A precision reducer is the core component in an industrial robot and largely determines the performance of a robot. Lost motion is a key index of the performance of precision reducers and its measurement and evaluation are the basis for improving the performance of precision reducers. The dynamic measurement of the lost motion is affected by load torque and speed, but there is no uniform standard or scientific basis for measurement speed, thus leading to the deviations in measurement results. In this paper, the principle of the dynamic measurement of the lost motion of precision reducers was firstly introduced and the method of separating the geometric lost motion based on optimal measurement speed was put forward. Then, according to the Stribeck friction effect of precision reducers, the method of determining the optimal measurement speed was proposed and the calculation model of the optimal measurement speed was established. Finally, taking the RV reducer as an example, the measurement experiment was carried out. The friction torque of the RV reducer was the least and the mean value of the measured geometric lost motion was the least under the optimal measurement speed, thus verifying the proposed method.

Our previously proposed method for arranging adhesive joints for an optical assembly was adopted to resist long-term exposure to severe environments. This method is expected to passively compensate for permanent optical misalignment due to long-term exposure to hot and humid conditions by arranging the bonding direction and position of each optical component in order to tend to displace or tilt a desired amount in a desired direction over time. We calculated the beam spot deviation on a detector, which is an indicator of misalignment of the optical assembly due to heat, humidity, and time effects on adhesive joints. We arranged to adhesively bond the laser diode to the holder from the top so that the laser diode could be lifted up due to the moisture expansion of the adhesive joint over time. The developed adhesive joints arrangement had a great advantage in that upward displacement of the laser diode compensates for beam spot deviation caused by tilt of the prism and displacement of the detector along the warp deformation of the housing. The calculated and measured results showed that the beam spot deviation was within about ±5 % after the optical pickup was left at 60 °C and 90 %RH for 336 hours. It is concluded that our method for arranging adhesive joints of each component effectively compensates for passively permanent optical misalignment that occurs after long-term exposure to hot and humid conditions.

The dynamic characteristics of a joint affect the machine tool operations notably. In this paper, an improved approach is proposed to identify the dynamic stiffness of the joint and to construct a relevant dynamic model. The theoretical dynamic model assembled of two beams is built using the frequency response functions. The identification formulas are derived based on the mechanical equilibrium condition and the displacement compatibility condition to describe the relationship between the assembled structures and substructures. Then an inconsistent equation containing the identification relationship is developed. The equivalent value of the dynamic stiffness is extracted employing the least square method. In the identification process, a part of frequency response functions, which is difficult to be measured, is considered as an intermediate variable to avoid introducing errors. According to the results, the calculated dynamic responses of the assembly show better agreement with the experimental measured data in comparison with the results derived from a typical method, which validates the feasibility. And the results demonstrate higher accuracy of the proposed method.

The need for automated distribution systems has recently grown due to labor shortages and an increasing preference among consumers to shop online. The process of depalletizing roll box pallets (RBPs) requires skillful work in a limited space, but an automated RBP depalletizing system with a speed equal to or exceeding that of human workers has yet to be developed. We solved this issue using two technologies. One is a system design and the other is an algorithm for prioritizing the packages to be removed from the stack. We designed a highspeed depalletizing system that can be installed in a narrow space at the distribution site. Equipped with a compact linear actuator mechanism, the system utilizes an end effector that can handle multiple packages simultaneously. Moreover, by optimizing the process of transferring packages on the system, we increased the speed. Our algorithm for prioritizing packages allows the proper selection of packages held simultaneously for faster depalletizing. However, if the height of complexly stacked packages is not known, depalletizing is difficult. To solve this issue, we developed an algorithm with special rules. Finally, we conducted depalletizing trials with the latest prototype system using different arrangements of stacked packages, and achieved an average speed of 683 packages per hour.

