Software Tool Maximizes Overall Equipment Effectiveness
Increasing networking in production also increases its complexity. Expertise in data analysis is required in order to predict system failures at an early stage and to identify the causes of loss of effectiveness. With MOEE, research teams from the Fraunhofer Institute for Manufacturing Engineering and Automation IPA recently presented a tool at the Hanover Fair that identifies the causes of productivity losses in linked systems and enables malfunctions to be eliminated quickly.
The overall system effectiveness (OEE) is a decisive indicator in production. It defines the percentage at which a plant produces quality products at a given speed. At the same time, it forms the basis for improving processes by identifying production losses. With MOEE, short for Maximize Overall Equipment Effectiveness, researchers at Fraunhofer IPA in Stuttgart have developed a software tool that detects production losses in relation to the three parameters of performance, quality and availability in complex, networked, automated systems.
Implemented algorithms automatically analyze the system’s behavior and then create an individual process model. The various process steps of a production cycle are visualized and assessed. “The algorithms calculate which processes take place when and in which order and how long they take. If process steps do not take place at the desired speed and if they are not optimally coordinated, this says something about the performance. »Short-term stops by the robots, for example, are usually not recognized, and their effects are difficult to quantify. If several such stops add up, however, this leads to errors”, explains Brandon Sai, head of the ‘Autonomous Production Optimization’ group at Fraunhofer IPA, using an example to explain how the software works. If machines were to stand still, this would say something about availability, another criterion for insufficient system effectiveness. In addition, the specially developed, self-learning algorithms provide information on the quality achieved. The aim is to assign the identified losses to the components and thus to identify the specific weak points.
Automatic Process Modeling Combined with Machine Learning Processes
Overly calculated safety buffers are a frequent cause of malfunctions. Inaudible to the naked eye, MOEE records minimal downtime as well as dynamic bottlenecks – caused by a production backlog. Malfunctions such as the jamming of a machine component or the inadequate application of a layer of grease are also registered as the software encodes each state in detail. “Using a combination of automatic process model creation and machine learning methods, we recognize productivity losses when they occur and thus contribute to the rapid elimination of the fault,” says the engineer. The worker should not be burdened with all this information, he is only notified directly in the event of problems.
Detect Productivity Losses at the Signal Level
MOEE uses the I/O interface of the controller for the analyzes. “The input / output interface is the brain of the machine. The system is monitored directly on the controller. Here, the behavior can be recorded in an optimal and fine-grained manner”, says Sai. This makes it possible to assign productivity losses at the signal level, increase availability, increase performance and identify quality deviations. Loss of performance and quality can be recorded down to the component level – for example a valve.
MOEE is already in use in manufacturing companies.
For more information: www.fraunhofer.de
