Unexpected Machinery Failures and Preventive Maintenance
Unexpected machinery breakdowns and their negative consequences, such as unplanned production stoppages and associated high costs, have long been major concerns for manufacturers and large-scale industry owners. Estimates suggest that maintenance expenses in various industries account for 15% to 40% of total production costs, with the highest expenses (up to 40%) reported in the steelmaking and rolling industries. However, these estimates exclude the costs associated with unplanned production stoppages and the resulting damages.
Given this issue, manufacturers have continually sought ways to reduce the high costs associated with repairs. Advances in technology in this field are largely driven by these efforts. Today, with the introduction of innovative maintenance methods and advanced techniques, it is often possible to reduce repair costs by 20% to 50%. In some cases, even greater cost reductions have been achieved.
The significant potential for increasing production and reducing costs through the implementation of new methods has made the maintenance unit in any industry one of its most important departments. Nowadays, there are few factories that do not prioritize strengthening and equipping this unit with experienced human resources and advanced fault detection and diagnostic equipment.
Maintenance systems are divided into three main categories: breakdown maintenance, preventive maintenance, and predictive maintenance. Studies show that predictive maintenance, compared to other methods, offers better results both economically and in terms of safety and productivity. This method relies on equipment condition monitoring, to the point where condition monitoring units have now become one of the key pillars in manufacturing industries. The main goal of this system is to detect early-stage failures in critical production equipment.
Various techniques are used to monitor equipment condition, selected based on the type of equipment and its operating conditions. Vibration analysis, thermography, oil analysis, and electrical testing are among the most common methods. Vibration testing is particularly effective for mechanical equipment, as it helps identify forces acting on the equipment in real operating conditions. This method is especially efficient for high-speed equipment, which is why maintenance for these machines has shifted from breakdown or preventive approaches to predictive maintenance.
However, for low-speed equipment, due to the lack of accurate condition monitoring techniques, suitable periodic maintenance programs are still unavailable. This leads to frequent breakdowns and high repair costs. Some researchers have suggested using acoustic tests, such as sound emission testing, as a potential solution for more accurate condition monitoring of low-speed machinery.
Torsional Vibration-Based Condition Monitoring
One of the widely used condition monitoring methods is frequency spectrum analysis of transverse vibration signals, which has been extensively studied by many researchers. In this approach, sensors are typically mounted on bearings, motors, gearbox housing, and the foundation. Due to the complexity of the frequency spectrum of transverse vibration signals—resulting from amplitude modulation frequencies from gears and environmental vibrations— torsional vibration analysis has also attracted researchers' attention.
Torsional vibration signals are not affected by the additional modulations mentioned above and have a simpler structure compared to transverse vibration signals. Consequently, advanced signal processing methods have been employed to detect faults in rotating machinery under both constant and variable operating conditions. Torque measurement sensors, primarily of the dynamic type, are commonly used in condition monitoring, as static sensors are less applicable. In this regard, a laboratory setup has been designed and implemented to simulate faults and dynamic loading on the system for experimental purposes.
Intelligent Condition Monitoring and Fault Diagnosis
In recent years, intelligent condition monitoring and fault diagnosis have gained significant attention as advanced approaches in predictive maintenance. These approaches utilize technologies such as machine learning, Internet of Things (IoT), big data, and especially advancements in artificial intelligence, to enable more accurate analysis and prediction of machinery status. In these methods, sensor data is collected and processed using advanced AI algorithms, which not only identify potential failures but also more precisely detect possible root causes. Intelligent fault diagnosis, in particular, leverages data analysis to uncover hidden patterns within equipment performance data, enabling fast and accurate identification.
The role of artificial intelligence is crucial in development of condition monitoring and fault diagnosis systems. Deep learning algorithms and artificial neural networks can process vast amounts of sensor data, detecting hidden patterns related to failures and operational issues. These technologies allow systems to automatically adapt to changing operating conditions of machinery without human intervention. Additionally, AI enables more advanced data analysis, allowing condition monitoring and fault diagnosis systems to predict potential issues with high accuracy, preventing unexpected downtimes. This capability leads to increased productivity, optimized maintenance processes, and reduced operational costs.
With the advent of Industry 4.0, the significance of intelligent condition monitoring and fault diagnosis has become even more prominent. Industry 4.0, with its integration of cyber-physical systems, IoT, and advanced automation, has contributed to the expansion of intelligent technologies. In this context, intelligent condition monitoring and fault diagnosis, as core components of Industry 4.0, enhance the accuracy and speed of failure detection while reducing maintenance costs, thereby boosting productivity and minimizing operational risks.