The orthogonal experiment was established to reduce the wrinkling of convex flange rubber forming with 2A12 aluminum alloy sheet. The impact factors consisted of the feed of master-cylinder, the hardness of the rubber, and the radius of the die. The rubber forming was simulated for the quantities, widths, and heights of the wrinkles in ABAQUS software, where the results was analyzed by gray correlation. Theoretically the greatest influence on the wrinkles is the feed of master-cylinder that the more the feed is, the smaller the wrinkles are. However, neither in the simulation nor in the experiment, all the wrinkles of the convex flanged parts still could not be flattened when the master-cylinder was fed to the rubber's compression limit. Aiming to further eliminate wrinkling, a new method called multi-stage rubber forming was proposed, in which four extra-cylinders were added to the broadside of the container to enhance the forming force. Both the numerical simulation and experiment were taken into the multi-stage rubber forming, proving that multi-stage rubber forming could effectively reduce the wrinkling in rubber forming of the convex flange, as the feed of the extra-cylinders increased risks of the wrinkles. When the feed of the extra-cylinders was set 8 mm while the master-cylinder 34 mm, the wrinkle of the convex curve bending edge could be eliminated smoothly.

Rolling bearing is an essential part of various rotating machines, and its signal is the typical nonlinear signal. Traditional fault diagnosis usually relies on manual experience to extract the features of signals first. Deep convolutional neural networks (DCNN) can make fuller use of time series than traditional convolutional neural networks (CNN). Because of the low accuracy rate, fault diagnosis using gated recurrent unit (GRU) alone is not unsatisfactory. In order to improve the temporality of one-dimensional convolutional neural networks (1D-CNN) and enhance the accuracy of GRU, a novel fault diagnosis method called deep convolutional neural networks - gated recurrent unit (DCNN-GRU) is first put forward, which combines DCNN with GRU. The original signals without preprocessing are input into the DCNN, and the outputs of DCNN are input into the GRU consequently. Then the faults of rolling bearing can be diagnosed effectively. As the post-processing method, the t-distributed stochastic neighbor embedding (t-SNE) method is applied to visualize the fault diagnosis results. Six different network models, including the DCNN-GRU, are used to train the same fault dataset for comparison. The simulation results show that the proposed method can reach more than 99.9% accuracy stably for the given dataset, which can verify the feasibility and effectiveness of proposed method. And the DCNN-GRU can also be verified with good generalization ability using different dataset.

Magnetostrictive electro-hydraulic actuator (M-EHA) is a highly integrated hydraulic servo device with very complex movement and conduction processes. The output performance could be affected by many factors. To design such actuator and analyze its performance with simple and efficient analysis method, an M-EHA impedance network modeling method considering the fluid inertia and the compressibility is proposed. The M-EHA components behavior were respectively described based on the fluid flow continuous equation, the pressure drop equation, and the dynamic motion equation. By employing the electro-mechanical and electro-hydraulic analogies, the impedance network of M-EHA is established and the output performance is analyzed. Through independent impedance analysis of different structures of the M-EHA, the natural frequency and dynamic characteristics of manifold, hydraulic cylinder, and accumulator are obtained respectively, and this provides the theoretical guidance for optimizing the actuator structure and performance. The influence of driving frequency, system bias, and load on the output performance of M-EHA are studied based on the impedance network. Simulations show that the best output performance is obtained under 593 Hz with 1.2 MPa system bias pressure, and the flow is 0.52 L/min.

Sustainable design is the theme of society at present. Hybrid electric vehicle (HEV) has higher fuel economy because of the assistant of electric motors. Customers’ psychological preferences can be satisfied through sustainable design of HEV, which can reduce environmental pollution and enhance customers’ desire to buy. However, the emotional needs of customers are often subjective and imprecise. Therefore, this paper makes a study on the sustainable design of HEV by combining continuous fuzzy Kano model (C-FKM) and fuzzy quality function development (QFD). Firstly, the emotional semantics related to HEV are identified by focus interviews and the C-FKM is used to directly quantify the importance of these emotional indicators. Secondly, the design elements of HEV are deconstructed by morphological analysis and the information entropy is used to screen out the key design elements that have a significant impact on customers’ satisfaction. Finally, the incidence relation between emotional semantics and key design elements is established by fuzzy QFD, thus obtaining the final weight of key design elements from the viewpoint of triangular fuzzy number. The research result provides guidance for sustainable design of enterprises in the HEV development stage, which not only reduces environmental damage, but also increases the social awareness and sales of HEV.

