A Systematic Review on Smart and Predictive Maintenance in Tool Condition Monitoring

The main goal in the field of reliability and maintenance is ensuring and enhancing the availability of assets. A decrease in the production capability of machines can be the outcome of untimely and inefficient maintenance planning. Unexpected and unscheduled machinery shutdown due to required maintenance reflects poorly on a business, resulting in damaged credibility and financial losses. This puts organizations in a position to decide between undertaking preventive replacement of parts that could have been used for some more time or running the machine till it dies (run to failure). On the other hand, organizations can improve their uptime by promptly replacing potentially good parts that could have been used for some more cycles. In addition to assisting enterprises in minimizing or preventing unplanned downtime, smart and predictive maintenance (SPM) extends the machinery’s remaining useful life (RUL). A crucial instance is the cutting tool in machinery used for milling, drilling, or turning. It is an ideal asset to apply tool condition monitoring (TCM) since a breakdown of this part will result in unexpected downtime, resulting in a downturn in productivity. In a situation like this, a well-planned SPM strategy involving monitoring real-time health of tools used for cutting is beneficial. In the industrial predictive maintenance domain of Industry 5.0, accurate prediction of RUL of machinery is highly desired. Much research has been done on this topic, but none of it has covered all the techniques that have been used or have the potential to be used in the future. This study aims to support a comprehensive and methodical review of studies on the data-driven approach for estimating the RUL of cutting tools used in various computer numerical control (CNC) machining processes, including drilling, milling, and turning operations. This paper is a summary of various methods for monitoring, feature extraction techniques, decision-making models, and sensors currently available in this domain. A comparison of the accuracy of different prediction models used for estimating tool wear in TCM is also presented in this paper. The study concludes with a discussion of recent advances, challenges, and limitations in RUL prognostic methods that use artificial intelligence (AI), as well as the potential for further research in this domain.