How does DIAdem 2010 calculate the RMS value in analysis?

[DIAdem2010RMS]

I have a time channel with a monotonic increasing time (20 seconds) and a data channel with 20 numbers. How does DIAdem calculate the RMS valua when i set the "One-sided interval width as % of the channel length" to 5%? Can i reproduce this values with Excel?

For example this is the data i am analysing. (time channel at the left, data channel at the right)

01/01/0000 00:00:00.0000 5
01/01/0000 00:00:01.0000 5
01/01/0000 00:00:02.0000 5
01/01/0000 00:00:03.0000 5
01/01/0000 00:00:04.0000 5
01/01/0000 00:00:05.0000 5
01/01/0000 00:00:06.0000 5
01/01/0000 00:00:07.0000 5
01/01/0000 00:00:08.0000 5
01/01/0000 00:00:09.0000 5
01/01/0000 00:00:10.0000 10
01/01/0000 00:00:11.0000 10
01/01/0000 00:00:12.0000 10
01/01/0000 00:00:13.0000 10
01/01/0000 00:00:14.0000 10
01/01/0000 00:00:15.0000 10
01/01/0000 00:00:16.0000 10
01/01/0000 00:00:17.0000 10
01/01/0000 00:00:18.0000 10
01/01/0000 00:00:19.0000 10

[Ralf_Mueller]

Hi,

There is a slight problem in the formula given in the DIAdem help (which practically only matters for short channels) and will be fixed with the next DIAdem version. The difference is that the first and the last value of the sum are devided by 2:

sqrt ( 1/(N-1) (0.5*y_1^2 + y_2^2 ....+...+ y_(N-1)^2 + 0.5*y_N^2)

In your example 5% of the channel lenght results in 1 neighbour value and leads to the following result:

5; 5; 5; 5; 5; 5; 5; 5; 5; 6.61437827766148; 9.01387818865997; 10; 10; 10; 10; 10; 10; 10; 10; 10

Let's have a look at the values in the middle:

RMS(10)= sqrt( 0.5* ( 0.5*5^2 + 5^2 + 0.5* 10^2)) = sqrt(175/4) = 2.5 sqrt(7) ~ 6.614

RMS(11)= sqrt( 0.5* ( 0.5*5^2 + 10^2 + 0.5* 10^2)) = sqrt(325/4) = 2.5 sqrt(13) ~ 9.014

I'm sorry for the mistake in DIAdem help, which came to our knowledge shortly after the release of DIAdem 2011 and will be fixed in DIAdem 2012.

Best regards,
Ralf

[xzhcong]

Hello,

     For calculating vibration RMS,what value of "the one-sided interval width of the average" should be filled in the RMS dialog box?

I don't know exactly what does this parameter maters.

Thank you!

Xu

[Ralf_Mueller]

Hello Xu,

The RMS function replaces each individual value of a channel by the RMS value calculated from the actual value and a specified number of neighbor values. The one-sided intervall width N specifies the number of neighbor values which DIAdem uses for averaging on each side of the actual value. In other words if you choose a "one-sided intervall width" of 10, the RMS is calculated on a moving window of the value itself and 10 values on each side, resulting in 21 samples. If the averaging window moves towards the beginning or the end of your data channel, there are less than 10 values on the edge of your channel. On the first (and the last) data point in your channel DIAdem would use 11 values, on the second 12 values and so on until the normal width of 21 data points is reached.

Hope this helps,

Ralf

[xzhcong]

Hello,

    Thank you! And what value would ususlly be used as the one sided interval with?10% or other value?

Xu

[xzhcong]

Hello Ralf,

   Good morning,

    As the difference of "One-sided interval width as % of the channel length" would result in the different RMS ruslt.

As in the attachment.So I would like to know what value will fit for my test data? How to decide it?

second, I would like to know about RMS calculation at the begin and last value.could decrisb them by fomular?

Thank you!

[Ralf_Mueller]

Hello Xu,

Please find the formula of the RMS calculation below:

The moving window includes N values. The first and the last value is divided by 2 and thus the overall sum is divided by (N-1) instead of N.

As you mentioned that you're trying to calculate the RMS for vibration signals I expected that your measurement data contains oscillations. I'm not sure whether the RMS calculation makes sense with your rather static data.

In general I don't know a rule of thumb but I guess one would choose the intervall length such that it includes one or several periods of the dominant oscillation.

Sorry that I couldn't be more helpful,

Ralf

[toybear]

hallo Ralf,

do you know the equation of the floating rms in the frequency weighting calculation?

i am a little confused with the result of the floating rms.

the floating rms result of my vibration acceleration signal with average time 0.25s is the same with the running rms result (specified in VDI 2057) with integration time 0.125s.

do you know the definition of the average time in the DIAdem?

thank you.