I.

1 Introduction
Industrial equipment can have various adverse consequences on operational safety (reliability,
availability, security, maintenance), personal safety, the environment, as well as the economy
and politics. All these consequences have driven the improvement of its performance level
(productivity, quality) and reduced its overall cost throughout its life cycle.
Implementing quality maintenance is the main economic challenge for production optimization.
Equipment maintenance allows the industry to take measures based on production regularity,
manufacturing costs, and competitiveness in order to achieve commercial success.
This chapter presents general definitions of industrial maintenance, monitoring of rotating
machinery, as well as vibro-acoustic diagnostics.

I.2 Generalities on maintenance

I. Generalities on maintenance and monitoring

I.2.1 Definition of maintenance

Maintenance refers to the set of technical, administrative, and management actions during the
life cycle of an asset, aimed at maintaining or restoring it to a state in which it can perform the
required function [1] "excerpt from standard NF-EN 13306."

I.2.2 Different types of maintenance

Maintenance actions vary depending on the implementation techniques, as shown in Figure I.1:
. Corrective maintenance
. Preventive maintenance

FIGURE I.1….

I.2.2.1 Corrective Maintenance

Corrective maintenance is defined by the standard [NF EN 13306] as the execution of a
maintenance operation after the detection of equipment failure, in order to restore it to a
functional state [1]. In this type of maintenance, the intervention occurs after the occurrence of
the failure.

There are two types of corrective maintenance: palliative maintenance and curative maintenance.
• Palliative Maintenance: defined as troubleshooting in the case of partial failure, its objective is
to eliminate the effects of the failure and allow the equipment to function until preventive
shutdown.
• Curative Maintenance: defined as repair in the case of total failure, the goal of this maintenance
is to get the equipment back up and running [1].
I.2.2.2 Preventive Maintenance
The objective of preventive maintenance is to reduce the probability of failure or degradation of
machine performance [2]. The operations of this maintenance are carried out through planned,
prepared, and scheduled interventions based on different criteria to avoid potential failure. There
are numerous operations involved in preventive maintenance, such as cleaning, lubrication of
friction points, equipment inspection and control both during operation and while stopped, partial
or complete revision, vibration signal analysis (measurement), oil analysis, and more [3].

Preventive maintenance is divided into three types:

• Systematic preventive maintenance: defined by the [AFNOR X 60-010] standard as a set of

activities performed according to a predetermined usage unit program. Changes to parts and
fluids occur regardless of their degradation states, which take place periodically [4].

• Conditional preventive maintenance: defined by the [AFNOR X 60-010] standard as a set of

maintenance activities performed based on predetermined criteria specific to the equipment's
degradation state. It is generally dedicated to monitoring sensitive components of the equipment
[4].

• Predictive preventive maintenance is used to determine the cause of problems in a machine or

process. This maintenance method provides experts with an effective way to reduce the total
equipment downtime. In fact, in this type of maintenance, the equipment operates with virtually
no unexpected downtime.

I.3 Surveillance Techniques

Surveillance is a necessary operation that allows for monitoring the real-time condition of the
asset. It can be calculated continuously, at predetermined or non-predetermined time intervals,
with time or the number of units used.
This function is ensured through the regular recording of degradation or performance indicators
to ensure the monitoring of machinery and equipment. Various analysis techniques exist (Figure
I.2), such as:
- Vibration analysis
- Acoustic emission
- Ultrasonic testing
- Thermography
- Oil and lubricant analysis.
- Resistance variation in an electrical circuit
- etc
FIGURE I.2 - Breakdown of different maintenance techniques [5]…

I.3.1 Vibration analysis

Vibration analysis of industrial equipment has been widely used in preventive maintenance
activities and effectively detects most anomalies in rotating machines. Defects caused by
bearings, worn clearances, or misalignment will manifest as a change in the internal forces of the
machine, resulting in the appearance of vibrations.

