Nagayoshi Kasashima

Online Failure Detection in Face Milling Using Discrete Wavelet Transform

Abstract This paper describes an application of the Discrete Wavelet Transform (DWT) to detecting tool failures in face milling operations. The wavelet transform uses an analyzing wavelet function which is localized in both frequency and time to detect subtle time localized small changes in the input signals. In this paper, the DWT is used to detect tool failures such as small chipping and breakage of an insert tip including eccentricity of the tool rotation center. The results indicate that the DWT can extract tool failures with much greater sensitivity than the FFT even when the amount of chipping is very small. In addition, the DWT enables the analyst to determine which insert tip failed, since it yields time localized signal information. On-line diagnosis of tool failures are demonstrated in both simulated and actual cutting force signals by using simple pattern recognition technique.

Laser displacement meter application for milling diagnostics

This paper presents the application of a laser displacement meter for direct multi-purpose sensing of milling tool conditions. Using a laser displacement meter, a laser beam is projected onto the cutting tool and subsequently reflected. The intensity as well as the angle of the reflected beam are measured. The signals are interpreted for identification of tool geometry, tool whirling, or vibration. Signal processing and analysis depend on the application. A prototype system has been developed to demonstrate the feasibility of various applications, namely (1) tool setting evaluation, (2) in-process measurement of milling cutter geometry and detection of tool failure, (3) continuous monitoring of milling cutter deterioration, (4) detection and measurement of chatter in milling, (5) measurement of milling tool bending and (6) thermal expansion.

An in-process direct monitoring method for milling tool failures using a laser sensor

Abstract An in-process tool geometry measurement system for milling has been developed using a commercially available laser displacement sensor. The system reconstructs and displays the 3-D image of a milling tool and evaluates the tool geometric failures via a hybrid laser sensor measurement method which uses displacement and intensity techniques simultaneously. The system is designed to compensate for interruptions in the laser path brought about by chips and coolant drops present under actual cutting conditions. In experimental tests, the reconstructed 3-D geometrical tool image and the intensity measurements enabled us to detect the size and the location of chipped part and to determine the length of the flank wear to an accuracy of 40 μn.

Diagnosing Cutting Tool Conditions in Milling Using Wavelet Transform

Until recently, detection of chipping and other small-scale breakages in the milling process has presented formidable difficulties because of limitations in the traditional Fourier signal processing techniques. This paper explores the use of the Discrete Wavelet Transform (DWT) in detecting small tooth chipping in the milling process. We demonstrate that the DWT surpasses the FFT in recognizing the subtle, time-localized changes in the cutting force caused by chipping. The DWT uses wavelet functions which are localized in time and in frequency, in contrast to the sines and cosines of Fourier analysis which are localized in frequency but spread throughout the time spectrum. Like the FFT, the it can also be implemented using fast algorithms based on standard matrix methods.

Torque Control Method of an Electromagnetic Spherical Motor Using Torque Map

This paper describes a new method for driving electromagnetic type spherical motors that has arranged permanent magnets on a rotor and arranged electromagnets on a stator. The most significant feature of this method is, in principle, applicable to any permanent magnets arrangement and electromagnets arrangement, and it can generate torque to any direction from any rotor attitude. In the spherical motor, cogging torque between permanent magnets and electromagnets causes not only instability of rotation but also causes rotor attitude error because the rotor has three degrees of freedom. The method of canceling the cogging torque is introduced. In the case of arranging four or more electromagnets, the combination of the electric current supply becomes infinite in number and generally not determined uniquely. However, we have developed a method for deciding the electric current distribution that generates maximum torque under the limited supply of each electromagnet. In the experiment of open-loop control, it was confirmed that the rotor was able to rotate about the axes that had been set by the calculation.

Driving technique of electromagnets to rotate spherical motors based on torque map

This study develops a technique for rotation control of spherical motors driven by electromagnetic forces. The technique uses the data of generated torque per unit current and cogging torque of an electromagnet that are obtained beforehand at relative positions between the electromagnet and the rotor. An electromagnet placed on the stator of a spherical motor gives two kinds of rotation torques to the rotor. One is the torque by flowing current to the electromagnet. The other is the cogging torque that a permanent magnet on the rotor attracts the iron core of the electromagnet. The torque map expresses the two kinds of rotation torques when the unit current flows to an electromagnet in its various relative position to the rotor. In order to obtain the currents of electromagnets in the case that there are more than three electromagnets on the stator, this paper proposes a technique applying the Lagrange multiplier method under a constraint condition. This paper shows an example to obtain the currents of electromagnets by the proposed technique in the case that five electromagnets are arranged at the vertices of regular pentagon.

Grinding mode classification of ground ceramics using 2-dimensional wavelet transform

A new method for characterizing surface topography using a 2-dimensional discrete wavelet transform (2-D DWT) has been developed. The wavelet transform (WT) is the basis for a wide range of techniques applied in image processing and pattern recognition. Its main advantages over other feature extraction methods are the space-frequency localization, and the multi-resolution view of the frequency components of a signal. In this paper, an automatic grinding modes classification technique using a 2-D DWT is introduced, and comparisons of automatic grinding modes classification and human eye inspection are also examined.