Predictive maintenance: reduce maintenance costs by predicting machine failures
Predictive maintenance: reduce maintenance costs by predicting machine failures
Predictive maintenance is a new way to schedule machine maintenance works. By monitoring the vibration pattern of machine parts, you can monitor the status of the machine. In this way, you know exactly when a component starts to show faults and can schedule the maintenance works accordingly.
MACHINE DOWNTIMES: EXPENSIVE AND AVOIDABLE
The costs connected to machine downtimes can hardly be overestimated. It concerns
- costs to repair the machine
- costs connected to delays in production
- costs of personnel that cannot do anything before the machine is repaired
It is obvious that the timely maintenance of machines to avoid failures is indispensable to any organisation.
PREVENTIVE OR PREDICTIVE MAINTENANCE: MAKE THE RIGHT CHOICE
PREVENTIVE MAINTENANCE
But when should you perform these maintenance works? It makes little sense to invest in maintenance when a machine is still in mint condition. Still, that is exactly the line of approach of traditional preventive or scheduled maintenance. In this approach, we assume that a component will age and thus start performing suboptimal after a given period of time. Every component has an expected life span and before that time has expired, the component in question is replaced. It is the same principle as changing the oil in your car after 10,000 km. Your car will not suddenly stop driving after 10.001 km but it is a sensible decision based on statistical, historical data.
PREDICTIVE MAINTENANCE
However, it is not always that simple in production environments. There are many variables that have an impact on the useful life span of machine parts, such as temperature or atmospheric humidity. These will be subject to variations and are difficult to assess. Besides, maintenance costs money. Persons responsible for a company’s machinery must therefore take strategic decisions and balance costs and benefits to make sure that the production environment continues to perform optimally as every machine breakdown will lead to soaring costs. So in fact it is best to service a machine before actual defects occur but after having observed dormant signals pointing at failing components. This is known as predictive maintenance.
Predictive maintenance is based on the current status or condition of the actual machine, in contrast to statistical data on the average useful life span of its components. As a result, useless maintenance works and corresponding costs can be avoided.
But how can you detect these early signs of an imminent failure if these signals are still dormant?
LISTEN TO THE MACHINE: VIBRATION ANALYSIS
To know whether or not a machine shows early signs of an imminent defect, Flanders Make has developed a measuring platform that is based on the vibration pattern of the machine. Machines constantly produce vibrations while they’re running. As long as a machine works normally, the pattern of these vibrations is normal and regular. As soon as there is a fault in the machine, no matter how small, this regular vibration pattern will start showing deviations. Thanks to sensors, the vibration pattern of machines is permanently monitored and the operator is immediately informed of deviations.
AMBIENT NOISE: APPLICABILITY OF VIBRATION MEASURING IN INDUSTRIAL ENVIRONMENTS
It is no doubt a fact that in an industrial environment there are a number of elements that could disturb the sensor signal, such as electromagnetic noise caused by the multitude of machines within a small space. This ambient noise is very specific for every production environment. When monitoring the vibration pattern of machines within the scope of a predictive maintenance approach, this should obviously be taken into account. To this purpose, the measuring platform that we developed makes use of algorithms that can be adapted to the environment. The measuring platform is set such that the measurements of the vibration pattern are very accurate. Only in this way, early prediction of machine maintenance is possible.