A new composite image sensor system is proposed. Two different type image sensors, area and line sensors, are used effectively to make dimension and shape measurements with high accuracy. An area sensor is used to obtain a whole image of the object at an accuracy of 1 mm and to pick up regions which are necessary to measure shapes and position of edges or holes, then a line sensor is used to obtain a high resolution image of picked-up regions by rotating and scanning the line sensor at an accuracy of 0.1 mm. This sensor system has an advantage that the acquisition time of image input is shortened and that the amount of data is much reduced as compared with an image input system using a line sensor only.
Experimental results show that this composite image sensor system is available in the field of not only parts'dimension and shape measurement but also three-dimensional geometrical measurement.

抄録全体を表示

If sufficient time is allowed for detection, the aerial ultrasonic pulsed radar sensor can detect reflecting objects without being unduly subject to the effect of wind. When used for safety assurance like obstacle detection of automated guided vehicles, the pulsed radar sensor must have such fail-safe characteristics that it can monitor an area wilder than that it is originally designed to monitor when it malfunctions and can produce a hazard-signifying output signal when it stops. The ultrasonic pulsed radar sensor discussed here is in its normal operating state when a short-distance transmitter/receiver signal is received as inspection signal for monitoring its operation. The pulse signal reflected from the obstacle in question, operation monitoring signal, and gate signal are added and converted into a ternary signal. The ternary signal is converted into a binary signal that indicates the normal state of the sensor (logic 1) or the faulty state of the sensor or detection of the obstacle (logic 0). A fail-safe pulsed radar can be realized in this way.

抄録全体を表示

This paper describes a train fire detection system which is set on Seikan tunnel as one of preventing disasters. Seikan tunnel is the longest one in the world, now. So that, two escape points are located in the tunnel to prepare for an emergency. These places are called “Submarine Station”. The train fire detection system inspects all train which pass the tunnel from the ground out of contact. In case this system detects a train on fire, emergency information is sent to Hakodate operation center. And that train is controled to stop at the “Submarine Station”.
This inspection system is an infrared (IR) detection system. The data is processed by computer on two steps which are high temperature check and abnormal hot zone check.

抄録全体を表示

In the reclaiming work in harbours, mud and sand are carried by barges and the measurement of the volume becomes important. A method now used is a simple one which almost depends on a perception of a skilled person. Therefore, it causes not only unexact measurement (and also time-consuming), but also a difficulty in the decision of an exact budget in large-scaled reclaiming.
This paper proposes an automatic measurement system which gathers the data of the surface of the carried mud and sand by using optical range sensors and measures the volume accurately even at night. This system first estimates the surface of the carried mud and sand as a linear combination of Gaussian type functions by least squares method, and next calculates the volume by integrating the estimated function of the surface. But, no numerical integration is required at the measurement, because integrals of the basis functions are a priori calculated. Finally, the validity of the proposed measurement system is verified by both simulations and experiments.

抄録全体を表示

As the glossiness of spherical specimens, we proposed previously the indistinctness-degree glossiness G_ID and fabricated the instrument for measuring G_ID. However, in this old instrument, fairly errors of the G_ID were caused by the deviation of the specimen holding position. Thus, an integration type indistinctness-degree glossiness G_ID-M was proposed in this paper. Next, an intelligent type sensor for measuring the G_ID-M was developed. Using the new sensor, we constructed a pearl glossiness measuring system. In this system, the measuring errors of the glossiness caused by the deviation of the pearl holding are reduced, and the measuring time per one pearl is as short as 1.5s.

抄録全体を表示

A parallel connected thyristor converter with an interphase reactor has been used widely as a low voltage high current DC power supply because of high efficiency, compared with a series one. It seems to be the most suitable for superconducting magnetic energy storage systems in particular. If the parallel converter consisting of two thyristor converters is operated under different firing angles in the two due to imperfect adjustment in the control circuit or something like that, unbalance may occur in the output DC current of each converter.
The purpose of this paper is to discuss the current unbalance of a parallel thyristor converter in detail. The main points are summarized as follows:
(1) It is clarified theoretically that the unbalance of DC output current may occur even when the difference of firing angles is less than 1°.
(2) It is verified experimentally that the proposed simple current feedback makes it possible to balance the DC output current of the parallel thyristor converter.
(3) An interphase reactor without gap is able to take the place of a conventional interphase reactor with gap which is used widely. Compared with the conventional one, the volume of the interphase reactor without gap is much small.

抄録全体を表示

To approximate a parts image as a polygon enables us to develop an advanced image processing method using the graphical processing methods, but it is not practical to use those methods in the real production lines because of the long processing time of polygonal approximation of a parts image.
To solve the problem, the authors have developed a high-speed polygonal approximation method of a parts image using look-up table. Its basic idea is based on the fact that the number of the pixel patterns to be approximated as a line is limited on the two-dimensional digital image. The all pixel patterns to be approximated as a line is listed on a template pattern table. In the case of the polygonal approximation of real images, it requires only a very short processing time to refer to the table.
The polygonal approximation of a parts image is a powerful expression for image processing. It takes only several mili-seconds to calculates most of feature parameters of parts images from the polygonal data. It also enables us to extract the local feature patterns such as corners or long parallel lines.

