Primary tabs

Predictive Maintenance: proven value in three steps

Predictive Maintenance is an upcoming discipline in which business intelligence, big data and manufacturing information processing meet. CGI organised an event to discuss possible applications of predictive maintenance and to share experiences.

The benefits of Predictive Maintenance for product manufacturing are huge, explains Henk van Haaster, Principal Big Data Consultant at CGI. ‘Nowadays, manufacturers are often forced to replace parts of a machine at fixed intervals, regardless whether the part is still functioning or not. This makes replacement very expensive. With big data applications, it is possible to accurately predict at what moment the part must be replaced. Thanks to this just in time maintenance, manufacturers can maximise the efficiency of their assets and reduce their costs dramatically.’

Exchange knowledge

Predictive Maintenance is a relatively new discipline, says Van Haaster. ‘But thanks to big data, it is developing very fast. We organised an event to exchange knowledge on how big data can be used for Predictive Maintenance. What is going on in the market and which steps have to be taken in information processing?’ Over thirty clients participated. Henk van Haaster quotes: ‘In a creative atmosphere, we discussed possible applications and exchanged ideas. Because clients from very different domains took part in the discussion it was very inspiring, both to our clients and to us.’

According to Van Haaster, the session was very useful. ‘We now have a better idea of how our clients view Predictive Maintenance. We now know what they regard as the most valuable application, which applications are relatively easy to realise and which applications would be more difficult to realise. Most of our clients appreciate the importance of Predictive Maintenance, but they do not know where to start. They immediately tend to think in terms of technical solutions, but that is not what predictive maintenance is about. It is about maximising asset value.’

Three stages

During the event, Van Haaster also explained CGI’s approach in this field. ‘We usually operate in three stages’, says Van Haaster. ‘We start with an ‘inspiration session’ about the goal of Predictive Maintenance and big data solutions. During a round-table discussion, we brainstorm about possible use cases for valuable application of Predictive Maintenance. The second step is a two-day deep dive for one of these use cases. In a multidisciplinary team, the use case is worked out to a functional solution. The multidisciplinary character of the team stimulates out-of-the-box thinking and an original solution to the client’s use case. The third step is a proof of value of this solution in CGI’s big data lab. A relevant part of the client’s data and possible other (external) sources are loaded into the lab and the value of the solution is assessed. The lab is fully equipped to handle data from all sorts of sources (e.g. OSIsoft and SAP). In no more than six weeks CGI’s three stage approach provides a client with a proof of the value of Predictive Maintenance for his assets.’ Currently, CGI is working on the third step for a Dutch water company. In three stages, we developed a solution for our client to predict and locate leakages. A perfect example of a specific (bespoke) solution for a specific problem.’

For more information about predictive maintenance, please watch this film (in Dutch).