Abstract
This chapter aimed to provide insights about the reliability of power electronic systems. We introduced the basics and metrics of reliability, new modern approaches such as Physics of Failure and Design for Reliability, and also detailed the step-by-step procedure to calculate the lifetime of power semiconductors and capacitors. With this knowledge in mind, it is possible to find the system reliability, depending on the association of the components in the reliability diagram block.
Nowadays, reliability is such an important topic to be considered in the design and project of new products, as important as their performance, efficiency, power density, costs, and scalability. However, there are a lot of major challenges associated with the task of designing more reliable components, such as the increasing complexity of power electronic systems, harsher operating conditions, cost limitations, inaccuracy and uncertainties on lifetime predictions, lack of information, and so on.
Additionally, there is a major inconvenience when applying accelerated degradation tests to obtain a lifetime model. In practice, the device under test may not be submitted to the same stressors as during the tests. Therefore, it becomes very difficult to accurately predict the failure of a device under real applications. To cover that fragility, several online condition monitoring methods can be applied to predict the real-time degradation of a certain failure mechanism. However, the application of these methods also inserts complexity into the system as one of the main challenges.
Recently, artificial intelligence is a newer tool that can be applied. The main advantages are to develop automatic optimizing processes, improve monitoring methods, and support the Design for Reliability decision-making procedure. In the coming years, this approach will support the development of fast reliability evaluation tools and advanced condition monitoring methods for power converters.
Therefore, it is clear the need for more efforts from academia and industry to work together in multiphysics domains and improve the knowledge about reliability. With that, it is possible to understand better the failure mechanisms, propose new physics-based lifetime models, develop new techniques to manufacture new products with fewer parameters variance, investigate new packaging technologies, improve methods and optimization, etc.
Nowadays, reliability is such an important topic to be considered in the design and project of new products, as important as their performance, efficiency, power density, costs, and scalability. However, there are a lot of major challenges associated with the task of designing more reliable components, such as the increasing complexity of power electronic systems, harsher operating conditions, cost limitations, inaccuracy and uncertainties on lifetime predictions, lack of information, and so on.
Additionally, there is a major inconvenience when applying accelerated degradation tests to obtain a lifetime model. In practice, the device under test may not be submitted to the same stressors as during the tests. Therefore, it becomes very difficult to accurately predict the failure of a device under real applications. To cover that fragility, several online condition monitoring methods can be applied to predict the real-time degradation of a certain failure mechanism. However, the application of these methods also inserts complexity into the system as one of the main challenges.
Recently, artificial intelligence is a newer tool that can be applied. The main advantages are to develop automatic optimizing processes, improve monitoring methods, and support the Design for Reliability decision-making procedure. In the coming years, this approach will support the development of fast reliability evaluation tools and advanced condition monitoring methods for power converters.
Therefore, it is clear the need for more efforts from academia and industry to work together in multiphysics domains and improve the knowledge about reliability. With that, it is possible to understand better the failure mechanisms, propose new physics-based lifetime models, develop new techniques to manufacture new products with fewer parameters variance, investigate new packaging technologies, improve methods and optimization, etc.
| Original language | English |
|---|---|
| Title of host publication | Power Electronic Converters and Systems : Converters and machine drives |
| Number of pages | 40 |
| Volume | 1 |
| Publisher | Institution of Engineering and Technology |
| Publication date | 1 Jan 2024 |
| Edition | 2 |
| Pages | 513-552 |
| Chapter | 14 |
| ISBN (Print) | 978-1-83953-767-7 |
| ISBN (Electronic) | 978-1-83953-768-4 |
| DOIs | |
| Publication status | Published - 1 Jan 2024 |