抄録全体を表示

This paper describes the role and meaning of the gauge condition of magnetic vector potential, and also a new formulation of the finite-element analysis. The relation between the Coulomb gauge and the condition for outer boundary, symmetrical boundary and the boundary of different materials are clarified. It is found that the gauge condition is essential for solving the three-dimensional magnetostatic field problems. The vector-Poisson's equation including the gauge is discretized using the Galerkin method which in turn yields a new formulation. In this paper, The three components of A(A_X, A_Y, A_Z) are formulated separately. By the separation of the components, memory size and calculating time can be reduced greatly. To confirm the theory and the formulation, the IEE model is examined. In the examination, it is found that the measured and calculated results are in close agreement and it is also confirmed that the divergence of A at every node which is on the boundary of different materials equals zero in the volume. These show that the theory and the formulation in this paper are valid.

抄録全体を表示

We examine a fault-diagnosis method with an artificial neural network which recognizes partial discharges or corona occurring before full breakdowns of power cables in our research. We use a feedforward type of a neural network with three layers, i. e., input, hidden and output layers. The input, hidden and output layers include 50, 4_??_16 and 2 neurons, respectively. Connection weights of the neural network are self-organized through back-propagation learning with normalized data of corona and noise waveforms in real power cables. It is shown that this system makes it possible to recognize partial discharges automatically with the correct rate of 95% at maximum. Also examined is the relationship between feature-extracting neurons self-organized in the hidden layer and the patterns of the input waveforms. Moreover, we discuss influence of temporal shifts in the input waveforms and methods to be tolerant to the influence. This research shows prospects of diagnosing breakdowns of power cables with neurocomputing.

抄録全体を表示

The commutatorless DC motor has been used in various fields as a variable speed motor since it can produce characteristics similar to those of a conventional DC motor without a mechanical commutator, and it is essentially important to know the dynamic performance of this variable-speed motor.
In previous papers the author and his colleagues derived various transfer functions for this motor and a set of nonlinear equations valid for large-signal transient state, and analyzed in detail transient performance of the motor by using these functions and equations. They also investigated the stability and the dynamic responses of speed control systems containing this motor, and proposed control strategies to improve the dynamic performance of the systems.
In this paper, the dynamic performance of the speed control system with air-gap flux and margin angle control loops as well as current control loop and a current limiter is analyzed in detail not only for the case of motoring but also for the case of regenerative braking operation. First, the dynamic responses of the system in the case of step-up change in speed reference is analyzed for two types of air-gap flux control circuits. It is clarified that when field current control loop is used for controlling the flux, the change in the margin angle of commutation during transients is reduced if the controller constants of the loop are selected so as to enlarge the increase in the field voltage in the transient state, and that satisfactory dynamic responses are obtained concerning the field voltage as well as the margin angle of commutation when the field voltage control loop is used as a flux loop. Second, the dynamic performance of the system in the case of regenerative braking operation is analyzed in detail and it is shown that the both types of air-gap flux control circuits in this operation are useful not only for a fast speed response but also for a safe commutation of the inverter, and that the transient increase in the field voltage for the case of the field current control is much larger than that for the field voltage control. Moreover, a guide for selecting controller constants for various moment of inertia is shown.

抄録全体を表示

In order to get compact and light-weight driving systems for rolling stock, the DC side's filter of the controlling inverter has a very high quality factor Q. This fact often causes an instability in the system.
A damping control, which modifies the inverter frequency according to the filter voltage, have been generally adopted to supress the instability phenomenon. This paper describes an analyzed and synthesized results on damping control, as follows.
(1) An approximated linear model of the system, which is composed of the filter, the voltage type inverter, induction motors and a digital controller, is obtained.
(2) According to the model, the existence of the following closed loop, which causes the unstable phenomenon, is proved.
(3) The damping control improves the characteristics of the closed loop. Then, the system can be stabilized. These results are verified by experiments.

抄録全体を表示

In this paper, we propose a novel modified equivalent circuit of an induction motor which can be directly applied to a distorted periodic power source such as an inverter output.
This modified equivalent circuit can be obtained only by adding a set of parallel resonance circuit of an inductor L and a capacitor C for the fundamental frequency to the conventional equivalent circuit. The voltage across this resonance circuit corresponds to the internal induced voltage of an induction motor. So it is useful to understand physically the operation of an induction motor from the equivalent circuit. The operating waveform of an induction motor can be directly obtained from the time domain analysis of the state variable equations of this equivalent circuit.
In this paper, at first, this modified equivalent circuit for two types of circuit models is proposed, and the analyzing method by the state variable equations of the circuit is described. After that, the digital simulation results for some output waveforms of the voltage fed and current fed type inverters are shown and verified from the presented literature. The analog simulator of this equivalent circuit and simulation results are also shown.
The modified equivalent circuit will be very useful to investigate and analyze the driving characteristics of an induction motor for any periodic wave power sources from the utility line voltage to any types of inverter outputs.

抄録全体を表示