Algorithm 3: CEEMDAN

Input: The original signal $x$, the amplitude of the added Gaussian noise $\sigma$, and the number of ensemble trials $m$.

Output: Several IMF and a residue, i.e., $IM{F}_k,\left(k=1,2,\dots ,p\right)$ and $r$.

1:  function CEEMDAN ($x$, $\sigma$, $m$)

2:   IMF$\leftarrow$0

3:    residue $\leftarrow 0$

4:   for i = 1 $\to i$= $m$ do

5:     $n_i\leftarrow N\left(\mathsf{\sigma }\ast \mathrm{std}\left(\mathrm{x}\right),1\right)$

6:     $x_i\leftarrow x+n_i$

7:    $IM{F}_i\leftarrow E_1$($x_i$)

8:     $IM{F}_1\leftarrow \frac{1}{m}{{\displaystyle \sum }}_{i}IM{F}_i$

9:   end for

10:    $r_1\leftarrow x-IM{F}_1$

11:   for k = 2 $\to k$ = p do

12:     $IM{F}_k\leftarrow mean\left(E_1\left(r_{k-1}+E_1\left(n_i\right)\right)\right)$

13:     $r_k\leftarrow r_{k-1}-IM{F}_k$

14:   end for

15:    $r\leftarrow r_p$

16:   return IMF and residue

17:  end function

18:

19:  function $E_1$($x$)

20:   $IMF\leftarrow$EMD ($x$)

21:    $IM{F}_1\leftarrow IMF\left(1,:\right)$

22:   return $\mathit{I}\mathit{M}{\mathit{F}}_{\mathbf{1}}$

23:  end function