In managing a product development project, one of the primary objectives is to minimize the gap between the resultant product quality and its predetermined target under the given due date constraint. Achieving this objective is not an easy task mainly because the fact that, due to uncertainties in the duration and the output of each activity, the project's lead-time and the product quality remain uncertain until the project is completed. In order to facilitate this task, the authors propose a new approach, called schedule FMEA. It classifies the possible outcome of each activity into several categories of results; success and one or some failure mode(s), and prepares a corrective action for each failure mode. If a failure mode is observed in running the project, the predetermined corrective action will be taken. In this article, only the truncation of each activity is considered as the optional corrective actions. In this case, the result of each activity is defined as success when it is completed with a shorter duration than the preset limit, otherwise it is regarded as failure and the corrective action will be taken, that is, the activity will be truncated. The problem of setting a limit to the duration of each activity is formulated as a dynamic decision making process, and a heuristic algorithm is developed to this problem. Some numerical examples are given to illustrate the power of the proposed approach.

抄録全体を表示

It is a difficult task to construct a product model that is a basis and a requisite for managing information about products totally and effectively through their life cycle. ISO 10303 or STEP that is an international standard for the computer-interpretable representation and exchange of product data takes a two-step approach that is, first to model application requirements and generic representation separately and then to make mapping between these two models. However, reviewing Application Protocol standards of STEP, we have found issues probably due to the insufficient mapping methodology. To remove deficits caused by the complications of mapping, we have developed a software tool that helps users who make mapping with graphics. The tool is developed based on mapping model that represents detailed mapping relation between models. In STEP, constraints can be added to the model using EXPRESS formal description language to make the two models consistent. We also have found the basic nature of the language disturb the correct description and understanding of constraints. To solve this problem, we propose a language dedicated for the correct mapping of cardinality that generates constraint description in EXPRESS. With this language, cardinality conditions and aggregate information can be described separately and explicitly.

抄録全体を表示

This paper reports on simulation results of multiple AGVs transportation in a virtual factory. AGVs drive on their ways autonomously. To do so, AGV learns paths between machining cells, and automated warehouse and machining cells by use of Q-learning before simulation. The fast learning methods, which take a consideration in failure learning, are proposed and the efficient learning method is found out. After the paths are learnt, transportation simulation for workpieces by autonomous AGV driving is performed. Results show that AGVs can transport the workpieces from the warehouse to machining cells and transport them back to the warehouse without any pre-programming and instructions, if enough number of AGVs are set in the factory. However, it is observed that deadlock phenomenon is observed very few times during transportation.

抄録全体を表示

An ultra-precision method of joining optical lenses in the cell has been developed by using a shrink fitter, in order to produce a laser scanner having a wide field of vision and high resolution. In a current method, optical axes of the lenses in the cell do not coincide with each other, when environmental temperature changes, because of great difference in thermal expansion coefficients between the lenses and the cell. Therefore, the laser spot cannot be converged to a designed size on an image plane. A shrink fitter, which has a higher thermal expansion coefficient than the lenses and the cell do, was inserted between them. The positioning accuracy of the lenses in the cell did not change at any temperature since the shrink fitter sufficiently expanded or contracted in the radial direction to compensate the change in interference. As a result, the lens unit assembled by the new method worked out to converge the laser beam over a wide scanning width. In this paper, we examined the shrink fitter theoretically and experimentally.

抄録全体を表示

We investigate the defect detection by making use of pre-made reference image data and classify the defects by using the artificial neural network. The approach is composed of three main parts. The first step consists of a proper generation of two references image data by using a low-level morphological technique. The second step proceeds by performing three times logical bit operations between two ready-made reference images and just captured image to be tested. In the third step, by extracting four features from each detected defect, followed by assigning them into the input nodes of an already trained artificial neural network we can obtain a defect class corresponding to the features. Experimental results show that proposed algorithms are found to be effective for flexible defect detection, robust classification and high-speed process by adopting a simple logic operation.

