Fraunhofer scientists and industry partners are collaborating on the development of a new maintenance technology that makes it possible to predict machine downtime before it occurs. This enables plant managers to fix faults before the machine breaks down. The system even fixes some faults automatically.
Unplanned breakdowns in machinery during production, it is a thorn in the side of anyone working on the production floor. Such incidents cause unwanted production downtime, increased costs and, ultimately, perhaps even a worsened competitive position. Often, the problem is just a minor defect. However, if the defect is not fixed, it can lead to major breakdowns and downtime of production.
Predict
What would be useful is a diagnostic procedure capable of monitoring the status of all components in the production line, identifying problems and weaknesses and informing the responsible employee in time. The latter can then take timely action to repair the defect, ideally without interrupting production.
This is the underlying idea behind the ambitious SelSus (Self-Sustainable Manufacturing System) project currently being worked on at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA. "The aim is not only to monitor the status of machines and components," says Martin Kasperczyk of Fraunhofer IPA. "By using intelligent software and sensor networks, the plan is to detect potential weak points in time so that the system can predict possible failures."
The diagnostic models developed also provide immediate suggestions or recommendations on how to solve the problem. One of Fraunhofer's project partners is Italy's Electrolux. They already use a so-called 'decision support' system in their production. The system was found to be able to predict with a high probability potential failures in a press for washing machine housing.
Sensors
The data needed to monitor machine status is partly provided by sensors. These measure values such as energy consumption, temperature, oil pressure and vibrations. The system is also capable of sending control pulses to individual machines. For example, a welding control on which a sensor has not done its job can continue working seamlessly and safely, without serious interruptions. The project also gets its name from this ability to repair itself without halting production.
However, there are still some technical hurdles to overcome before SelSus can be deployed in factories. Kasperczyk: "One of the biggest challenges was analysing the huge flow of data. After all, we are talking here about accurately and reliably predicting machine failures and defects. You don't get there by simply programming a few algorithms."
Mathematical model
The experts rely on so-called probabilistic networks. A probabilistic network is a mathematical model that can be used to calculate the probability of a given event happening. The model represents a group of variables and their conditional dependencies. Using the data collected by the sensors, the software calculates, for example, the probability of a specific high-voltage cable breaking in the near future and, where applicable, signals that the cable should be replaced.
The SelSus software also takes into account the technical characteristics of the machine and its performance parameters. This data must be taken during installation and configuration of the system. In addition, a comprehensive test tells the system how the machine and its components behave when the machine is in continuous operation and under load. Only then is it ready for use. The software also records new data. e.g. as a result of machine upgrades allowing the system to learn by itself.
Repairing itself
A UK project partner of Fraunhofer has created a system with self-repairing properties. In a motor manufacturing plant, a dispenser is attached to a robotic arm by vacuum. When the dispenser encounters resistance, it reacts flexibly instead of breaking down. It loses the grip created under vacuum and sinks a few centimetres until it is stopped by springs. The springs then pull the dispenser back to its original position. Subsequent calibration ensures that the tool is in the correct position and the work process continues.
Besides collaborating with several major companies, SelSus also received funding of nearly €5.4 million from the European Commission.
By: Kelly Bakker
Source: Fraunhofer
