A - CBM in Rotating Machines Chapter 15

CHAPTER 15

MEASUREMENTS IN ROTATING MACHINERIES

In chapters 13 and 14, we studied some of the experimental methods to estimate unbalances and bearing dynamic parameters, respectively. These methods involve measurement of input (forces) and output (vibration responses) in time domain. For subsequent processing often these measurements are required in frequency domain. In the present chapter, we would describe the overall measurement and analysis systems. General terminologies associated with measurement systems are presented. Sensitivity analyses of the estimated parameters due to errors involved in the measurements are presented. Various kinds of transducers, the conditioning and analyzing instruments, and vibration exciters are described especially those are suited for the measurement in rotating machineries. Transducers include the displacement, velocity, acceleration, force and acoustic transducers. In subsequent chapter, the focus would be to describe the basic techniques of the signal processing and associated error involved.

From experiments quantities that are desired may be the velocity, acceleration, displacement, force, and its phase. These quantities may be useful in predicting the fatigue failure of a particular machine element of a machine or may play important role in analyses, which are used to reduce the structure vibration or noise level. It may be useful in estimating system parameters, while the force is also measured which causes the vibration, by the modal analysis or model updating. The central problem in any type of motion or vibration measurement concerns a determination of the appropriate quantities in reference to some specified state, i.e., the velocity, displacement, or acceleration with reference to the ground. A vibration transducer is connected to the machine element in motion, and it gives an output signal proportional to the variational input. The transducer should be independent of its application, i.e., it should function equally well whether it is connected to a vibrating structure on the ground, in an aircraft or in a space vehicle. Sound may be classified as a vibratory phenomenon, and we shall discuss some of the important parameters used for specification of sound level. The measurement and analysis of sound levels (or signals) is very specialized subject, which are becoming increasingly important in modern rotating machinery design.

Machinery acoustics and vibrations are measured to monitor the condition of the machine. It enables to detect the machine fault so that it can be corrected as soon as possible. High levels of noise and vibration are indicative of high levels of component stress, high noise levels, and reduced machine fatigue life. Measurements are usually taken of the system acoustics and vibration amplitude, its phase and its frequency. The acoustics and vibration may be composed of several sinusoidal signals all at

899

different frequencies and it is necessary to distinguish the components signals from each other. These measurements can be processed and displayed in such a way as to enable judgments to be made about the condition of the machine. It will help in the diagnosis of some fault conditions by estimation of dynamic parameters of machine components and of faults.

When we investigate the causes of vibration, we first investigate the relationship between frequencies and the rotational speed. We can do such spectrum analysis using spectrum analyzer equipments (i.e., by the Fast Fourier transformation). Spectrum analysers have various convenient functions, such as the tracking analysis, Campbell diagram, and waterfall diagram. In tracking analysis, dynamic characteristics of a rotating machine are investigated by changing the rotational speed. A Campbell diagram is the variation of whirl frequency with respect to the rotor spin speed. A waterfall diagram is a 3-dimensional plot of the spectra at various speeds.

Vibration testing of rotors involves availability of various hardware and software components such as one schematically shown in Figure 15.1, which shows a typically layout for a simple measurement system. Basically, there are three main measurement mechanisms: (i) the exciter mechanism, (ii) the sensing mechanism and (iii) the data acquisition, conditioning and processing mechanism. In the present chapter all of these modules (except test rigs) will be described in detailed.

Figure 15.1 A simple measurement system

15.1 Specifications of Measuring Instruments An important part of the performance of rotating machinery depends on the efficiency of the vibration (displacement, velocity, acceleration, and force) sensors that are used. In order to measure the position of a rotating rotor, contact-free sensors must be used which, moreover, must be able to measure a rotating surface. Consequently, the geometry of the rotor, i.e., the surface quality, and the homogeneity of the material at the sensor location will also influence the measuring results. A bad surface will thus produce noise disturbances, and geometry errors may cause disturbances with the rotational frequency or with multiples thereof. Some of the terms which are used in specifications of measuring systems are described below:

Readability: of an instrument indicates the closeness with which the scale of the instrument may be read; an instrument with a 10 cm (or 120o) scale would have a higher readability than an instrument

Exciter mechanism

Test rig

Sensing mechanism

Data acquisition, conditioning

and processing mechanism

900

with a 5 cm (or 60o) scale for the same range of a measuring parameter (e.g., 100 m or 10 m/s or 100 m/s2, or 1 kN). With modern digital display (e.g., liquid crystal display: LCD) readability is related with the display size and its brightness relative to the varied ambient brightness (e.g., in the aircraft cockpit).

Least count: It is the smallest difference between two indications that can be detected on the instrument scale. It depends upon the scale length, spacing of graduations, size of pointer, and parallax effects. For the digital display it is the lowest decimal point of the measured physical quantity that can be seen in the display.

Hysteresis: An instrument exhibit hysteresis when there is a difference in readings depending on whether the value of the measured quantity is approaches from above or below. It may be the result of mechanical friction, inertia, magnetic effects, elastic deformation, or thermal effects.

Accuracy: For an instrument it indicates the deviation of the reading from a known input (true values). The deviation is called the error. In many experimental situations we may not have a known value with which one can compare instrument readings and yet we feel fairly confident that the instrument is within the plus or minus range of the true value. In such cases the plus or minus range expresses the

uncertainty of the instrument. Accuracy is usually expressed as a percentage of full-scale readings, for

example, for a 100 m displacement dial gauge with an accuracy of 1 percent would be accurate

within 1 m over the entire range of the gauge.

Precision: The precision of an instrument indicates its ability to reproduce a certain reading with a given accuracy. The difference between the instruments reported values during repeated measurements of the same quantity. As an example of the distinction between precision and accuracy, consider the measurement of a known speed of 1000 rpm with a tachometer. Five readings are taken, and the indicated values are 1040, 1030, 1050, 1030 and 1050 rpm, which has maximum deviation from the actual value of 50 rpm, average value of 1040 rpm, and the maximum deviation from the measured mean value is 10 rpm. From these values, it is seen that the instrument could not be

depended on for an accuracy of better than 5 percent (50100/1000), while a precision of 1 percent (10100/1040 = 0.96 1%) is indicated, since the maximum deviation from the mean reading of 1040 rpm is only 10 rpm. It may be noted that the instrument could be used to dependably measure speed

within 10 rpm. Hence, the accuracy gives the measure of absolute error, and the precision gives that

of the relative error. This simple example illustrates that the accuracy can be improved up to but not beyond the precision of the instrument by calibration.

901

Resolution: In addition to the useful signal, each sensor system produces noise disturbances in the output signal. The smallest increment of change in the measured value that can be determined from the instruments readout scale is called the resolution. Typically this value is often on the same order as the precision; sometimes it is smaller. The minimum value of the useful signal, which can be distinguished from the noise disturbance (mostly peak-to-peak value of the noise disturbance), is called resolution. The resolution is usually indicated in absolute values - for instance in m for a displacement sensor. Noisy signal cannot be improved by resolution, however, can often be improved by low-pass filters at the expense of the frequency range. For spectrum analyzer a resolution of 1 Hz is very common.

Sensitivity: The change of an instrument or transducers output per unit change in the measured quantity. A more sensitive instruments reading changes significantly in response to smaller changes in the measured quantity. Typically, an instrument with the higher sensitivity will also have finer resolution, better precision, and higher accuracy. The sensitivity indicates the ratio of the signal over

the quantity to be measured: for a displacement transducer, for instance, it is indicated in mV/m. For example, a 7.8-mV proximity voltage is equivalent to 1-mm displacement then its sensitivity would be 7.8 mV/mm; here it is assumed that the measurement is linear for the give displacement. Similarly, for the velometer, accelerometer and force transducers the sensitivity are indicted in mV/m-s-1,

mV/m-s-2 (or C/m-s-2) and mV/N, respectively. Generally, a operating (linear) range of the measurement is specified with the transducers. The sensitivity can be enhanced by the electronic amplification of the output signal.

Calibration: The calibration of all instruments is important, since it checks the instrument against a known standard (comparing with another instrument of known accuracy, i.e., the accuracy of the instrument must be specified by a reputable source) or known input source (direct calibration with a primary or alternative measurement procedure) and subsequently to reduce error in accuracy. It is the calibration, which firmly establishes the accuracy of the instruments. In principle, the calibration has to be performed before taking important measurement or at least periodically to ensure quality of the measured data. Since during operation due to improper handling of instrument, there is possibility of damage.

Measuring range: The output signal of a sensor changes according to a physical effect as a function of the measured quality. The range in which the output signal can be used often corresponds to that range having an approximately linear correlation between measured quality and output signal (i.e., the specified sensitivity is valid in this range of operation). This linear measuring range can be considered smaller than the physical one, where nonlinear effects also will be there. For example, the proximity

902

sensor can have linear measuring range as 0 to 1 mm, and for general purpose accelerometers the range would be up to 1000 g (1 g = 9.81 m-s-2).