Mobile manipulators are widely used to transport and manipulate objects in industrial settings. In this paper, we propose a mobile manipulator that consists of a parallel mechanism and two two-wheel-drive carts. The planar motions of the carts are transmitted to a platform through three screw pairs of the parallel mechanism, allowing the pose of the platform to be controlled by only four motors. Kinematic analysis for such a two-cart mobile manipulator gives a Jacobian matrix, reveals the effects of nonholonomic constraints, and demonstrates that the yaw angle of the platform must be limited to avoid singular and failure configurations, and that the pitch angle is quite sensitive to uncertainties. Based on these analysis results, we present a custom path planning method for the carts. This method provides a non-optimal but easily realizable path planning algorithm with low computational cost, since the complex constraint conditions of this two-cart mobile manipulator have little influence on the proposed path generation process. The path planning process consists of four steps. We describe the motions of the carts in each step and establish a path tracking control system for the carts. Some simulations are conducted to show the motions of the carts, investigate the changes in the pose of the platform, and quantitatively evaluate the sensitivity of the platform’s pitch angle. Moreover, we construct an experimental prototype and conduct experiments to verify the validity and usefulness of the proposed mechanism and path planning method.

Prostate Cancer (PCa) is one of the most frequently diagnosed type of cancer in men. Cancerous regions in the prostate gland are stiffer compared to surrounding normal prostatic tissue. The digital rectal examination (DRE) which is performed manually by a medical examiner in order to assess the presence of hard lesions is a readily available primary clinical screening method for PCa, since the growth of tumors in the prostate alters the morphology and stiffness of the prostate. This method, however, is subjective, lacks repeatability, and the results depend on the skill and experience of the medical examiner. Alternative objective palpation modalities capable of obtaining quantitative spatial-temporal information about the mechanical properties variations of the prostate as a result of a pathological condition are required. In this study we investigated the use of a robotic palpation sensor system for PCa screening based on the tissue mechanical properties characterization. We evaluated the performance of the sensor system through indentation palpation tests on silicone elastomers and prostate phantoms with embedded hard inclusions which mimicked different PCa conditions. We compared the results obtained by the proposed sensor system with those obtained by using a commercial indentation machine. From the force-indentation depth quantitative information obtained at specific palpation points on the samples, the proposed sensor system was able to obtain a good estimation of the elastic modulus (E) and was capable of distinguishing stiff regions embedded with hard inclusions from soft regions based on these values, thus, presenting potential in palpation applications for PCa screening.

In recent years, automatic driving and advanced driving support systems have become more widespread; however, control systems do not consider the driving characteristics of each driver. This study constructed a personalized driver model for following forward vehicles on highways using LSTM (Long Short-Term Memory). In particular, the setting parameters for training with LSTM was focused upon and consequently, the relationship between the setting parameters and model accuracy was evaluated. The LSTM parameters considered were the number of hidden units, learning rate, number of epochs, and number of LSTM layers. Further, the driving data were measured using a driving simulator. Moreover, the model was evaluated in terms of accuracy using the coefficient of determination, and 125 combinations of parameters were compared using the mean value of the coefficient of determination and its significant difference. The order of parameter comparisons was performed considering the accuracy and learning time of the model. Consequently, the parameter set that is likely to allow efficient building of a highly accurate driver model was determined. Applying the parameter evaluation in this study contributes to the development of personalized driver models with high accuracy.

The geometric characteristic of tooth surface is the key factor to its performance. In this paper, a geometric generation method of point contact gear pair was proposed. By given a tooth surface, a smooth contact trace was determined according to the transmission demand, and the conjugated contact trace can be obtained. A new tooth surface of mating gear was derived by sweeping along the conjugated contact trace. The calculation of sliding coefficient of this type of gear pair was studied. The internal gear pair was taken as an example by taking involute gear as known gear and gear with quadratic curve profile as the mating gear. The gear pair was manufactured and performance experiment was carried out. The results show that the gear pair maintains point contact during the whole meshing process, and the contact point moves along the contact trace. The variation range of the sliding coefficient mainly depends on the involute parameter values at the meshing-in and meshing-out point. The actual contact trace of gear pair is consistent with the theoretical contact trace, and the efficiency is stable in 97.3% ~ 98.4%.