I.3.2 Infrared thermography

Infrared thermography assumes that any object with a temperature above 0 Kelvin emits energy
in the form of electromagnetic radiation. This energy is closely related to its surface temperature.
Infrared thermography involves capturing and quantifying this radiation and, if properly
calibrated, reconstructing it with a temperature mapping of the object's surface (6]. Thermal
cameras are commonly used to monitor live electrical equipment, heating systems, or certain
industrial processes (such as welding and rolling). Its application program for monitoring the
operating condition of rotating machines is not significant. One of the main reasons is that it does
not require any obstacles between the camera and the object being observed.

1.3.3 Acoustic Emission

Acoustic analysis is a technique used in condition-based preventive maintenance to detect
defects by studying sound waves and vibrations generated by electromechanical systems. A
sound level meter is used to measure the sound or noise by evaluating the acoustic pressure level.
Consequently, the sound is converted into an electrical signal, where the amplitude and
frequency are comparable to acoustic phenomena. The electrical signal can be conditioned,
sampled, and processed to characterize the measured noise. This allows for determining acoustic
values expressed in dB, performing frequency analyses, statistical analyses, and integrating the
signal over a specific duration [7].

I.3.4 Analysis of oils and lubricants

The main function of lubricants is to ensure smooth contact between different components of the
machine, thereby reducing the risk of wear. They can also perform other tasks, such as cooling,
rust and corrosion prevention, and non-destructive monitoring of machine operating conditions
[8].
In the event of a machine part failure, monitoring the physical and chemical properties of the
lubricant can provide information about the type and degree of degradation (especially in the
case of degradation) of the lubricant, which is related to the lubricant itself, or contamination due
to the presence of wear debris. This information can be obtained from inspections and daily
visits, conducted through simple tests based on visual inspections, odor, and color of the
lubricants, or more in-depth laboratory analyses such as chromatography, photometry, or
spectroscopy [9].

Table I.1 shows the uses of each type of analysis:

1.4 Vibro-acoustic diagnosis
Vibro-acoustic diagnosis is a maintenance tool that is currently well-known thanks to
developments in the field of computer science and signal processing. This technique uses a
system that combines vibration measurement (accelerometers) and acoustic measurement
(microphone) to analyze the dynamic behavior of rotating machines and establish a diagnosis of
their operating condition [11].
The different methods for establishing vibro-acoustic diagnosis are divided into three classes:
- Class 1: A set of methods involving modifications of the machine's operation and a comparison
with simple acoustic measurements.
- Class 2: Combines methods based on signal processing technology, frequency, time, or space,
and applies these methods to signals obtained during normal operation of the device.
- Class 3: Mixed use of various methods, while using simple measurement values obtained
during normal operation (acoustic pressure, vibration velocity) and calculating structural
vibrations or sound radiation, so that causal calculations (from force to vibrations or from
vibrations to noise) can be performed.

I.4.1 Noise generation mechanisms

These mechanisms produce a mechanical force Fmot or an acoustic pressure Wmot. The
mechanical efforts can be attributed to all balancing and alignment failures of rotating parts,
inherent shocks from gear engagement or multipliers, electromagnetic forces related to the
operation of motors or transformers, and clearances. There are impact fluctuations or cutting
efforts, etc [11].
The acoustic pressures can come from turbulent flows in pneumatic or hydraulic circuits (in
valves, bends, at the exhaust of actuators, etc.) or from the high-speed displacement of parts (fan
blades, circular saw blades, etc.) [11].

I.4.2 Transfer mechanisms

These mechanisms vary depending on the nature of the source, and there are three types, namely
[11]:
• Solidiens: when they are due to the Fmot efforts that cause deformations of structures and
vibratory velocities Vstry
• Aerial: when they are due to acoustic pressures Wmot that radiate directly into the external
environment, for example through openings,
• Mixed: in the presence of conversions of vibratory energy into acoustic energy (or vice versa).
An aerial source placed in a closed cavity creates an acoustic pressure Pint that generates wall
vibrations: this constitutes a conversion of acoustic energy into vibratory energy. These wall
vibrations are converted outside the cavity into acoustic energy. The energy conversion
mechanism, resulting in a mixed transfer, is called transparency.

