Modeling and identification of power electronic converters
Nowadays, many industries are moving towards more electrical systems and components. This is done with the purpose of enhancing the efficiency of their systems while being environmentally friendlier and sustainable. Therefore, the development of power electronic systems is one of the most important...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Estado: | Versión publicada |
| Fecha de publicación: | 2022 |
| País: | España |
| Institución: | CBUC, CESCA |
| Repositorio: | TDR. Tesis Doctorales en Red |
| OAI Identifier: | oai:www.tdx.cat:10803/673581 |
| Acceso en línea: | http://hdl.handle.net/10803/673581 https://dx.doi.org/10.5821/dissertation-2117-363405 |
| Access Level: | acceso abierto |
| Palabra clave: | Àrees temàtiques de la UPC::Enginyeria elèctrica 621.3 |
| Sumario: | Nowadays, many industries are moving towards more electrical systems and components. This is done with the purpose of enhancing the efficiency of their systems while being environmentally friendlier and sustainable. Therefore, the development of power electronic systems is one of the most important points of this transition. Many manufacturers have improved their equipment and processes in order to satisfy the new necessities of the industries (aircraft, automotive, aerospace, telecommunication, etc.). For the particular case of the More Electric Aircraft (MEA), there are several power converters, inverters and filters that are usually acquired from different manufacturers. These are switched mode power converters that feed multiple loads, being a critical element in the transmission systems. In some cases, these manufacturers do not provide the sufficient information regarding the functionality of the devices such as DC/DC power converters, rectifiers, inverters or filters. Consequently, there is the need to model and identify the performance of these components to allow the aforementioned industries to develop models for the design stage, for predictive maintenance, for detecting possible failures modes, and to have a better control over the electrical system. Thus, the main objective of this thesis is to develop models that are able to describe the behavior of power electronic converters, whose parameters and/or topology are unknown. The algorithms must be replicable and they should work in other types of converters that are used in the power electronics field. The thesis is divided in two main cores, which are the parameter identification for white-box models and the black-box modeling of power electronics devices. The proposed approaches are based on optimization algorithms and deep learning techniques that use non-intrusive measurements to obtain a set of parameters or generate a model, respectively. In both cases, the algorithms are trained and tested using real data gathered from converters used in aircrafts and electric vehicles. This thesis also presents how the proposed methodologies can be applied to more complex power systems and for prognostics tasks. Concluding, this thesis aims to provide algorithms that allow industries to obtain realistic and accurate models of the components that they are using in their electrical systems. |
|---|