Pneumatics is widly used in the industrial implementation, in the operation of air-powered actuators and the control devices needed in their operation. However, there are many empirical factors in pneumatics design, it might cause excessive design of pneumatic components. For that reason, we developed VDES(Vacuum Design Expert System) for the economic design of pneumatic vacuum equipment. VDES is achieved with CLIPS(C Language Integrated Production System) and knowledge base that contains a number of facts and rules for pneumatic vacuum design. Forward chaining and depth first search technique are used in this system. Appling VDES to the actual field, this system is verified to be a good efficiency and could be applied to the field of pneumatic vacuum equipment design.

This paper presents results on dynamic analysis of the missile initial motion arising from the missile detent force. Using ADAMS (Automatic Dynamic Analysis of Mechanical Syatem) software, a non- linear46-DOF (Degree of Freedom) model is developed for the launcher system including missile and lunch tube contact problem. From the dynamic analysis, it is found that initial angular velocity of the missile incre- ases when the missile detent force increases and also when rocket exhaust plume is taken into account. To achieve the missile launching stability, it needs to reduce the missile initial detent force and exhaust plume area of the lancher. Results of the dynamic analysis on the system natural frequency agree well with those obtained from experimental modal tests. The overall results suggest that the proposed method is a useful tool for prediction of initial missile stability as well as design of the missile launcher system.

This thesis is concerned with the study on the characteristics of the tool failure occuring at the beginning of cutting in finish machining of hardened steels such as carbon tool steel and alloy tool steel by a ceramic tool (Al$_{2}$O$_{3}$+TiC) with nose radius. In the machining of hardened carbon steel STC3, the wear mechanism on the flank face of the ceramic tool is abrasion wear. The mode of tool failure is developed into catastropic fracture with flaking. It is thought that the fracture caused by FeO and TiO$_{2}$ results from the oxidation of Fe in the workpice and TiC in the ceramic tool and the deposit of Fe formed on the surface of the ceramic tool. In the machining of hardened alloy steel STD11, the wear mechanism on the flank face of the ceramic tool is that abrasion and adhesion wear exist simultaneously. The mode of tool failure at the beginning of cutting features is DOC notch wear. It is thought that the DOC notch wear caused by FeO and TiO$_{2}$results from the oxidation of Fe and TiC in the workpiece and ceramic tool, respectively.

A new dynamic model for the cutting process inb the end milling process is developed. This model, which describes the dynamic response of the end mill, the chip load geometry including tool runout, the dependence of the cutting forces on the chip load, is used to predict the dynamic cutting force during the end milling process. In order to predict accurately cutting forces and tool vibration, the model which uses instantaneous specific cutting force, inclueds both regenerative effect and penetration effect, The model is verified through comparisons of model predicted cutting force with measured cutting force obtained from machining experiments.

This paper presents the development of intelligent performance management of computer communication networks for larger-scale integrated systems and the demonstration of its efficacy using computer simula- tion. The innermost core of the performance management is based on fuzzy set theory. This fuzzy perfor- mance manager has learning ability by using principles of neuro-fuzzy model, neuralnetwork, genetic algo- rithm(GA). Two types of performance managers are described in this paper. One is the Neuro-Fuzzy Per- formance Manager(NFPM) of which learning ability is based on the conventional gradient method, and the other is GA-based Neuro-Fuzzy Performance Manager(GNFPM)with its learning ability based on a genetic algorithm. These performance managers have been evaluated via discrete event simulation of a computer network.

Sealing of lubricant-air mixture in the high performance machining center is one of most the important characteristics to carry out enhanced lubrication. High speed spindle requires non-contact type of sealing mechanism. Evaluating an optimum seal design to minimize leakage is concerned in the aspect of flow control. Effect of geometry and leakage path are evaluated according to variation of sealing geometry. Velocity, pressure, turbulence intensity of profile is calculated to find more efficient geometry and variables. This offers a methodological way of enhancement seal design for high speed spindle. The working fluid is regarded as two phases that are mixed flow of oil phase and air phase. It is more reasonable to simulate an oil jet or oil mist type high speed spindle lubrication. Turbulence and compressible flow model are used to evaluate a flow characteristic. This paper considers a design effect of sealing capability of non- contact type seals for high speed spindle and analyzes leakage characteristics to minimize a leakage 7 on the same sealing area.

Objective of calibration is to find out the accurate kinematic relationships between robot joint angles and the position of the end-effector by estimating accurate model parameters defining the kinematic function. Estimating the model parameters requires measurement of the end-effector position at a number of different robot configurations. This paper studies the implication of measurement configurations in robot calibration. For selecting appropriate measurement configurations in robot calibration, an index is defined to measure the observability of the model parameters with respect to a set of robot configurations. It is found that, as the observability index of the selected measurement configurations increase the attribution of the position errors to the parameter errors becomes dominant while the effects of the measurement and unmodeled errors are less significant; consequently better estimation of parameter errors is expected. To demonstrate the implication of the observability measure in robot calibration, computer simulations are performed and their results are discussed.