A digital hydraulic cylinder that uses mechanical closed-loop feedback regulation to realize double closed-loop control of displacement and velocity is designed in this study. Based on the working principle, the transfer function model and AMESim simulation model of digital hydraulic cylinder were established. The influence of pre-opening distance of the four-way reversing valve, ball-screw lead, inner diameter, and load on the displacement error, adjustment time, overshoot was analyzed. The standard regression analysis about the results was carried out as well. The results show the pre-opening distance of the four-way reversing valve has the greatest influence on the dynamic response characteristics of double closed-loop control digital hydraulic cylinder, and three Pearson coefficients are all greater than 0.9. The non-inertial load has a certain influence on the displacement error of the digital hydraulic cylinder. The optimal value of the ball-screw lead and inner diameter which have less effect on the dynamic characteristics of the system are 10mm and 90mm, respectively. Furthermore, the performance tests of the digital hydraulic cylinder sample were carried out.

In this paper, the effects of rigid-flexible coupling knees on the performance of the exoskeleton developed by us for power assist are mainly concerned. The work was conducted through a series of numerical simulation comparisons between rigid-flexible coupling models and pure rigid structure models of the exoskeleton under different loads. Before the dynamic analysis, a flat walking gait plan in sagittal plane was designed for the exoskeleton, and the coupling numerical model of the novel knee was verified through comparing with the experimental result. Compared with the pure rigid knees, the simulation results show that the rigid-flexible coupling knees exert a mixed influence on dynamics of the exoskeleton under different loads. Via the current study, a simple and practical gait plan is acquired for biped exoskeleton to imitate human level walking, and flexible parts utilized in the exoskeleton joints could lead to a vibration of joint driving torque without load, and on the contrary, it contributes to the lower impact ground reaction force and undemanding driving torque requirement with heavy load.

The aim of this paper is to research the synchronization of two excited motors in an anti-resonance system. After establishing the dynamic equations of the system by Lagrange equation, the displacement responses of the system and the ability of vibration isolation are determined. And the synchronous condition and synchronous stability criterion of the system are derived by using the average small parameter method and Hamilton's principle, respectively. Moreover, the influences of the structural parameters on synchronous characteristics between the rotors are discussed by numerical analysis. Then, the correctness of theoretical analysis is verified by numerical simulations. The research results show that, the ability of vibration isolation is mainly influenced by frequency ratios of the system; the synchronous ability of the system is improved with the increase of the distance between two motors; meanwhile, the synchronous behavior of the system is closely related to the mounting position and rotation direction between the motors. The research result can provide theory guidance for designing the anti-resonance drilling fluid shakers.

In haptic virtual reality (VR), there are two important challenges: one is the difficulty of reproducing a surface feeling with haptic devices, and the other is the presentation method used for the display. We developed a tactile mouse using palm presentation for the former challenge to tackle the issue, “which tactile perception is superior for texture recognition, active or passive touch?” In a psychophysical experiment, two oblique edges of dot-patterns are presented consecutively as simple textures, and human subjects compare them to determine which is larger. We evaluated the precision of perception using the difference threshold of the edge angles calculated by the constant stimuli method. Experimental results show that the perception precision at low edge movement speeds (45 and 90 mm/s) was higher than that at high speeds (130 and 170 mm/s), and moreover, that there was no significant difference in precision of the edge angle perception between active and passive touches. The former finding seems to be caused by how well the mechanoreceptive unit handles uneven surfaces. The latter finding is due to the mechanism of the efference copy of a motor signal provided from the motor cortex not having a significant influence on texture recognition. The tactile image deterioration induced by the object movement might be compensated for with information processing in the central nervous system, which keeps stable tactile image without help from the efference copy. This work contributes to developing the haptic device which provides visually impaired persons with the tactile map.

To improve machining efficiency, it is necessary to know the machining status and optimize cutting conditions. Cutting force prediction is one method of determining the machining status. The instantaneous rigid force model is widely used and can be easily applied to cutting force prediction. However, this model requires six parameters called cutting coefficients, which have to be determined in advance through a preliminary experimental cutting test. Therefore, in this study, a new cutting force prediction method that does not require a cutting test is proposed to enable the practical understanding of the milling process in a factory setting. For this purpose, the conventional instantaneous rigid force model was revised based on the oblique cutting model and the orthogonal cutting theory to reduce the number of cutting parameters required for cutting force prediction. In the proposed model, only the shear angle is required for cutting force prediction. In practical situations, the shear angle can be identified immediately from the measured spindle motor torque, which can be monitored without any additional sensors at the start of milling operation, and the milling forces can then be predicted. In the proposed force model, the effect of tool runout can be expressed by considering the rotational radius deviation at each cutting edge. In addition, tool chipping can be detected by comparing the monitored and predicted torques. To validate the effectiveness of the proposed model, cutting experiments were conducted. The predicted force showed good agreement with the measured one. The similarity between the monitored and predicted torques was decreased by tool chipping. These results indicate that the in-process machining status can be understood and tool chipping can be detected practically without any experimental milling to determine the required parameters for prediction or any additional force sensors.