Linearity: The linearity is usually represented as a percentage of the maximum measuring range. It shows to what extent the measured quantity deviates from a linear relationship between measured quantity and output signal.

Frequency range: A linear frequency response, i.e., a sensitivity independent of the frequency, is necessary in some applications, especially when working with the displacement and accelerometer transducers. The frequency with a sensitivity reduced by 3 dB is usually called cut-off frequency. One must consider here that the output signal at the cut-off frequency, depending on the transducer, may already show a significant phase lag. For general purpose accelerometers the frequency range of operation is usually up to 1.2 kHz with resonance frequency of the accelerometer of the order of 40 kHz.

Impedance matching: In many experimental setups it is necessary to connect various items of electrical equipment in order to perform the overall measurement objective. When connections are made between electrical devices, proper care must be taken to avoid the impedance mismatching.

Figure 15.2 Two-terminal instrument with internal impedance Ri and external load of R

The input impedance of a two-terminal device may be illustrated as in Figure 15.2. The device behaves as if the internal resistance Ri is connected in series with the internal voltage source E. The connecting terminals for the instrument are designated as A and B, and the open circuit voltage presented at these terminals is the internal voltage E. Now, if an external load R is connected to the device and the internal voltage E remains constant, the voltage presented at output terminals A and B will be dependent on the value of R. The potential presented at the output terminals is

11 ( / )AB i i

RE E ER R R R

= =

+ + (15.1)

903

The larger the value of R, the more closely the terminal voltage approaches the internal voltage E.

Thus, if the device is used as a voltage source with some internal impedance, the external impedance (or load) should be large enough that the voltage is essentially preserved at the terminals. If one wish to deliver power from the device to the external load R. The power is given as

2ABEPR

= (15.2)

The value of the external load that will give the maximum power for a constant internal voltage E and internal impedance Ri can be obtained as follows. On substituting equation (15.1) into equation (15.2), we get

( )22 2

2i i

E R E RPR R R R R

= =

+ + (15.3)

and the maximizing condition 0dPdR

= is applied. It results in

( ) ( ) ( ) ( )2 2 2

2 2 32 0; 2 0; i

i i i

dP d E R E E R R R RdR dR R R R R R R

= = + = + = + + +

iR R= (15.4)

That is, the maximum amount of power may be drawn from the device when the impedance of the external load just matches the internal impedance. This is the essential principle of impedance matching in electric circuits. The internal impedance and external load of a complicated electronics device may contain the inductive and capacitive components that will be important in AC transmission and dissipation. However, the basic idea is the same. The general principle of impedance matching is that the external impedance should match the internal impedance for maximum energy transmission (minimum attenuation), and the external impedance should be large compared with the internal impedance when a measurement of internal voltage of the device is desired.

The impedance matching can be important in mechanical systems also. Consider a simple spring-mass system as a mechanical transmission system. From the frequency response function describing the system behaviour, it is seen that frequencies below the natural frequency are transmitted through the system, i.e., the force is converted to displacement with a little attenuation. Near the natural frequency undesirable amplification of the signal is performed and above this frequency severe attenuation is

904

present. It is a case of a system that exhibit behaviour characteristics of a variable impedance, which is the frequency-dependent. When it is desirable to transmit mechanical motion through a system, the natural frequency and damping characteristics must be taken into account so that good matching is present.

15.2 Uncertainty Analysis of Estimated Parameters Uncertainty of the test data is a result of the individual uncertainties inherent with each instrument. The method described by Holman (1978) is briefly described here to estimate the uncertainty in rotor dynamic parameters (RDPs). The method is briefly stated as follows. Let the results R (e.g., RDPs) is given as the function of independent variables 1 2, , , nx x x (e.g., the rotor speed, inlet pressure, pressure drop, diameter, length, clearance, temperature, force, excitation frequency, displacement, acceleration, etc.). Thus,

( )1 2, , , nR R x x x= (15.5)

Let Rw be the uncertainty in the result and 1 2, , , nw w w be the uncertainties in the independent

variables. Then the uncertainty in the result is given as

1/ 222 2

1 21 2

R nn

R R Rw w w w

x x x

= + + +

(15.6)

with

( ) ( )1 1 11 1

R x x R xRx x

+ =

;

( ) ( )1 1 11 1

R x x R xRx x

+ =

; (15.7)

where 1 2, , , nx x x are the small perturbations of the independent variables. It should be noted

that the uncertainty propagation in the results Rw predicted by equation (15.7) depends on the squares of the uncertainty in the independent variables ( )1,2, ,kw k n= . This means that if the uncertainty in one variable is significantly larger than the uncertainties in the other variables, then it is the largest uncertainty that predominates and other may probably be negligible. The relative magnitude of uncertainties is evident when one considers the design of an experiment, procurement of instrument in force, excitation frequency, displacement, and acceleration measurements on the rotor dynamic parameters.

Example 15.1: A voltage is impressed on the resistor R and the power dissipation is to be calculated in two different ways (i) from P = E2/R and (2) from P = EI. In (1) only voltage measurement will be

905

made, while both current and voltage will be measured in (2). The register has a nominal stated value of 5 1 percent. Calculate the uncertainty in the power determination in each case when the

measured values of E and I are: E = 50 V 1% (for both cases) and I = 5 A 1%.

Solution: The schematic of the power measurement across a register R is shown in Figure 15.3. The

uncertainty of voltage and current would be 50 0.01 0.5Ew = = V and 5 0.01 0.05Iw = = A.

Figure 15.3 The power measurement across a register

Case (1) For the first case P = E2/R, we have two independent parameters to be measured that is E and R, which will have the uncertainty. Hence, the uncertainty in the power measurement would be

( ) ( ) 1/ 22 22 2, ,p E R

P E R P E Rw w w

E R

= +

(a)

with

( ), 2P E R EE R

=

and

2

2( , )P E R E

R R

=

(b)

On substituting equation (b) into equation (a), the uncertainty in the power could be written as

1/ 222 22 2

22

p E RE E

w w wR R

= +

or

1/ 22 2

4p E Rw w w

P E R

= +

(c)

Inserting numerical values for the uncertainty, we get

( ) ( ) ( ) ( )1/ 2 1/ 22 2 2 2100 100 4 5/ 50 0.5 / 5 100 4 0.01 0.01 2.24%pwP

= + = + =

(d)

(2) For the second case P = EI, we have two independent parameters to be measured that is E and I, which will have the uncertainty. Hence, we have

( ),P E II

E

=

and ( , )P E I E

I

=

(e)

906

On using equation (e), the uncertainty in the power could be written as

( ) ( ) 1/ 22 22 2, ,p E I

P E I P E Iw w w

E I

= +

or ( ) ( ) 1/ 22 22 2P E Iw I w E w = + or

1/ 22 2P E Iw w w

P E I

= +

(f)

On substituting numerical values for the uncertainty, we get

( ) ( ) 1/ 22 2100 100 0.01 0.01 1.414%PwP

= + =

(g)

Since, calculations are based on percentage that is why actual values of various parameters have no effect on the final uncertainty. However, the second method of power determination provides considerably less uncertainty than the first method, even though the primary uncertainties in each quantity are the same. In this example, the uncertainty analysis is that it affords the individual a basis for selection of a measurement method to produce a result with less uncertainty. It should be noted that from transducers generally we get these electrical parameters only and with the sensitivity

subsequently it is converted to vibration parameters.

Example 15.2 In most of the practical voltmeter an internal resistance Rm is always present. The power measurement in Example 15.1 is to be conducted by measuring the voltage and the current across the resistor with circuit shown in Figure 15.4. Calculate the nominal value of the power

dissipated in R and the uncertainty for the following conditions: R 120 , Rm = 1200 5%, I = 5

A 1% and E = 500 V 1%.