The face gear is a crossed gear similar to a hypoid gear. Owing to its compactness and good assemblability, it is used in geared motors and transmissions for helicopters. Studies on various aspects such as design, processing, and tooth modification methods have been conducted for face gears, which are typically used in fishing reels as spinning reels owing to their easy fabrication and small vibration during reel handle rotation. Because face gears without tooth flank modification are sensitive to assembly errors, the fishing reel sensation becomes unstable during mass production. Therefore, assembling and adjusting face gears to achieve a good fishing reel sensation is time consuming. Previously, tooth flank modification with a transmission error (TE)-controlled curve was used to solve this problem. The TE-controlled curve stabilized the fishing reel sensation for assembly. However, in mass production, the manufacturing error of a pinion and face gear must be considered in addition to the assembly error. In this study, we investigated the assembly and manufacturing errors when the pinion and face gear come in contact. Furthermore, we elucidated the meshing mechanism of face gear pairs under various assembly and manufacturing errors. As a result, it was confirmed that the TE-controlled curve was affected by manufacturing errors only. Moreover, a new interesting finding was obtained regarding the meshing of the face gear with a TE-controlled curve. The TE occurrence ought to be attributed to the pressure angle deviation. Hence, it is probable to replace the main cause of the transmission error with the pressure angle deviation. This consideration is the same for a cylindrical gear pair. The meshing of face gear, which was considered complicated, can be reconsidered as the meshing of a simple cylindrical gear pair. Consequently, we revealed that the TE-controlled curve is significantly useful for considering the meshing and design of a face gear pair, which is an advantageous feature.

This study aims to establish a simple method for determining tool replacement timing based on tool wear progress. First, the relationship between machining conditions and tool wear progress was investigated via dry cutting experiments conducted using a 8.0dia. cobalt high-speed steel square end mill as the tool and a piece of SS400 steel as the work material under different spindle speeds, feed rates, and radial depths of cut. From the obtained results, a linear relationship was found between the cutting edge/workpiece contact length, and the flank wear width regardless of different cutting conditions. In this paper, a simple method for determining tool replacement timing based on this relationship and tool wear progress information is proposed. In this method tool replacement timing is determined based on correlations between the cutting edge/workpiece contact length as calculated by a machining simulation and the tool wear progress measured during machining operations. Since the correlation is determined during actual machining operations, tool wear prediction and tool replacement timing determinations are performed simultaneously, which means prior experiments are not necessary. In order to verify the validity of our proposed method, cutting experiments were conducted based on the premise of customized production, in which the machining conditions vary from one product to another. From these results, it was confirmed that the tool can be used up to its tool life based on tool wear predictions and tool replacement timing determinations.

The present paper describes an inspection method for gear nicks at the final assembly process for automotive transmissions. This method applies multi-scale time synchronous averaging and optimal filtering to ensure the quality of transmission gears against abnormal noise due to gear nicks. At present, a highly reliable method has been developed for measuring gear noise in the final transmission assembly inspection process. However, the inspection system has realized not to evaluate nick-induced noise quantitatively. Therefore, the nick noise evaluation still relies on skilled workers who can estimate it subjectively. Eventually, it may reveal nick-induced noise on rare occasions after mounting the transmission on a vehicle. To avoid that requires the mechanism of nicks-induced vibration to be apparent. This paper proposes a method that can detect nick-induced vibration with high reliability. It also enables itself to identify a nicked gear even among several gears on a shaft in a transmission. The proposed method was implemented in the test equipment for the final assembly inspection and used to measure the noise of CVT units built with nicked gears. The results confirmed that this method allows itself to detect nicks that were hard to discern in subjective estimations and leads to making it possible to identify nicked gears. This method secures quality assurance against abnormal noise induced by transmission gear nicks.

Hard whirling is a highly efficient and green process for precision ball screw machining. Due to the intermittent cutting characteristics and the helical motion of the tool, the cutting forces are periodic and time-varying, making it extremely challenging to establish an accurate cutting force model. This paper proposes a cutting force modelling method based on the transient geometric characteristics of uncut chips, subject to the essential link between uncut chips and cutting forces. The time-varying features of the chip are revealed by building an accurate 3D digital model of the uncut chip. An equivalent cutting simulation model is developed using the discrete cross-section of the chip, and the effect of cutting parameters and tool edge on the cutting forces are investigated. A combination of simulation and experiment is presented to identify cutting force model coefficients, and the theoretical cutting force model for ball screw hard whirling is established. The results show that the uncut chip cross-sectional area and edge length involved in cutting first increase and then decrease, and the equivalent numerical simulation method based on the transient geometry of uncut chips can obtain more accurate cutting forces. The combined simulation and experimental strategy achieve fast and low-cost identification of cutting force coefficients, and the established cutting force model has good prediction accuracy. This study supports the investigation of the cutting mechanism of hard whirling, the optimisation of process parameters and the improvement of machining quality.

To study the influence of eccentricity on the dynamic characteristics of grinding electric spindle, firstly, the nonlinear oil film force (NOFF) expression of sliding bearing is established based on the boundary condition of adhesion pressure. Then, the expression of the unbalanced magnetic pull (UMP) under the eccentricity of the air gap is derived. Finally, the bearing-rotor dynamics equation of the grinding electric spindle considering the NOFF and UMP is established, and the dimensionless processing is carried out. The nonlinear dynamics study of the bearing-rotor system under the influence of geometric eccentricity and mass eccentricity is carried out through the shaft center trajectory and the shaft center vibration spectrum. The research results can provide reference for the optimal design of grinding electric spindle.

