Google Scholar
Publications in reversed chronological order.
2022
-
A generic battery-cycling optimization framework with learned sampling and early stopping strategies
Changyu Deng, Andrew Kim, and Wei Lu
Patterns 2022
There are many parameters to optimize for a battery, in both simulations and experiments, from design to manufacturing. It is time consuming and costly to evaluate the lifetime performance of batteries since it takes a long period to cycle them. We introduce a generic framework leveraging machine-learning algorithms. The framework is designed to optimize battery parameters to enhance cycling performance in a systematic and efficient way, which allows parallel cyclers, stops unpromising cycles, and automatically yields new configurations of parameters. The framework could reduce the average cycling time per battery from years to months/weeks for cycling experiments or from weeks to days/hours for cycling computations. This method is flexible to scale up for many applications, from fundamental research to industrial development in batteries and other similar fields.
-
Physics-encoded deep learning in identifying battery parameters without direct knowledge of ground truth
Bin Wu, Buyi Zhang,
Changyu Deng, and Wei Lu
Applied Energy 2022
We show a method to embed physical laws and on-line observation into machine learning so that irrelevant low-cost battery data can be utilized to identify complex system parameters by machine learning without knowledge of their ground truth as the training data. Lithium diffusivity, a complicated function of lithium concentration, is a crucial parameter for battery performance but difficult to measure directly. We take diffusivity as an example and show that it can be obtained from easily measured sequence of battery voltage over time. In simulations, our results show that this method accurately quantifies not only the diffusivities of both positive and negative electrodes, but also as complex non-linear functions of lithium concentration, purely based on the cell voltage data requiring neither diffusivity nor concentration measurement. Notably, it can accurately predict non-monotonic, many-to-one relations such as “w” shape functions. Moreover, this method is immune to measurement noise and capable of simultaneously estimating multiple parameters. In experiments, our method demonstrates more robust diffusivity estimation than a pure physics-based parameter fitting method and a widely used experimental technique. Our results suggest that the approach enables identifying physical parameters and their interdependence without direct measurements of those parameters.
-
Self-Directed Online Machine Learning for Topology Optimization
Changyu Deng, Yizhou Wang, Can Qin, Yun Fu, and Wei Lu
Nature Communications 2022
Topology optimization by optimally distributing materials in a given domain requires non-gradient optimizers to solve highly complicated problems. However, with hundreds of design variables or more involved, solving such problems would require millions of Finite Element Method (FEM) calculations whose computational cost is huge and impractical. Here we report Self-directed Online Learning Optimization (SOLO) which integrates Deep Neural Network (DNN) with FEM calculations. A DNN learns and substitutes the objective as a function of design variables. A small number of training data is generated dynamically based on the DNN’s prediction of the optimum. The DNN adapts to the new training data and gives better prediction in the region of interest until convergence. The optimum predicted by the DNN is proved to converge to the true global optimum through iterations. Our algorithm was tested by four types of problems including compliance minimization, fluid-structure optimization, heat transfer enhancement and truss optimization. It reduced the computational time by 2 5 orders of magnitude compared with directly using heuristic methods, and outperformed all state-of-the-art algorithms tested in our experiments. This approach enables solving large multi-dimensional optimization problems.
2021
-
A Facile Process to Fabricate Phosphorus/Carbon Xerogel Composite as Anode for Sodium Ion Batteries
Changyu Deng, and Wei Lu
Journal of The Electrochemical Society 2021
Sodium ion batteries have become popular due to their low cost. Among the possible anode materials for sodium ion batteries, phosphorus has great potential owing to its high theoretical capacity. Previous research that yielded high capacity and long duration for a phosphorus anode used expensive materials such as black phosphorus (BP) and phosphorene. To take advantage of the low cost of sodium ion batteries, we report a simple and low-cost method to fabricate the anode: condensing red phosphorus on carbon xerogel. Even with a large particle size (∼ 50 m) and a high mass loading (2 mg cm2), the composite cycled at 100 mA g-1 yielded a capacity of 357 mA g-1 or 2498 \rmmAh\,\rmg_\rmP^-1 based on phosphorus after subtracting the contribution of carbon. The average coulombic efficiency is as high as 99.4%. When cycled at 200 mA g-1, it yielded a capacity of 242 mAh g-1 or 1723 \rmmAh\,\rmg_\rmP^-1, with average degradation rate only 0.06% in 80 cycles. Our research provides an innovative approach to synthesize anodes for sodium ion batteries at extremely low cost, with performance exceeding or comparable to state-of-the-art materials, which will promote their commercialization.
-
A Minimal Physics-Based Model on the Electrochemical Impedance Spectroscopy of Solid-State Electrolyte
Changyu Deng, and Wei Lu
arXiv preprint arXiv:2110.00551 2021
Solid state batteries have emerged as a potential next-generation energy storage device due to safety and energy density advantages. Development of electrolyte is one of the most important topics in solid state batteries. Electrochemical Impedance Spectroscopy (EIS) is a popular measurement technique to obtain the conductivity and diagnose the electrolyte. Current interpretation mainly uses the semicircle part of the curves and discards other information revealed by EIS such as the slope of the curve at low frequency. What is worse, some features on the curve are not fully interpreted. To better understand the transport mechanism and interpret EIS curves, we introduce a continuous model to quantify the ion transport and current flow in the electrolyte. The produced EIS curves from the model are compared with experiment data to show good agreement.
