An in-depth study on the BTA (Boring and Trepanning Association) deep hole drilling process.
This project involves analyzing time-series data from various sensors using advanced techniques such as Fast Fourier Transform (FFT) and periodograms. The goal is to develop a detection mechanism to identify early signs of issues like chattering and spiraling during the drilling process.
One of the most important applications of drilling is to produce extremely precise and smooth holes, such as axial bores in turbines or compressor shafts. However, drilling is a complex and dynamic process that can be affected by various disturbances, such as chattering and spiralling. Chattering is a self-excited vibration that occurs when the drill bit interacts with the workpiece material, resulting in irregular cutting forces and poor surface quality. Spiralling is a deviation of the drill bit from its intended path, resulting in a helical-shaped hole that does not match the desired geometry. Both of these phenomena can compromise the performance and safety of the drilled components, and cause damage to the tool or the machine. Therefore, it is important to analyze the machine’s behavior and detect the problem earlier, before it leads to catastrophic failure or costly rework.
This project utilizes data from 10 experiments that are part of Project C5 and are measured by the Institut für Spanende Fertigung, FB Maschinenbau, University of Dortmund. This data contains 7 sensors to measure different parameters of the machining process. The purpose of the analysis is to select the suitable sensor in order to provide an early warning before chattering or spiralling occurs. The experiments data is provided in two types of files: header files and data files. The header files are text files that contained information about the recording conditions. The data files are binary files that stored the A-to-D converted data from each recording, obtained through the GX-1 device. The first step of the analysis is to read the data files in a readable format, using the header files as a reference. Then, time series plots are generated to visualize the patterns of each sensor in an experiment and to identify changes.
A custom function is developed that analyzes the pattern changes in the "Moment" sensor and the dominant frequencies in the periodograms. The function compares the values of these features with a predefined threshold and flags the phase that exceeds it. This way, it can detect the problematic phases.
The given data is part of Project C5 and are measured by the Institut für Spanende Fertigung, FB Maschinenbau, Technical University of Dortmund.
- R Programming Language: Used for data analysis, visualization, and statistical modeling.
- RStudio: Integrated development environment (IDE) for R.
- R Libraries:
- tuneR: For handling audio data and waveform manipulation.
- Techniques Utilized:
- Fast Fourier Transform (FFT): Applied to transform time-domain signals into frequency-domain to analyze dominant frequencies.
- Periodograms: Used to visualize the power spectral density of the signals and detect patterns indicative of drilling issues.
- Custom Functions: Developed functions to read header and data files specific to the sensor data format.