Flexible assembly: production in motion

Mass production has historically not been the most flexible way of manufacturing. Production lines were set up to build a single product and switching to another product required a lot of effort. However, products are increasingly customised as customers increasingly expect personalisation. This is the case with consumers, who expect their products in multiple colours, sizes and finishes. But also in the corporate world, there is a high demand for specialised products, such as air compressors, looms and feed harvesters, products produced in small batches. This evolution makes flexible assembly increasingly important in industry and is therefore at the forefront of Flanders Make's agenda. Our new site in Kortrijk will focus even more on this trend.

To make specialised and personalised products at mass production prices, we need a high degree of flexibility but still at high productivity. This puts us at the fault line of flexible, human labour and fast, automated production lines. So to achieve our goal, we need to improve either human productivity or the flexibility of automated production lines. Research at Flanders Make naturally focuses on both avenues and on hybrid solutions where people work together with automated production lines.

To demonstrate these concepts in an industrially relevant context, we built a robotised flexible assembly cell, which we named Infraflex. With this, we incorporate a whole range of concepts around flexible assembly in a very small area. For example, the central robotic arm can be quickly programmed with a recipe that combines a range of skills into a task. Around the robot arm there are also a series of add-ons that take over small tasks, prepare kitting or prepare new parts. The setup shows concrete situations where the cell assembles a compressor, for example, but temporarily pauses this task for a rush order, preparing new parts for a completely different assembly process. Afterwards, the cell can resume its original task.

When we talk about the automated assembly of kits, many naturally raise the legitimate concern around the picking up of these parts. After all, a kit can contain very different parts, from a heavy electric motor to the tiniest of screws. If we want to move these in an automated way, it has to be done with the right gripper and each part has to be gripped in a specific way. However, a specific gripper for each part is not possible, so there is a need for flexible grippers that can grip a whole range of parts.

First of all, we have to make an important choice between some types of grippers. For example, there are mechanical grippers with moving fingers that grab objects, as well as grippers that function with electromagnets or vacuum suction cups. Depending on the gripper type chosen, we then look at the optimal gripping points to grab each object. These can be sides of the object that grippers can clamp, but also surfaces that suction cups or electromagnets can clamp to. Finally, we create a final design for the gripper, further adapting the gripper to the objects if necessary. This can be done, for example, by adjusting the length of a gripper's fingers, but also by correctly aligning suction cups. Flexibility is also possible here by making grippers interchangeable or adaptable so that adjustments in production can also be smoothly accommodated here.

Of course, everything we have discussed so far mainly applies to smaller products and components. Therefore, the next challenge for Flanders Make is to make the leap towards larger products, such as trucks or looms. When we work at that scale, we can no longer speak of supplying small parts to a robot arm, but rather, in many cases, move the robot arm to the product. That brings a lot of additional challenges, such as keeping a shop floor safe where employees also walk around or accurately determining where all the parts and machines are in a larger factory hall.

To explore this further, we are currently developing a new lab in our new building in Kortrijk, where we are taking the concepts of this blog post to a larger scale. Mobile robots, automated cranes and multiple assembly cells working together form the basis for this. But certainly also the software and data that will link it all together.