Table 10 Performance indices of libs with Mcmi profile for batch ‘2018-04-12’.

From: Hybrid optimized remaining useful life prediction framework for lithium-ion batteries with limited data samples

LIB

Executed

algorithms

Performance indices

 

RMSE

MSE

MAPE

RUL error

Training Time (s)

Inference latency (ms)

C33

FNN

0.0720

5.18*10− 5

0.0545

−3

0.5038

0.35

JFO-FNN

0.6973

0.0049

0.4801

−20

151.74

0.45

PSO-FNN

0.9507

0.0090

0.8606

+ 60

118.31

0.42

C34

FNN

0.4069

0.0017

0.1409

−4

0.6628

0.36

JFO-FNN

0.2877

8.27*10− 4

0.1066

0

223.55

0.47

PSO-FNN

0.8363

0.0070

0.2211

−4

170.26

0.44

C35

FNN

0.8413

0.0071

0.4271

+ 1

0.8019

0.38

JFO-FNN

0.2692

7.24*10− 4

0.1252

+ 1

150.33

0.48

PSO-FNN

0.3791

0.0014

0.1711

+ 4

93.39

0.43

C36

FNN

0.1324

1.75*10− 4

0.1115

+ 6

0.4523

0.34

JFO-FNN

0.0949

9*10− 5

0.0862

+ 12

190.57

0.46

PSO-FNN

0.1385

1.91*10− 4

0.1074

+ 4

144.37

0.41

Back to article page