Grinding has always been an effective precision machining method for hard-to-machine materials. The knowledge to grinding-induced material properties variation is important for promotion of grinding efficiency and quality, especially for High Speed Grinding (HSG) process. This paper is devoted to investigate the HSG induced material toughening mechanism with Smoothed Particle Hydrodynamics (SPH) simulation method and experimental analysis. In this paper, single grit SPH simulation work is conducted to reveal the indentation and scratching mechanism under different process parameters. Then grinding experiments with grinding force, microscopic topography and surface roughness are given to investigate the grinding wheel speed and chip thickness effects on grinding quality and efficiency. The results show that the SiC ceramic material gets toughened with extended plastic deformation under a higher indentation speed, which could directly avoid generating brittle fracture cracks after elastic deformation. Combined with the scratching results, it is believed that the transition of material removal mode occurred under different process parameters. Through the grinding experiments, it is suggested to grinding of SiC ceramics under a higher grinding wheel speed while moderate chip thickness to keep a desired materials removal rate.

A dynamic model of a high-speed instrument rotor bearing is established, in which the parched oil lubrication is considered. The machine for dynamic test of the cage is designed to measure the cage centroid trajectory, and there comes a good agreement between the experimental results and the theoretical results. Based on the dynamic model, the force and motion of bearing cage under different conditions is analyzed, for instance, inner ring rotates with inner ring guidance (IRIG), inner ring rotates with outer ring guidance (IROG), outer ring rotates with inner ring guidance (ORIG), and outer ring rotates with outer ring guidance (OROG). Considering of ring guidance mode and rotation mode, the cage movement is investigated through the decomposition of the force on the cage. It is found the direction of cage-guide ring friction force has a significant effect on the cage whirl. When the cage is guided by rotated ring, the direction of cage-guide ring friction force is same as that of the cage whirl, and the cage movement is relatively stable; When the cage is guided by fixed ring, the direction of cage-guide ring friction force is opposite to that of the cage whirl, and the instability of the cage increases. The cage is more stable when the inner ring rotation is relative to the outer ring rotation. The instability of the cage whirl speed will cause fluctuations in the whirl radius of the cage.

To avoid the occurrence of the hub tooth tip edge contact with the sleeve tooth root flank and the sleeve tooth tip edge contact with the hub tooth root flank, a novel crown gear coupling composed of a hub with profile crowning, longitude crowning, and tip sphere teeth, and a sleeve with straight internal teeth is presented in this paper. The generating procedure of the hub with profile parabolic modification for the novel crown gear coupling is proposed. And the meshing model with angular misalignment of the novel crown gear coupling is developed. Tooth contact analysis (TCA) based on the theory of gearing and loaded tooth contact analysis (LTCA) based on finite element method (FEM) are employed to simulate contact path and to calculate contact stress distributions of the novel crown gear coupling. An example crown gear coupling is presented. The effects of profile parabolic modification coefficient and misalignment angle on contact path are investigated both for the conventional and novel crown gear couplings. Also, contact stress distributions are analyzed both for them with the maximum misalignment angle. The performed research proves that an appropriate amount of profile modification for the hub of the crown gear coupling permits the avoidance of edge contact in presence of angular misalignment, and can significantly reduce the maximum contact stress, compared with the conventional one.

