巻号一覧
17 巻, 1 号
選択された号の論文の25件中1～25を表示しています
- |<
- <
- 1
- >
- >|
-
橋本 修輔1981 年 17 巻 1 号 p. 1-7
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーIn this paper, the degradation of performance index value caused by quantization errors incurred in a digital control system having the A.D-D.A converter of finite quantization figure number is analyzed. It seems that between this degradation and the hardware cost of the A.D-D.A converter exists a tradeoff of the digital control design problem.
It is shown that the degradation in a linear quadratic optimal control problem is formulated as the sum of quantization errors occurring at each sampling time in a quadratic form.
In case the initial states are unknown random variables, the expected value of degradation is derived under the assumption that the probability density function of quantization errors is uniform. The optimal quantization figure number with which the expected value of degradation becomes equal or less than the specified allawable value of degradation, is determined.
Digital simulation results of a second order control system are presented to demonstrate the feasibility of the approach.抄録全体を表示PDF形式でダウンロード (801K) -
加藤 誠, 嘉納 秀明, 増淵 正美1981 年 17 巻 1 号 p. 8-15
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーFor an nth order continuous time polynomic nonlinear system with m independent available outputs described by using high order form vectors, an observer of order n-m which is itself a polynomic nonlinear system driven by the available outputs and inputs of the original system is presented.
A method, by which the difference of two higher order form vectors is factorized by the difference of two lower order form vectors, is given, and the error equation is described as a linear time varying system by using this method. It is shown that the time-varying matrix contained in the equation can be separated into a constant matrix and a time-varying vector. Then, sufficient conditions for the global uniformly asymptotical stability of the state estimation error are derived by Lyapunov methods on the basis of this separation. The results are very practical stability criteria since the definiteness of the constant matrices and the boundedness of the input, output and state vectors only have to be checked.
The observer is further improved by using the linear combination of outputs for the reconstruction of unknown states in order to be applied to larger systems.
A numerical example is included to show the performance of this polynomic nonlinear observer.抄録全体を表示PDF形式でダウンロード (974K) -
藤井 省三, 水野 直樹1981 年 17 巻 1 号 p. 16-22
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーThis paper presents a new method for designing a discrete model reference adaptive control (DMRAC) system by input synthesis adaptation. The controlled plant is a single-input and single-output, continuous-time, linear plant. The control is performed by a digital computer. Liapunov's direct method is used to establish a parameter-adjustment law which ensures an asymptotic stability of the error signal. A sufficient condition of stability for the DMRAC system is clarified. The derived adaptive control algorithm can be implemented through a simple and systematic design procedure. In the implementation only available input-output data of the plant are used and no parameters of the reference model or of the plant are used.
Moreover, this paper presents the equivalent parameter adaptation of this method. The results of computer simulation of the adaptive control applied to a stable plant and to an unstable plant are included to illustrate the effectiveness of the proposed method.抄録全体を表示PDF形式でダウンロード (921K) -
荒木 光彦, 加藤 敏夫1981 年 17 巻 1 号 p. 23-28
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーDiscrete-time composite systems which consist of several subsystems interconnected with each other by gain-bounded time-varying nonlinear elements are treated. Each subsystem is assumed to contain one time-varying nonlinear element which satisfies a sector condition. A sufficient condition for asymptotic stability of such systems is given in terms of the pulse transfer functions of the subsystems. A method to construct a Lyapunov function and to calculate an estimate of the stability region is also given. The results are useful in a theoretical study of large-scale discrete-time systems.抄録全体を表示PDF形式でダウンロード (836K) -
田沼 正也, 諸岡 泰男1981 年 17 巻 1 号 p. 29-34
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーAn optimal preview control incorporating integral control function is studied for a linear system with transport delay, in order to obtain a practical control system which not only can respond to a change in the desired values and disturbances but also can eliminate offset errors in the output variables.
The problem is formulated as that of minimizing a quadratic performance index containing new state variables, intergral of deviation of the output variable from the desired value during the preview period. The problem is solved by the regulator theory.
The developed control system consists of two controllers. One is a feedback controller which posseses proportional control function and integral control function. The proportional control function can improve process stability by using state variables in the transport delay process. The integral control function can eliminate offset errors in output variables. The other is feedforward controller whose manipulating variables are calculated on the basis of the prediction of the desired values and process inputs, This function can contribute to improve the response speed.
