Traditionally, Facility Management companies have followed preventative maintenance programs. These allocate a time period or amount of usage to a piece of equipment. At the end of this period of time, or after the usage limit has been hit, the equipment is attended to and predefined preventative measures are undertaken. Examples might be replacing air filters and inspecting fan belts monthly.
The theory behind preventative maintenance is sound. After all, regularly looking after expensive and / or critical equipment is much better than neglecting it. Moreover, conservative setting of Preventative Maintenance schedules means that equipment failure is minimised and staff rotas can be set appropriately.
However, Preventative Maintenance can also lead to expensive and unnecessary downtime, changing parts that might not actually be ready for change and causing operational disruption whilst doing so. Furthermore, it doesn’t take into account external variables that can affect the useful life and performance of equipment.
Predictive maintenance addresses some of the issues that pure Preventative Maintenance programs face, by seeking to identify the optimal time to replace or repair particular elements of a piece of equipment or system, based on the analysis of certain variables. This data might show fluctuations in temperature, light, water pressure or whatever else is relevant to that particular item and often makes use of algorithms that are refined through machine learning in order to predict the need for intervention.
As well as maximizing the life of parts and simultaneously mitigating against the likelihood of unforeseen failures, predictive maintenance also allows FM providers to detect changes in performance where the equipment is not actually nearing full failure but is running inefficiently and draining resources.
Smart sensors are a part of the Internet of Things (IoT) revolution and play an important role for the data capture that is the backbone of predictive maintenance. Key examples include the monitoring of temperature, humidity, light, movement / presence and air quality. There are also many applications for ultrasound and infrared sensors.
Ultrasound sensors, as an example, can measure the level of solids or liquids in closed containers. These are useful for waste management and are used on smart bins.
Humidity sensors, on the other hand, can be crucial for clients that manufacture sensitive items such as Food and Beverage. Being able to quickly identify excess humidity and its source is vital as mistakes can be very costly.
In many cases, implementing a predictive maintenance regime requires investment in both infrastructure and training. It is not necessarily the case that predictive maintenance is always the better choice for every situation. Maintenance carries a cost, which includes time and service disruption. Therefore, if you are dealing with expensive and mission critical equipment, then predictive maintenance is likely to be a cost saving option. However, if you are considering non-critical and easily replaced items (such as items that can be quickly rented at low cost), then the cost of investing in predictive maintenance capex and processes may not be worth it. In practice, most FM providers will opt for a mix of preventative and predictive maintenance depending on these kinds of considerations. What is important for clients is that their providers are able to explain the choices that they do make.