Comparison of Machine Learning Models on Performance of Single- and Dual-Type Electrochromic Devices

Abstract

This study shows that the model fitting based on machine learning (ML) from experimental data can successfully predict the electrochromic characteristics of single- and dual-type flexible electrochromic devices (ECDs) by using tungsten trioxide (WO 3 ) and WO 3 /vanadium pentoxide (V 2 O 5 ), respectively. Seven different regression methods were used for experimental observations, which belong to single and dual ECDs where 80% percent was used as training data and the remaining was taken as testing data. Among the seven different regression methods, K -nearest neighbor (KNN) achieves the best results with higher coefficient of determination ( R 2 ) score and lower root-mean-squared error (RMSE) for the bleaching state of ECDs. Furthermore, higher R 2 score and lower RMSE for the coloration state of ECDs were achieved with Gaussian process regressor. The robustness result of the ML modeling demonstrates the reliability of prediction outcomes. These results can be proposed as promising models for different energy-saving flexible electronic systems.

Physical Simulation Model of WO3 Electrochromic Films Based on Continuous Electron-Transfer Kinetics and Experimental Verification

“…13 Based on the framework of electrochemical theory and the ionic diffusion model (E−D model), coupling of multiphysics was an effective way to expand the electrochromic model, and the finite element method (FEM) can provide a powerful calculation for complex multiphysics electrochromic models. 14 In addition, some other electrochromic models were also proposed based on assumptions and empirical equations, such as the transmittance versus potential (T−E) empirical model, 15 adaptive neuro-fuzzy inference model, 16 charge versus current empirical (Q−I) model, 17 machine learning model, 18 and so forth. These empirical models can agree well with the experimental data by adjusting the model parameters, but the physical nature behind the parameters was usually unclear.…”

A Machine Learning Approach for Metal Oxide Based Polymer Composites as Charge Selective Layers in Perovskite Solar Cells

“…[14][15][16] For an ML based data-driven approach, it is necessary to couple experimental data with the modelling. [17] An extensive amount of data is not required and computing time is fast compared to other complicated models. [18] By employing ML technique in solar cells, material properties, optimized device architects and fabrication processes can be predicted, and data reconstruction is attracting significant interest for research and development.…”

“…It will replace the traditional trial and error method which demands longer time and resources to predict the performance and reliability of PSCs (stability, PCE, fabrication techniques and material synthesis) [14–16] . For an ML based data‐driven approach, it is necessary to couple experimental data with the modelling [17] . An extensive amount of data is not required and computing time is fast compared to other complicated models [18] .…”

Toward the Prediction of Electrochromic Properties of WO3 Films: Combination of Experimental and Machine Learning Approaches

“…Moreover, a limited number of samples can be prepared regarding the time and resources that it takes. In this context of designing properties of materials, machine learning can be a useful tool to help experimentalists. − This strategy has already been used for the discovery of new compounds in material science , and in the prediction of material properties , while having been scarcely explored in the field of electrochromic materials . The aim of this study is to demonstrate that the traditional experimental approach can be nicely complemented by the use of machine learning algorithms to help find sputtering conditions leading to the desired electrochromic properties: color persistence (i.e., retention of color while no potential is applied) as well as good reversibility.…”

“…16−18 This strategy has already been used for the discovery of new compounds in material science 19,20 and in the prediction of material properties 21,22 while having been scarcely explored in the field of electrochromic materials. 23 The aim of this study is to demonstrate that the traditional experimental approach can be nicely complemented by the use of machine learning algorithms to help find sputtering conditions leading to the desired electrochromic properties: color persistence (i.e., retention of color while no potential is applied) as well as good reversibility. Such a combination allows for a significant decrease of the experimental research duration.…”