Simulation results show that the proposed control system is superior to the conventional preview control system with respect to steady characteristics and is superior to the optimal regulator with respect to dynamic characteristics.抄録全体を表示PDF形式でダウンロード (689K) -
相吉 英太郎, 米田 良樹, 志水 清孝1981 年 17 巻 1 号 p. 35-42
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーThis paper is concerned with an optimization-satisfaction problem to determine an optimal solution such that a certain objective function is minimized subject to satisfaction conditions against uncertainties of any opponents' decisions or disturbances. Such satisfaction conditions require that plural performance criteria are always less than specified values against any opponents' decisions or disturbances. Therefore, this problem is formulated as a minimization problem with the constraints which include max operations with respect to the opponents' decision variables. The most general class of those problems where the objective and constraint functions depend on the maximizers as well as the minimizers is studied. A new computational method is proposed such that a series of approximate problems transformed by applying a penalty function method to the max operations within the satisfaction condition are solved by usual nonlinear programming. It is proved that a sequence of approximated solutions converges to the true optimal solution. The proposed algorithm may be useful for the systems design under unknown parameters, the process control under uncertainties, the study of general approximation theory, the strategic weapon allocation program etc.抄録全体を表示PDF形式でダウンロード (1156K) -
李 哲榮, 寺野 寿郎1981 年 17 巻 1 号 p. 43-48
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーIn evaluating complex systems, we encounter two difficulties, one of which is how to put multidimentional evaluating items together and the other is how to choose reasonable measures of these items. The artistic creation is typical one of those complex systems, of which evaluation can be done by human but its process is not yet clarified.
In this paper, the authors try to evaluate the contens of a story by the cooperation of computer and human.
The story is very complicated and fuzzy in its contents and expression, and the results of evaluation are affected by the subjectivity of human. Therefore, fuzzy integral is adopted as the mathematical model of evaluation in which human can intervene by choosing the importance of viewpoints subjectively and also by making final decision considering the calcu lated membership-function.
Six aspects, namely, variety, unexpectedness, freshness, enrichment, density and readability are adopted as the items of evaluating the contents of story. The fuzzy measures identified by using some sample stories are applied to the evaluation of candidated stories.
The validity of this evaluating process is examined through the experiments by four subjects.抄録全体を表示PDF形式でダウンロード (869K) -
Yoshiteru ISHIDA, Norihiko ADACHI, Hidekatsu TOKUMARU1981 年 17 巻 1 号 p. 49-55
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーIn this paper, we discuss the sign stability and the potential stability for the linear time invarient system described by
dx/dt=Ax, x∈Rn, A∈Rn×n.
The necessary and sufficient conditions for sign stability are obtained. The sufficient conditions for potential stability are also discussed with the concept of cycle decomposition. The results obtained are more general than those so far known. The concept of inertia preservation, which is the generalization of sign stability, is introduced to classify matrices from the structural point of view. Some system structures to preserve inertia are shown.抄録全体を表示PDF形式でダウンロード (788K) -
松尾 強, 新家 茂1981 年 17 巻 1 号 p. 56-63
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーIn this paper, we establish the realization theorems of Continuous Time Affine Dynamical Systems which are defined as the subclass of (general) Dynamical Systems.
The Affine Dynamical Systems are dynamical systems whose state spaces are the Affine Representations of the concatenation Monoid Ω, called Affine Ω-module, and whose readout maps are affine maps.
Though the definition of Affine Dynamical System is given in the form of (general) dynamical systems and affine spaces, we convert it to a new form of Affine Dynamical Systems, called Normal Affine Dynamical Systems, which consist of linear space parts of affine spaces and the initial states 0. By this new form of Affine Dynamical Systems the realization theory of this class becomes almost the same as the realization theory of the Linear Representation Systems, except its initial object of the category of linear Ω-modules with affine input map. Then we obtain the existence and the uniqueness theorems of canonical Affine Dynamical Systems, i. e., any input response map can be realized by canonical Affine Dynamical Systems and they are isomorphic in the category of Affine Dynamical Systems.
Affine Dynamical Systems are quasi-reachable iff their state sets are the affine hulls of the strict reachable sets, and are canonical iff they are quasi-reachable and distinguishable. The definition of canonicality of Affine Dynamical Systems is different from that of Linear Representation Systems, hence we need to establish the realization theorems for Affine Dynamical Systems.
