Study on Performance Evaluation of Small Axial Fan

In order to establish the design methodology of a small axial fan, the axial fan with impeller diameter of 36 mm was designed, manufactured and tested. Especially, in order to investigate the influence of difference in blade cord length and blade thickness on performance characteristics, the performance characteristics obtained by the designed axial fans with difference blade shapes were examined. Also, by using CFD, the same flow field as the experiment was visualized. It was found that the lift of the blade was increased and the performance was improved in high flow rate region by thinning of the blade thickness and by extending the blade chord length.

Development of Multi-Fingered Prosthetic Hand Using Shape Memory Alloy Type Artificial Muscle

A new prosthetic hand with a fixed thumb and four fingers actuated by Shape Memory Alloy （SMA） type artificial muscle has been developed in this paper. Different from typical geared motor, SMA actuator is lightweight and silent, however shows a little short stroke and small attracting force per each unit. In order to achieve enough output force and motion range of each finger, multiple SMA type artificial muscles with special device which facilitates enough length are equipped in the hand. The fundamental properties of the SMA type artificial muscle including output force and electrical response were determined experimentally and considered for the design of hand mechanism. Besides, the structure of each finger and whole system has been designed based on observation of human hand. The electrical hardware to control multiple shape memory alloy type artificial muscles has been also developed. Finally, the usefulness of the prosthetic hand has been investigated through experiments for grasping several types of objects.

Training of Multi-layered Neural Network for Data Enlargement Processing Using an Activity Function

In this paper, we present a study on activity functions for an MLNN (multi-layered neural network) and propose a suitable activity function for data enlargement processing. We have carefully studied the training performance of Sigmoid, ReLu, Leaky-ReLu and L & exp. activity functions for few inputs to multiple output training patterns. Our MLNNs model has L hidden layers with two or three inputs to four or six outputs data variations by BP (backpropagation) NN (neural network) training. We focused on the multi teacher training signals to investigate and evaluate the training performance in MLNNs to select the best and good activity function for data enlargement and hence could be applicable for image and signal processing (synaptic divergence) along with the proposed methods with convolution networks. We specifically used four activity functions from which we found out that L & exp. activity function can suite DENN (data enlargement neural network) training since it could give the highest percentage training abilities compared to the other activity functions of Sigmoid, ReLu and Leaky-ReLu during simulation and training of data in the network. And finally, we recommend L & exp. function to be good for MLNNs and may be applicable for signal processing of data and information enlargement because of its performance training characteristics with multiple teacher training patterns using original generated data and hence can be tried with CNN (convolution neural networks) of image processing.

Effect of Double Air Injection on Instability Phenomena in Centrifugal Compressor

In the operation of a centrifugal compressor of turbocharger, instability phenomena such as rotating stall and surge are induced at a lower flow rate close to the maximum pressure ratio. In this study, the compressed air at the exit of centrifugal compressor was re-circulated and injected to the impeller inlet by using two injection nozzles in order to suppress the surge phenomenon. The most effective circumferential position was examined to reduce the flow rate at the surge inception. Moreover, the influences of the injection on the fluctuating property of the flow field before and after the surge inception were investigated by examining the frequency of static pressure fluctuation on the wall surface and visualizing the compressor wall surface by oil-film visualization technique.

Improvement of Algorithm of Diagnostics for PEFC Performance

Recently,polymer electrolyte fuel cell(PEFC)begins to be used in a lot of fields such as automotive industry and power generation etc.Therefore our laboratory developed the simple maintenance equipment which can diagnose a performance degradation of PEFC instantaneously by analyzing transient response at a current interruption.However,because it was not able to diagnose some operating conditions,we determined that it was necessary to improve the diagnostic algorithm.Therefore we developed a new algorithm which sequentially approximates diagnostic parameters by the on-line algorithm of the method of least squares.The adequacy of a new algorithm was verified using a PEFC single cell of 25 cm^2 applied about two Membrane Electrode Assemblies(MEAs)((a)10 mg,(b)7.5 mg)with different amount of platinum.As a result,we judged that this algorithm was able to apply to the new diagnostics because it distinguished the difference of two MEAs in each parameter.

Learning Performance of Linear and Exponential Activity Function with Multi-layered Neural Networks

This paper presents a study on the improvement of MLNNs(multi-layer neural networks)performance by an activity function for multi logic training patterns.Our model network has L hidden layers of two inputs and three,four to six output training using BP(backpropagation)neural network.We used logic functions of XOR(exclusive OR),OR,AND,NAND(not AND),NXOR(not exclusive OR)and NOR(not OR)as the multi logic teacher signals to evaluate the training performance of MLNNs by an activity function for information and data enlargement in signal processing(synaptic divergence state).We specifically used four activity functions from which we modified one and called it L&exp.function as it could give the highest training abilities compared to the original activity functions of Sigmoid,ReLU and Step during simulation and training in the network.And finally,we propose L&exp.function as being good for MLNNs and it may be applicable for signal processing of data and information enlargement because of its performance training characteristics with multiple training logic patterns hence can be adopted in machine deep learning.

Effect of Double Air Injection on Performance Characteristics of CentrifugalCompressor

In the operation of a centrifugal compressor of turbocharger, instability phenomena such as rotating stall and surge are induced at a lower flow rate close to the maximum pressure ratio. In this study, the compressed air at the exit of centrifugal compressor was re-circulated and injected to the impeller inlet by using two injection nozzles in order to suppress the surge phenomenon. The most effective circumferential position was examined to reduce the flow rate at the surge inception. Moreover, the influences of the injection on the fluctuating property of the flow field before and after the surge inception were investigated by examining the frequency of static pressure fluctuation on the wall surface and visualizing the compressor wall surface by oil-film visualization technique.

Effect of Double Air Injection on Performance Characteristics of Centrifugal Compressor

In the operation of a centrifugal compressor of turbocharger, instability phenomena such as rotating stall and surge are induced at a lower flow rate close to the maximum pressure ratio. In this study, for the suppression of surge phenomenon resulting in the extension of the stable operating range of centrifugal compressor to lower flow rate, the compressed air at the compressor exit was re-circulated and injected into the impeller inlet by using the double injection nozzle system. The experiments were performed to find out the optimum circumferential position of the second nozzle relative to the fixed first one and the optimum inner diameter of the injection nozzles, which are able to most effectively reduce the flow rate of surge inception. Moreover, in order to examine the universality of these optimum values, the experiments were carried out for two types of compressors.