The competitive nature of the modern manufacturing industry coupled with increasing consumer demand and a dynamic economy, lead manufacturers to operate their production computer numerical control (CNC) machines beyond their capable operational limits. This unsustainable behaviour leads to the rapid deterioration and breakdown of the CNC’s critical subsystems and components. Consequently, increasing the amount of unplanned downtime and cost of maintenance. To put it into perspective, downtime in the automotive manufacturing sector can reach upwards of USD 2 million per hour. The linear axis is one of the CNC’s critical subsystems, its robust and accurate positioning capabilities drive the operational and geometric performance of CNC machines. Failure to promptly address developed faults in the linear axis, may lead to poor geometric performance and system sluggishness, overall affecting the production quality of manufactured products. Therefore, it is crucial to monitor the health condition of this subsystem and its components. The principal aim of the presented research was to develop a condition-based monitoring (CBM) system to monitor the health condition of the linear axis and its various components. To achieve this aim, the research work was divided into four objectives. Firstly, objective one consisted of designing an industrial level linear axis testbed that resembles a CNCs linear axis to conduct all experiments. Furthermore, the testbed’s main contribution is to serve as a data generation platform for the research community. Objective two focused in the design of a framework to establish a reference baseline dataset for CBM systems. The framework’s contribution consisted of building an understanding of the healthy condition of the linear axis and its components using vibration monitoring and time domain statistical feature analysis. It was found that under a known healthy condition the time domain features exhibit low variability, there is a negligible difference between a forward and reverse stroke, and a robust baseline can be established by collecting data for approximately an hour of operation rather than a 6-hour operational shift. The third objective of the research consisted in conducting a comprehensive health assessment of the linear axis and its components through a multi-sensor approach, when a root-cause failure fault (FF) is present. Additionally, the health assessment’s contribution was further enhanced by analyzing the repair condition of these faults and comparing the results to the original baseline. The assessment demonstrated that the most frequently occurring root-cause FF, carrier block raceway blockage, can easily be detected through the system’s internal data. Moreover, the repair state condition exhibited less than 10% error when compared to the baseline state. Finally, the fourth objective was centered in the development and application of a novel signal segmentation technique to detect and localize the leading root-cause of failure in the linear axis, misalignment. The technique’s main contribution rests in its functionality as a localization and verification tool in linear axis maintenance. The findings from the conducted studies revealed that the technique increased the usability of time domain features such as RMS, by approximately double. Lastly, the evaluation of both stroke directions aided in localizing the misalignment in the linear axis. / Thesis / Doctor of Philosophy (PhD) / Computer numerical control (CNC) machines are pivotal assets for the manufacturing sector, yet the demand from the modern industry often drives the operation of these machines to exceed their capable operational limits. Consequently, causing the breakdowns of their critical subsystems, resulting in high maintenance costs and machine downtime. The linear axis is one of these critical subsystems, primarily tasked with the positioning function in the CNC machine. In the presented research, a condition-based monitoring (CBM) system was developed, to monitor the health state of this critical subsystem and its components. The development of the system first focused on understanding the healthy condition of a linear axis and its components by implementing a statistics-based baseline dataset framework. Then, a comprehensive health assessment was conducted to evaluate multiple data streams and detect signal variation when root cause failure faults (FFs) are present in the system. Unlike previous health assessments, the repair condition of the evaluated components was also included in the analysis. Finally, a novel signal segmentation algorithm was developed to complement the CBM system, specifically serving as a localization tool to pinpoint an artificially induced misalignment on the guide rail component of the linear axis.