Filtering Curve Data
You can filter curve data to eliminate noise on time signals or to emphasize a specific frequency content of a time signal. Adams/PostProcessor supports two different types of filters:
|
•
|
Butterworth filter - butter() in MATLAB developed by The MathWorks, Inc.
|
|
•
|
Transfer function - A filter you define by directly specifying the coefficients of a transfer function.
|
|
•
|
Analog filtering - The numerical procedure for analog filtering is equivalent to:
|
|
•
|
Digital filtering - Digital filtering operates directly on the time signal. The filtered signal at a certain time step is a linear combination of previous input and output signals, with the discrete transfer function defining the coefficients.
|
You use Filter command on the Plot menu to create or modify a filter function. You can create a Butterworth filter or a transfer function. For the transfer function, you can define the coefficients manually or by defining a Butterworth filter.
|
1.
|
From the Plot menu, point to Filter, and select either Create or Modify. If you selected Modify, select the name of the filter to modify from the Database Navigator.
|
|
Shortcut for creating a filter:
From the Curve Edit toolbar, select the Filter Curve tool  . Right-click the Filter Name text box, point to filter_function, and then select Create. |
|
2.
|
If you are creating a filter, in the Filter Name text box, enter the name you want to assign to the filter function.
|
|
3.
|
Select Butterworth Filter.
|
|
5.
|
Select OK.
|
|
1.
|
From the Plot menu, point to Filter, and select either Create or Modify. If you selected Modify, select the name of the filter to modify from the Database Navigator.
|
|
Shortcut for creating a filter:
From the Curve Edit toolbar, select the Filter Curve tool . Right-click the Filter Name text box, point to filter_function, and then select Create. |
|
2.
|
If you are creating a filter function, in the Filter Name text box, enter the name you want assigned to the filter function.
|
|
3.
|
Select Transfer function.
|
|
5.
|
Enter the numerator and denominator coefficients as explained in the table below. You can enter the coefficients manually or use a Butterworth filter to define them.
|
|
6.
|
To check the format and plot the filter's gain and phase, select Check Format and Display Plot.
|
If you have not defined the filter correctly, an error message appears. If you've defined the filter correctly, a plot appears in which you can switch between the filter's gain and phase plots and change scales.
|
7.
|
To associate comments with the filter function, select the Comments tool
 , and then enter the comments. |
|
8.
|
Select OK.
|
|
To enter the coefficients:
|
|||||||||
|
Using a Butterworth filter
|
|
|
2.
|
|
3.
|
In the Filter Name text box, enter the name of the filter function.
|
|
4.
|
To check if compensation for filter lag should be used, select Zero Phase. (Not available for analog filters.)
|
Zero-phase digital filtering filters the input data in both forward and reverse directions. The resulting sequence has precisely zero-phase distortion and double the filter order.
|
5.
|
Select OK.
|
The following example shows how you can define a transfer function by defining a Butterworth filter. The Butterworth filter that you will create is a Basspand filter, based on specifying two values each for passband and stopband corner frequency.
|
1.
|
|
2.
|
|
3.
|
In the Filter Name text box, enter bandpass_filter1.
|
|
4.
|
Select Transfer function.
|
|
8.
|
Select the Generate Filter Order _ Frequency checkbox.
|
Adams/PostProcessor loads the appropriate values in the Order and Cutoff Frequency text boxes at the top of the dialog box. See the dialog box below for how the values appear.
The following values appear in the Create Filter Function dialog box for numerator and denominator coefficients.