抄録全体を表示

The aim of this research is to develop an efficient and stable dynamic analysis approach for multibody systems. In this first report, an efficient extension of Rosenthal's order-n algorithm, which is originally limited for only multibody systems with open-loop topology, to multibody systems containing closed-loops is presented. Relative coordinates are employed for modeling systems. Closed topological loops are handled by cut joint technique. Equations of motion and cut joint constraint equations establish a set of differential-algebraic equation (DAE). Then, Rosenthal's order-n algorithm is extended to solve this DAE. Dynamic analysis of a multibody system is carried out to demonstrate efficiency of the proposed approach. Finally, number of arithmetic operations required by the proposed method and other well-known recursive formulations for analyzing a single closed-loop multibody system are counted in order to examine the efficiency of the proposed method compared to other well-known recursive methods

抄録全体を表示

This paper describes a high-precision free-form measurement machine named "A-Ruler." It will be used to measure the shape of freely curved surfaces with a precision of 5 nmRMS at a measurement volume of 350×80×50mm³. A contact-type probe guided by a pair of parallel springs is employed for its capacity to easily measure oblique surfaces. The probe's five motion errors are measured and compensated for using six laser interferometers and three mirrors arranged perpendicularly to each other. This is the core structure of the device, and it ensures accuracy of measurement. On the other hand, other portions such as XYZ-slides are made with rather ordinary components. Studies on margin of error indicate that the measurement of a spherical surface is influenced by almost all types of errors. Therefore, the performance of A-Ruler can be assessed by measuring a spherical surface. A concave sphere work-piece made of Zerodur^TM was prepared for this purpose. Its curvature radius is 104.508 mm, the evaluation area is φ120 mm and the maximum angle of obliqueness is 35 degrees. The repeatability, or the difference between two consecutive measurements, is 3.8 nmRMS. The error of the probe's sphericity is found to be sufficiently small compared to this repeatability. Precision is assessed by comparison with an interferometer. The difference between A-Ruler and the interferometer is 3.5 nmRMS per φ120mm area.

抄録全体を表示

Most plastic surfaces are textured to suppress specular reflection and to enhance their aesthetic quality. Those surface textures are generated in the injection molding process. First of all in this paper, some significant 3-D geo-stastistical parameters for assessing the textured surfaces were selected and their measuring conditions were introduced. Geometrical transfer from the mold to the plastic surfaces could be evaluated by the parameter values. Experimental results and multi-variate analysis showed that reflection luminance distribution of the isotropic surface texture can be well correlated with the 3-D local slope distribution of the facets and some aesthetic indices of visual inspection can be expressed by a combination of those parameters.

抄録全体を表示

More than 80% of the pattern defects in a semiconductor manufacturing process are caused by the adhesion of particles on a wafer. In a factory, the particles larger than 0.1 μ m in diameter on the wafer on which a Si0₂ thin film was deposited, have to be inspected. In this study, a system for detecting the light intensity scattered by particles 0.1-0.2, μ m in diameter on the wafer on which a Si0₂ thin film was deposited, without being influenced by the thickness variation of the thin films, is developed by simulation. The following results were obtained from the simulation. (1) P-polarized laser light with the incident angle of 75°is suitable for the illumination. (2) A method of detecting the light scattered perpendicularly upward onto the wafer is suitable for the detection. (3) The optimum numerical aperture of the detection lens is approximately 0.6. The variation of the detected light intensity is 61% when the light scattered by a particle 0.1 μ m in diameter on a wafer on which a 0-300-nm-thick Si0₂ thin film was deposited, is detected. The correlation factor between the scattered light intensity detected by a photo-detector and the standing wave intensity on the Si0₂ thin film is greater than 0.7. The good correlation between these two intensities is verified.

抄録全体を表示

SCARA robots are used in the assembly tasks because of its excellent characteristics. The repeatability of the assembly robot is about 0.02mm in average but the absolute positioning accuracy is not so good. The lack of absolute positioning ability causes the obstruction of the effective off line teaching by CAD. To improve the absolute positioning accuracy, machine vision system especially stereo camera system is useful. In this study, the 3D error model of the SCARA type robot is proposed and calibration method of the error parameters using stereo CCD camera measurement is presented. The positioning accuracy is improved up to 0.12mm in the root mean square and less than 0.20mm in maximum value.

抄録全体を表示

This paper presents link layout optimization for coordinate measuring machine (CMM) based on a spatial three or six degree-of-freedom (DOF) parallel mechanism. The optimization using genetic algorithm (GA) minimizes the measurement error caused by all joint runouts in whole measuring space. Moreover the optimization improves the measuring resolution for the scale resolution better than the critrion (<1). The measurement error applied to the objective function is also calculated using GA so the directions of joint runouts maximizing measuring error are showed and consequently the measurement error can be calculated simply. As the result of optimization, the obtained optimal link layouts for three and six DOF mechanism reduce the measurement error to 62% and 83% of former mechanisms respectively. Moreover the measuring resolutions also have been improved.

