Tool replacement based on pattern recognition with LAD While traditional maintenance cost optimization is based on finding the reliability, and thus the probability of failure over time, in this paper, we show how to exploit condition monitoring data in machining operation in order to extract intelligent knowledge, and use this knowledge to determine the tool replacement time. This work is motivated by the increasing use of sensors in general, and specifically in condition monitoring. We show how the large volume of data that is now available in many industrial sites can give indications to the machining’s operator in order to replace the tool. We use a methodology called Logical Analysis of Data (LA D). This methodology enables us to extract meaningfulpatterns that describe the state of the tool’s wear, based on monitoring and measuring the cutting forces. Unlike the traditional experts’ rule-based methods, the extracted patterns are not based on experts’ opinion, but on information and hidden relations between the monitored forces. We apply our methodology on data obtained from experiments that are conducted in the laboratory. The experimental data are collected during a turning process of titanium metal matrix composites (TiMMCs). These are new generation of materials which have proven to be viable in various industrial fields such as biomedical and aerospace, and they are very expensive. In order to validate our methodology, we compare the results obtained when applying LAD to those obtained by using the well-known statistical Proportional Hazards Model (PHM). Findings and conclusion are given in the paper.