Smarter maintenance means cheaper maintenance

The fourth industrial revolution is in full swing: big data analytics, the internet of things, cloud computing, 3D printing, augmented reality… machines are increasingly being connected with each other and with the internet. As a result, you can collect remote data relating to the condition of a machine. However, this gives rise to the question: what can you do with this data? Together with several colleagues, Professor Robert Boute developed algorithms which help companies plan their machine maintenance more smartly. This also reduces maintenance costs and allows them to optimise their stock of spare parts.

Corrective or preventive? It depends on the condition!

Cars, hospital scanners, production machines, trains, nuclear reactors – if a component of these kinds of expensive machines breaks down, it generally makes good economic sense to repair the broken part rather than replace the whole machine. Companies will maintain expensive assets in order to extend their service life, preferably as cheaply as possible.

Corrective maintenance (CM) repairs the faulty component, whereas preventive maintenance (PM) takes place at set times and is not dependent on any damage. A significant disadvantage of CM is that you can't use your machine until it has been repaired. This is precisely why most companies opt for PM. The disadvantage is that you end up throwing away part of the expensive service life of a component, but you are doing so in order to avoid far more expensive downtime.

Collecting and analysing data about the condition of a device or component now allows you to predict more accurately when the device or monitored component will fail and plan your preventive maintenance more smartly: not too soon and not too late. Known as condition-based maintenance (CBM), the internet of things is now actually making this possible.

Two thresholds

With CBM, you therefore replace a component once a predetermined degradation threshold has been reached. There are two thresholds which make it possible to maximise the benefits of CBM:

  • Opportunity threshold: the degradation has reached a point at which PM would be beneficial as long as it can be combined with PM or CM for other components or machines, thus avoiding extra transport costs.
  • Intervention threshold: the degradation has reached the point at which maintenance can no longer be postponed if you wish to avoid a breakdown.

If you have a machine with various components of which one (the most critical) is monitored – or several of these machines – you will choose a first degradation threshold (the opportunity threshold) and a second threshold (the intervention threshold), which you fall back on if the opportunity to combine maintenance interventions has not arisen.

Optimal combination

It is important to select these thresholds in such a way as to minimise the average maintenance costs per operating hour.

Robert and his colleagues developed an algorithm which determines the optimal combination of threshold values taking into account the cost price of the component in question, the costs resulting from damage, transport costs and various other parameters such as the time required to plan extra maintenance, the progress of the degradation and the probability density function of the service life of the monitored component.

They worked on the basis of a scenario in which CBM is used for the monitored component of a machine and PM and CM for the other components. “At the moment, most companies are only using PM and CM,” explains Robert. “Switching over to CBM for all the machines and components is still a step too far. For the time being, it's better to restrict this to the critical components of your most important machines.” 

Transport costs versus downtime costs

The simulations showed that this kind of maintenance scenario can reduce costs significantly. This applies all the more to components which wear out slowly but have a highly unpredictable chance of failure and high costs resulting from damage to the component.

“The algorithm also shows that while it is appropriate to work with two thresholds in some cases, in others one is enough,” says Robert. “If the transport costs are high, it is mainly the opportunity threshold which is useful whereas the intervention threshold is important in situations involving high breakdown costs, certainly given low transport costs. For components with a long service life, the economic value of the opportunity threshold is higher and for components with a short service life, the value of the intervention threshold is higher.”

Business case for digital technology

Although the algorithm was developed and validated for an OEM which specialises in compressed air tools and compressors and the parameters were therefore developed specifically for this OEM, given the necessary modifications it could also be used in other situations. Would every organisation benefit from CBM? “The economic value is higher for some organisations than others and our algorithm allows you to quantify this value,” replies Robert.

“Not only does the algorithm help companies to plan their maintenance in a smarter way, it also provides an answer to the question ‘should we invest in these new digital technologies, such as the internet of things’? After all, connecting machines with each other and with the internet, monitoring components and using the data... none of it is free.”