抄録全体を表示

For the purpose of realizing not only positioning but also orientating accuracy of a robot, a position and orientation calibration method is newly proposed and applied to a 6-DOF articulated robot (PUMA 560). The existence of gear transmission errors and gear backlashes of the 4-6th axes of this robot is detected by using a high precision gyroscope and it is proved that these non-geometric errors cause the orientation error of robot arm's tip. A position and orientation measuring system using both a laser tracking system and this gyroscope is proposed. A calibration method, which uses both position and orientation data and considers non-geometric errors of 4-6th axes as parameters to be calibrated in the forward kinematics model, is proposed. The positioning accuracy of 0.2mm and roll, pitch, yaw orientating accuracy of (0.17°, 0.10°, 0.10°), which is near the limit value of calibration predicted by the orientating repeatability, are successfully achieved and it is experimentally proved that the proposed measuring system and the calibration method are very effective for improving the orientating accuracy while keeping the high positioning accuracy. The optimal weights of position and orientation data in the least squares estimation process of parameters are also discussed.

抄録全体を表示

A simple and economical ultrasonic measurement system for 3-D orientation of a moving object was developed. In this system three transmitters are attached to the object and their absolute positions related to the coordinate system, which is based on receivers' positions, are calculated by using the measured transit times of ultrasonic pulses. The absolute orientation is calculated from these three positions. An electric spark and a condenser microphone of a small diameter were adopted as the transmitter and the receiver respectively. Since both of them are approximated to be non-directional points, the system composition was simplified while keeping high measuring accuracy. Since three transmitters must spark successively one by one in order to avoid the interference of ultrasonic waves, the dynamic orientation measuring method which predicts two non-sparking transmitters' positions from their previous measured ones was newly proposed. It was proved by using a developed orientation standard machine that this system can measure the dynamic orientation by error within ±0.2°when Z coordinates of transmitters keep almost same value and within ±0.4° when they change. As an example of applications, the vibration and position control of the suspended object of a prototype crane was executed by using orientation information obtained by this system. The experimental results proved the good practicability of this system.

抄録全体を表示

This paper proposes a novel probe system to reduce contact force and to keep a workpiece from deformation. In this probe system, a stylus is supported by an elastic hinge. The displacement of stylus tip is measured by an optical detector. The optical detector consists of two QPDs, two lasers and a ball lens. The ball lens is attached to the end of stylus. As the displacement of the ball lens is proportional to that of the stylus tip, the displacement of the ball lens is measured by two QPDs and two lasers to measure three-dimensional displacement of the stylus tip. In this paper, the structure of the probe system and the principle of the optical detector are described, and the experimental results in X-Y plane are shown. In a result of using the optical detector, this probe system can detect the contact with the workpiece with the resolution of 0.1μm.

抄録全体を表示

Grinding has drawbacks where the attrition wear of grain deteriorates the accuracy and the quality of workpiece under the condition of high speed and high pressure. So far it has been made clear that a chemical reaction in manufacturing is the cause of the attrition wear of grain. In other words, the chemical reaction control should make possible to manufacture with the sharp grain. This study pays attention to the chemical reaction caused by grinding into steel materials with aluminum oxide grain. The purpose of this study is to reduce a chemical reaction by removing oxygen that is caused by it in the grinding point and to propose a new grinding method that maintains sharp grain on wheel surface for long time. The method used is a kind of processing method where argon gas passes through a grinding wheel. According to this method, the chemical reaction should be caused between workpiece and grain controlled, and the workpiece surface should be manufactured as good one with the sharp grain. As a result obtained it was clear that the manufacture efficiency of the new processing method was improved better in comparison with a conventional method of it. This paper shows the valuation of grinding phenomena and effect of the new processing method.

抄録全体を表示

In order to reveal the mechanism of the electrochemical processes of cathode surface in ultrapure water, first-principles molecular-dynamics simulations of hydrogen-terminated Si(001) surfaces interacting with H atoms and H20 molecules were carried out on the basis of the Kohn-Sham local-density-functional formalism. A plane-wave basis set was used, and the cut-off energy is 327eV(24Ry). A norm-conserving pseudopotential was also used. We adopt the standard molecular-dynamics method for the optimization of the ionic system and the preconditioned conjugate-gradient (CG) method for the quenching procedure of the electronic degrees of freedom. We determined the optimized atomic configurations and electronic distributions for H and H2O chemisorbed hydrogen-terminated Si(001) surfaces. It was confirmed that an H atom reacts with H20 molecules on the hydrogen-terminated Si(001) surface to produce an OH molecule, and chemisorption of two OH molecules to the hydrogen-terminated Si(001) surface atom breaks the back-bonds and the Si surface atom is etched with forming an SiH2(OH)2 molecule. From these simulation results, we deduced that the electrochemical machining of Si(001) cathode surface in ultrapure water is possible.

