Challenge:
Manufacturing Image Processing
Introduction
As the result of this challenge we want an application to be realized, that analyzes a video image stream of an existing camera next to the cutting insert of a horizontal turning lathe. The insert including the cutting process itself is captured from above, close to the tool. The processing software should run on a Raspberry Pi device or similar. Input of the device will be the stream itself. The expected output of the device is a single signal on each cutting tool insert change and a repeating signal for the current chipping shape. The signal is sent to a server via MQTT protocol.
Problem
- Documenting cutting insert changes on heavy machining is so far a manual process and suffers on time accuracy.
- Defining the chipping shape (flow chip, discontinuous chip, etc.) is a manual process as well, and based on subjective clustering.
- The actual chipping shapes of the process are not documented yet.
Keywords
Challenge Owner
Siemens Energy is focused on helping customers navigate the world’s most pressing energy problems, both for today and tomorrow. Essential applications include providing products, solutions and services that make fossil energy greener; delivering decentralized, flexible power solutions; managing the complexities of the grid; improving and de-risking aging assets; and connecting supply and demand through storage technologies such as grid-scale batteries and Power-to-X technologies.