Figure 15.4 Effect of the meter impedance on the measurement

Solution: The uncertainty in various parameters are: 500 0.01 5Ew = = V, 5 0.01 0.05Iw = = A, and

1200 0.05 60mR

w = = . Let I1 and I2 are currents flowing through registers R and Rm, respectively. A

current balance on the circuit gives

907

1 2I I I+ = m

E E IR R

+ = 1

m

EI IR

= (a)

The power dissipated in the resistor R is give as

2

1m

EP EI EIR

= = (b)

so that

2m

P EIE R

=

,

P EI

=

,

2

2m m

P ER R

=

(c)

The nominal value of the power is thus calculated as

2500500 5 2292 W1200

P = = (d)

In terms of known quantities the power has the functional from ( , , )mP f E I R= and so the uncertainty for the power is now written as

( )1/ 2 1/ 22 2 22 2 2

22 2 2 2 2 22

2m mp E I R E I R

m m m

P P P E Ew w w w I w E w w

E I R R R

= + + = + +

(e)

On substituting the given numerical values in equation (d), we get

( ) ( ) ( )

[ ]

1/ 222 22 2 22

2

1/ 2

2 500 5005 5 500 0.05 601200 1200

434 625 108.5 34.2 W

pw = + +

= + + =

(f)

or

34.2100 100 1.49%2292

Pw

P= =

(g)

From equation (f), the order of influence on the final uncertainty in the power are as follows: (i) the uncertainty of the current determination, (ii) the uncertainty of the voltage measurement, and (iii) the uncertainty of the knowledge of internal resistance of the voltmeter. However, it should be noted from

908

equation (15.1) that this results from the fact that mR R ( 10 )mR R= . Moreover, if the uncertainty

in one variable is significantly larger than the uncertainties in the other variables, say, by a factor of 5 or 10, then it is the largest uncertainty that predominates and others may probably be neglected. The relative magnitude of uncertainties is evident when one considers the design of an experiment, procurement of instrumentation, etc. Very little is gained to reduce the small uncertainties. Because of the square propagation it is the large ones that predominate, and any improvement in the overall experimental technique connected with these relatively large uncertainties.

A simple device for the measurement of the vibrational frequency is shown in Figure 15.5(a). The small cantilever beam mounted on the block is placed against the vibrating surface to provide a base excitation (Figure 15.5(b)). Provision is made to varying the beam length, which in term is expected to vary its natural frequency. When the beam length is attuned so that its natural frequency is equal to the frequency of the base excitation, the resonance condition as shown in Figure 15.5(c) will result, which can be visualize by naked eye also. The aim would be to measure the length of the beam each time we attune the resonance condition to obtain the natural frequency. However, due to uncertainty in measurement of the beam length would lead to uncertainty in the measurement of the frequency. It

should be noted that there could be so many other uncertainty (e.g., uncertainty in attuning the resonance, etc.) that might affect the uncertainty of frequency measurement, however, for brevity only a single uncertainty have been considered.

(a) No excitation (b) Excitation other than n (c) Excitation at n Figure 15.5 Cantilever beam used as frequency measurement device

Considering the beam as a continuous system, the fundamental natural frequency of the beam is given by

43.52nfEI

mL =

(15.8)

where n is the natural frequency in rad/s, E is the Youngs modulus in N/m2, I is the moment of

inertia in m4, m is the beam mass per unit length in kg/m and L is the beam length in m. We use

909

equation (15.8) to determine the allowable uncertainty in the length measurement in terms of the uncertainty in the frequency measurement. From equation (15.8), we have

37.04nf EI

L L m

=

(15.9)

The uncertainty in the natural frequency is given by

1/ 222

nfnf

Lw wL

= (15.10)

where Lw is the uncertainty in length measurement in m. On substituting equation (15.9) into equation (15.10), and after simplification we obtain

3

7.04 /nf

L

w Lw

EI m

= (15.11)

Now with an example the above method will be illustrated.

Example 15.3 A 0.6 mm diameter spring-steel rod to be used for a vibration-frequency measurement as shown in Figure 11.5(a). The length of the rod may be varied between 60 mm to 200 mm. The mass density of this material is 7800 kg/m3 and the modulus of elasticity is 112.1 10 N/m2. Calculate

the range of frequencies that may be measured with this device and the allowable uncertainty in L at 200 mm in order that the uncertainty in the frequency is not greater that 2 percent. Assume the material properties are known exactly.

Solution: We have

112.1 10E = N/m2; 4 4 3(0.6) 6.362 1064 64

I dpi pi = = = mm

4 156.362 10= m4

( )2 62 37800 0.6 10 2.205 104 4

dm

pipi

= = = kg/m (a)

For L = 60 mm, from (15.8), we have

910

( ) ( ) 1/ 211 154 3 4

2.1 10 6.362 103.52 3.52 761.1

2.205 10 0.06nfEI

mL

= = =

rad/s (b)

Similarly, for L = 200 mm, we will have 68.50n = rad/s. Hence, the range of the frequency is from

68.50 rad/s to 761.1 rad/s.

For L = 200 mm, we have 0.02 68.50 1.37nf

w = = . From equation (15.11), we have the allowable

uncertainty in the measurement of length

3 33

11 15 3

1.37 0.2 1.999 107.04 / 7.04 2.1 10 6.362 10 / 2.205 10

nfL

w Lw

EI m

= = =

m = 2.0 mm (d)

Hence, the uncertainty of 200 1% would be allowable.

15.3 Transducers: A large number of devices transform values of physical variables into equivalent electric signals and such devices are called transducers (e.g., LVDT (linear variable differential transformer) gauges, eddy current, inductive, capacitive, piezoelectric, photoelectric, photoconductive, pressure transducers, nuclear radiation detectors, etc.). For measuring motion, there are two basic types of transducers; the first being the seismic that produces a signal proportional to the absolute motion in space; and the second a signal proportional to the relative motion between a reference point and the point of interest. Most of displacement sensors are based on the relative motion, and most of accelerometers are based on the absolute motion.

15.3.1 Displacement Sensors a. Potentiometer. The simplest form of displacement transducer is the potentiometer. Although they are available for measurement of the linear and rotational displacements, they tend to be noisy and are only suitable for relatively low frequency and large displacement applications.

b. Linear Variable Differential Transformers (LVDT): This is another form of displacement transducer that has been used successfully for vibration measurements for many years. The principle of operation of an LVDT is that a freely-moving magnetic core is used to link the magnetic flux between a surrounding primary coil and two secondary coil as shown in Figure 15.6. A schematic cross-section of an LVDT is shown in Figure 15.7.

The primary coil is energised by an external AC source. The alternating magnetic flux induces voltages at the null position are equal in magnitude but opposite in phase. When these two coils are

911

connected together, the net output of the transducer at the central position is zero. As the magnetic core is moved away from the central position the induced voltage in one of the secondary coils increases. At the same time, the induced voltage in the other coil decreases, resulting in a differential voltage output that varies linearly with the position of the magnetic core. In moving from one side of the central position to the other, the polarity of the demodulated output changes instantaneously. The

core has a small rod and is separated from the coil structure by a low friction lining that produces an almost frictionless device that is insensitive to radial motion of the core. For vibration measurements, the core is usually connected to the structure via a push-rod (stinger). The push-rod has two functions: to decouple lateral motion and to provide a convenient method of attachment to the structure. To maintain the calibration of the device, the push-rod should be nonmagnetic.

Figure 15.6 Schematic diagram of a differential transformer

Figure 15.7 A typical construction of a linear variable differential transformer (LVDT)

c. Rotary Variable Differential Transformers (RVDTs): It is used for measurement of the angular displacement. RVDT is an electromechanical transducer that provides a variable alternating current (AC) output voltage that is linearly proportional to the angular displacement of its input shaft. When

912

energized with a fixed AC source, the output signal is linear within a specified range over the angular displacement. RVDT utilizes brushless and non-contacting features to ensure long-life and reliable, repeatable position sensing with very high resolution.

d. Proximity transducers (Relative motion transducers): Proximity transduers use sensors that are able to detect the presence of nearby objects without any physical contact. A proximity transduers often emits an electromagnetic or electrostatic field, or a beam of electromagnetic radiation (infrared, for instance), and looks for changes in the field or return signal. The object being sensed is often referred to as the proximity sensor's target. Different proximity transduers targets demand different sensors.

For example, a capacitive or photoelectric sensor might be suitable for a plastic target; an inductive proximity sensor requires a metal target. The relative-motion transducers are the proximity probe type, which sense the gap (i.e., the displacement) between the mounting point (usually the bearing housing) and the point of interest (usually the rotating shaft). Proximity probes are widely used on the turbo-machinery as the sensor for permanent monitoring systems. They are particularly suitable for such machineries, where there are small internal clearances.

Displacement sensors are necessary to detect the radial (and sometimes axial) movement of the rotor. The requirements for a displacement sensor are as follows: (i) wide frequency response, (ii) low noise, (iii) low interference noise, (iv) low temperature drift, (v) good linearity, (vi) compactness, and (vii) reliability. Displacement sensors detect the linear position during the movement of an object without a mechanical contact. When selecting the displacement sensors, depending on the application, measuring range, linearity, sensitivity, resolution and frequency range are to be taken into account as well as: temperature range, temperature drift of the zero point and sensitivity; noise immunity against other sensors, magnetic alternating fields of the electromagnets, electromagnetic disturbances from switching amplifiers, dust, aggressive media, or vacuum.

