Evaluation of Neural Networks to Predict Target Impedance Violations of Power Delivery Networks

Generating AI modules for decoupling capacitor placement using simulation

“…In recent years, an alternative way to achieve an accurate and optimized design of PDN that does not violate the TI, in the academic as well as in the industrial field, are artificial intelligence (AI) methods that have made it possible to redesign and optimize the design with little effort in time. The approach of Schierholz et al (2019) illustrates the prediction of TI violation with ANN using a simple board design with respect to decoupling capacitors (decaps) placement. A deep learning approach is proposed by Zhang et al (2021), where a boundary element method (BEM) is used to predict the impedance of the arbitrary irregular generated board.…”

Fast impedance prediction for power distribution network using deep learning

“…In recent years, the success of deep learning for complex and non-linear problems like computer vision, 6 natural language processing, 7 and strategy games 8 has also impacted many other fields. There has been some research [9][10][11][12] in applying machine learning in PDN modeling and optimization. However, most of these works do not have a welltrained and generalized machine learning model for PDN impedance prediction at the PCB level.…”

“…However, most of these works do not have a welltrained and generalized machine learning model for PDN impedance prediction at the PCB level. In the work of Schierholz et al, 9 an artificial neural network has been adopted to predict target impedance violations for PDN by considering the variations of IC location, decap placement, and target impedance. However, their task is just a simple [Corrections added on 9 November 2021, after first online publication: Figures 10(c) and 11(c) have been corrected in this version.]…”