1. Introduction

Vibration can be a tell-tale sign of damage in rotating equipment. In industrial settings, its reliability and increased performance are vital to gain competitive advantage and smooth operation. Mechanical failures are a nightmare to repair and cause considerable down-time in operations, not forgetting risk to life and limb. According to the Vibration Institute, up to 70 per cent of mechanical failures are due to misalignments and imbalances. If such failures are caught early, considerable downtimes and costs can be saved. Traditional maintenance strategies often involve periodic inspections, which are likely to detect developing issues only when failure is almost at the doorstep. This reactive approach often results in unplanned downtimes and bloated operational costs. Recent advances in sensor technology have begun to open new possibilities for improving machinery monitoring. Piezoelectric sensors are a case in point. Because piezoelectric sensors can measure vibration over a wide frequency range, they could be used to detect imbalance and wear developing in rotating equipment. Capturing vibrational modes and stress distributions within machines would let operators make informed decisions regarding maintenance interventions, thereby lengthening the useful life of critical components. Integrating these sensors with internet of things (IoT) technologies and machine intelligence algorithms will create further opportunities to improve the predictive maintenance of machines for smooth operations [1]. The paper outlines various applications of piezoelectric sensors on rotating equipment and turbines and also presents the findings of an experimental investigation of piezoelectric sensors implementation in those machines.

2. Vibration Monitoring in Rotating Equipment

2.1. Detection of Imbalance

Rotating machinery is not properly balanced, which causes excessive vibration, which can destroy expensive equipment if left untreated. Vibration Institute suggests 70% of mechanical breakdowns result from misalignment and imbalance. Piezoelectric sensors pick up vibrations at very high frequencies, usually in the range of 1-1000 Hz, and can spot imbalances quickly. When a piezoelectric sensor is attached to a motor shaft, for example, it can detect vibrations that reflect imbalance states, such as a 5-10% shift in resonant frequency from operating specifications. With a monitoring system measuring vibration amplitude and frequency, maintenance personnel can decrease vibration by as much 40%, resulting in a longer service life of equipment and greater operational efficiency [2]. The following table gives an overview of experimental evidence on how piezoelectric sensors can effectively detect imbalance in moving machinery, in the form of frequency shifts and vibration cancellation.

Table 1: Experimental Results of Imbalance Detection in Rotating Machinery Using Piezoelectric Sensors

2.2. Shaft Misalignment Monitoring

Shaft misalignment increases the load on bearings and gearboxes, causing premature failure. Studies show that misalignment reduces bearing life by up to 50 per cent. Vibration analysis, provided by piezoelectric sensors, can detect shaft misalignment by detecting characteristic frequency patterns associated with misalignment, a phenomenon sometimes called the ‘1X frequency’, after the fundamental rotational frequency of the equipment. One study, published in the Journal of Sound and Vibration, showed that piezoelectric sensing allowed operators to detect misalignment changes as small as 0.2 mm, which could be corrected before more catastrophic failures occurred and before costly repair and replacement needs arose [3]. Catching misalignment early can allow for maintenance interventions to avert costly repairs and replace outages, saving companies up to 30 per cent of operational costs by reducing downtime significantly.

2.3. Bearing Wear Detection

Though bearings are essential to reduce friction and wear for rotating equipment, they’re also the most vulnerable components of most machines. Bearing wear will increase vibration levels dramatically, as a study done by a Japanese research foundation showed that a 10 per cent increase in vibration amplitudes heralds bearing failure. Continuous monitoring of the vibration spectra of the bearings will pick up changes in characteristic frequencies due to wear. If a bearing is in good vibrating health, it’ll exhibit a fundamental frequency of, say, 1,000 hertz. A worn bearing will exhibit harmonics and sidebands associated with bearings in rough shape [4]. Depending on how the frequency transforms due to wear, the fundamental frequency can range from, say, 1,050 Hz to a higher frequency. Such features can be characterised and studied through machine learning algorithms to predict a bearing failure with greater than 85 per cent accuracy, allowing operators to perform scheduled maintenance and head off the failing bearing without an unplanned shutdown.

3. Stress and Strain Monitoring in Turbines

3.1. Blade Fatigue Detection

Turbine blades are subject to cyclic stress that can lead to fatigue failure, which can account for up to 40% of turbine failures in some contexts. Piezoelectric sensors embedded in the tips of the turbine blades can collect crucial data on vibrational modes and overall stresses. Fatigue can be detected from changes in the natural frequencies of the blades; deviation from expected patterns may reveal fatigue. In one study, the researchers found that piezoelectric sensors integrated into a fatigue monitoring system allowed for early fatigue detection with a 90% reliability. This type of early warning allowed the blade to be removed before catastrophic failure and, hence, expensive outages and safety risks. Real-time maintenance interventions based on the data can extend blade life by up to 20%, thereby increasing turbine efficiency. Table 2 shows the experimental experimental results showing the superiority of piezoelectric sensors in detecting fatigue in turbine blades, with key data such as detection reliability, blade life and turbine efficiency indicated in the Table 2 [5].