There are basically three types of displacement transducers: electromagnetic, capacitive, and optical.

Brief outline of each of these transducers will be discussed now.

913

Figure 15.8 Principle of a displacement sensor Figure 15.9 An equivalent circuit of the displacement sensor

(i) Electromagnetic displacement transducers: Electromagnetic displacement transducers are of two types. First is the inductive while the other is the eddy current type. Figure 15.8 shows the structure and the principle of operation of an electromagnetic displacement sensor. An E-shaped magnetic core has a winding with two terminals. A target (i.e., the rotor shaft) is drawn as a rectangular solid having air gap. The input impedance, Zin, at the terminals varies with the air gap. When input terminals are

excited by a high frequency voltage then the coil impedance will be dominated by the inductance (which is the variable part of the impedance: An electric current i flowing around a circuit produces a magnetic field and hence a magnetic flux through the circuit. The ratio of the magnetic flux to the current is called the inductance, or more accurately self-inductance of the circuit); and it is obtained by detecting the terminal voltage and the current. Figure 15.9 shows the equivalent circuit of a sensor winding. The inductance L0 is a constant while the inductance L1 is dependent on the length of air-gap.

Inductive displacement transducers: An inductive transducer is an electronic proximity sensor, which

detects metallic objects without touching them. An inductor coil placed in a ferrite core is a part of an oscillating circuit (Figure 15.10). The excitation frequency is in the range of 20-100 kHz and the inductance varies as a function of air gap length (approximately inversely). If the air gap is small then there is a high impedance. When a ferromagnetic object (of high permeability such as laminated silicon steel, ferrite and carbon steel) displacement to be measured approaches the coil the inductance changes and the oscillating circuit is detuned. The signal is demodulated and linearised and becomes proportional to the gap between the sensor and the object of which the displacement to be measured. Two sensors opposing each other are frequently arranged on a rotor (Figure 15.11). They are operated differentially in a bridge circuit with a constant bridge frequency, producing a nearly linear signal.

Inductive sensors are operated with modulation frequencies from approximately 5 kHz up to 100 kHz.

914

The cut-off frequency of the output signal lies in a range between one tenth and one fifth of the modulation frequency.

Figure 15.10 Inductive displacement sensor Figure 15.11 Differentially measuring sensors

Eddy-current transducers: The transducer function by detecting changes in the eddy current loss as the gap between the probe and the target surface varies (Fig. 15.12). The high-frequency alternating current runs through the air-coil cast in a housing. The electromagnetic coil section induces eddy currents in the conducting object (of low resistance such as copper, non-magnetic stainless steel, aluminum, carbon steel and other metallic material) to be measured, thus absorbing energy from the oscillating circuit. Depending on the clearance, the amplitude of oscillation varies. This amplitude variation will provide a voltage variation proportional to the clearance, once it is demodulated,

linearised and amplified. The usual modulation (excitation) frequency lies in a range of 1-2 MHz and have measuring frequency ranges of approximately 0 Hz up to 20 kHz.

Figure 15.12 An eddy current displacement sensor

915

Precautions and limitations: In-homogeneities in the material of the moving rotor cause disturbances (noise) and reduce the resolution accordingly. If the target is close to the sensor core then eddy currents are induced into the target which reduces the flux (almost as a short-circuit transformer) and produces a low input impedance. As the target moves away, the coupling decreases which increases the input impedance (which is opposite to the inductive type). Manufacturers usually indicate the sensitivity used on aluminum. When measuring steel, sensitivity is smaller. Shielded sensors must be used for applications where high frequency magnetic field occurs. Sensors may cause mutual interference. Therefore, the minimum clearance between sensors is mostly defined in the mounting

guide. Within the linear range, which typically extends from 250-2250 m gap, current standards

require either a 4 mV/m or 8 mV/m proportionality between gap and voltage. Thus, a 250 m

change in gap should produce a voltage change of 1 volt at 4 mV/m or 2 volts at 8 mV/m (some times instead of mm the mils; is used and the mils is one thousands of an inch). The standard sensitivity for these transducers is 8 mV/m (or 8.0103 mV/mm) for the normally used target materials (e.g., steel).

The extension cable and oscillator demodulator of the transducer make up a turned resonant circuit. In order to establish and maintain a constant ratio between gap and voltage, the transducer, oscillator demodulator and extension cable must be properly matched and calibrated. Most manufacturers will specify the type of probe, generally the tip diameter, and the total electrical length of the extension and probe cables, which must be used, with each oscillator demodulator.

The slope of the curve, the linear range, and the DC output corresponding to a given gap will vary with changes in a targets conductivity and permeability. If a probe and oscillator demodulator calibration for 4140 steel are used without recalibration on a material as such stainless steel, the curve shifts to the left, producing a higher-output voltage for a given gap. Due to this shift and potential inaccuracies, a non-contact probe system calibrated for one material should not be used with another without recalibration.

Temperature may also affect the range limits of a non-contact probe and the DC output at a given gap;

however, the shift is generally small across the temperature range experienced within a bearing housing. Elevated pressures may affect the sensitivity of a non-contact probe. If the probe is installed in an area of high or fluctuating pressure, its response should be tested in actual environment to determine what changes in sensitivity or output will occur. With everything else equal, the maximum linear range obtainable with a non-contact displacement measurement system will increase with increasing probe tip diameter and likewise increase with an increasing supply voltage. At a sensitivity

of 8 mV/m , linear range of typical non-contact measuring systems observing 4140 steel will vary

916

from approximately 1525 m with 5 mm tip diameter and -18 VDC supply to 2160 m with a 8 mm tip diameter and 24 VDC supply.

When the target is moving surface such as the periphery of a shaft, the displacement measuring system cannot distinguish between shaft motion or vibration and defects such as scratches, dents and variations in conductivity or permeability. As a result, the output, rather than being pure vibration, is the sum of vibration and all surface variations passing beneath the probe. Since the magnetic field of the probe penetrates the surface of the observed material, any repair which results in an interface between two materials (when the shaft is plated or metal sprayed) will introduce distortion in the output signal measured by an eddy current displacement transducer.

Eliminating excessive runout is often a very difficult task. The first and obvious step is during manufacture when every effort must be taken to ensure the surface which will be observed by the shaft probe is concentric with the journal, has a smooth finish, and is protected from damage during handling and assembly. Produced when the shaft is machined, ground or degaussed incompletely following a magnetic particle inspection, electromagnetic runout can generally be eliminated by degaussing the shaft surface observed by the probe. If after degaussing, runout persists despite a smooth and concentric shaft surface, it is likely due to a changing permeability or conductivity around the circumference of the shaft. Often a problem with high-alloy, precipitation-hardened shafts, this type of runout has been successfully reduced by burnishing the area- running it on balancing machine

rollers or producing similar effect in a lathe with a special roller tool.

Should all these steps fail or be impossible to implement for one reason or other, runout can be eliminated electronically with a runout subtractor on-line or off-line. The runout subtractor digitally memories a phase-referenced shaft motion at slow roll when all motion is assumed to be runout then automatically subtracts the slow-roll waveform from the raw waveform observed by the probe to produce a corrected waveform representative of actual shaft motion. The off-line procedure is explained in more details in chapter related to balancing of rotors.

Inductive sensors are not as sensitive as the eddy current sensors and it is advantageous. However, there are disadvantages with inductive sensors : (a) There are very few variety of inductive sensors available in the market as it is very costly, (b) The shaft target ring has to be made from laminated silicon steel.

In eddy current sensors, further improvements of the output voltage are possible: (a) Two eddy current sensors can be placed on one axis and the output voltages are subtracted to give differential operation. The second harmonic and even harmonics are reduced and temperature drift is decreased.

917

(b) The target material can be replaced with non-magnetic stainless steel (or copper) to avoid magnetic imperfections. However, sensor amplifier linearity should be taken care of. (c) The sensor head diameter should be small so that the target ring does not interfere with the two-axis movement. For example, the sensor head diameter should be 5 mm or less for a target ring diameter of 50 mm. On the other hand, the sensor head diameter should be large compared to the air gap length for better

sensitivity and linearity. For example, a head diameter of 5 mm should be used for 1 mm air gap length or less, and (d) the excitation frequency of the x-, y- and z-axis sensors should be set far apart to avoid interference. The difference in excitation frequency should be greater that the frequency range of the sensor. For example, for a sensor response range of 20 kHz, when the x-axis sensor is

excited at 2 MHz the frequencies of the other sensors should be less than 1.96 MHz or higher than 2.04 MHz to provide enough frequency separation.

(ii) Capacitive displacement transducers: In capacitive proximity sensors, the sensed object changes the dielectric constant between two plates. The capacity of a plate capacitor varies with its clearance.

