There has been considerable confusion on the most suitable way to inspect the components and machinery in an industry for the presence of a pre-specified failure mode. The various forms of inspection used in this are sensory inspection, quantitative inspection and/or some technique of condition monitoring. The last technique is also known as predictive maintenance and an easy predictive maintenance definition would be a system which predicts the requirement of maintenance with the help of data gathered via various IOT devices and analysing them using machine learning and artificial intelligence (also known as Künstliche Intelligenz (KI) in German).
The various techniques used in defect inspection:
- Sensory inspections – This is perhaps the method with the maximum number of loopholes in the process of inspecting defects. This was traditionally considered to be the main pillar of any programme of inspection. The belief was that sending personnel regularly to inspect for malfunctioning would reveal the defective parts with enough time to take the corrective measures to prevent unplanned downtime. The person inspecting would depend on vision, hearing and touch to detect any changes from the previous inspection. If there are any changes, those would be investigated by the specialist on the upcoming scheduled date. Despite the benefits of sending someone around to look for defects, it should never be the main inspection technique because it can only detect the most prominent problems. Detecting early internal defects in machinery is impossible using sensory inspection.
- Enhanced Sensory Inspection – This is a combination of sensory and quantitative techniques of inspections with some features of condition monitoring. This method uses certain devices like handheld vibration pens and simple ultrasonic metres to enhance the power of human sensory inspection enabling to detect defects which earlier was not possible. However, it must be kept in mind that this is not a replacement of a full-fledged PDM.
- Quantitative inspections – This method proves useful in collecting data to identify any trends or features in any failure mode. Generally measurements in this form of inspection include the measurements of temperature, clearance and so on. This helps the technician in deciding if any further maintenance activity is necessary. A technician measuring something should have the lower limit, upper limit and average measurement of what he is measuring. That would ensure that he is able to find out if there are any faults with the components. However, if quantitative inspection is done in a regular basis, the chances of finding some abnormal measurements are minimal.
- PDM – The use of predictive maintenance in the inspection of defects in an industry has increased over time. This is undoubtedly the most advanced of all the other inspection methods. With the use of machine learning, a part of artificial intelligence (also known as Künstliche Intelligenz (KI) in German) to analyse the data received from various IOT devices, this method is the most efficient way of detecting problems with any components of the machinery. The owners have the power to make the decision to make or not make any maintenance depending on the current situation. As the definition of PDM suggests, it helps predict the need of a maintenance to avoid unforeseen downtime.
The PDM is a way to find out defects which were untraceable with the previous mechanisms. Moreover, this method of inspection takes place while the machines are operating with the least amount of hindrance. The use of Infrared Thermography, vibration sensors and other tools gives it an edge over the others. The owners should identify the exact inspection methods which would help them to achieve what they want. They should aim for the balance between risk taking and cost for designing the appropriate strategy.