Shupei Yu | Searcharxiv

arXiv2024

Solving Multi-Group Neutron Diffusion Eigenvalue Problem with Decoupling Residual Loss Function

In the midst of the neural network's success in solving partial differential equations, tackling eigenvalue problems using neural networks remains a challenging task. However, the Physics Constrained-General Inverse Power Method Neural Network (PC-GIPMNN) approach was proposed and successfully applied to solve the single-group critical problems in reactor physics. This paper aims to solve critical problems in multi-group scenarios and in more complex geometries. Hence, inspired by the merits of traditional source iterative method, which can overcome the ill-condition of the right side of the equations effectively and solve the multi-group problem effectively, we propose two residual loss function called Decoupling Residual loss function and Direct Iterative loss function. Our loss function can deal with multi-group eigenvalue problem, and also single-group eigenvalue problem. Using the new residual loss functions, our study solves one-dimensional, two-dimensional, and three-dimensional multi-group problems in nuclear reactor physics without prior data. In numerical experiments, our approach demonstrates superior generalization capabilities compared to previous work.

math.NA↗

arXiv2022

Active-carbon based supercapacitors with Au colloids: the case for placing the colloids at the electrolyte/electrode interface

Supercapacitors (S-C) are short-term energy storage elements that find many applications, e.g., electronic charging devices and suppressors of power fluctuations in grids that are interfaced with sustainable sources. The capacitance of an ordinary capacitor increases when dispersing metallic colloids in its dielectric. A similar strategy for S-C means a deployment of nano-scale metal colloids (in our case, Au nano particles, or AuNPs) at the very narrow interface between an electrolyte and the porous electrode (here, active-carbon film on a grafoil current collector). This is achieved by making the ligand that is coating the AuNPs negatively charged. We demonstrated a very large specific capacitance increase with a minute addition of functionalized AuNPs to the slurry. For example, C-V data at a scan rate of 20 mV/s indicated a specific capacitance amplification by a factor of 10 when 30 micro-g of AuNPs were incorporated with 200 mg of active carbon while using a 1 M Na2SO4 electrolyte and a 5% cellulose acetate butyrate as a binder. We make the case that the adhesion of the AuNPs to the surface of the electrode was strong: upon replacing the electrolyte, from 1 M Na2SO4 to 1 M KOH and retaining the same set of electrodes, the enhancement capacitance factor decreased as compared to 1 M Na2SO4 electrolyte but remained large, ~3, as determined by C-V traces at the same scan rate of 20 mV/s.

physics.chem-ph↗