Although superficially monitoring a closely related quantity, velocity transducers are fundamentally distinct from accelerometers. The velocity transducer is an electromagnetic device in which an electrical coil moves through a permanent magnetic field. Although the usage is declining, it does have some advantages over the piezoelectric accelerometer. It remains sensitive to lower-frequency motion and is rather less sensitive to temperature and interference signals.
Of course, signals only need to be integrated once to yield the displacement and this may be a consideration in some occasions.
The disadvantages are that they tend to be larger and heavier than accelerometers, and they suffer more cross noise among different directions. However, they have a role, albeit a fairly minor one currently.
The introduction of proximity sensors in the 1950s was indeed a major step forward in the monitoring of rotating machinery. For the first time, engineers were able to follow the motion of a rotor directly rather than inferring information from data gathered at the stator.
Donald Bently developed the proximity transducer for use on rotating machines. It is important to note that these instruments measure displacement rather than velocity or acceleration. There are two main types: the eddy current type (as introduced by Bently) and the capacitive probe.
While seemingly similar, these two operate on quite different principles and consequently have rather distinct applications. Whilst eddy current probes rely on the magnetic field between the instrument and the target, the capacitive probe examines the electric field and consequently there are differences in both sensitivities and performance.
The eddy current probe is more sensitive to variations in the material properties of the rotor and consequently there will often be an "electrical run-out" (or baseline variation) observed, although this is usually very small. These probes also need calibration with respect to a particular rotor. This is not the case with capacitive instruments which are much less sensitive to rotor (target) material.
Against this, eddy current probes have much higher tolerance to hostile environments and higher temperature. They tend to have a higher power usage which can be an issue in some vacuum installations.
It is true to say that the introduction of the proximity transducer represented a major step forward in measurement and hence understanding of rotating machinery.
Linear Voltage Differential Transformer (LVDT)
These devices have a much less significant role in machinery monitoring but they can provide useful insight on occasions. The instrument is simply a variable resistor (in a structure similar to a bicycle tire pump) in which the motion of the target moves the contact varying the resistance of a measurement circuit. Owing to the mechanical system involved, these instruments are only useful at very low frequencies (1 Hz or less). They can be helpful, for instance, in evaluating the influence of slow pipework vibration/motion. An interesting case arose some years ago: during the investigation of problems on a feed pump, an LVDT was mounted to measure the relative motion of the pump casing to the rotor. Note that on this pump the bearings were mounted on separate pedestals. As the rotor accelerated to full speed from rest, the relative vertical position changed by 40 pm. Initially, this seemed very surprising but the explanation soon became clear. As the pump's body comprised a rigid cylinder with a hole cut in the top (connected to a flexible pipework system), a substantial force was effectively pushing the pump's body downward as a pressure was developed. This insight proved valuable in resolving the plant problems, and no other instrumentation would have been so convenient.
In the context of studies on machinery, strain gauges are rather specialized and are used for specific investigations rather than as a standard tool. One case in which strain gauge was invaluable is discussed in Section 9.4. In more general structural dynamics, the strain gauge has been in use for many years and is the fundamental sensing element for many types of sensors, including pressure sensors, load cells, torque sensors, position sensors, etc. The majority of strain gauges are foil types, available in a wide choice of shapes and sizes to suit a variety of applications.
They consist of a pattern of resistive foil which is mounted on a backing material. They operate on the principle that, as the foil is subjected to strain, the resistance of the foil changes in a defined way. The strain gauge is connected into a Wheatstone bridge circuit with a combination of four active gauges (full bridge), two gauges (half bridge), or less commonly a single gauge (quarter bridge). In the half and quarter circuits, the bridge is completed with precision resistors.
Although installation is a somewhat demanding process, strain gauges have benefits including lightness, wide frequency response, and temperature stability. A problem with respect to use on rotors is recovering the data. This is most conveniently achieved with a telemetry system as described in Chapter 9.
Acoustic Emission Sensors
Acoustic emission (AE) is the transmission of elastic waves in a material generated by some change such as growth of a crack. High frequencies arise from local resonance of the wideband spectrum generated by the breaking of molecular bonds and these are usually in the region of 25 kHz to 1 MHz. Joseph Kaiser, in 1950, was the first researcher to record that engineering materials in general emit low-amplitude, high-frequency clicks when stressed, which we now call AEs. Kaiser's most fundamental finding was that material does not emit unless the applied stress exceeds any previously applied stresses. This is known as the "Kaiser effect."
The sensors are piezo-electric high-frequency accelerometers. In many cases their frequency response is less smooth than those of standard accelerometers, but in many instances, this is of little importance. This is because many AE investigations aim to identify the activity just by counting so-called events. It is only in recent years that investigators have been examining frequency content and its implications. This trend may well dictate some changes in the emphasis.
Storage of Data
It was, at one time, common practice to make analogue recordings of plant data and store on magnetic tape for subsequent analysis. Whilst there are occasions where an analogue record can be useful, in the vast majority of cases, recordings are now made digitally. It is important to appreciate, however, that in making a digital record, some important decisions have to be made. The most obvious is the sampling frequency which must be at least double the maximum signal frequency of interest.
Another point to be considered is the possibility of leakage (which is discussed more in Chapter 2). On the assumption that the data at some point will undergo Fourier transformation, it is important that higher frequencies in the data are not allowed to yield ghost "aliased" frequencies in the range of interest. If it is suspected that any high frequencies are present in the signal, appropriate antialiasing filters must be applied. The issues related to sampling and aliasing are discussed in Chapter 2.