Skip to main content
. 2021 Feb 4;21(4):1054. doi: 10.3390/s21041054
AE acoustic emission
b-AE burst-type AE
c-AE continuous-type AE
CNN convolution neural network
LSTM long short-term memory
x(t) AE signal
ω frequency factor
Xhalf(ω) Fourier transform of x(t) (the positive-frequency part)
e the base of natural logarithms
δ the delta function
u(t) the unit step signal
α the damping factor
A the amplitude of b-AE
ht the output of LSTM cells
ct the internal state of LSTM cells
xt the input of LSTM cells
ft the output of the Sigmoid gate of forgetting
it the output of the Sigmoid gate of input
ot the output of the Sigmoid gate of output
c˜t the output of the tanh gate
bf the bias parameter of the Sigmoid gate of forgetting
bi the bias parameter of the Sigmoid gate of input
bc the bias parameter of the tanh gate of the LSTM cell
bo the bias parameter of the Sigmoid gate of output
Wf the forgetting weight matrix
Wi the input weight matrix
Wo the output weight matrix
Wc the state weight matrix of the LSTM cell
σ the Sigmoid function
σf the Sigmoid gate of forgetting
σi the Sigmoid gate of input
σo the Sigmoid gate of output