Fault detection to optimize processes in tobacco industry

Manual handling allows machine operators to spot filling errors in the tobacco-processing industry. But when automatic solutions are used, sensors are required to perform this function. Hauni, the most successful supplier of machines, technologies and solutions for tobacco processing, filter and cigarette production, together with Omron, has now come up with a solution that is astonishingly simple.

Mission impossible?

Around 4,000 filter or cigarette rods fit into a tray, depending on the diameter. Viewed from above, the rods present a lattice-like honeycomb appearance. The layers are staggered by precisely the radius of one filter, thus resulting in a stable pack that can be handled without the rods falling out. In order to regulate the flow of filters or cigarettes, a COMFLEX unit handles the full trays, which are emptied by turning them upside down.

However, problems can occur when filling or emptying. If trays are overfilled, the rods may be damaged, which can have undesirable consequences since the process was developed to process geometrically uniform rods. If, on the other hand, the container is not full, the rods may become jumbled. And if rods in the top layer are not parallel with the others, they can cause blockages in the fill chute. If this occurs, the machine operator has to intervene manually and quickly. Failure to do so will stop the machine.

Production speeds of around 16,000 product rods per minute

With today’s production speeds of around 16,000 product rods per minute, four tray changes are required, leaving the machine operator only 7.5 seconds to change each tray, which is why automation of the feed and transfer processes is becoming increasingly important.

But automation requires robots that are able to recognize the above-mentioned problems. The usual solution is to provide optical-recognition sensors that furnish the entire automated system with the necessary information so that machine operators can sort out problem trays manually. But paper rip-off can obscure optical sensors.

Ingenious solution with ZX sensors solution

Instead of sophisticated (and expensive) camera systems, the solution employs two readily available laser probes (ZXLDA41 with ZX- LD100L) attached to the robot’s arm. These scan the clearance of the uppermost layer of filter rods while the robot’s arm grasps the tray. The scanned area is so wide that it takes in the entire height of the tray.

A simple processor (ZX-CAL) subtracts the clearances measured by the two sensors from each other. If the rods are lying correctly, the clearance signals should be exactly the same and the processor should register O, meaning everything is okay. A rod that isn’t perfectly parallel, however, will cause asymmetrical signals. The first sensor’s signal starts metering the number of motor impulses that occur when the robot’s arm transports the tray, while the second sensor’s signal stops this procedure. The number of pulses indicates the disorderliness of the top tray.

If a certain value is reached, the tray is earmarked for further processing. The clearance signal also serves to monitor how full the trays are, thus making it possible to monitor even empty trays.

A further solution (F160) in the HCF tray filter identifies areas (nests) in trays where the uniform, honeycombed structure is interrupted, while an additional sensor (E3C) on the robot’s arm detects product rods that are either bent or stick out of the container and could lead to a blockage or spoil the parallel arrangement.