Vibration Data Collection Error Sources and their Effects

From Imrtwiki

Return to Vibration Monitoring

Vibration Data Collection Error Sources and their Effects

The most important concern in vibration data collection is the accuracy of the data collected. If information collected is incorrect, at best, the route will have to be collected again and at worst, incorrect judgements may be made with large financial consequences. This section explains way errors can occur and how they can be avoided.

Measuring the wrong equipment - This can be an easy error to make with equipment that is similar and with confusing naming systems but must be avoided. Ensure you have a diagrammatic layout of each area identifying machines. It is also useful to have a more detailed text descriptions of equipment location and naming details. If while measuring on a machine, a number of measurement points show change, immediately re-evaluate if it is the correct machine.

Measuring the wrong machine point or direction - Understand the machine numbering and direction conventions. Look for the transducer location spot markers and repaint these if becoming obscured. Marking systems are not always perfect, so you must know how to find the standard transducer locations on the standard machine housing arrangements. If a measurement is changed from the previous reading, immediately re-evaluate if you have selected to correct transducer location. If you are not absolutely certain that the correct location has been selected, make a note indicating the uncertainty on the route record sheet.

Poor cable connections & cable movement - The electrical signals produced by transducers have quite small voltages and currents. Small changes in the electrical resistance of connectors & connecting cables can substantially affect measured values. Connectors should be regularly cleaned with electronic cleaning spray and connectors and cables checked for damage. Some connectors have sealing ‘O’ rings that should be inspected. It is important to check cables and connectors before going onto site.

Connect the standard transducer and measure overall Velocity Vibration data while moving the cable and wiggling connector. Ensure that the cables and connectors movements do not significantly affect the measured values.

Regularly check for cable effects while collecting data. Some causes of these effects are:

• Damaged cable
• Damaged connectors
• Moisture or dirt entry
• Ineffective shielding

Poor accelerometer to magnet attachment - Both the accelerometer and magnet connection surfaces should be correctly machined and free of damage. The faces should be cleaned, a connection fluid added and tightly coupled by hand. Coupling tension should be regularly checked.

Poor magnet to machine attachment - A strong connection is required between the machine and the magnet with a good vibration transmission path between the vibration energy source and the accelerometer. Both the machine point and the magnet have to be clean. A flat magnet base requires a flat surface. A slight sliding action on attachment is recommended to test that it is corrected seated.

A 2 pole magnet can be used on more rounded surfaces. A slight twisting or sliding attachment is recommended to test that each pole is well attached

A final check should be made by gently holding the top of the accelerometer & trying to wobble it in a circular pattern to ensure attachment strength is good in all directions.

Inadequate measurement stabilization - When an accelerometer is attached, it requires a short time for the reading to stabilise. Longer if attached with an impact. Place the edge of the magnet down first and pivot down to full contact to minimise the impact. Look at the instrument reading and ensure it is stable before recording the value.

Most data collectors display a trend line or a bar to observe the current measured value. Do not press store until the measured value is stable.

Fluctuating vibrations - Vibrations in some situations can vary in a slow cyclic or other pattern. Record vibrations when vibrations are at the greatest magnitude if practical. Describe the vibration pattern on your route record sheet.

Change in background vibrations - Large vibrations from nearby equipment (eg. Crusher or materials chute) can often affect vibration levels on the machine being measured. If these background vibrations vary, collect vibrations when the background interference is lowest. Note the presence, level and source of the background vibrations to assist in future trend analysis.

Vibration signal high amplitude clipping - High frequency, high amplitude vibration signals can cause the accelerometer amplifier to saturate and become overloaded. An accelerometer has an acceleration value over which it will not measure correctly. Some reasons for such high amplitude vibration causing accelerometer overload are severe pump cavitation, fluid release, impacts from loose or reciprocating parts and even gear mesh noise. If you record a very high value, measure readings away from this point to check that the original reading was sensible and to determine how localised the high amplitude vibration is.

This problem can be checked if you can do a vibration spectrum analysis or time waveform of the vibration. The overload problem will produce truncated waveforms and large ‘ski-slope’ spectrums (see below). The example shows how sensor overloading destroyed the frequency spectrum on a gearbox. If high acceleration amplitudes or impacts are expected then the spectrum should be checked to ensure it has no significant ‘Ski Slope’ (see lower spectrum)

The specification of accelerometer or magnet type being used will have a significant influence on amplitude clipping. Different magnet designs and accelerometer designs have different frequency responses and different accelerometers will have maximum amplitude specifications. This will have a marked effect on high frequency parameters such as acceleration and ultrasonic noise. Where a change in accelerometer to one with a higher frequency response is made, it will often give a significant step change in amplitude of high frequency parameters. Change from a flat magnet to a hand held probe can significantly reduce high frequency amplitude (Figure below).

