Predictive maintenance of pneumatic systems with edge calculation

Pneumatic systems are actuators that provide movement by compressing the air and gaining pressure. This movement can be linear, angular or circular. Pneumatic systems are used in many areas in the industry due to its abundance in the atmosphere, easy to obtain and clean. The widespread use of pneumatic systems in the industry began in the 1950s, and today, while new factories are being built, electricity and water installations as well as pneumatic installations are made. In other words, pneumatic systems are generally used in almost all businesses that work on production in the industry. When we analyze the concepts of pneumatic systems and Industry 4.0 together, we see that there is a lot of data to be collected from the field and stored for analysis in order to increase system efficiency. At this point, while it is possible to analyze the data collected from the field with Cloud Computing in the cloud, processing the data collected from the field with the Edge Computing approach in the closest place to the source is another solution management. In this way, it is predicted that the data obtained from the production will be analyzed faster and the production efficiency will increase by getting a faster reaction. On the other hand, it is assumed that the operating costs will decrease by transferring only meaningful data to the Cloud. In this study, the general concepts of pneumatic systems are explained, and predictive maintenance approach with edge computing are handled as a real-life problem. The Fully Automatic Bread Processing and Arranging Line of Brotmas company operating in Konya and producing bakery machines was selected for the application. In the system, data has been collected for important issues such as Detection of Leaks in the Air Installation, Compressor Oil Level Control, Manometer Pressure Value Control, Compressor Belt Tension Control and Inverter Compressor Motor Health Control. By comparing the data collected on the production line with the newly obtained data, it was tried to prevent malfunctions that may occur in the system. As a result of the thesis study, it has been seen that the notifications of compressor off, compressor pressure drop, maintenance-free compressor and pneumatic piston failures are reduced by 67%.