lifetime estimation and modeling of power converter in wind turbine using digital twin and machine learning

نوع: Type: thesis

مقطع: Segment: masters

عنوان: Title: lifetime estimation and modeling of power converter in wind turbine using digital twin and machine learning

ارائه دهنده: Provider: Ali mohammadian

اساتید راهنما: Supervisors: Dr. Abbas Ramezani

اساتید مشاور: Advisory Professors: Dr. Mohammad Mehdi Shahbazi

اساتید ممتحن یا داور: Examining professors or referees: Dr. Yunus Selgi and Dr. Mohsen Hasan Babai Nozadian

زمان و تاریخ ارائه: Time and date of presentation: 2023

مکان ارائه: Place of presentation: Department of Electrical Class 3

چکیده: Abstract: ue to the rapid advances in technology in recent years, one of the problems of human progress is the supply of energy sources. Wind energy is one of the available and reliable sources of energy. Wind turbines are an important source of intermittent renewable energy and are used in many countries to reduce energy costs and reduce reliance on fossil fuels. Converters play a very important role in wind turbines, so the malfunction of wind turbine converters can stop the operation of wind turbines and cause problems in the network. In wind turbines, it is accepted that the converter must have high reliability and fault tolerance. If the failure can be diagnosed, unplanned stops can be avoided. is to IGBT, that is, if the lifetime of the IGBT can be obtained, it means that the lifetime of the entire converter has been obtained. The NASA Reliability Laboratory has conducted a degradation test. In this research, we used the dataset of the NASA Reliability Laboratory, which is and from: collector emitter voltage, collector-emitter current and gate-emitter voltage. Among these data, we considered the collector-emitter voltage as the leading parameter, which has seven phases and the seventh phase is failure. Finally, MLP and CNN-LSTM networks were trained and tested by choosing the appropriate features to reduce the computational complexity and increase the accuracy of the results. Finally, the results of the two networks were compared, and the MLP detection accuracy was more appropriate

فایل: ّFile: Download فایل