Transducer temperature - Temperature has an effect on the electrical characteristics of an accelerometer (see diagram). If equipment measured is very hot, try to minimise the sensor contact time to reduce heat soak into the accelerometer. Even with termperature effects, repeatable reading can be taken if accelerometer contact time with equipment is kept consistent. A thermally insulating pad can be placed between the accelerometer and the magnet to minimise the effect. A 5% increase in acceleration amplitude at 120DegC would be typical.

Touching a transducer after contact with a hot surface can cause burns. Care is required and gloves should be used if the high temperature is normal.

Effects of cable length - The portable accelerometers typically used with data collectors have internal amplifiers and so cable length should not be a major factor. Where very long cable length (>50meters) are used for permanent accelerometers there may be an influence on outputs. The same is the case at much shorter lengths when non-internally amplified accelerometers are used.

Damaged transducer - A transducer that is damaged internally can produce substantial errors. If a transducer consistently measures Velocity Vibrations that are significantly different than that felt by the hand, then compare the readings measured by another vibration instrument. When measurements are being taken for a vibration monitoring system, a damaged accelerometer will cause many or all of the collected data values to change. Vibration Technicians which should wary of this.

Electrical interference - The signal going from a vibration transducer to the vibration instrument is very small and can be effected by magnetic and electric fields. The electrical shielding on cables is supposed to eliminate this effect. If you notice unusual measurement variations around electric motors or similar electrical equipment, the transducer cable shielding may be damaged and may require cable replacement.

Accelerometer contact force and direction - The contact force between an accelerometer and the machine can have a significant effect on measurements, especially high frequency vibrations. If very high frequency vibrations are important for monitoring, the machine may be drilled and tapped for stud mounting of an accelerometer. With very high vibrations the momentum of a transducer may overcome magnetic force and cause transducer movement & measurement error. If this is suspected, keep hand on magnet. If a handheld accelerometer probe is used then it is important to hold the probe steadily onto the equipment with a reasonable force (0.5 to 1 kg force). Typically the probe is held in the direction indicated (H, V, R or A) and usually perpendicular (90o) to the surface. If that is not possible then the details should be recorded and the same angle should be maintained for each future measurement.

Identifying, Correcting & Rejecting Bad Data

Measurement Fluctuations

Regular sinusoidal variation in vibration amplitude - If there are two vibrations in a machine that are almost the same frequency, but not quite, a vibration ‘Beat’ will occur. These vibrations go in and out of phase, which varies the amplitude. This is not usually dangerous but should be noted. If the variation is slow, measure vibration when the amplitude is highest. Sometimes variable speed machines can also produce vibration variations, if the speed control system is slightly unstable. Note this if suspected.

Irregular variation in vibration amplitude - Check magnet or transducer connection to the machine. If due to high vibration or surface shape, reposition magnet or stabilise with force applied by your hand. Check cable and cable connections by manipulating and seeing the effect on vibration output. Clean connections and if still a problem, minimise cable movement while collecting data, note issue and if it is a problem, organise to get cable replaced.

Check for presence, level and possible source of the background vibrations that could be affecting vibration amplitudes. Large vibrations from nearby equipment can be a cause eg. Crusher or materials chute. Collect vibrations when the background interference is lowest and note the issue.

If acceleration levels are very high and there is sharp amplitude jumps, there may be transducer overload causing. If overloading is confirmed try different attachiment arrangements or try placing rubber material under the transducer.

Measurement Changed from Previous Value –

• Check if wrong machine, wrong point location or wrong direction has been measured.
• Check transducer and magnet attachment and double check measurement
• Comparing measured values to 5 senses (hand) measurements
• If measured value is low and variation is less the 30%, treat it as normal.

Measurement lower than expected - It is unusual for equipment vibrations to reduce, without a repair or other changes to the causes of the vibration eg greasing, realignment, machine balance etc. Check for evidence of maintenance. Other issues that can effect vibrations are load changes, speed changes, process variations or nearby machines not running. These are also relevant for the item below.

Measurement higher than expected - Check for possible causes of higher vibration by 5 senses inspections. More detailed 5 senses inspections to be performed if the increase is more than 100%, the original value was high, a trend (or other pattern) is recognised or a defect was already identified.