A novel full-parameter ageing modelling approach for capacitors based on complex impedance analysis

Abstract

Capacitors are ubiquitous in electronic devices and play a critical role in the devices' electromagnetic compatibility performance. The ageing of capacitors, due to intrinsic degradation mechanisms and external thermal and electrical stresses, has drawn a wide attention in the reliability field. The variations of the capacitance and equivalent series resistance are typically seen as representative for the ageing of capacitors, while the other parasitic parameters are usually neglected. In this study, a new full-parameter ageing modelling approach of capacitors is proposed based on complex impedance analysis. The ageing characteristics of all the proposed 7 parameters, including all parasitic parameters, are identified by carrying out an electrical overstress accelerated ageing test.

Long-Term Prediction of Multistress Accelerated Aging of Capacitors by Long Short-Term Memory Network

“…For instance, film capacitors are widely used in EMI (Electromagnetic Interference) suppression circuits. The failure criteria are usually as follows: 20% decrease in capacitance values or 2 times the ESR values [2]. Therefore, based on the proposed LSTM-Based DNN architecture, we can provide a more accurate and reliable prediction of the ageing conditions or the failure of the capacitors, in this way providing a better view on the normal operation of the circuits.…”

“…APACITORS are extensively used to suppress electromagnetic interference (EMI) in circuits [1,2]. Unfortunately, the long-term operation, especially in harsh environments with high temperatures and over-voltages, can cause considerable degradation in the capacitance values and the equivalent series resistance (ESR) values, which are two representative parameters for capacitors.…”

“…Unfortunately, the long-term operation, especially in harsh environments with high temperatures and over-voltages, can cause considerable degradation in the capacitance values and the equivalent series resistance (ESR) values, which are two representative parameters for capacitors. The prediction of the capacitors' ageing is generally based on a model-driven method [2] or a data driven method [3][4][5]. The classic regression and artificial neural networks (ANN) have been widely used to forecast the ageing conditions and the end-of-life (EOL) of capacitors.…”