Innovation Boosting for Deceuninck

Innovation Boosting for Deceuninck

As a profile manufacturer, Deceuninck hardly needs to be introduced. When Benari Deceuninck started producing buttons, buckles and combs from plastic sheets as a small one-man business in 1937, he undoubtedly did not dream that the company would grow into the multinational we know today. Meanwhile, the company operates production plants in 17 countries, employs some 4,000 people worldwide and can be counted among the top three PVC window profile producers overall. This makes them the local leader in the design, development and extrusion of plastic systems and profiles for the construction industry and their recycling.

Like all companies active in Flanders, Deceuninck must also take into account the high cost of labour and a shortage of suitable workers. An additional trend in building materials is the increasing complexity and growing diversity of the products. In the production environment, this translates into more and smaller batches, so shorter production runs.

That is why Deceuninck has been working for years on process optimization to optimize productivity and cost efficiency without having to sacrifice flexibility. This flexibility is mainly needed to be able to produce smaller quantities of a certain profile within a short period of time.

Reducing changeover times

For example, the LEAN methodology was applied years ago to reduce changeover times, a major cost driver in profile manufacturing. They have also been working together with external specialist consultancies for some time now, for example by SMED (single minute exchange of die), to keep changeover times as short as possible. Other, more innovative projects have also been launched in recent years. For example, there is currently an initiative with support from VLAIO to convert existing machines in order to accomplish further mechanical and process optimizations.

An important next step that Deceuninck has in mind is the further optimization of the planning as well as an even greater focus on the transition from a way of working based on experience to a way of working based on knowledge. For example, in the future, Deceuninck wants to group profiles better on the basis of formalized product and process knowledge. This grouping allows for better selection of the follow-up of profiles in the delivery line, which reduces the total changeover time. At the moment this is already done on the basis of the (subjective) judgment of a number of experienced employees, which also involves a risk with regard to continuity. By doing this on the basis of product data and according to a dynamically formalized and computerized process, Deceuninck can further increase the ratio productivity/time of delivery without having to experience the limitations of a rigid planning process.

Innovation Boosting project

In order to test the feasibility of this new way of working, Flanders Make was asked to carry out a short preliminary study (Innovation Boosting) to investigate how a data-driven approach (e.g. by assigning parameters to setting positions and taking this as the starting point for drawing up shape sequences) could result in shorter and more efficient set-ups without compromising flexibility.

Flanders Make has helped Deceuninck with these challenges by programming two concrete building blocks. A first building block is an algorithm to determine the optimal profile sequence for the crowning line. Based on measured or estimated changeover time and a desired collection of profiles, an algorithm determines the profile order that causes the lowest changeover time. A basic requirement for a good profile sequence is that changeover times are known or can be estimated well, which is not always the case. A second building block supports this: a piece of software that analyzes production drawings (2D DXF) of the forming operation and, in a matter of seconds, extracts distinguishing characteristics that are directly related to the forming. For example, the software takes into account the area to be adhered, the symmetry and the number of operations in the process. These properties allow to determine which profiles are similar to each other for the concrete process of adhesion. The degree of similarity between profiles is useful information for estimating the time required for conversions that do not take place often. Even conversions that have never taken place, such as with new profiles, can be estimated this way.

At the jaw line, the decorative film is pressed onto the profile. The photos illustrate the complexity of positioning the wheels to achieve good adhesion.

Decueninck clamping wheels

Insights from Deceuninck

Deceuninck has recently started various projects that should allow us to process the increasing complexity and diversity of PVC profiles more efficiently. The focus here is on a data-driven approach, while at the same time also looking for ways to secure the valuable knowledge and experience of our employees and to include them in future optimizations. In concrete terms, the collaboration has resulted in the necessary tools to optimize the sequence of profiles in production, taking into account determining factors that are increasingly well mapped out (such as changeover times, but also product geometries and machine properties).

The algorithm provides a first step towards the optimization of profile sequences in production, which reduces setup times and thus increases active production time, but also provides insights that can be used to redesign production lines and group settings. These optimizations also create room for the employees to follow up the production processes more smoothly, and optimize them further where necessary. Currently, research is being done to test the adapted profile sequences for a selection of product families and production lines, and to smoothly secure the related data and information. In this sense, the Innovation Boosting project is certainly not the end point of the collaboration; the data-driven approach will be further expanded to include new product families and guarantee flexibility in a rapidly changing production environment.

More about Innovation Boosting

Innovation Boosting is an initiative of Flanders Make with the support of VLAIO (Flanders Innovation & Entrepreneurship) to support companies with an establishment in Flanders in developing their innovative ideas. Via a feasibility study, Flanders Make helps companies to obtain better founded insights into the possibilities and feasibility of an innovation envisioned by the customer. The focus of such a short trajectory is on preparatory research and possibly on testing. In a later follow-up trajectory, the idea can be further developed into a working product or process.

More information?

Would you like to learn more about the Innovation Boosting initiative and what it can do for your business? Don't hesitate to contact us!

Jan Stroobants, Research engineer

Jan holds a MSc degree in mechanical engineering from KU Leuven (2008). In his current position at Flanders Make as senior research engineer, he has led a research project on multi-material topology optimisation and he is active in simulation & validation in the context of lightweighting and structural optimisation and in developing tools that give designers more insight during the early design phase.