Note that the class of Affine Dynamical Systems contains Linear Systems and Inhomogenous Bilinear Systems as a subclass, and the realization theory of these classes can be discussed in the framework of Affine Dynamical Systems.抄録全体を表示PDF形式でダウンロード (1132K) -
小林 伸明, 中溝 高好1981 年 17 巻 1 号 p. 64-70
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーA procedure is presented in this paper to construct the low order minimal integral left inverse of linear multivariable systems. The approach is based on the design of an observer to estimate a linear function of state variables of a linear system. Some simplified criteria for obtaining the minimal order of the minimal integral inverse are shown. Under a mild condition, it is shown that the minimal order is uniquely determined from the knowledge of ranks of appropriate augmented matrices. Then the stability of the reduced order minimal integral inverse is examined, and the results are also used to make a comparison between the integral inverse and the Silverman type inverse systems.抄録全体を表示PDF形式でダウンロード (859K) -
藤井 隆雄, 奈良崎 優1981 年 17 巻 1 号 p. 71-78
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーSimplification of linear time-invariant discrete systems with stochastic noises is discussed. The “stochastic equivalence” is newly defined. Two systems are stochastically equivalent when their outputs have the same statistical properties in the steady state. The following three problems are considered: a) Find necessary and sufficient conditions for system equivalence. b) Obtain a simplifying procedure for finding one of those systems of simplest possible structure which are equivalent to a given system. c) Investigate how all these simplified systems are related with each other. The newly defined “minimum phase fundmental system” is shown to play an important role in this investigation.抄録全体を表示PDF形式でダウンロード (1205K) -
両部 實, 椹木 義一1981 年 17 巻 1 号 p. 79-84
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーSince their first publication, A. G. Ivakhnenko and his group have shown the usefulness and applicability of the basic GMDH (Group Method of Data Handling) in the construction of most complex models. But those models are often criticized as being based on a vulnerable and insufficient theory. And so the basic GMDH is not made full use of, though this new method has many advantages for the analysis of a complex system.
In this paper we propose GMSM (Group method of Structural Modeling) including the basic GMDH method in it. This is a process which is structured by “Control Matrix” imposing some restrictions on the GMDH calculation, “ISM” (The Interpretive Structural Modeling; this was originated by J. N. Warfield to understand easily total relations among variables of a complex system) and “Man's Judgement” by a theory of a researching object by accumulated experience.
In the following a modeling process by GMSM for “The Business Forecasting of Kyoto City” supported by economic judgement is introduced. In this process, the comparatively short time required and the ingenious and easy combination of calculation and economic judgement show the practicality of GMSM.抄録全体を表示PDF形式でダウンロード (915K) -
松井 剛一, 日外 雅人, 井上 俊彦, 有本 卓1981 年 17 巻 1 号 p. 85-91
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーIn the design and operation of two-phase flow systems, it is essential to know the two phase flow pattern, on which the hydro- and thermo-dynamic data such as pressure drop, void fraction and quality are dependent. It is therefore important to develop a technique to evaluate the invisible flow pattern in a pipe or duct.
Pressure signals generated by flowing are easily accesible to measurement and also they may contain sufficient information on peculiar features of the flow pattern and flow condition for the two-phase mixture. In this study we propose a technique of estimating the representative flow pattern of gas-liquid two-phase flow in a vertical tube, based on the processing of measurement data on four static pressure signals. Differential pressures are measured in order to eliminate the influence of pressure fluctuations occurring outside of the measurement section. Differential pressure signals at both small and large intervals are examined at the same time. Two small intervals are chosen to be respectively equal to the inside radius of pipe in order to discriminate a bubble (i. e. a spherical bubble) with a length half of the inside diameter of pipe. The large one is chosen to be about ten times the inside diameter in order to evaluate a fully developed gas slug.
Experimental results with vertical pipe flows of nitrogen gas-water mixtures show that in a regular flow pattern the differential pressures exhibit a characteristic statistical property different from that of another flow pattern. Therefore it is possible to estimate the flow pattern from the configuration of probability density function, the presence of periodicities, the order of variance and the average value of differential pressures. The slug-length, rise velocity and characteristic frequency in slug-like flows are also obtainable.抄録全体を表示PDF形式でダウンロード (882K) -
増田 好治, 一條 文二郎1981 年 17 巻 1 号 p. 92-99
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーThis paper presents an electronic circuit capable of measuring a small resistance and capacitance change of high and low loss materials as well as its absolute resistance and capacitance. The measuring capability has been obtained by the use of 180° vector method.