In the temperature analysis of hot metal forming process, the heat transfer conditions between the work-piece and the tool have improtant influences upon the temperature distribution. The accuracy of thermal analysis depends on the proper description of boundary conditions including heat transfer. At the contact surface of two materials with different temperatures, this requires the knowledge of the overall heat transfer coefficient. In order to determine the overall heat transfer coefficient, a technique is developed. The technique involves temperature measurement by using thermocouples during hot upsetting operations and finite element computation. The overall heat transfer coefficient is determined using a non-linear optimization technique.

In machine tools, frictional force exists between the table and the guideways, and in ballscrews. In this paper, feed motor current measured by a hall sensor is used to calculate the motor torque. Some frictional phenomena are studied in feed drive systems, such as the relationship between feedrate and frictional torque, and chip cover effects on frictional torque. Considering frictional phenomena, the relation- ship between the feed froce and the feed motor current id obtained. Feed force can be well estimated by feed motor current measurement considering frictional behavior. The relationship between the cutting force and the feed motor current is slightly different between up milling and down milling due to the effect of y direc- tional cutting force on frictional torque.

This paper suggests a new non-linear friction compensation for digital control systems. This control adopts a hysteresis nonlinear element which can introduce the phase lead of the control system to compensate the phase delay comes from the inherent time delay of a digital control. A proper Lyapunov function is selected and the Lyapunov direct method is used to prove the asymptotic stability of the suggested control.

Designing a feedback controller requires a dynamic model of the process to be controlled. But the GMA welding process models have not been fully developed for many reasons such as complexity of the welding process and lack of reliable sensors. Because of the complexity of the welding system, we obtained a dynamic model for control using system identification routines, rather than derived a model from fundamental physical laws. The controller was designed based on the experimentally derived linear dynamic model of the welding process. In order to demonstrate application of the designed controller, the simulation was carried out.

A robust controller design to corrdinate a robot manipulator under unknown system parameters and bounded disturbance inputs is presented in this paper. Generally, robust controllers require high input torque so that they may face input saturation in actual application due to the power limitation of the actuator. To solve this problem, an improved robust controller with saturated input torque using a fuzzy logic control is proposed. Numerical examples are shown to validate the proposed controller using two degree-of-freedom planar arm.

Multi-layer neural networks with error back-propagation algorithm have a great potential for identifying nonlinear systems with unknown characteristics. However, because they have a demerit that the speed of convergence is too slow, various methods for improving the training characteristics of backpropagition networks have been proposed. In this paper, a fuzzy division method is proposed to improve the convergence speed, which can find out an effective fuzzy division by the tuning of membership function and independently train each neural network after dividing the network model into several parts. In the simulations, the proposed method showed that the optimal fuzzy partitions could be found from the arbitray initial ones and that the convergence speed was faster than the traditional method without the fuzzy division.

The drilling experiment of carbon fiber epoxy composite material with WC-drill has been done under the various cutting conditions in order to minimize the problems occurred in the material while being drilled. It has been confirmed by a frequency analysis of the cutting force signals that the variation of cutting force resulted from the periodic variation of the angle between the ortating drill and the stacking angle of the carbon fiber. By the drilling experiment with several drills having different point angles, the drilling char- acteristics, which show the relations between the change in the point angle and cutting force or external surface condition, were analyzed.

A sensorless motion planner which succeeds in grasping a polygonal part firmly into a desired orientation has been developed through the dynamic analysis. The analytical results on the impact process with friction are used for modeling the contact motionduring the parallel-jaw grasp operation, which is com- posed of the pushing and the squeezing process. The developed planner succeeds in grasping a part into a specified orientation in the face of uncertainties of initial position and orientation of the part, motion direction of the finger, and the physical parameters such as the coefficients of friction and restitution. The motion planner has been fully implemented into a viable package on the computer system, and verified experimentally. The motion of parts is recorded using a high-speed video camera, and then compared to the results of the planner and the graphic simulation results that illustrate the simulated motion of the grasp operation.

The characteristics of fatigue crack growth subject to out-of-plane bending fatigue are studied in terms of crack opening behavior by using pre-cracked smooth specimens. Crack opening stress is measured by an elastic compliance method which may precisely and continuously provide many date using strain gages during experiment. The results of the short crack and the long crack arranged by crack closure concept show that the effective stress gange ratio of short crack is grester than that of long crack, and ano- malous growth behavior of short crack may be elucidated by the variation of crack opening stress. When the variation of fatigue crack growth rate is arranged versus effective stress intensity factor range. Iinear relation is held also for the short crack. It shows that growth behavior of short crack can be quantitatively represent- ed by the fracture mechanics parameter using effective stress intensity factor range.

A computer-aided process design system for axisymmetric deep drawing products has been developed. An approach to the system is based on the knowledge based system. The hypothesized process outline of the deep drawing operations is generated in the geometrical design module of the system. In this paper, the module has been expanded. The rules of process design sechems for complex cup drawings are formulated from handbooks, experimental results and empirical knowhow of the field experts. The input to the system is final sheet-metal objects geometry and the output from the system is process sequence with intermediate objects geometries and process parameters, such as drawing load, blank holding force, clearance and cup-drawing coefficient.