Therefore, a good isolation between the sensor and the shaft is necessary. In addition, the air must be clean, and the oil and other particles should not be present because this will affect the dielectric. Using the capacitive measuring method, the sensor and the opposing object to be measured form one electrode of a plate capacitor each (Figure 15.13). Within the measuring system, an alternating current with a constant frequency runs through the sensor. The voltage amplitude at the sensor is proportional to the clearance between the sensor electrode and the object to be measured, and is demodulated and amplified by a special circuit.

A proximity sensor has a range, which is usually quoted relative to water. Because changes in

capacitance take a relatively long time to detect, the upper switching range of a proximity sensor is about 50 Hz. The proximity sensor is often found in bulk-handling machines, level detectors, and package detection. One advantage of capacitive proximity sensors is that they are unaffected by dust or opaque containers, allowing them to replace optical devices. In addition, the air must be clean, and the oil and other particles should not be present because this will affect the dielectric. A typical capacitive proximity sensor has a 10-mm sensing range and is 30 mm in diameter. The proximity sensor incorporates a potentiometer to allow fine tuning of the sensing range and can repetitively detect objects within 0.01 mm of the set point. Switching frequency is 10 Hz, and operating temperature range is 30 to 70C. A proximity sensor that measures current flow between the sensing

electrode and the target provides readouts in appropriate engineering units. Usually, one side of the voltage source or oscillator connects to the sensing electrode, and the other side connects through a current-measuring circuit to the target, which generally is a metal part at earth or ground potential.

918

Figure 15.13 Capacitance displacement sensor

Probes used with a capacitive proximity sensor have either a flat disc or rectangular sensing element surrounded by a guard electrode that provides electrical isolation between the proximity sensor and its housing. The guard also ensures that the lines of electrostatic field emanating from the probe are parallel and perpendicular to the surface of the proximity sensor. Capacitance proximity sensor

systems can make measurements in 100 sec with resolutions to 0.001 micron, however, the

ccommercially available capacitive displacement measuring systems are expensive. The bandwidth of the output signal ranges between approximately 5 kHz and 100 kHz. The electrostatic charging of the contactless rotor may cause interferences too. The sensors are sensitive to dirt which modifies the dielectric constant in the air gap.

(iii) Optical displacement transducers: The simplest principle of an optical displacement sensor consists of covering a light source opposite to a light-sensitive sensor by the object to be measured (Figure 15.14). The resulting difference in the light intensity is converted into an electric signal and serves as a measurement for the position of the object. By selecting appropriate light sources, light sensors and suitable apertures, we obtain a nearly linear displacement signal. A similar approach consists of reflecting light by the object to be measured. The fraction of light received by the sensor changes according to the motion of the object (Figure 15.15). For this kind of system photo diodes, photo transistors, photo resistors, and photo-electric cells can be used as sensors. The wavelength of the light source should be adjusted to the sensor.

919

Figure 15.14 Light barrier principle Figure 15.15 Light reflecting principle

Another possibility is the application of an image sensor. Charge-coupled device (CCD) is an electronic memory that records the intensity of light as a variable charge. Widely used in still cameras, camcorders and scanners to capture images, CCDs are analog devices. Their charges equate to shades of light for monochrome images or shades of red, green and blue when used with colour filters. Take for example a line array camera (CCD sensor) in a rotor system (Figures 15.16). The rotor image is reflected both for the x- and the y- direction over a mirror on a CCD sensor. The picture of the rotor, tinted black in front of a lit-up background, is converted into a video signal. By counting the pixels (light-sensitive dots) until the light-dark boundary is reached one obtains a digital displacement signal. However, optical displacement measuring systems are not appropriate for many application fields, since they are very sensitive to dirt, and the resolution is limited due to defraction effects.

Figure 15.16 An optical displacement sensor

Since the advent of the laser in the early 1960s the field of optical metrology has provided accurate experimental data in situations in which, previously, it would have been considered unattainable. The technique of laser Doppler velocimetry (LDV) is now well established and was initially applied to obtain non-intrusive measurements in fluid flows by laser Doppler anemometry (LDA; Durst et al.,

920

1981 and 1988). Although the use of LDV for solid surface velocity measurement was recognized at an early stage, its development in this area received little attention compared with the effort in fluid mechanics. Accordingly, the measurements of vibration was still extensively achieved with accelerometers or other forms of transducer which rely on contact with the measurement surface for successful operation. There are, however, many cases of engineering interest where this approach is

either impractical or impossible. Typical examples are the measurement of very hot or light surfaces, such as exhaust pipes or loudspeakers, and measurement on rotating surfaces which prevent their use.

In the area rotating surfaces the measurement of torsional vibration of rotating components presented a particularly difficult measurement problem. When designing rotating machinery components, an engineer must be careful to suppress torsional oscillations, since incorrect or insufficient control may lead to fatigue failure, rapid bearing wear, gear hammer, fan belt slippage and can produce associated excessive noise problems. Torsional oscillations are a particular problem in engine crankshaft design where torsional dampers are commonly used to maintain oscillations at an acceptable level over the

working speed range of the engine. Torsional transducers have formerly included optical, seismic and mechanical torsiographs, strain gauges and slotted discs. The latter system has found common use in the automotive industry and consists of a slotted disc fixed to the end of the crankshaft. A proximity transducer monitors the slot passing frequency, which is then demodulated to provide a voltage analogue of the crankshaft speed and hence torsional oscillations, but within a limited frequency range. Strain gauges and associated telemetry or slip ring systems are disreputably difficult to fix, calibrate and use successfully. In summary, the measurement of torsional oscillations presented difficult problems for contacting transducer technology and, of course, necessitated machinery downtime and special arrangements being made for fitting, calibration, etc. Very often, the cost of this

machinery stoppage would prevent a measurement being attempted, even though the vibration engineer had concluded that it was vital if a design improvement is to be made. There was therefore a real need for a torsional vibration transducer which was user friendly and could provide data immediately in on-site situations. It was not until the advent of laser technology that a solution was found. It allows the engineer to point low powered laser beams at a rotating target component and obtain torsional vibration information (Halliwell, 1996).

Laser optical range transducers: It operates on the principle triangulation (Figure 15.17). A laser light beam reflected from the surface of a structure is focused onto an internal photo-sensitive device.

As the structure moves, the position of the focused spot on the photo-sensitive device moves. The photo-sensitive device generates a signal according to the position of the focused spot. This output is then conditioned and linearised to give an analogue signal proportional to the range of the surfaces motion. The displacement is detected by reflecting laser light so that a uniform target surface is required to prevent the noise.

921

Figure 15.17 The basic principle of the laser optical sensors

A laser vibrometer is an optical system that can be used to measure the instantaneous velocity of a

point (or points) on a structure. The instrument is a non-contact device in which the velocity measured is the velocity components in the direction of incident laser beam. The velocity is measured by the

detection of the Doppler frequency shift (is the change in frequency and wavelength of a wave for an observer (e..g, the surface of shaft) moving relative to the source of the waves.) of light scattered from the moving surface. Sophisticated optics and signal processing mean that these devices are expensive. Scanning systems are now available in which the laser beam can be moved rapidly over a grid of measurement points on a structure. It is possible to make finely detailed measurements on complex structures that are not amenable to, or accessible for, conventional transducers. Further development in laser measurement techniques now enable the measurement of rotational responses (Tiwari et al., 2005).

15.3.2 Accelerometers: The most widely used types of seismic transducers give an output signal proportional to the acceleration. Accelerometers contain usually piezo-electric crystals, which are

loaded with a small inertia weight and rigidly mounted in a casing. They produce a voltage output, which is proportional to the acceleration over a wide frequency range, up to the point where the output/(unit acceleration) starts to rise due to natural frequency of the inertia weight supported on the crystal.

The primitive type of accelerometer had very high source impedance (1010 Ohms) with all the cabling problems associated with this. In recent years, accelerometers have become available with the matching charge amplifier inbuilt within the accelerometer casing. Often the size of these low-

impedance accelerometers is not very much larger then the original version. They are no longer self-generating and need a dc power supply to drive them (typically 18 V DC).

922

The primitive type of velocity transducer in the form of a spring-mounted coil (resonant frequency in the region of 10 Hz), producing a signal proportional to velocity, has become virtually absolute. This is because of their limited frequency range, the relatively large size and weight, and mounting problems, together with problems of maintaining optimum damping necessary to obtain a flat frequency response. Instead of this inductive type of velocity transducer are available, some

manufacturers supply a piezoelectric velocity transducer with the internal integration and the low impedance output (the impedance is defined as the ratio of applied SHM force to resulting velocity).