-
Thermal Conductivity of 1, 2-Ethanediol and 1, 2-Propanediol Binary Aqueous Solutions at Temperature from 253 K to 373 K
Changyu Deng, and Ke Zhang
International Journal of Thermophysics 2021
1,2-Ethanediol, 1,2-propanediol and their aqueous solutions are widely used as heat transfer fluids. Their thermal conductivity is a vital physical property, yet there are only few reports in literature. In this paper, thermal conductivity of binary aqueous solutions of the two glycols was measured using the transient hot wire method at temperature from 253.15 K to 373.15 K at atmospheric pressure. Measurement was made for six compositions over the entire concentration range from 0 mol to 1 mol fraction of glycol, namely, 0.0 mol, 0.2 mol, 0.4 mol, 0.6 mol, 0.8 mol, and 1.0 mol fraction of glycol. The uncertainties of temperature and concentration measurement were estimated to be 0.01 K and 0.1 %, respectively. The combined expanded uncertainty of thermal conductivity with a level of confidence of 0.95 (k=2) was 2 %. The second-order Scheffé polynomial was used to correlate the temperature and composition dependence of the experimental thermal conductivity, which was found to be in good agreement with the experiment data from the present work and other reports.
2020
-
Geometry Optimization of Porous Electrode for Lithium-Ion Batteries
Changyu Deng, and Wei Lu
ECS Transactions 2020
We report a method to optimize the topology of porous electrodes without a prior assumption of the pattern. We take a two-dimensional NMC-Li cell as an example. The domain of NMC is discretized by a 5×5 mesh and the solid volume fraction is represented by 25 nodal variables. We use a deep-learning-boosted optimization algorithm to find the optimal material distribution that gives the maximum specific energy. The method produces an electrode pattern with 18% higher specific energy than that of a uniform electrode. This generic geometry optimization approach provides a powerful tool for the design and optimization of porous electrodes.
-
Consistent diffusivity measurement between Galvanostatic Intermittent Titration Technique and Electrochemical Impedance Spectroscopy
Changyu Deng, and Wei Lu
Journal of Power Sources 2020
Galvanostatic Intermittent Titration Technique (GITT) and Electrochemical Impedance Spectroscopy (EIS) are two popular methods to measure the diffusivity of lithium ions in electrode particles. What has puzzled the community for a long time is that these two techniques often give an order of magnitude difference in the results. By analyzing the diffusion profile and approximation error for various particle geometries, we show that these two techniques are consistent only when the current excitation does not impact deep inside the particles, which corresponds to the condition of short pulse time for GITT or high frequency for EIS. GITT does not depend on particle size by its assumption while EIS does. Thus we propose an innovative approach of using EIS to determine diffusivity accurately independent of particle size or geometry. We further demonstrate experimentally that the two techniques yield identical results under the right measurement conditions. This work may provide insight on the scattering of measured diffusivity data in the literature.
-
Integrating Machine Learning with Human Knowledge
Changyu Deng, Xunbi Ji, Colton Rainey, Jianyu Zhang, and Wei Lu
iScience 2020
Machine learning has been heavily researched and widely used in many disciplines. However, achieving high accuracy requires a large amount of data that is sometimes difficult, expensive, or impractical to obtain. Integrating human knowledge into machine learning can significantly reduce data requirement, increase reliability and robustness of machine learning, and build explainable machine learning systems. This allows leveraging the vast amount of human knowledge and capability of machine learning to achieve functions and performance not available before and will facilitate the interaction between human beings and machine learning systems, making machine learning decisions understandable to humans. This paper gives an overview of the knowledge and its representations that can be integrated into machine learning and the methodology. We cover the fundamentals, current status, and recent progress of the methods, with a focus on popular and new topics. The perspectives on future directions are also discussed.
2019
-
Ultralight and Flexible Monolithic Polymer Aerogel with Extraordinary Thermal Insulation by A Facile Ambient Process
Man Li, Zihao Qin, Ying Cui, Chiyu Yang,
Changyu Deng, Yunbo Wang, Joon Sang Kang, Hongyan Xia, and Yongjie Hu
Advanced Materials Interfaces 2019
High performance thermal insulation materials are desired for a wide range of applications in space, buildings, energy, and environments. Here, a facile ambient processing approach is reported to synthesize a highly insulating and flexible monolithic poly(vinyl chloride) aerogel. The thermal conductivity is measured respectively as 28 mW (m K)-1 at atmosphere approaching the air conductivity and 7.7 mW (m K)-1 under mild evacuation condition. Thermal modeling is performed to understand the thermal conductivity contributions from different heat transport pathways in air and solid. The analysis based on the Knudsen effect and scattering mean free paths shows that the thermal insulation performance can be further improved through the optimization of porous structures to confine the movement of air molecules. Additionally, the prepared aerogels show superhydrophobicity due to the highly porous structures, which enables new opportunities for surface engineering. Together, the study demonstrates an energy‐saving and scalable ambient‐processing pathway to achieve ultralight, flexible, and superhydrophobic poly(vinyl chloride) aerogel for thermal insulation applications.
-
Numerical study on equilibrium stability of objects in fluid flow — A case study on constructal law
Changyu Deng, Xin Qi, and Yingwen Liu
Case Studies in Thermal Engineering 2019
Stability of objects in fluid flow is an interesting and significant subject in many fields. From thermodynamics point of view, the constructal law by Bejan et al. proposed that in some scenario, objects tend to be stable when the drag reaches maximum. To investigate the relation between drag and stable positions, we analyzed two simple cases by the finite element method: an elliptical cylinder and a rectangular cylinder immersed in 2D uniform laminar flow. The ellipse is stable when its major axis is perpendicular to the mainstream and the drag reaches maximum; yet the rectangle is stable when the length is perpendicular to the mainstream, with drag ranging from minimum to maximum depending on its aspect ratio. Our results show that there is no universal relation between drag and stability.