Towards Real-Time Machining Tool Failure Forecast Approach for Smart Manufacturing Systems

Abstract

Industry 4.0 is characterized by a dynamic market that constantly looking for new methods to optimize and integrate manufacturing processes. In this context, Artificial Intelligence has gained prominence in problem-solving, such as failure prediction and decision making, thus improving product quality, and consequently bringing competitiveness to the company. Aiming to contribute to this scenario, this research develops a data treatment system that, from an intelligent tool and an interoperable ontological model, automates the prediction and detection of failures in machining machines lines. The system was developed for the prediction of faults in machining lines includes an artificial intelligence formed from prediction algorithms and inferences, it is possible to guarantee the correct treatment and communication of data at different stages of the process. For the experimental research, was used data collected from a machining line of a public dataset. The information is collected and classified by Artificial Intelligence that supports a decision system. The prediction of tool wear would enable the system to infer the type of problem that is causing this wear, a possible root cause, and the needed maintenance based on the ontological inference tool. By this classification of data, it is possible to achieve, through inferences, a reduction in the decision scope, bringing the possible problems caused by the incoming value. The semantic interoperability ensures correct data exchange and processing, which generates a more assertive view of production failures. The system may help companies to increase their productive process by helping them identify future failures in production if applied in a real scenario.