I.4.3 Principle of diagnosis and vocabulary

The diagnosis of rotating machines follows a logical sequence for fault localization, evaluation
of its severity, and decision-making at the end, as shown in Figure 1.3, the steps for
implementing a diagnostic system:

FIGURE 1.3 General principle of vibro-acoustic diagnosis.

Step 1: System Analysis This is a crucial step that involves defining an objective for the system
to be implemented. In order to do so, it is necessary to investigate the system's characteristics
(such as the number of gear teeth to monitor), the defect and the symptoms of the defect to
identify, etc. It is this step that led to the selection of the method to be implemented, specifically
the signal processing method to be used [12].

Step 2: Instrumentation and measurement This step involves choosing the type of sensor to
use (accelerometer, current sensor, microphone, etc.) based on the measurement conditions
(contact surface, machine type, etc.), as well as the technical characteristics of the sensor
(permissible frequency range, sensitivity, etc.). The second point is the choice of the acquisition
card, which depends on the types of faults to be monitored. This allows for the definition of the
acquisition frequency, the type of recording (continuous, periodic, etc.).

Step 3: Monitoring and detection Once the sensor is installed and the acquisition begins, the
monitoring operation consists of extracting statistical indicators from the acquired signals that
reflect the state of the machine. The indicators used are generally "simple" and are usually only
used for detection purposes, for example, to indicate when the indicator exceeds a certain
threshold, which indicates an abnormal state of the machine. These thresholds are usually
determined by empirical or statistical signal processing methods. Many signal processing
methods are better suited for "detection" [13].

Step 4: Diagnostic The diagnostic process involves determining the faulty part of a complex
system. The term "must be isolated" is often used by Anglo-Saxons. This is where signal
processing can play its biggest role. It involves extracting from the signal, with knowledge of the
system and the symptoms of the faults, the parameter(s) related to these symptoms. In addition to
signal processing methods, this step can use "decision" methods, such as pattern recognition.
[14]

Step 5: Decision This is the final step in the diagnostic procedure. In an industrial environment,
it involves determining the actions to be taken, initiating maintenance operations, stopping the
machine, etc. [12].

I.5 Vibration Sensors

Vibration is characterized by three kinematics: displacement expressed in µm, velocity expressed
in [mm/s], and acceleration expressed in [m/s]. The first step to obtaining a vibration reading is
to convert the mechanical vibration generated by the machine into an equivalent electrical signal.
This operation is performed by the vibration sensor. Among the most commonly used sensors are
proximity probes (for displacement measurement), velocimeters (for velocity measurement), and
accelerometers (for acceleration measurement). Depending on the physical laws implemented by
the manufacturer, there are several types of vibration sensors: eddy current sensors, displacement
sensors with capacitive or inductive probes, Doppler effect velocimeters, and piezoelectric
accelerometers. In all of these sensors, piezoelectric technology is widely used due to its qualities
of metrology, compatibility, and ease of use. Figure 1.4 shows a piezoelectric accelerometer.
I.6 Microphone
The primary function of a microphone is to capture sound waves and transform them into an
electrical signal called an audio signal. In other words, a microphone is an energy transducer that
converts acoustic energy into electrical energy. There are several types of microphones (electret,
condenser, ribbon, moving coil, dynamic, etc.), and the system used for energy conversion is
generally specified by the microphone's name. The choice of microphone depends on various
factors, such as the sound source (instrument, solo voice, etc.), the recording location (studio,
multipurpose hall, outdoor, etc.), and the microphone's placement relative to the source, among
others. Figure 1.5 illustrates the diagram of a piezoelectric microphone.

1.7 Conclusion
In this chapter, we have studied maintenance in general, with a particular focus on maintenance
through vibrational and acoustic analysis. Various tools can be used, such as oil and lubricant
analysis, infrared thermography, acoustic emission, and vibrational analysis, to ensure effective
preventive maintenance. Vibrational and acoustic analysis is well-known and widely used, as it
allows for the detection of almost all possible faults in rotating machinery. This technique offers
several advantages, including ease of measurement, early fault detection, and in-depth
diagnostics to identify the root cause of the failure. The next chapter will be dedicated to
presenting the main faults of rotating machinery, specifying the characteristic frequencies of
each fault.