Estimation of State of Charge of Lithium-Ion Batteries Based on Wide and Deep Neural Network Model

Table 2

Performance comparison between the wide and deep model and BPNN model.

Model	Stage	RMSE (%)	MAPE (%)	Error (%)
Wide and deep	Cell 1 charge	0.115	0.419	[−0.42,+0.38]
	Cell 1 discharge	0.312	0.640	[−0.86,+0.38]
	Cell 2 charge	0.116	0.528	[−0.34,+0.50]
	Cell 2 discharge	0.150	0.327	[−0.48,+0.56]
BPNN	Cell 1 charge	0.128	0.9	[−0.50,+0.45]
	Cell 1 discharge	0.369	0.875	[−1.39,+1.1]
	Cell 2 charge	0.131	0.848	[−0.41,+0.47]
	Cell 2 discharge	0.171	0.412	[−0.77,+0.41]