Table 2: Experimental Results of Blade Fatigue Detection Using Piezoelectric Sensors

3.2. Rotor Stress Monitoring

Monitoring rotor stress plays an essential part in turbine performance. High output applications will put a staggering amount of stress on the rotor. If operators don’t monitor stress levels then the high centrifugal forces can lead to catastrophic failure of a turbine. In several studies it has been shown, that rotor stress can increase by as much as 30% under excessive loading. This is why it is essential for operators to monitor stress in real-time in order to take corrective action. Using sensors with piezoelectric properties, operators are now able to monitor stress by collecting strain gauge data that provide information on the operational condition of the rotor [6]. For example, the exemplary monitoring system shown in Figure 1 will typically show a shift from 0.5% rotor strain to 0.8% under normal operation. This would alert the operator that maintenance might be required. Using the piezoelectric sensor has also been shown to reduce the failure rate of rotors through the application of a real-time monitoring system by as much as 60%. Figure 1 below shows both the rotor stress under normal and excessive load conditions as well as the reduction in rotor failure rate after implementing piezoelectric sensors for rotor stress monitoring [7].

Figure 1: Rotor Stress Levels and Failure Reduction with Piezoelectric Sensor Monitoring

3.3. Creep and Thermal Stress

Under the high temperatures of the turbine environment, creep can cause gradual deformation that compromises component integrity. Studies show that components of high-temperature combustion turbines (a subtype of gas turbine) can exhibit creep rates doubling every 10°C increase in temperature. Piezoelectric sensors can measure both temperature and mechanical strain. An example of the damage that can be avoided by proactive maintenance based on piezoelectric feedforward is creep in the turbine section of a jet engine [8]. Data from the piezoelectric sensors can reflect creep deformation rates and enable a preemptive adjustment in operation conditions before failure occurs. The monitoring strategy including piezoelectric sensing can reduce maintenance costs associated with thermal stress failures up to 50 per cent by avoiding extensive damage [9].

4. Predictive Maintenance and Data Integration

4.1. Real-time Data Analysis

Predictive maintenance would rely on real-time data from piezoelectric sensors; studies have indicated that once such systems are in place, unplanned downtime can be reduced by as much as 75 per cent. Through continuous data collection, operators can analyse the sensor feeder data in real time and quickly detect deviations from regular operation that warrant prompt attention [10]. For instance, an increase in vibrations above 15 per cent can trigger an alert for immediate action, which would prevent the problem from worsening and allow the operator to take corrective action. Incorporating advanced analytics, such as Fourier Transform, would enable operators to decompose the vibration signal into different frequency components, thereby enabling the identification of specific failure modes. Studies have shown that incorporating these analytics could improve the fault detection rates by 90 per centstyle='width:.05pt;height:0;mso-left-percent:-10001;mso-top-percent:-10001;mso-position-horizontal:absolute;mso-position-horizontal-relative:char;mso-position-vertical:absolute;mso-position-vertical-relative:line;mso-left-percent:-10001;mso-top-percent:-10001' />style='width:.05pt;height:0;mso-left-percent:-10001;mso-top-percent:-10001;mso-position-horizontal:absolute;mso-position-horizontal-relative:char;mso-position-vertical:absolute;mso-position-vertical-relative:line;mso-left-percent:-10001;mso-top-percent:-10001' />.

4.2. IoT Integration

Integrating piezoelectric sensors with IoT enables enhanced monitoring using networked systems. Combining sensors with a centralised IoT system that can be accessed remotely can provide real-time data to allow for rapid decision-making. McKinsey found that IoT-enabled predictive maintenance can cut maintenance costs by up to 30 per cent and reduce downtime by 25 per cent. For example, remote monitoring systems can trigger an alert based on sensor data reading that shows a sharp rise in vibration level, prompting maintenance teams for corrective measures in advance. Moreover, IoT integration can allow data-sharing across systems, offering a multi-dimensional perspective of the operational state and support for cross-departmental collaboration.

4.3. Machine Learning for Predictive Maintenance

Machine learning algorithms maximise the value of predictive maintenance by mining large data sets produced by piezoelectric sensors. These algorithms learn from past failure data to identify complex failure patterns and predict parts that might fail. Machine learning can improve prediction accuracy to more than 90 per cent compared with traditional methods [11]. For instance, a company that used machine learning to analyse vibration data was able to reduce unnecessary maintenance activities by 40 per cent, ensuring they covered only equipment that showed signs of distress. In addition to maximising use of resources, the reduced disruption from maintenance operators helped cut costs.

5. Conclusion

The drastic implication of this study on the effectiveness of measurements taken by piezoelectric sensors to determine early failure of rotating equipment is the undeniable fact that at an early stage of this detection, proactive maintenance will become possible, allowing downtime of turbines to be kept at a minimum and extending the lifetime of the machinery even more. The experimental results shown in the figure above display significant advantages for operational efficiency: reduction in vibration levels and increase in the life of the blades used. The introduction of IoT technologies into manufacturing processes is providing new horizons to predictive maintenance. Using machine learning tools, it has become possible to develop a more comprehensive approach to machine health In light of the ongoing trend of the Fourth Industrial Revolution in which digital technologies are transforming industry and merging the physical, digital and biological spheres towards smart manufacturing and automation, the incremental adoption of piezoelectric sensors in machine health monitoring is going to be a revolutionary step in an industrial world where the focus is becoming more and more to create safer and more efficient environments. Future research on this topic will need to refine these technologies and test their applicability to different industrial sectors to achieve more robust and effective predictive maintenance.