**2. Ceramic Grinding Overview**

Ceramic components are susceptible to damages, such as cracks and residual stress due to their extreme hardness and high brittleness, which may affect the surface properties of the material [31]. Ceramic machining using diamond cutting tools is the primary technique for achieving specified dimensions and acceptable surface finish. Depending on the application of the component, the machining of the synthesized ceramics can represent more than 50% of the production cost, compared to 5% to 15% for metallic components. Among all machining processes, grinding represents more than 80% of all ceramic machining [32]. According to Brinksmeier et al. [33], the results of a grinding process can be subdivided into characteristics concerning the geometry and surface integrity of the workpiece. The essential macro geometric characteristics are dimension, shape, and waviness, while surface roughness is the main micro geometric characteristic. The surface integrity can be described by residual stresses, hardness, and material structure.

In Nascimento et al. [34], the viability of ceramic grinding with minimum quantity lubrication (MQL) with water was studied. A total of 45 grinding tests were performed with different MQL water concentration, depth of cut, and feed rate. The measured output variables were surface roughness, power, and scanning electron micrographs. The results show that MQL water–oil (1:1) was superior to conventional lubrication in terms of surface quality. Thomas et al. [35] proposed a new mathematical model to predict the surface roughness of ground ceramics. The effectiveness of this model was proved by the comparison of the experimental results with the predicted results. The author concluded that the optimization of the surface roughness can be done by controlling the grinding parameters. Liu et al. [36] studied the effects of the grinding parameters in the silicon nitride ceramic grinding. The influence of the grinding parameters, such as grain size, wheel speed, workpiece speed, and grinding depth, were analyzed regarding their effects on the grinding force, surface roughness, and subsurface damage. The ceramic grinding process can be optimized by the correct choice of grinding parameters. It is worth mentioning that the cited works used the surface roughness as the main variable in the evaluation of the quality of the workpiece surface.

The workpiece surface roughness is usually the most significant evaluation indicator in assessing the quality of ground surfaces. The evaluation of the competitiveness of the system can be performed through the estimation of the surface roughness, allowing productivity improvements and reducing costs [37]. The mean surface roughness (*Ra*) is defined as the arithmetic average of the absolute values of the deviations of the surface profile height from the mean line within the sampling length *l* [38]. Another tool used to evaluate the surface characteristics is the confocal microscopy. Thus, surface changes can be evaluated by reconstructing the topography of the surface from optical sections and light reflection [39]. In the ceramic grinding process, confocal microscopy can be found in the study of the cutting tool modeling [40] and the estimation of the maximum depth of cut during MQL grinding [41].

The monitoring of the ceramic grinding process by sensors has barely been studied, and only a few studies have been published. Nakai et al. [42] performed the estimation of the surface roughness of ceramic components through artificial neural networks (ANN). Four artificial models were tested with input statistics derived from the AE and power signals. The results obtained an accuracy of over 90%, which contributes to the automation of the ceramic grinding process. Feng et al. [43] monitored the cutting tool wear during ceramic micro-end grinding using AE, vibration, and force signals. The feasibility of monitoring the wear of the ceramic micro-grinding tool without knowing the machining characteristics was verified. Junior et al. [44,45] performed a feature extraction using the frequency domain spectrum and time-domain analysis of vibration signals to monitor the advanced ceramic grinding process. The statistics calculated by the application of digital filters in the chosen frequency bands were correlated with the workpiece surface roughness. Finally, the statistical analysis indicated similarity and confidence among the results presented.