An accelerometer can have several parameters, which can be used for the selection of the transducer. For example, sensitivity, frequency range, residual noise level in the measuring range, temperature range, maximum operational and shock levels, weight, connectors, mountings, type of out put (charge /voltages), etc. Accelerometers are available based on applications, e.g., the general purpose, high sensitivity, high temperature, high frequency (very small size), shock, human vibration, under water, modal analysis, industrial, aerospace and flight test, special purpose like the tri-axial and rotational

measurements, etc.

Figure 15.18 Simple spring-mass Figure 15.19 Schematic of typical seismic instrument -damper system

The seismic instrument is a device that has the functional form of the system shown in Figure 15.18, which is a single-DOF spring-mass-damper system with the support motion. A schematic of a typical instrument is shown in Figure 15.19. The mass is connected through the parallel spring and damper arrangement to the housing frame. This frame is than connected to the vibration source (e.g., bearing housing) whose characteristics are to be measured. From Figure 15.18 using Newtons law of motion, we have

2 2 2 1 1my cy ky cy ky+ + = + (15.12)

923

where 1y and 2y are the absolute displacements of the housing and the suspended mass, respectively.

It is assumed that the damping force is proportional to the velocity. We assume that a harmonic motion is applied on the instrument such that

1 1 cosy Y t= (15.13)

where Y is the displacement amplitude. The aim is to obtain an expression for the relative

displacement ( )2 1y y in terms of this base motion. The relative displacement is that which is detected by the transducer shown in Figure 15.19. On substituting equation (15.13) into equation (15.12) and rearranging it gives

2 2 2 cos cosc k k cy y y Y t tm m m m

+ + =

(15.14)

The solution to equation (15.14) is

( ) ( ) ( ) ( )( ){ }2

122 1 1/ 222 2 2

coscos sin

d d

cm

nf nfmY t

y y e A t B tk m c

= + +

+

(15.15)

where the damped natural frequency is given by

1/ 22

2dnfk cm m

=

for 1.0c

c

c . For low values of damping ratio the displacement

amplitude may become quite large. The output becomes essentially a linear function of input at high-

frequency ratios (curve becomes relatively straight). Thus a seismic-vibration pickup for measurement of displacement amplitude should be utilized for measurement of frequencies substantially higher than

its natural frequency. The instrument constants ccc and nf should be known or obtain from the

calibration. The anticipated accuracy of measurement may than be calculated for various frequencies.

The acceleration amplitude of the input vibration is

21 1 1a y Y= = (15.22)

We may thus use the measured output of the instrument as a measure of acceleration. However, there

are restrictions associated with this application. In equation (15.18) the bracketed term is the one that governs the linearity of the acceleration response, since nf will be fixed for a given instruments. In

Figure 15.21 we have a plot ( ) 22 1 1/nfY Y a versus / nf , which indicates the non-dimensionalised acceleration response.

Figure 15.21 The acceleration response of a seismic instrument as given equation (15.22)

Thus by measurement ( )2 1Y Y , we can calculate the input acceleration 1a . Generally inadequate performance is observed at frequency ratio above 0.4. Thus for acceleration measurements we need to

operate at frequencies much lower than the natural frequency, in contrast to the desirable region of

operation for displacement measurements. In view of instrument construction we need to have a low

0/ =ccc

Frequency ratio

Acc

eler

atio

n pa

ram

eter

926

natural frequency (soft spring, large mass) for displacement measurements and a high natural frequency (stiff spring, small mass) for acceleration measurements in order to be able to operate over a wide range of frequencies and still linear response. The seismic instrument may also be used for

velocity measurements by employing a variable-reluctance magnetic pickup as the seismic transducer.

The output of such a pickup will be proportional to the relative velocity amplitude, i.e., the quantity

( )2 1V V . From the above discussion it may be seen the seismic instrument is a very versatile device that may be used for measurement of a variety of vibration parameters. This encourages to operate

many commercial vibration and acceleration pickups on the seismic instrument. The seismic

instrument may be used for either the displacement or acceleration measurements by proper selection

of the mass, spring and damper combinations. In general, as large mass and soft spring are desirable

for vibrational displacement measurements, while a relative small mass and stiff spring are used for

acceleration indicator.

The transient response of the seismic instrument is governed partially by the exponential decay term

in equation (15.15). The time constant for this term could be taken as

2mTc

= (15.23)

or, in terms of the natural frequency and critical damping ratio

1n

T = (15.24)

The specific transient response of the seismic-instrument system is also a function of the type of input

signal, i.e., whether it is a step function, harmonic function, ramp function, etc. The linearity of a

vibration transducer is thus influenced by the frequency-ratio requirements that are necessary to give

linear response as indicated by equations (15.15) and (15.18). The design of a transducer for particular response characteristics must involve a compromise between these two effects, combined with a

consideration of the sensitivity of the displacement sensing transducer and its transient response

characteristics.

927

Figure 15.22 A stud mounting on an accelerometer on the vibrating surface

Example 15.4 For measurement of displacement using the amplitude ratio equation (15.18) and 0.72 = , calculate the value of / nf such that ( )2 1 1 0.98Y Y Y = ; that is, the measurement error is

2 percent.

Solution : We have

( ) ( ){ }2

1/ 22 220.98

1 2 0.72

=

+ (a)

Rearranging this equation gives the quadratic relation

4 21.78 24.25 0 = (b)

which yields / 2.427nf = . It is evident from this example that the natural frequency of the

instrument should be low.

Example 15.5 For measurement of acceleration the amplitude ratio equation (15.18) and 0.72 = , calculate the value of / nf such that ( ) ( )2 22 1 1 0.98nfY Y Y = ; that is, the error is 2 percent.

( )( ) ( ){ }

2 11/ 22 2 221

10.981 2 0.72

Y YY

= =

+ (a)

Rearranging this equation gives the quadratic relation

928

4 20.0736 0.0412 0 + = (b)

which yields / 0.412nf = . It is evident from this example that the natural frequency of the

instrument should be high.

15.4 Signal Conditioning & Analysis Equipments: The raw signal from the vibration transducer may need to be transformed into the right form, e.g., signals from accelerometers may need to be

integrated to provide a velocity or displacement signal. Furthermore, signals may need to be amplified

before being fed to the metering and alarm circuits, or in some cases passed through a filter system to

eliminate unwanted portions of the frequency spectrum, and finally the system impedance may to be

reduced. All of these processes are known as signal conditioning and this can be defined as the

transformation of the transducer signal into a form, which is suitable for the analysis, metering, or

feeding into an alarm or advance signal processing system.

15.4.1 Filters: Filters are probably the most widely used of all vibration analysis equipment once the signal is

available from transducers. It can be there in-built in the conditioning amplifier (which amplify the weak signal usually available from transducers) or as a separate device in the form of hardware or software. A filter limits a vibration signal in some predictable fashion such that a single frequency or

group of frequencies may be isolated for the measurement or study. Filters can be classified mainly

two different ways:

(i) Frequencies passed or passband: Under this category filters are further classified based on the frequencies that to be allowed or rejected.

(a) High pass: It passes all frequencies above some specified frequency; generally it is required whenever the signals from accelerometers are double integrated to displacement.

(b) Low pass: It passes all frequencies below some specified frequency; it is often used with shaft displacement signals to eliminate high frequencies generated by shaft scratches.

(c) Bandpass: It passes a band of frequencies while eliminating all frequencies both above and below the desired passband.

(d) Band (notch) reject: The reverse of a bandpass filter, eliminating all frequencies within a specified band while allowing all others both above and below to pass; it permits a rapid

assessment of the total vibration energy present, exclusive of a specific frequency.

929

(ii) Method of tuning: Under this categories filters are further classified based on the method of tuning.

(a) Manual tracking: In manual tuning filters are of two types namely, the constant bandwidth (pass a constant frequency band regardless of where the filter center frequency is positioned hence it provides uniform resolution) and the constant percentage bandwidth (the frequencies passed are some fixed percentage of the filter central frequency such that as the filter is tuned

to higher frequencies, the bandwidth becomes larger with a corresponding reduction in

resolution) filters are in use. (b) Automatic tracking: In the automatic or tracking filter the tuning signal is generated by and

synchronized with the shaft under study (i.e., at rotating frequency or multiple of running frequency). It is widely used in balancing applications, and for tracking phase and amplitude response during a startup or coast-down of heavy rotating machineries (e.g., turbines and generators).

15.4.2 Measurement amplifier Generally, an amplifier is any device that changes, usually increases, the amplitude of a signal. The

signal is usually voltage or current. The relationship of the input to the output of an amplifier

usually expressed as a function of the input frequency is called the transfer function of the

amplifier, and the magnitude of the transfer function is termed the gain. A The measurement amplifier

is used to convert the charge signal output from the transducers (accelerometer or force transducers) to voltage signal. The amplifier can be used for amplification of the one signal and the sensitivity of

the transducer has to be matched (or fed) with the amplifier. Different level of amplification could be achieved depending upon the requirement and quite often the amplifier also have provisions for

filtering.

