In this last post we take a closer look at the Interval Servicing maintenance model.
Interval Servicing has a lot of similarities with Preventive Maintenance. The philosophy is to identify and solve small problems before they become big ones. Only with Interval Servicing, parts are not replaced after a specific time. Instead, parts are frequently inspected to estimate when it needs replacement . A good example is your car, which in most countries needs annual inspections. During these inspections, parts such as your brake pads, suspension and steering alignments are checked as they inherently have a strong varying lifetime.
This maintenance model has some great advantages. By regular inspections you know relatively well whether or not a part might fail in the next inspection interval. Therefore, you mostly stay ahead of unexpected failures. Similar to Preventive Maintenance, this increases the operational safety. Compared to Preventive Maintenance, Interval Servicing has the additional benefit of increasing the lifetime. Instead of replacing the asset after a certain interval, the asset is checked whether it is most likely to survive or fail in the next interval. This is especially beneficial for assets with a very large variance in their failure ratio. However, as with Preventive Maintenance, it’s not all sunshine and roses. Although you increase asset lifetime with Interval Servicing, there are more check-ups and inspections. This results in even more employees, maintenance costs and challenges for your maintenance management.
Remember, even identical assets will not fail after a similar operating interval. Instead, this failure time follows a probabilistic bell-shaped curve. Your mechanics take quantitative measurements which approximate the assets location on this curve and execute maintenance accordingly. Therefore, Interval Servicing is a decent maintenance model when the variance of failure probability is large, as the increased lifetime by regular inspections is significant.
We like the philosophy of Interval Servicing, as it tries to capture the probabilistic failure curve and prolongs the operational time and avoid unexpected failure. However, the approach can be much more efficient if the information of sensor data is constantly analysed. Our machine learning algorithms can make this translation and provide you with failure probability information. This allows you the optimal benefits of Interval Servicing without having to pay the high price of frequent physical inspections.
Here we conclude our mini series about maintenance models. Next post we’ll post about our approach, how we actually translate your sensor data into valuable maintenance information.