Process automation means handing work that used to be done by a person over to a machine: a robot, a PLC, a camera or software. In manufacturing this usually means one of three things: the machine moves or processes a part, the machine inspects or measures, or the machine collects and passes on data. This article covers what that looks like in practice and where to start.
What to automate first
The best candidates are jobs that repeat every shift and follow a clear rule:
- Feeding and handling. Loading blanks into machines, moving boards, palletising. A robot works at the same pace across every shift.
- Quality inspection. A camera checks every part the same way and records everything, so manual logs are no longer needed.
- Measurement and sorting. Thickness, length, diameter, colour: vision measures, and the PLC directs the part to the right place.
- Data collection. Readings from meters, test benches and equipment go automatically into a spreadsheet or database.
Jobs where the task is different every time and there is no clear rule are less suitable. Partial automation often helps here: the machine does the heavy part, and a person makes the decision.
What an automated system consists of
- Control: PLC (e.g. Delta), servo drives, frequency inverters, sensors.
- Motion: industrial robots (KUKA, Yaskawa Motoman), collaborative robots (FAIRINO), linear axes.
- Vision: 2D cameras, 3D scanners (e.g. Photoneo), lighting and image processing software.
- Communication: industrial networks: EtherCAT, CANopen, Modbus, Profinet.
- Operator screen (HMI): start, stop, faults and statistics in plain language.
- Safety: guarding, light curtains, a safety controller. Motion is always permitted by the safety circuit, not by software.
How a project runs
- Task definition. What the part is, what the cycle time is, what happens before and after the operation, what tolerances apply.
- Trial. In more complex cases (especially with vision), we first test with your real parts. That way the risks become clear before any major investment.
- Design. Electrical schematics, control cabinet, robot cell layout, safety solutions.
- Programming. PLC, robot and vision programs, operator screen.
- Commissioning and training. Tuning in production, training operators and maintenance staff.
- Support. Remote assistance, changes when the product changes.
The most common mistakes
- Automating a disorganised process. If blanks arrive chaotically, the robot will need vision, or the feeding will have to be organised.
- Forgetting the operator. If the screen is hard to understand, operators will switch the system off and work by hand.
- Leaving safety until the end. Safety zones and stop logic must be part of the project from day one.
- No trial with real parts. A sample part and the real production flow often differ.
Examples from our work
- A KUKA robot picks a log from a pile based on a 3D camera image and feeds it into a machine.
- A machine vision system on a metrology bench reads meter displays every second and writes the results to a spreadsheet.
- Sorting logs by thickness in a sawmill using a camera.
More on the projects page. If you are considering automating a specific job, write or call us: we will discuss the task and tell you where it makes sense to start. Our services are described here.