Double-decked protuberant foil bearing is a type of foil bearing that uses two stacking-up layers of protuberant foils as supporting structure. In preliminary experiments, wear failure of foil thrust bearing occurs occasionally due to excessive axial load in the bearing. Further improvement of the bearing performance is necessary for its better engineering application. In this paper, a numerical model of this kind of foil bearing is applied to predict its static characteristics. Reynolds equation and Kirchhoff equation are coupled via successive iterations of film pressure and foil deformation. A parametric study is conducted to improve the static characteristics of the bearing, which is conducive to load capacity improvement. The load capacity of the bearing is represented by the bearing load with the given nominal film clearance. Due to point and surface contact between the foils, distribution of protuberances plays an important role in the stiffness distribution of the protuberant foil thrust bearing. Effect of radial location of protuberances on bearing load is analyzed. Furthermore, parametric studies with different foil thickness and diameter are conducted for further optimization. The results indicate that large unevenness of stiffness distribution will deteriorate its static characteristics. Through adjustment of the radial locations of protuberances, static performance of the double-decked protuberant foil thrust bearing can be improved prominently under different operating conditions.

Additive manufacturing is an advanced manufacturing technology that allows building any complex 3D geometry of product by adding layer by layer of material. One of the most important effectiveness of additive manufacturing is able to manufacture lattice structure inside product space in order to reduce the usage of material and the weight of the product but still ensure its mechanical properties. The lattice structure is a network of bars linked each other in a space of product and it can be quickly manufactured due to the development of additive manufacturing technologies. However, the design of a model of the lattice structure in the computer-aided design environment has many difficulties. The current modeling technologies do not support product designer to automatically create a model of the lattice structure with the different type of configurations. Therefore, the paper will present new approaches to automatically generate a 3D model of the lattice structure in the design space of the product. These approaches allow creating different configurations of the periodic or non-periodic lattice structure.

The manufacturing method of two spatial-orthogonal rotary axis linkage is usually used to cut hourglass worm. However, the traditional processing method of roughing tooth groove depends on the special machine, specific tool, and its processing efficiency is lower. To avoid these defects, a processing method of the rough cutting hourglass worm on general NC lathe without C-axis functions was presented in the paper. Based on the forming principle of the hourglass worm and the traditional processing steps, the machining principle of the new approach was elaborated. Key techniques such as the shape of the blade, the contours of tooth groove, tool path loop and the variable lead of worm, etc. were discussed. Macro-program of the new method was programmed based on Fanuc oi mate-tc CNC system, and tooth groove of the hourglass worm were trial manufactured on general NC lathe. Addendum axial pitch and a normal tooth thickness of the trial model were measured. It is noted that measurement results can satisfy the technological requirement. In conclusion, the new processing method of roughing hourglass worm is feasible and efficient.

This paper proposes a shape optimization method of local shape features on a flexible beam modeled with the absolute nodal coordinate formulation (ANCF). A fully parameterized ANCF beam element is used to model the global geometry for flexible body dynamics analysis. The local shape feature of the ANCF element surfaces can be described by a nonuniform rational B-spline (NURBS) without the need for mesh refinement by changing the distribution of the integration points. A mapping between the ANCF element local coordinates and the NURBS localized geometric parameters is used for the consistent implementation of the overlapping methods. The selected shape parameters of the localized surface geometry are taken as the design variables for beam shape optimization, and the weighted summation of the compliance of the beam is taken as the objective to evaluate the influence of the dynamic response. Different weighting schemes are used to compare the effects of weighting. An example of an ANCF beam with a rib is given to validate our method. A two-stage shape optimization method based on SQP (Sequence quadratic program ) is run to reduce the compliance of the beam. Because few design variables are used to represent the local and global shapes of the beam geometry, our shape optimization method based on the localized surface geometry has a significantly reduced computational burden and improved optimization efficiency.

In order to tackle the incongruous error of the normal vector of globoidal cam in non-equal diameter machining, the tool axis vector with minimum machining error is preliminarily fitted through the principle of ruled surface generation by using spatial linear-regression algorithm. The vector error of the initial tool axis is gradually reduced by linear regression iteration. The improved tool axis is used as the generatrix of NURBS ruled surface to reconstruct the theoretical tool axis surface. Furthermore, the least square method is carried out to optimize the tool path, and the optimization model of flank milling error of the tool axis path is constructed. Subsequently, a real-coded artificial immune algorithm for solving the optimization model is proposed. The validity of the algorithm is verified by the results of simulation and calculation of tool axis vector of globoidal cam machining.

