You have a distribution of sample times. What does that distribution look like? Characterize the distribution you got using all of the usual descriptive statistics for central tendency, Mean, median, range, SD. Graph it out - is it normally distributed?Hi!
Ive aqcuired dig. values from a transmitter, with freq. 12Hz for 49sec.
I got asked to compare sampl. times with theoretical variations in measurement points.
What exactly does it mean and how is it best done?
Thanks alot! I have about 32 samples, where the data is not changing alot (almost static, small changes due to vibrations) and each of the data have a timestamp.You have a distribution of sample times. What does that distribution look like? Characterize the distribution you got using all of the usual descriptive statistics for central tendency, Mean, median, range, SD. Graph it out - is it normally distributed?
Find out what the "theoretical variations" are, e.g., does it say something like 49sec +/- 1 sec?
Compare your observed results with the "theoretical results" - are they in line or not?
That is how I interpret the question. Also, you mention only time and not frequency, I would think that you would want to do both but if you only have duration values then that is all you can work with.
You said you have about 32 samples...of what exactly? What does a line of data for one of your samples look like?
samplenumber timestamp xxx ?
if so, what exactly is xxx
It's confusing because you started out saying "
I've aqcuired dig. values from a transmitter, with freq. 12Hz for 49sec.
I got asked to compare sampl. times with theoretical variations in measurement points."
But, I also see that you are taking 10 samples/sec and that you have ~ 32 samples. If you could provide more details, I think that it would help.
Time s1 s2 s3 s4 s5 s6 s7 s8
2018-08-21 12:56:13,738 18,20511 18,15024 18,12095 18,19822 18,305 18,30021 18,24143 18,34312
2018-08-21 12:56:14,414 18,25105 18,20457 18,12859 18,22646 18,30282 18,24065 18,19152 18,24398
.
.
.
It sounds like you have 8 pressures sensors and you sample all 8 at different times. (calling them transmitters is a little confusing).Sorry, I mixed up my setups.
So: I have a couple (8) of transmitters (measure pressure) connected in serie and that are forming a bus network. I am taking measurements from each of the transmitter and as fast as the transmitters let me. The measurements are taken continuously for about 20 sec and from this I acquire 30 data points from each transmitter. I save the data and the corresponding timestamp (year-month-day hour:minute:sec,msec) for each transmitter.
This is an example of what I store in an excel sheet:
Time s1 s2 s3 s4 s5 s6 s7 s8
2018-08-21 12:56:13,738 18,20511 18,15024 18,12095 18,19822 18,305 18,30021 18,24143 18,34312
2018-08-21 12:56:14,414 18,25105 18,20457 18,12859 18,22646 18,30282 18,24065 18,19152 18,24398
.
.
.
Yes, 8 sensors. Unit is in m.H2O.It sounds like you have 8 pressures sensors and you sample all 8 at different times. (calling them transmitters is a little confusing).
So, what is the unit of measurement? Something like millibars or ?
"connected in serie and that are forming a bus network" What is the significance of the network? Are the sensor just a few inces from each other or is there some significance to their placement? For example, is sensor 1 situated such that the pressure is slightly less?
2018-08-21 12:56:14,414 is the data & time [yyyy-mm-dd hh:mm:ss,msec] that the measurement was taken.Also, what is 12:56:14,414 - what does the 14,414 mean? I see the 8 sensor readings that follow but I don't know if that is something important (I think it is) like a standard, or is it somehow related to time?
what is 2018-08-21 12:56:14,414 18,25105 18,20457 18,12859 18,22646 18,30282 18,24065 18,19152 18,24398
ok so then you want to characterize the error in the sensors. One approach is to convert each of the raw measures (calculated flowrate) to a difference from true flowrate, e.g. the error. Then perhaps calculate a root mean square error for each sensor...in addition to the aforementioned descriptive statistics.I am calc. flowrate for each timepoint with the sensors and compare this calculated value with a true flowvalue.
by Jake Hertz
by Luke James
by Jake Hertz
by Robert Keim