBCBM a la detección de fallas de balanceodinámico y estático a rotores, que es otro aspecto muy relevante en la mantención de equiposrotatorios. Este caso permite demostrar que la aplicación de la estrategia MBCBM esaplicable no solo con curvas de potencia y rendimiento, sino que es flexible para usarse enotros casos. El Capitulo 6 continue las conclusions generales de este trabajo.
In industrial production systems, maintaining high levels of operability is necessary as a fundamental premise to achieve satisfactory economic results. Therefore, maintenance schemes are needed to maximize equipment operational time, minimizing downtimes due to failures and operating at maximum capacity. This work addresses condition-based maintenance (CBM), which aims to extend machine operation time as much as possible, provided that the equipment is capable of doing so, to avoid catastrophic failure. This scheme offers advantages over traditional approaches by avoiding unnecessary scheduled downtimes. Condition monitoring involves continuously monitoring productive equipment using sensors to assess the status of relevant operational variables, thereby evaluating operational continuity, detecting any anomalies, and taking appropriate protective actions as needed. However, often the sensors required for CBM of a particular equipment are difficult to install and maintain, making its implementation challenging. Hence, a variant of this scheme known as Model-Based Condition-Based Maintenance (MBCBM) emerges. This variant allows operating similarly to CBM but with a reduced set of sensors, through which the remaining variables of interest can be derived using physical-mathematical models. The advantage is that it can be applied with fewer sensors, albeit requiring high-fidelity or accurate physical-mathematical models. This thesis proposes applying MBCBM using power and performance curves of productive equipment obtained in real-time using minimal sensor sets with an appropriate physical-mathematical model. The evolution of capacity and performance curves of equipment and machinery over time enables the detection and diagnosis of a wide range of relevant faults in all types of equipment. These faults are detected early, and through their evolution over time, it is possible to evaluate and decide whether the equipment can continue in operation or if it needs to be stopped before the failure becomes catastrophic. This work formulates the relevant aspects of MBCBM using power and performance curves and describes three cases where the student worked on the development of their thesis: a) small wind turbines, b) synchronous permanent magnet generators, and c) timber logging carriage with claws. In these three cases, special emphasis is placed on obtaining the corresponding physical-mathematical models and their experimental validation, as a fundamental basis for the success of this maintenance trategy, indicating in each case what type of faults can be detected. It should be noted that in each case, the physical-mathematical model is of different nature, but the approach to fault detection is addressed in the same manner. This text is organized into 6 chapters. Chapter 1 contains a discussion of the state of the art regarding MBCBM. Chapter 2 presents the application of this scheme to a wind turbine, where the model nature is a combination of Fluid Mechanics and Electromechanics principles. Chapter 3 applies MBCBM to a synchronous electric generator with permanent magnets, where the model is of an electromechanical nature. Chapter 4 applies it to a forestry logging car, where the model nature corresponds to principles of Oleo hydraulics. Chapter 5 focuses on applying the MBCBM approach to detect dynamic and static balancing faults in rotors, which is another very relevant aspect in the maintenance of rotating equipment. This case demonstrates that the MBCBM strategy can be applied not only with power and performance curves but is also flexible for use in other cases. Chapter 6 contains the general conclusions of this work.
In industrial production systems, maintaining high levels of operability is necessary as a fundamental premise to achieve satisfactory economic results. Therefore, maintenance schemes are needed to maximize equipment operational time, minimizing downtimes due to failures and operating at maximum capacity. This work addresses condition-based maintenance (CBM), which aims to extend machine operation time as much as possible, provided that the equipment is capable of doing so, to avoid catastrophic failure. This scheme offers advantages over traditional approaches by avoiding unnecessary scheduled downtimes. Condition monitoring involves continuously monitoring productive equipment using sensors to assess the status of relevant operational variables, thereby evaluating operational continuity, detecting any anomalies, and taking appropriate protective actions as needed. However, often the sensors required for CBM of a particular equipment are difficult to install and maintain, making its implementation challenging. Hence, a variant of this scheme known as Model-Based Condition-Based Maintenance (MBCBM) emerges. This variant allows operating similarly to CBM but with a reduced set of sensors, through which the remaining variables of interest can be derived using physical-mathematical models. The advantage is that it can be applied with fewer sensors, albeit requiring high-fidelity or accurate physical-mathematical models. This thesis proposes applying MBCBM using power and performance curves of productive equipment obtained in real-time using minimal sensor sets with an appropriate physical-mathematical model. The evolution of capacity and performance curves of equipment and machinery over time enables the detection and diagnosis of a wide range of relevant faults in all types of equipment. These faults are detected early, and through their evolution over time, it is possible to evaluate and decide whether the equipment can continue in operation or if it needs to be stopped before the failure becomes catastrophic. This work formulates the relevant aspects of MBCBM using power and performance curves and describes three cases where the student worked on the development of their thesis: a) small wind turbines, b) synchronous permanent magnet generators, and c) timber logging carriage with claws. In these three cases, special emphasis is placed on obtaining the corresponding physical-mathematical models and their experimental validation, as a fundamental basis for the success of this maintenance trategy, indicating in each case what type of faults can be detected. It should be noted that in each case, the physical-mathematical model is of different nature, but the approach to fault detection is addressed in the same manner. This text is organized into 6 chapters. Chapter 1 contains a discussion of the state of the art regarding MBCBM. Chapter 2 presents the application of this scheme to a wind turbine, where the model nature is a combination of Fluid Mechanics and Electromechanics principles. Chapter 3 applies MBCBM to a synchronous electric generator with permanent magnets, where the model is of an electromechanical nature. Chapter 4 applies it to a forestry logging car, where the model nature corresponds to principles of Oleo hydraulics. Chapter 5 focuses on applying the MBCBM approach to detect dynamic and static balancing faults in rotors, which is another very relevant aspect in the maintenance of rotating equipment. This case demonstrates that the MBCBM strategy can be applied not only with power and performance curves but is also flexible for use in other cases. Chapter 6 contains the general conclusions of this work.
Description
Tesis (Doctor en Ciencias de la Ingeniería)--Pontificia Universidad Católica de Chile, 2024
Keywords
Model Based Condition Based Maintenance, Condition Basedmaintenance, Small Wind Turbine Power Curves, SPMG characteristic curves, HydraulicSystem Power Curves, Dynamic and Static Rotor Balancing