The cell formation problem which arises in cellular manufacturing can be formulated in graph theoretic terms. The input for a cellular manufacturing problem consists of a set X of m machines and a set of Y of p parts and an m × p matrix A = (a_ij), where a_ij = 1 or 0 according as the part p_j is processed on the machine m_i. This data can be represented as a bipartite graph G with bipartition X,Y and m_i is joined to p_j if a_ij = 1. Let G₁,G₂,...,G_k be nontrivial connected subgraphs of G such that V(G₁),V(G₂),...,V(G_k) forms a partition of V(G). Then π = {G₁,G₂,...,G_k} is called a k-cell partition of G. Any edge of G with one end in G_i and the other end in G_j with i ≠ j represents an inter cellular movement of a part. One of the objectives in cellular manufacturing problem is to minimize the inter cellular movements of parts. Let β(G,π) denote the number of edges in G with one end in V(G_i) and other end in V(G_j). Let β(G,k) = min_πβ(G,π), where the minimum is taken over all k-cell partitions π of G. In this paper we propose a graph theoretic algorithm using Depth-First-Search to solve the cellular manufacturing problem for the case when k = 2. Comparison of the results that we have obtained with solutions obtained by other known algorithms shows that our algorithm gives a better solution.

Recently, in order to reduce energy consumption while maintaining productivity in lines, industries require pre-evaluation methods and production management methods that consider these aspects simultaneously. A pre-evaluation method that considers both productivity and energy consumption and simultaneously calculates productivity and energy consumption during the planning of a manufacturing system has already been proposed. In addition, with production management methods that consider productivity and energy consumption, the lot-size dependence of energy consumption per unit of production throughput has been formulated and verified. On the other hand, because the needs of consumers have become more diverse, mass customized or high-mix low-volume manufacturing systems, in which lot sizes vary, have been developed. We defined the condition in which lot size is not varied as “constant lot-size”, whereas the condition in which lot size is varied as “variable lot-size”. We define the variable lot-size that lot sizes are not constant during a certain continuous production time. However, production management methods that consider variable lot size have not developed. Therefore, we propose a formulation for the lot size dependence of the energy consumption per unit of production throughput for variable lot sizes. The key parameter of this formulation is the average lot size, as the throughput and energy consumption per unit of production throughput are independent of the lot size distribution. We also carry out case studies using a discrete event simulation to verify the proposed formulation.

In design and manufacture of machine structures and elements, the characteristics and sizes are randomly changing factors because of the change of loads, the mechanical properties of materials, and the quality of manufacturing process (size tolerances). Thus, the reliability of structures is extremely important to be considered carefully. This paper presents the application of Second-Order Reliability Method (SORM) to analyze and design of CNC milling machine structure. Many examples are utilized to validate the use of the proposed SORM approach. Then, the analytical findings are verified by comparing with First-Order Reliability Method (FORM) and Monte Carlo Method (MCS). The obtained findings highlighted that the SORM method is more reliable than the FORM method. Besides, the result deviations between the SORM method and the MCS method are negligible. Therefore, SORM method is a potentially effective approach in calculation and design with the aims of improving the reliability, increasing safety, and working efficiency of machine structures.

The new precision transmission system-Anti-backlash roller enveloping hourglass worm gears are now widely employed in many industrial sectors such as robot joint design, packaging production line, and CNC (computer numerical control) machine tools. Therefore, it is of great interest to have knowledge of their corresponding transmission principles that are seldom discussed in the literature. Using the theories of differential geometry and gear meshing, the objective of this paper is to analyze and compare meshing characteristics of anti-backlash single- and double-roller enveloping hourglass worm gears in terms of their transmission principle, engagement equation, induced normal curvature, lubrication angle, autorotation angle, entrainment velocity, helix angle, and distribution of contact curves. The major differences between these two roller enveloping hourglass worms are theoretically investigated in this work. Our results show that the anti-backlash single-roller enveloping hourglass worm (ASEHW) gears provide better performance with respect to gear meshing and transmission than the anti-backlash double-roller enveloping hourglass worm (ADEHW) gears. In addition, compared with the ADEHW gears, the ASEHW gears are easier to fabricate, install, and calibrate. However, the ADEHW gears are more auto-adjustable in eliminating gear backlash. Our study puts forward a theoretical background for futuristic design, application, and promotion of the anti-backlash roller enveloping hourglass worm gears.

