Ski-slope misunderstanding

Ski slope is a graphical pattern representative to the received data at the end of the data acquisition and signal processing.

Typically it's associated with a fault in the hardware of the data acquisition system.

However, there are many instances where the instrument loop is completely healthy and the data shows a ski slope. In general if the object being monitored suddenly (and significantly) change its position, the ski slope can be generated depending on the waveform collected. For example, assume a shaft mounted on fluid film bearings is monitored by proximity probes. When the shaft encounters a sudden force (maybe a process upset), the waveform will not be symmetric about the x-axis and this will fool the FFT resulting in a high peak at the low frequency side of the spectrum.

I hope my understanding is correct.

Regards- Ali M. Al-Shurafa

Shurafa posted:

Ski slope is a graphical pattern representative to the received data at the end of the data acquisition and signal processing.

Typically it's associated with a fault in the hardware of the data acquisition system.

However, there are many instances where the instrument loop is completely healthy and the data shows a ski slope. In general if the object being monitored suddenly (and significantly) change its position, the ski slope can be generated depending on the waveform collected. For example, assume a shaft mounted on fluid film bearings is monitored by proximity probes. When the shaft encounters a sudden force (maybe a process upset), the waveform will not be symmetric about the x-axis and this will fool the FFT resulting in a high peak at the low frequency side of the spectrum.

I hope my understanding is correct.

Regards- Ali M. Al-Shurafa

Hi dear all

Because of eliminate this transient vibration of process and other noise of machine at low frequency (usually below 15 hz) ,recommended to set high pass filter from 10 Hz   

Best regards

I have seen many real world instances where "ski slope" was a result of high frequency sensor overload and not a hardware issue. High frequency sources such as pump cavitation and paper machine dryer can flooding (turbulence), high speed gearmesh frequencies etc... can give sensor saturation effects. Google the terms Nyquist, hard and soft clipping, intermodulation distortion and signal saturation as well as accelerometer ski slope for more information.

ckyjohn posted:

I have seen many real world instances where "ski slope" was a result of high frequency sensor overload and not a hardware issue. High frequency sources such as pump cavitation and paper machine dryer can flooding (turbulence), high speed gearmesh frequencies etc... can give sensor saturation effects. Google the terms Nyquist, hard and soft clipping, intermodulation distortion and signal saturation as well as accelerometer ski slope for more information.

If a machine has the capability of malfunctions that exhibit high frequency, and the device installed to monitor that high frequency goes into overload, then I see that as a hardware issue that failed to properly detect a machinery malfunction.  Someone has incorrectly made a transducer choice for that specific malfunction.  For instance, on many a high speed gearbox I recommend a 25 mV/g accelerometer be used as opposed to the typical 100 mV/g accelerometer.

Yes of course changing accelerometers to adjust sensitivity and adding high pass filters in software can help cut this off. The question was asked if the ski slope represents "bad" data and the answer is not always. For a high speed gearmesh, of course change to a lower sensitivity. Would one change some 400 - 100 mv/g accelerometers on a paper machine because excessive dryer can flooding causes ski-slopes...not a chance. Would they add a high pass filter in the collection spec, most likely. In this specific case however it was seen as a useful tool in diagnosis of a process condition which in itself is an interesting special case. There are many possibilities to many issues and not one correct answer for this and most questions the vibration field. Better answers come from supplying the background to the question and information surrounding the issue.

Add Reply

×
×
×
×