FIND: A Platform for Hydrogen Storage Alloys

Forward Design: Users can predict key thermodynamic and capacity-related properties, such as plateau pressures (for hydrogen absorption and desorption), enthalpy, entropy, and hydrogen capacity, based on alloy chemical compositions and test temperatures. The platform also supports rapid predictions for multicomponent systems with flexible and undefined elemental ratios, making it suitable for high-throughput screening and composition optimization. 正向设计：用户可利用化学式和测试温度来预测储放氢双侧的平台压力、焓、熵和容量；高级功能支持用户对不定元素摩尔比下进行多组元成分快速预测。
Inverse Design: Powered by a variational autoencoder (VAE) model, FIND enables users to input target performance metrics (desired plateau pressure, hydrogen capacity, and operating temperature) and outputs the top 10 candidate alloy compositions most likely to meet these criteria. This reverses the traditional trial-and-error approach, facilitating performance-driven materials discovery. 逆向设计：基于变分自编码器(VAE)模型，根据用户提供的温度、平台压力和容量等性能目标，逆推出可能满足性能目标的前十名材料化学式。

Training Set Element Distribution 训练集元素分布

H

He

Li

Be

B

C

N

O

F

Ne

Na5

Mg1186

Al417

Si

P

S

Cl

Ar

K

Ca100

Sc11

Ti1567

V636

Cr1099

Mn1383

Fe1199

Co305

Ni1502

Cu108

Zn18

Ga10

Ge

As

Se

Br

Kr

Rb

Sr

Y506

Zr955

Nb93

Mo85

Tc

Ru3

Rh5

Pd13

Ag32

Cd4

In37

Sn20

Sb

Te

I

Xe

Cs

Ba

La938

Hf15

Ta1

W12

Re4

Os

Ir

Pt

Au

Hg

Tl

Pb4

Bi7

Po

At

Rn

Fr

Ra

Ac

Rf

Db

Sg

Bh

Hs

Mt

Ds

Rg

Cn

Nh

Fl

Mc

Lv

Ts

Og

Ce342

Pr138

Nd125

Pm

Sm149

Eu

Gd48

Tb

Dy

Ho6

Er15

Tm

Yb

Lu

Th

Pa

U

Np

Pu

Am

Cm

Bk

Cf

Es

Fm

Md

No

Lr

count 1000+ 100-999 10-99 1-9 0

Model Performance Metrics 模型精度与性能指标

Target	Properties	Feature Number	MAE	RMSE	MAPE	R²
Absorption	ln(P_eq)/MPa	132	0.286	0.676	26.02%	0.940
	ΔH/kJ·mol⁻¹	132	0.844	1.735	3.16%	0.993
	ΔS/J·mol⁻¹·K⁻¹	132	3.168	7.128	3.88%	0.939
	Capacity/wt%	132	0.100	0.181	5.15%	0.984
Desorption	ln(P_eq)/MPa	132	0.310	1.192	25.80%	0.895
	ΔH/kJ·mol⁻¹	132	0.832	2.628	2.15%	0.983
	ΔS/J·mol⁻¹·K⁻¹	132	3.104	8.626	3.19%	0.890
	Capacity/wt%	132	0.092	0.178	5.11%	0.984