Ruxu Du

Fault diagnosis using support vector machine with an application in sheet metal stamping operations

This paper presents a new method for fault diagnosis using a newly developed method, support vector machine (SVM). First, the basic theory of the SVM is briefly reviewed. Next, a fast implementation algorithm is given. Then the method is applied for the fault diagnosis in sheet metal stamping processes. According to the tests on two different examples, one is a simple blanking and the other is a progressive operation, the new method is very effective. In both cases, its success rate is over 96.5%. In comparison, the success rate of the popular artificial neural network (ANN) is just 93.3%. In addition, the new method requires only few training samples, which is an attractive feature for shop floor applications.

Modelling Machining Dynamics Including Damping in the Tool-Workpiece Interface

Machining instability, namely chatter, occurs due to the interaction between the structural dynamics and the cutting process. The process damping generated at the tool-workpiece interface is an important parameter of that interaction. A significant enhancement to the chatter simulation model in milling is presented. It includes tracking of the interference between the tool flank and the generated wavy surface, which is the source of process damping. Results of simulation runs performed to determine the limits of stability are presented for sharp tools as well as for tools with various amount of flank wear. The phase relationship between the ploughing force and tool vibrations is explained using these simulations. It is also shown that the improved model accurately predicts the increase in the limit of stability due to tool wear, as well as the effect of the wave length of the machined surface undulations on process damping. Cutting tests of aluminum confirmed the simulation results.

The Design of a New Metal Forming Press with Controllable Mechanism

Metal forming press is one of the most commonly used manufacturing machines. Every day millions of parts are produced by metal forming ranging from battery caps to automotive body panels. Therefore, even a small improvement may add to significant corporative gain. Currently, the metal forming presses can be divided into two categories: mechanical presses and hydraulic presses. The former is fast (high speed presses may reach up to several thousand shots per minute) and energy efficient (the large flywheel eases the impulsive force), but lacks flexibility. On the other hand, the hydraulic presses are flexible (their motions can be programmed) and accurate, but are expensive to build and to operate. Recently, there are mechanical presses driven by seryomotors. They could perform as flexible as hydraulic presses with high speed. Nevertheless, they are even more expensive to build and to operate. This paper introduces a new design of mechanical press whose performances are programmable, including the trajectory and the velocity of the stroke, and yet, it is relatively inexpensive to build and to operate. The key idea of the new design is a 2-degree-of-freedom seven-bar linkage mechanism driven by a large constant speed motor and a small servomotor. First, the kinetics and kinematics of the design are presented including the feasibility conditions, mechanical advantage, as well as the torque and power distribution between the two motors. Next, a number of simulation results are given. The design (parameter) optimization is also carried out using Genetic Algorithm (GA). Based on computer simulation, it is shown that the new design is indeed very attractive.

Fuzzy estimation of feed-cutting force from current measurement-a case study on intelligent tool wear condition monitoring

It is very important to use a reliable and inexpensive sensor to obtain useful information about manufacturing processing, such as cutting force for monitoring automated machining. In this paper, the feed-cutting force is estimated using inexpensive current sensors installed on the ac servomotor of a computerized numerical control (CNC) turning center, with the results applied to the intelligent tool wear monitoring system. The mathematical model is used to disclose the implicit dependency of feed-cutting force on feed-motor current and feed speed. Afterwards, a neuro-fuzzy network is used to identify the cutting force with current measurement only. This hybrid math-fuzzy approach will reduce the modeling uncertainty and measurement cost. Finally, the estimated cutting force is applied in the tool-wear monitoring process. Successful experiments demonstrate robustness and effectiveness of the suggested method in the wide range of tool-wear monitoring applications.

Hidden Markov Model based fault diagnosis for stamping processes

Metal stamping process plays a very important role in the modern manufacturing industry. Owing to an ever-increasing demand for better quality at reduced cost, a practical on-line monitoring and diagnosis system is of much appeal. However, the stamping process is a complicated transient process involving a large number of variables. It is rather difficult to monitor and diagnose by classical methods such as statistical classification. In this paper, a new method for fault detecting the stamping process is developed. First, it uses a number of autoregressive (AR) models to model the monitoring signal in different time periods of a stamping operation and uses the residues as the features. Then, it uses a Hidden Markov Model (HMM) for classification. The experiment results indicate that the new method is effective with a success rate between 80% and 90%.

Bispectral analysis for on-line monitoring of stamping operation

Abstract Owing to the ever-increased demand for product quality improvement and production cost reduction, on-line monitoring of stamping operations has become a common practice in shop floors around the world today. Based on a market survey, for most stamping processes, monitoring systems use tonnage signals and/or strain signals. Recently, a few attempts have been reported using acceleration signals as they contain rich information and are relatively inexpensive. However, it is known that acceleration signals are vulnerable to the noise disturbances, and hence are less robust. This paper presents a study that uses bispectrum to analyze the acceleration signals. It is shown that the bispectrum can suppress Gaussian color noise to boost the signal-to-noise ratio. It also extracts the features of the signal that are related to the defective parts (such as material too thick or slug). The experimental results demonstrate that the method presented is effective and has a good potential for applications in shop floor. We also present new method for reducing the computation load in the process.