Rolling bearings operating in harsh environments are extremely exposed to the risk of exogenous dust or wear debris particles, so there is a strong need to investigate the effects of particulate contamination on bearing operation and to identify problems early. The lubricated oil containing silica particles of various sizes and concentrations are prepared and introduced into thrust ball bearings to replicate the severe operating conditions. Vibration and acoustic emission (AE) signals were measured by accelerometer and AE sensor in a thrust ball bearing test rig. The experimental bearing raceway was observed by scanning electron microscope (SEM) and confocal laser scanning microscopy, and the evolution of the surface topography was analyzed to probe the wear mechanism. The results show that the most severe wear areas occur at the bottom of the bearing raceway, and its degree of wear is significantly influenced by the size and concentration of the silica particles. As the experimental time increases, the raceway surface undergoes an accelerated wear process originating from slight dimples, which leads to bearing failure. Furthermore, the introduction of particles into the lubricant increases surface root mean square (RMS) roughness, vibration and acoustic emission for different experimental durations. The vibration and AE signals are affected by a combination of solid particles and surface wear. The measured monitoring results can provide valuable information about the damage morphologies and subsequently about the state of the bearing during operation.

Pedaling by right and left legs will pass through two dead zone and power zone during cycling. The power zone is an area where the downstroke power can be transferred effectively, while the dead zone is an area where the leg power cannot be delivered maximally. Therefore, building a noncircular chainring shape considering the power phase and dead zone on the chainring could be a strategy to improve pedaling performances. This paper introduces a method to connect four superellipse curves to be a chainring shape with flexibly adjustable vertical and horizontal axis lengths. Thus, the noncircular chainring design conceptualized can be constructed easier. Here, the introduced two novel chainring prototypes are designed using the proposed method to maximize the pedaling on the power zone by enlarging the chainring radii around that area and reducing the chainring radii around the dead zone to overcome the dead zone effect. Verification is performed by comparing the pedaling test results between the two novel prototypes, the circular chainring and the popular commercial elliptical chainring (Q-ring). The results show that pedaling with the two novel prototypes is more efficient than circular chainring. Moreover, it can generate a cycling speed faster than the other chainrings tested on the same number of teeth.

Level 3 automated driving requires a driver to operate a vehicle in response to a takeover request according to the Society of Automotive Engineers; therefore, the driver should maintain a state of wakefulness. The study aimed to evaluate and verify whether the presented stimuli, i.e., saccade-inducing stimuli (SaS) as visual stimuli, scent stimuli (ScS) as olfactory stimuli, and a combination of saccade-inducing and scent stimuli (CoS), can maintain the wakeful state of a driver. A driving simulator (DS) that can reproduce the actual movement of a motor vehicle using a turn table and six-axis motion device is used to simulate driving under four different conditions: no stimulation, SaS, ScS, and CoS. The percentage of eye closure (PERCLOS) and the electroencephalogram (EEG) analysis were used for the quantitative evaluation, while the participants were subjected to subjective evaluation. The results showed that the average increase in arousal time for each condition was 30.2, 53.7, and 82.3% for the SaS, ScS, and CoS cases, respectively. The average reduction rate in the EEG readings was 39.9, 41.0, and 71.4% for the SaS, ScS, and CoS cases, respectively. All participants indicated that the CoS had the highest effect of arousal as a driver stimulation tool.

To solve the problems of over-reliance on personal experience, lack of thinking about users' perceptual preferences, and neglect of the overall coupling characteristics of CMF design, the construction and application methods of product CMF design knowledge adjacency network are proposed based on the multi-layer network, intuitionistic fuzzy set, and gray correlation method theory. Firstly, the construction of the product CMF design multilayer network is completed by extracting and selecting the design resources of product families with similar imagery. Secondly, considering the product CMF design characteristics, a multi-layer network association division method based on leader nodes is proposed, and the construction of the CMF design knowledge adjacency network is further completed. Thirdly, based on intuitionistic fuzzy set theory, user preferences are considered to further improve the science of evaluation data. Finally, the calculation of association coefficients is combined with the gray correlation method, and the optimal matching between the adjacency network and the product surface assignment solution space is completed to assist designers to improve the design efficiency. The proposed method is applied to the CMF design work of a smart tractor, and perceptual experiments are set up to evaluate the design results, and the results show that this design meets the target requirements, which proves the usefulness of the method to assist the designer's work.