Monitoring in precision metal drilling process using multi-sensors and neural network

Estimating high precision hole diameters of aerospace alloys using artificial intelligence systems: a comparative analysis of different techniques

“…So, a comparative analysis of the performance of three artificial intelligence techniques to estimate the diameter of holes machined by high-precision drilling in two metallic materials widely employed in the aircraft industry. This investigation also contributes to extend the work of Cruz et al [5], who employed multilayer perceptron artificial neural networks (MLP ANNs) to analyse the relationship between the signals obtained in monitoring the drilling process and the hole diameters obtained in the process.…”

“…This section discusses the implementation of artificial intelligence systems, and the methodology used to define their configurations. All the information about the tests and signal acquisition and processing was obtained from Cruz et al [5].…”

“…The drilling process was carried out in a Zocca U-30 universal milling machine that is driven by a 6 CV motor drive, a variable spindle speed from 45 to 2800 rpm and a variable feed rate ranging from 4 to 250 mm/min, with axis position control via HMI. The milling machine also has electronic rulers with micrometric precision in each axis, which was used to ensure that the coordinates of the holes in the Zocca workpiece specimen would be standardized [5].…”

“…2), which is fundamental for the performance of drilling; • The collected signals were analysed by computational routines in MATLAB ® to eliminate the portions before and after the cutting operation was performed when the tool was not in contact with the workpiece. The absolute average value of the region of each signal was calculated and used as an input for the ANN, in addition to the cutting speed and feed rate employed in the condition under analysis, as previously reported by Cruz et al [5]. Examples of a typical ANN input data can be found in the works carried out by Cruz et al [5] and Ferreiro et al [25] and Grzenda et al [26].…”

“…Drill wear is generally monitored using direct methods such as optical devices. To enable such inspection, the process must be stopped and both drill and machined part should be carried out to a specific gate in line production what increases production time and costs [5]. In addition, drill point is hidden inside the hole during operation and chip evacuation is hampered when drilling blind holes, what are unfavorable conditions compared to turning and milling operations.…”