freqresp_relative_errorΒΆ
Relative error between a reference signal and a second signal.
The error metric is defined as in [1] to be:
\[\varepsilon_{rel}[\mathbf{Y}_1, \mathbf{Y}_2] = \frac{\max_{i,j} (\sup_{w\in[0, \bar{w}]}[\mathbf{Y}_2 -
\mathbf{Y}_1]_{i, j})}{\max_{i,j}(\sup_{w\in[0, \bar{w}]}[\mathbf{Y}_1]_{i,j})}.\]
param y1: | Reference signal frequency response. |
---|---|
type y1: | np.ndarray |
param y2: | Frequency response matrix. |
type y2: | np.ndarray |
param wv: | Array of frequencies. Required when specifying a max and min value |
type wv: | Optional [np.ndarray] |
param **kwargs: | Key word arguments for max and min frequencies. See below. |
- Keyword Args
- vmin (float): Lower bound value to find index in
wv
. vmax (float): Upper bound value to find index inwv
.
returns: | Maximum relative error between frequency responses. |
---|---|
rtype: | float |
References
Maraniello, S. and Palacios, R. Parametric Reduced Order Modelling of the Unsteady Vortex Lattice Method. AIAA Journal. 2020