Pneumatic anti-vibration apparatus (AVA) has been used in order to suppress vibration from a floor. In this paper, central pattern generator (CPG) composed of reciprocal inhibition oscillator (RIO) model is applied for suppression of flow disturbance caused by pressure fluctuation of compressed air supplied to the pneumatic AVA. Concretely, the disturbance suppression is confirmed by implementing it to the control system of the AVA. In addition, this paper shows that the control system using internal information in CPG can realize the stable levitation and the disturbance suppression simultaneously. In the industry situation, the flow disturbance is not periodic because the consumption flow rate of the compressed air changes. Therefore, the band-pass filter in CPG using RIO model is redesigned to be broadband, and the disturbance can be suppressed without declining performance even when the disturbance period changes. Moreover, this paper illustrates the roles of the sigmoid function and the connection between two neurons in CPG using RIO model. When the sigmoid function is replaced with the proportion function and the saturation function, the influence is verified by experiments.

In press processing, quality inspection of a product is often carried out for each lot in the post process stage. When a failure occurs, it may result in a large number of defective products due to the fast processing speed. In order to prevent this, it is ideal to immediately stop the processing just after the defect occurs. Therefore, confirming the processing state in-process is required. This study proposes a new quality inspection method for deep drawing processes by using the count rate of acoustic emission (AE) signals. To analyze the AE count, deep learning, which is a multilayered neural network, is employed to recognize defects during a deep drawing process. The material used was a ductile material of cold rolled steel plate and is relatively difficult to find cracks during the deep drawing process. Characteristics were clarified by analysis of AE counts at plastic deformation and brake of material, and performing the forming state recognition experiment by a multilayer neural network (deep learning) showed a maximum recognition rate of 97.3%. High recognition rate was obtained despite the small number of data used.

This paper evaluates a simple anti-sway control algorithm for fixed-speed cranes. Anti-sway control of suspended loads under cranes has been widely studied, and input shapers have been found as powerful tools. Whereas typical input shapers require variable-speed control for cranes, unity-magnitude (UM) shaper can be applied to fixed-speed cranes. However, UM shaper introduces a start-up delay and over-travel, and thus might not be user-friendly if applied to human-operated cranes. In this work, much simpler anti-sway algorithm for fixed-speed cranes with non-damped oscillations is implemented and evaluated in terms of its robustness. The algorithm is based on the conventional zero-vibration (ZV) shaper, and consists of two sub methods. In the first sub method, the command from an operator is subdivided into a pulse train, which is then fed to ZV shaper. Resulting shaped command is a sum of the original command and one short pulse to suppress the residual oscillation, which is called suppression pulse. Although the command cannot suppress sway during transportation, the crane operated based on the command does not cause start-up delay. On the other hand, the command creates over-travel due to the suppression pulse. To eliminate the over-travel, the second sub method adds a rewind pulse, which is also fed to ZV shaper before being added. The resulting total command exhibits no start-up delay and no over-travel. Experiments using a miniature experimental crane verified the suppression of the residual oscillation, as well as no start-up delay and no over-travel. The paper analyzes its sensitivity against the modeling error. The analyses show that the method is more sensitive to the modeling error when the operation command has a longer duration. Therefore, the algorithm would be effective for cranes with a short-traveling time and/or a long oscillation period.

We propose an energy harvesting device driven by the contraction of an electrically-stimulated skeletal muscle that can be used as an alternative to batteries for implantable medical devices. In order to realize a durable generator, this device has a contactless plucking mechanism comprising parallel leaf springs and magnets, with which the generator can be driven without friction. By utilizing this mechanism, the generator can be driven not only in the contraction phase of the muscle, but also during relaxation. For the design we optimized the stiffness of the springs, the gap between the magnets, and the magnetic circuit in order to maximize the power generated. The power generated by a prototype in a benchtop experiment was 35.8 μW, which is sufficient to drive an implantable medical device. Furthermore, we evaluated the power generated in an ex-vivo experiment in which the gastrocnemius muscle of a toad weighing 193.4 g was electrically stimulated to drive the mechanism. In this experiment, 18.1 μW of power was generated from a skeletal muscle weighing only 3.5 g. It was also confirmed that the power generated exceeded the power required to electrically stimulate the skeletal muscle. The results showed the feasibility of an energy harvesting system utilizing the proposed mechanism.