The maintenance strategy determines the demand

It will be obvious that your maintenance strategy affects the required stock of spare parts and your stock management:

  • Corrective maintenance (CM): you have no idea when a component will fail and will therefore need to be repaired. The demand for spare parts is virtually unpredictable as a result, which makes stock management extremely difficult.
  • Preventive maintenance (PM): the demand for spare parts is perfectly predictable which makes stock management much easier, apart from the demand for unexpected interventions for CM.
  • Condition-based maintenance (CBM): the demand for spare parts is less predictable than in the case of PM, but because you are monitoring the degradation you have a rough idea of when the component will fail.

With PM and CBM, because you are actually replacing the component early, i.e. before it is broken, the demand and therefore also the stock of spare parts will be higher than in the case of CM. Simulations showed that the stock costs for PM are an average of 20 to 84% higher than for CM, whereas the stock costs for CBM remain limited to less than 20%.

All lumped together

Robert: “But what really happens in practice? Even if people try to predict the use of spare parts in one way or another, they hardly ever record the reason why spare parts are being used. The maintenance strategy is not taken into account at all, although this definitely affects the demand and therefore the stocks required. And of course it doesn't help that two separate departments are usually responsible for maintenance and stock management.”

“All the same, if a part is required for PM or CBM then you will be aware of this in advance. It would therefore be interesting to use this advanced demand information (ADI) for your stock management.”

The demand determines the stocks

On the basis of an algorithm developed for the same OEM, Robert and his colleagues determined the impact of the use of the advanced demand information (ADI) on various maintenance strategies, taking into account all kinds of different factors such as installed base, the progress of the degradation, maintenance costs and downtime costs. In their simulations, they optimised both the stock management and the maintenance strategy (length of the maintenance interval and the intervention threshold for CBM).

They demonstrated that the use of advanced demand information reduces or even eliminates the disadvantages of PM and CBM, namely the higher consumption of spare parts: “According to our simulations, ADI allows you to reduce your stock costs to 15% for PM, although they are still higher than in the case of CM. But if you use CBM, the use of ADI leads to savings of 14%. To put it even more strongly, the stock costs can even be lower than in the case of CM.”

Breaking down silos

This algorithm is also perfectly applicable in other organisations. “It shows how you can use data from one department to help the other to work more cost-efficiently – you break down the silos. And even more importantly, you can reduce your stock costs with no reduction in service levels.”

He concludes: “It should be clear that it is well worth optimising both your maintenance strategy and the management of your stock of spare parts. Our algorithms help to create a business case for both scenarios.”

Source: The papers ‘A hybrid condition-based maintenance policy for continuously monitored components with two degradation thresholds’ and ‘Numerical study of inventory management under various maintenance policies’ were published in renowned scientific journals. They can be requested from the authors.

  • Poppe J, Boute R, Lambrecht MR (2018). A hybrid condition-based maintenance policy for continuously monitored components with two degradation thresholds. European Journal of Operational Research, Vol. 268, Issue 2, pp. 515-532.
  • Poppe J, Basten R, Boute R, Lambrecht MR (2017). Numerical study of inventory management under various maintenance policies. Reliability Engineering & System Safety, Vol. 168, December, pp. 262-273

About the authors:
Robert Boute is a Full Professor in Operations Management at Vlerick Business School and the Faculty of Business and Economics at the University of Leuven. Rob Basten is an Associate Professor at the School of Industrial Engineering at Eindhoven University of Technology. Marc Lambrecht is a Professor (Emeritus) of Operations Management at the Faculty of Business and Economics at the University of Leuven. Joeri Poppe is a doctoral researcher who is affiliated with the Research Center for Operations Management at the Faculty of Business and Economics at the University of Leuven.

Discover our expertise in digital transformation

Want to know more on what the digital future of your company could look like? Not everything you read on digital transformation can be taken for granted. Let us guide you through the clutter.

& Rankings

Equis Association of MBAs AACSB Financial Times