15.4.3 Oscilloscope, Spectrum analyzer and Data Acquisition System An oscilloscope (commonly abbreviated to scope or O-scope) is a type of electronic test equipment that allows signal voltages to be viewed, usually as a two-dimensional graph of one or more electrical

potential differences (vertical axis) plotted as a function of time or of some other voltage (horizontal axis). Oscilloscope can have several functions that helps in capturing and analysis the vibration signal. Depending upon the level of the signal can be amplified or reduced. The time base can also be varied

to have better visulaising of the signal on the screen before capturing. One important feature of the

trigger level setting provides capturing of the singal when it exceeds certain level. This feature helps

in capturing relevant signal expecially during modal testing using the impact hammer to synchronise

the time of hitting and the caturing of the signal.

930

A spectrum analyzer is an instrument that displays signal amplitude (strength) as it varies by signal frequency. The frequency appears on the horizontal axis, and the amplitude is displayed on the

vertical axis. A spectrum analyzer looks like an oscilloscope and, in fact, some instruments can

function either as oscilloscopes or/and as spectrum analyzers. In spectrum analyzer various in-built

functions for statistical processing of periodic or random signals are available. It includes FFT, power

spectrum, autocorrelation, cross-correlations, spectral density, probability density function, ensemble

or temporal averages, etc. Multi-channel spectrum analyzers are very expensive and that has led to

the development of various software for the analysis of the vibration signal. Such multi channel

analyzer system consists of a PC with LAN interface and data acquisition hardware. The system can

possess time capture and FFT analyzers. It also has provision for setting of a project is to ensure that a measurement is set up exactly according to individual specifications.

A data acquisition system is a device designed to measure and log some parameters. The purpose of

the data acquisition system is generally the analysis of the logged data and the improvement of the

object of measurements. The data acquisition system is normally electronics based, and it is made of hardware and software. The hardware part is made of sensors, cables and electronics components

(among which memory is where information are stored). The software part is made of the data acquisition logic and the analysis software (and some other utilities that can be used to configure the logic or to move data from data acquisition memory to a laptop or to a mainframe computer).

15.5 Vibration Exciter Systems: In order to apply a test item (e.g., a rotor system) to a specific vibration, a source of motion is required. Devices used for supplying vibrational excitation are usually referred to simply as shakers or exciters. In most cases, simple harmonic motion is provided, but systems supplying complex

waveforms (two or multi-frequency, random, impulse, etc.) are also available. There are various forms of shakers and the variation is depending on the source of driving force. In general, the primary source

of motion may be electromagnetic, mechanical, or hydraulic-pneumatic or in certain cases, acoustical,

aerodynamic. Each is subjected to inherent limitations, which usually dictate the choice.

15.5.1 Electromagnetic Systems A sectional view of an electromagnetic exciter is shown in Figure 15.23. This consists of a field coil, which supplies a fixed magnetic flux across the air gap and a driver coil supplied from a variable-

frequency source. Permanent magnets are also sometimes used for the fixed field (or the biased field), which reduces the power consumption. The support of the driving coil is by means of springs, which

permit the coil to reciprocate when driven by the force interaction between the two magnetic fields. It

can be seen that the electromagnetic driving head is very similar to the field and voice (moving) coil arrangement in the ordinary radio loudspeaker.

931

Figure 15.23 A sectional view of the electromagnetic exciter

An electromagnetic shaker is rated according to its force capacity, which in turn is limited by the

current-carrying ability of the moving coil. Temperature limitations of the insulation basically

determine the shaker force capacity. The driving force is commonly simple harmonic (complex waveforms are also used) and may be thought as a rotating vector. The force used for the rating is the vector force exerted between the moving and field coils. The rated force is never completely available

for driving the test item. It is the force developed within the system, from which must be subtracted

the force required by the moving portion of the shaker system proper. It may be expressed as

rt t aF F F= (15.25)

in which Frt is the net force available to shake the test item, Ft is the manufacturers rated capacity, or

the total force produced by the magnetic interaction of the moving and field coils; and Fa is the force

required to accelerate the moving parts of the shaker system, including the moving coil, table and

appropriate portions of the moving coil flexure beam. Specification for a typical electromagnetic

exciter systems contain (i) maximum rated force, (200 2105) N (ii) frequency range, (0 10000) Hz (iii) Peak-to-peak amplitude (up to 25 mm), (iv) cooling requirement and (v) weight of the moving armature (0.35 110 kg), (vi) type of excitation (sinusoidal, multi-frequency, random, impulse, sine-sweep, etc.). While using sine-sweep excitation, it is often required to cross the resonance condition. In advanced electromagnetic exciters a feedback control based on the vibration level provide variable

force so that at the resonance the force applied is very small to avoid catastrophic failure of the test

item.

932

15.5.2 Mechanical-Type Exciters There are two types of mechanical shakers: the directly driven and the inertia. The directly driven

shaker consists of a test table that is forced to reciprocate by some form of mechanical linkage. Crank

and connecting rod mechanisms, Scotch yokes, or cams may be used for this purpose. Another

mechanical type uses counter-rotating masses to apply the driving force. The force adjustment is provided by relative offset of the weights and the counter-rotation cancels shaking forces in one

direction, say the x-direction, while supplementing the y-force. The frequency is controlled by a

variable speed motor. There are two primary advantages in such inertia systems. In the first place,

high force capacities are not difficult to obtain. Secondly, the shaking amplitude of the system

remains unchanged by frequency cycling. Therefore, if a system is set to provide a 1 mm amplitude at

30 Hz, changing the frequency to 40 Hz will not change the amplitude (since the stroke of the linkage remains the same for a particular setting, however, it could be changed by changing the linkage

dimensions). It should be noted that both the available excitation force and the required accelerating force are harmonic functions of the square of the exciting frequency; hence as the requirement

changes with frequency, it also changes the available force.

15.5.3 Hydraulic and Pneumatic Systems Important disadvantages of the electromagnetic and mechanical shaker systems are limited load

capacity and limited frequency, respectively. As result, the search for other sources of controllable

excitation has led to investigation in the areas of hydraulics and pneumatics. In this arrangement an

electrically actuated servo valve operates a main control valve, in turn regulating flow to each end of a

main driving cylinder. Large capacities (up to 2 MN) and relatively high frequencies (to 400 Hz), with amplitudes as great as 46 mm, have been attained. Of course, the maximum values cannot be attained simultaneously. As would be expected, a primary problem in designing a satisfactory system of this

sort has been in developing valving with sufficient capacity and response to operate at the required

speeds.

Relative Merits and Limitations of Each System Excitation Frequency: The upper frequency ranges are available only through use of the

electromagnetic shaker. In general, the larger the force capacity of the electromagnetic exciter, the

lower its upper frequency will be. However, even the 2105 N shaker boasts an upper frequency of

2000 Hz. To attain this value with a mechanical exciter would require spin speeds of 1 20 000 rpm.

The maximum frequency available from the smaller mechanical units is limited to approximately 120

Hz (7200 rpm) and for the larger machines to 60 Hz (3600 rpm). Hydraulic units are presently limited to about 2000 Hz with very high peak load.

933

Force Limitations: Electromagnetic shakers have been built with the peak force ratings of 2 MN.

Variable-frequency power sources for shakers of this type and size are very expensive. Within the

frequency limitations of mechanical and hydraulic systems, corresponding or higher force capacities

may be obtained at lower costs by hydraulic shakers.

Maximum Excursion: The upper limit of peak-to-peak displacement for the electromagnetic exciter

may be considered as 25 mm or slight more. Mechanical type may provide displacement of the order of 150 mm, whereas the hydraulic exciter can provide displacement of the order of 450 mm.

Magnetic Fields: Because the electromagnetic shaker requires relatively intense fixed magnetic field, special precautions are sometimes required in testing certain items such as solenoids or relays, or any

device in which, induced voltages may be a problem. Although the flux is rather completely restricted

to the magnetic field structure, relatively high stray flux is nevertheless present in the immediate

vicinity of the shaker. Operation of items sensitive to magnetic fields may therefore be affected,

degaussing coils are sometimes used around the table to reduce flux level.

Non-sinusoidal Excitation: The shaker head motion may be sinusoidal or complex, periodic or completely random. Although sinusoidal motion is by far the most common, other waveforms and

random motions are sometimes specified. The electromagnetic shaker offers most of waveforms.