抄録全体を表示

Mechanical vibration plasma etching system has been improved by an enforced vibration mechanism. In the new system, two types of vibrators have developed for higher frequency and larger specific energy than the previous system. They have resonant frequencies of 28kHz and 40kHz, and vibration is amplified by step horns resulting in amplitudes of 12 μ m and 5μm, respectively. Etching experiments are performed by using silicon chips with SiO₂ masks. A maximum etching rate improvement of 78% is obtained in an SF₆+O₂ gas mixture with a vibration frequency of 40kHz, however, anisotropy is not so improved as expected by the vibration. This paper describes expecting effects, construction of the vibration mechanism and experimental results, which indicates usefulness of the new system.

抄録全体を表示

According to the roadmap of the semiconductor industry, the current Φ200mm (8") silicon wafer will be replaced by Φ300mm (12") by the year 2003. The present manufacturing process for silicon wafer includes lapping, etching and polishing. Such free slurry methods can achieve better surface roughness only when smaller abrasives are applied, therefore, require multi-stage equipments and are energy consuming. In addition to the low throughput, the current pressure-controlled process is also unable to fulfil the requirement on global flatness as the wafer size increases. To address the problems of throughput rate, post-process cleaning, environmental impact as well as to achieve the total surface integrity for Φ300mm silicon wafer, the semiconductor industry was looking for a fixed abrasive solution as the alternative. This research has developed a one-stop manufacturing system for Φ300mm silicon wafer, using fixed abrasive instead of conventional free slurry, to provide a totally integrated solution for achieving the surface roughness Ra<1nm (Ry<5-6nm) and the global flatness<0.2μm/Φ300mm. In addition to the space saving, this one-stop system also significantly reduces the total energy consumption, compared with the current process used for Φ200mm Si wafer. Three core technologies: the hybrid process mechanics, the GMM actuated positioning/alignment device and the ecologically friendly coolant circulation system are described in this paper. The system performance and results are then presented and discussed.

抄録全体を表示

Fine finished steel ball set into a bearing have been needed in precision machinery field. Spin angle controlled ball lapping method developed recently, is to have an excellent performance to lap the ball in which V-groove lapping plates are separated into two parts and three plates are able to rotate independently. Present paper focuses on an effect of lapping plate geometry on finishing characteristics of the steel ball. It is found that Rcurved lapping plates constituting of V-groove have yielded greater stock removal than flat ones. Experiments show that the stock removal is dependent upon spin angle θ which is closely related to slipping distance of ball. Measurements of a sphericity of lapped ball indicate that the geometry of lapping plate has small effect with respect to improvement of it, but causes larger stock removal rate and very smooth surface.

抄録全体を表示

Abrasive jet machining (AJM), a specialized form of shot blasting, is considered one of the most hopeful micro machining methods for hard and brittle materials such as glasses and ceramics. This paper describes an elementary approach to the characterization of fundamental AJM properties for silicon nitride. In dimpling of the Si₃N₄ surface, the material removal was identified by mild wear within the grain particles in a manner resembling ductile behavior. Moreover, effects of several AJM operating parameters, e.g. jet pressure, jet distance, abrasive type and size, on the machinability have been empirically evaluated. A proper selection of the jet distance can maximize the machining efficiency, although it is recommended to reduce jet distance as much as possible in order to decrease the dimple diameter. The machinability of the AJM process is strongly dependent on the abrasive particles; with harder and/or larger jet particles, the material removal rate increases and the resulting surface roughness deteriorates.

抄録全体を表示

Owing to their mechanical isotropy, structural homogeneity and good formability for three-dimensional micro structure, thin film metallic glasses (TFMGs) are considered to be one of ideal materials for micro-electro-mechanical systems (MEMS). However the physical properties of thin film metallic glasses have not yet examined sufficiently. In this paper, physical properties of two kinds of TFMGs; Zr-based (Zr₇₅Cu₁₉Al₆, atomic %) and Pd-based (Pd₇₆Cu₇Si₁₇), were examined to provide an important guide for the analysis and design of their MEMS applications. The TFMGs were fabricated using RF-magnetron sputtering system. Firstly, an influence of sputtering condition on amorphicity and toughness were examined. Secondly, amorphicity, composition and supercooled liquid region of the as-sputtered TFMGs were verified. Thirdly, the thermal, chemical, mechanical and electrical properties of the TFMGs were experimentally examined. Finally, an influence of annealing of the Pd based TFMG on its Young's modulus was clarified. As a result of this study, it became clear that the TFMGs are suitable for elastic MEMS.

抄録全体を表示