Electrical modulus analysis on the Ni/CCTO/PVDF system near the percolation threshold

A type of Ni/CCTO/PVDF three-phase percolative composite was prepared, in which the filler content (volume fraction) of Ni and CCTO was set at 60?vol%. The dependence of permittivity, electrical modulus and ac conductivity on the concentration of Ni and CCTO fillers near the percolation threshold was investigated in detail. The permittivity of the composites dramatically increased as the Ni content approached 24?vol%. This unique physical mechanism was realized as the formation of conductive channels near the percolation threshold. Analysis on the electrical modulus showed that the conductive channels are governed by three relaxation processes induced by the fillers (Ni, CCTO) and PVDF matrix, which are the interfacial polarization derived from the interfaces between fillers (Ni, CCTO) and PVDF matrix, and the polarization of CCTO ceramic filler and PVDF matrix. The conductivity behaviour with various Ni loadings and temperature suggested that the transition from an insulating to a conducting state should be induced by charge tunnelling between Ni?Ni particles, Ni?CCTO fillers and Ni?PVDF matrix. These findings demonstrated that the tunnelling conduction in the composite can be attributed to the unique physical mechanism near the percolation threshold.

Predicting the Parts Weight in Plastic Injection Molding Using Least Squares Support Vector Regression

To achieve the desired quality in plastic injection molding, advanced monitoring techniques are often recommended in the workshop. Unfortunately, the signal in plastic injection modeling process such as nozzle pressure that is relevant to part quality is not easy to obtain because of the cost of sensors. The sensor-based modeling idea is therefore adopted. In this paper, a new method for predicting the parts weight in plastic injection molding using least squares support vector regression (LS-SVR) is proposed, which is composed of two steps. The first step is to estimate the nozzle pressure with the hydraulic system pressure using an LS-SVR model. The second step is to predict product weight using the estimated nozzle pressure, which is done using another LS-SVR model. The experimental results show that the new method is very effective.

Design and Analysis of a Biomimetic Wire-Driven Robot Arm

ABSTRACT Inspired by the octopus and snakes, we designed developand built a wire-driven serpentine robot arm. The robot arm is made of a number of rigid nodes connected by two sets of wires. The rigid nodes act as the backbone while the wires work as the muscle, which enables the 2 DOF [3]bending. The forward kinematics is derived using D-H method, while the inverse kinematics and its workspace can be solved by geometric analysis. To validate the design, a prototype is built. It is found that the positioning error of the robot arm is generally less than 2%. The advantage of this robot arm is that with several nodes fixed the rest nodes are still controllable. The positioning error is smaller when the fixed node is closer to the end effector. 1. INTRODUCTION Nowadays, robot arms are increasingly used both in industry and daily life. In general, robot arms There are mainly four types of artificial muscles: SMA (Shape could be divided into three categories: discrete, serpentine, and continuum [1][12]. Traditional sequential and parallel robots are of discrete type. They are made Ceramic)of several rigid links and joints, and can position accurately with large payload. However, they are usually big in size and have limited DOF, which constrain their performance in confined environment. Continuum robots are compact and theoretically have infinite DOF. As a result, thethey may be more suitable for confined environment. However, because of the lack of high rigidity, they are limited by its positioning accuracy and payload capacity. According to literatures, existing continuum robots are mainly in the ment stage with few practical applications [2]. Serpentine robots have higher rigidity than continuum robots and are more flexible than discrete robots. It has found many applications, such as endoscope in minimally invasive surgery .This paper presents our serpentine robot armIt is well known that traditional robots are driven by servo motors arranged at each joint, however nature works differently.For example, a human finger is made of three or four rigid bones and a number of flexible muscles and tendons [4]. As muscles can only produce tension to realize reciprocate motion, a group of muscles would work together: some extend whileothers contract. In the nature, such a bending mechanism is widely seen, such as human spines, fishes, snakes, etc. According to literatures, various artificial muscles have been developed for bending robots, continuum or serpentine. Memory Alloy) wires [5], EAP (Electro Active Polymer) [6][7], PAM (Pneumatic Artificial Muscle) [8], and PZT (Piezoelectric[9]. For SMA artificial muscles, as heating and cooling process is necessary, the response is slow. Besides, high temperature makes it unsuited for bio-related applications. For EAP artificial muscles, due to the low rigidity of the polymer ir payload capacity is rather limited. PAM artificial muscles, typically McKibben actuator, can produce large force and areeasy to control. However, they rely on air compressor system,

A CAD Approach for Designing Customized Shoe Last

AbstractThis paper presents a novel CAD framework for designing customized shoe last. The new approach consists of two parts. The first part is a global deformation approach that can automatically form a customized shoe last based on a scanned foot data and an existing shoe last data. It uses the modified distant map method giving the design being similar to the last and fit with the foot shape. The second part is an interactive local deformation method. It allows the shoe last be modified in specific regions to meet various needs. Preliminary simulation results show that the new approach is effective.