Although, the hydraulic type may produce non-harmonic motion, precise control of a complex

waveform is not easy. Here again, future development of valving may alter the situation. The voice

(moving) coil of the ordinary loudspeaker normally produces complex random motion, depending on the sound to be reproduced. Complex random shaker head motions are obtained in essentially the

same manner. Instead of using a fixed-frequency harmonic oscillator as the signal source, either a

strictly random or a predetermined random signal source is used. Electronic noise sources are

available, or a record of the motion of the actual end use of the device may be recorded on magnetic

tape and used as the signal source for driving the shaker. As an example, the electronic gear may be

subjected to combat-vehicle motions by first tape-recording the output of motion transducers, then using the record to device a shaker.

15.5.4 Impact hammer One of the popular methods of excitation is through use of an impact or hammer. It is a relatively

simple means of exciting the structure into vibration. The equipment consists of an hammer, usually

with a set of different heads and tips, which serve to extend the frequency and force level ranges for

testing a variety of different structures. The hammer tips can be of rubber, aluminum, steel, etc. Using

different sizes of hammer may also extend the useful range. Integral with the impact is a force

transducer, which detects the magnitude of the force, felt by the impact, and which is assumed to be

934

equal and opposite to that experienced by the structure. The impact incorporates a handle to form a

hammer as shown in Figure 15.24, so that impact can be applied manually. Basically, the hammerhead and the acceleration with which it is moving when it hits the structure determine the

magnitude of the impact. The frequency range, which is effectively excited by this type of device, is

controlled by the stiffness of the contacting surface and the mass of the impact head. The stiffer the tip

materials, the shorter will be the duration of the pulse and the higher will be the frequency range

covered by the impact. It is for this purpose that a set of different hammer tips and heads are used to

permit the regulation of the frequency range to be encompassed. Care should be taken while

impacting so that multiple impacts or hammer bounce does not occur, otherwise these would create

difficulties in the signal processing stage.

Figure 15.24 Impact hammer

A typical specifications of the impact hammer are as follows: Sensitivity at output of hammer: 0.95 pC/N, Rubber tip specifications: Frequency range 0-500 Hz, Duration range 5-1.5 ms, Force range 0-700 N, Physical: Weight of the hammer 280 g, Materials of Tips: Anodized aluminum, stainless

steel, titanium, neoprene rubber. When using the above type of the hammer, the actual impact force

applied to the test structure will always be greater than the force measured across the transducer

because of the inertia of the tip. These forces are related as follows

( )a m tF F m m m= (15.26)

where Fa is actual force input to structure, Fm is measured force, m weight of the hammer plus tip and

mt weight of the tip.

Determining Natural Frequencies of the Rotor Bearing System Using Impact Hammer Test Natural frequencies of the rotor bearing system are important parameters to be determined prior to any

investigation. For a two rotor system two natural frequencies are obtained by using the impact test.

Impact is applied at one of the rigid disks while the rotor is stationary (non-rotating). Displacement to impulse force is measured at the bearing end both in the horizontal and vertical directions using

935

proximity probe transducer. The FFT of the measured impulse response then gives frequency domain

impulse response. In the frequency domain response natural frequencies appear as higher amplitude

peaks. Figures 15.25 and 15.26 show the absolute value of the FFT of the measured impulse response in the horizontal and vertical directions, respectively. These plots indicate the first and second natural

frequencies, and these are equal to 38 Hz and 125 Hz, for the present configuration of the rotor bearing system.

Figure 15.25 Natural frequencies of the rotor bearing system in the horizontal direction

Figure 15.26 Natural frequencies of the rotor bearing system in the vertical direction

Example 15.6 An electromagnetic-type sinusoidal vibration exciter has a rated force capacity of 25 N is to be used to excite a test item weighing 3 kg. If the moving parts of the shaker have a mass of 0.75 kg and the amplitude of the vibration is 0.15 mm (peak-to-peak amplitude = 0.30 mm), determine the maximum excitation frequency that can be applied.

Solution: We have, the amplitude of vibration a = 0.15 mm, the rated force capacity Ft = 25 N, mass of the test item m = 3 kg, mass of the moving parts of shaker mc = 0.75 kg, and the maximum

frequency of excitation is to be determined. The dynamic force can be expressed as

936

( ) 2t cF m m a= + (a)

On substituting into equation (a), we have

25 = (3 + 0.75) 2 0.00015 (b)

which gives the maximum excitation frequency = 210.82 rad/s = 33.6 Hz. It could be seen that

because of relatively heavy mass as compared to the capacity of exciter, the maximum frequency of

excitation is relatively low.

Example 15.7 Suppose a vibration test requires a sinusoidal excitation force for a 10-kg test item at 100 Hz with a displacement amplitude of 2 mm (peak-to-peak amplitude = 4 mm) What will be the capacity of the exciter required? If support fixtures are required, they too must be shaken along with

the moving coil of the shaker itself. Suppose these items (the fixture and moving-coil assembly) have a mass of 5 kg.

Solution: The maximum force will correspond to the maximum acceleration, and the maximum acceleration can be calculated as follows:

The circular frequency = = 2 pi 100 = 628 rad/s

The maximum acceleration = the displacement amplitude 2 = (2 10-3) (628)2 = 789 m/s2

The maximum force = m a = 10 789 = 7890 N

This, of course, is the force amplitude required to shake the test item only. An additional force of 5

789 = 3945 N is required towards the fixture and moving-coil assembly. The rated capacity of the

shaker must therefore be a maximum of about 7890 + 3945 = 11835 N 12 kN.

15.6 Sound Measurements Sound waves are a vibratory phenomenon. Acoustic effects also give rise to harmonic pressure

fluctuations that they produce in a liquid or gaseous medium. They also characterized by an energy

flux per unit area and per unit time as the acoustic waves moves through the medium. A mathematical

description of different acoustic will be given in this section. It is standard practice in acoustic

measurements to relate the sound intensity and the sound pressure to certain reference values 0I and

0p , which correspond to the intensity and mean pressure fluctuations of the faintest audible sound at

a frequency of 1000 Hz. These reference levels are

937

120 10I

= W/m2 (15.27) and

50 2 10p

= N/m2 (15.28)

The intensity and pressure levels are measured in decibels. Thus

Intensity level (dB) = ( )010log /I I (15.29) and

Pressure level (dB) = ( )020log /p p (15.30)

When pressure fluctuations and particle displacements are is phase, such as in plane acoustic wave,

these levels are equal,

0 010log 20log pI I p=

The magnitudes of the particle velocity and pressure fluctuations created by a sound wave are small.

For example, a plane sound wave having an intensity of 90 dB is considered the maximum

permissible level for extended human exposure. In many circumstances we shall be interested in the

sound intensity that results from several sound sources. This calculation could of course, be

performed with equation (15.29) and (15.30). Sound pressure is the local pressure deviation from the ambient (average, or equilibrium) pressure caused by a sound wave. Sound pressure can be measured using a microphone in air and a hydrophone in water. The SI unit for sound pressure is the Pascal

(symbol: Pa or N/m2).

In machinery analysis the dB scale is used to express the ratio between two voltages - an output to an

input, for example (mathematically a dB is a 20-log voltage ratio). When the output equals the input, 20 log 1 = 0 or 0 dB. Thus, the zero or reference on the dB voltage scale occurs when input and

output are equal. It follows that perfect reproduction of an input signal could be stated by specifying a

0 dB permissible tolerance or deviation, which is an ideal case. In practice, it is customary to specify

the output of an instrument as a tolerance, 20 dB for example (ten-to-one attenuation across the instrument: 20log10 = 20; similarly a 100-to-1 attenuation corresponds to 40 dB and 1000-to-1 = 60 dB), over a given frequency range as a measure of the instruments deviation in voltage output compared to the voltage input.

938

The dB scale can be used equally well for the gain or an increase in the voltage. Using previous

example a gain of 20 dB equals an increase in voltage by a factor of 10, 40 dB equals an increase by a

factor of 100 etc. The dB scale is nothing more than a method to express the ratio between two

quantities. As pointed out earlier, 20 dB represents a voltage ratio of 10 with multiple of 20 dB (e.g., 20, 40, 60, 80, 100) equal to powers of 10 (e.g., 10, 102, 103, 104, 105). Similarly voltage ratio, 2 dB = 1.26, 3 dB = 1.41, 5dB =1.78. Next, 6 dB represents a voltage ratio of approximately 2:1 while 10 dB approximates 3:1 and 14 dB, 5:1. As a voltage ratio 50 dB = 40 dB (nearest multiple of 20) + 10 dB. Since adding algorithms is equivalent to multiplying numbers, 50 dB equals a voltage ratio of 100

(40dB) 3 (10 dB) 300 (actual 316).

Example 15.8 Calculate the total sound intensity from two sound sources at 40 and 50 dB.

Solution: T