Use Case: Introducing OCME

Based in Parma, Italy, OCME is an Original Equipment Manufacturer (OEM) operating in several different markets ranging from food and beverage to tissue and oil. Its products comprise filling machines, secondary packing units, palletizers as well as logistics with LGV.

Traditional packing machines are either built for a single product or a bundle of explicit products. They are typically efficient and reliable, but changing the product range is difficult and requires high engineering efforts. Due to shorter product life-cycles, both performance and flexibility became an important issue. Therefore, OCME is currently examining ways to improve the performance of modular lines, so-called compact lines. As illustrated in the figure below, the compact line consists of integrated modules which work with the efficiency of a single machine while producing as a completely synchronised line.

There are, however, some challenges that need to be tackled in order to benefit from compact lines. With reduced buffers between the modules, single modules have to interact in a synchronised way in order to be efficient. Due to the shorter buffers, energy can be saved while line efficiency, on the other hand, is strongly affected by failures of single modules which lead to whole lines having to be stopped almost immediately. It is therefore important to predict any machine wear in advance in order to enable the maintenance team to buy the needed spare parts, and, more importantly, to plan their substitution during the normal maintenance activities.

  • THE STUDY

    Taking this into consideration, OCME uses the project lifetime to focus on the wear prediction of the plastic film cutting module, which forms part of its shrink-wrapping machine named VEGA. In particular, the objective of the model learning algorithms is the cutting blade which physically cuts the plastic film from the unwinder reels at the right length to be wrapped around the producing packs.

    Although this is obviously not the only critical component of the machine, surveys conducted with final customers have revealed that it is considered a component perceived as a weak spot of the machine.

    This is mainly due to the following reasons:

    • The blade is, for safety reasons, situated in a mechanical cage which makes it invisible to exterior inspection
    • Visual evaluation of the blade life time it is almost impossible and it is also difficult to estimate the cutting quality of the plastic film
    • Placed under the main machine conveyor, the cutting unit is not easily accessible which affects not only the normal maintenance duties, but also the time required to replace the blade

    Moreover, OCME aims to have a reliable indicator for machine wear which doesn’t rely on additional sensors to be placed in the unit which, from a mechanical point of view, is already quite complex in itself.

  • THE SAVINGS

    The prediction for the blade replacement could create a financial benefit for the OCME customers. The savings are not related to the cost of the blade itself or to the fact that a completely worn blade could damage other machine components. The real advantage for the customers lies in the reduced production losses which are due to the ability of arranging the component replacement at the right time within the normal maintenance activities (which commonly take place on a weekly basis)

    One example: A classical soft drink production line which is working 24/7 with a production nominal speed of 65,000 bottles/hour can normally generate revenues of 0.05 euros per bottle for the customer. Assuming the worst case scenario of the blade replacement taking four hours, avoiding the action during the normal production process could, in fact, lead to savings of 13,000 euros.

    Savings = 65000 * 4 * 0.10 = 13,000 euros

  • THE IMPACT

    If these study results prove to be reliable, OCME is expecting to display this information on the machine HMI in order to provide their customers with an additional service. Moreover, the study is important to demonstrate the viability of such artificial intelligence techniques applied to real industrial cases. If successful, the same approach could be applied to other machine components whose replacement is now simply driven by end user experiences and predefined time tables.

Further information on OCME and its role in IMPROVE can be found here.