The 180° vector method has a advantage that small changes of circuit constants due to a temperature change are cancelled each other because of its completely symmetrical circuit construction, and then the circuit shows very stable and highly sensitive characteristics.
In using the circuit as a capacitance meter, the observed stability was within less than ±0.25mV in terms of zero drift during 6 hours continuous operation at temperature in the range from 5°C to 20°C when the sensitivity was 7mV/10-3pF. When used as a resistance meter, the circuit's stability is superior to that observed when used as a capacitance meter. The measuring frequency ranging from 1MHz to 30MHz can be used.抄録全体を表示PDF形式でダウンロード (1012K) -
片面ヨーク法の適用とその限度吉谷 豊, 大野 二郎1981 年 17 巻 1 号 p. 100-104
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーThe strong correlation between the strength and the coersive force of steel has been observed by many scholars of magnetism. The non-destructive measurement of steel strength by this correlation has been taken up from time to time by several students. However, the use of this measurement of steel strength has not been found practical in steel mills. Needs for such measurement of steel strength have been increasing in the steel industry because of the increasing processing speeds and the adoption of continuous process.
The aim of this paper is to confirm the limitation and applicabilities of this measuring principle. The paper presents important design considerations for different grades and shapes of steel products.抄録全体を表示PDF形式でダウンロード (663K) -
三浦 守, 千葉 茂樹, 志田 純一, 樋口 龍雄1981 年 17 巻 1 号 p. 105-110
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーIn general, the hardware realization of digital signal processing elements have been based upon the binary arithmetic concept and PCM (pulse code modulation) signals. But the pulse-train signal processing is superior in hardware construction to the conventional digital signal processing. On the other hand, in the residue number system, arithmetic operations can be performed with complete independence between digits and without the need for carry.
This paper discusses a universal arithmetic module for pulse-train signal processing based on the residue number system (UAMR), realized by expanding the function of multi-level magnetic scaler. In order to implement digital signal processing systems, three arithmetic operations (addition, coefficient multiplication and delay) are required. The proposed UAMR has adder, coefficient multiplier and delay functions, all in one unit.
So it is expected that the pulse-train signal processing systems will be implemented by using only UAMR's as basic building blocks. And this will result in the high reliability, the simplicity and low-power consumption with respect to the construction of the systems.抄録全体を表示PDF形式でダウンロード (1986K) -
安藤 繁1981 年 17 巻 1 号 p. 111-117
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーThe quantized picture with insufficient quantizing levels often shows spoiling spurious effects such as “false contours” crossing awkwardly through a smooth region in the picture. These effects are usually prohibitive for a realization of the simplest low level quantizing scheme which aims at efficient picture encoding. These spurious phenomena which in a sense are derived artificially from a careless reproduction of encoded picture can however be reduced by realizing a smooth reproduction scheme which carefully assigns continuous but consistent graynesses to every pixel in the picture.
In this paper a new reproduction method of quantized picture with these properties is presented based on the concept of the statistical estimation. Using a Gaussian model of the picture and the quantizer error, the equation of maximum aposteriori probability (MAP) estimation is derived. Although this equation has high dimensions and also is nonlinear, a computer implementation of the equation is achieved by an efficient iterative algorithm based on the direct maximization of posterior density by using the steepest descent method. Several examples of reproduction show usefulness and efficiency of the presented method.抄録全体を表示PDF形式でダウンロード (2391K) -
内海 権三, 平林 庄司, 渡辺 忠勝1981 年 17 巻 1 号 p. 118-124
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーVarious nondestructive and breakdown tests were conducted with the two types of traction motors of electric locomotives of JNR which had been in service for many years. This paper describes nondestructive and breakdown characteristics. As the insulation resistance and tan δ are subject to environmental conditions such as humidity and temperature, we should be very careful to prevent their effect. On the other hand, the partial discharge characteristics are relatively little affected by the surrounding, also a considerable correlation is observed among the partial discharge characteristics and the breakdown voltage. This leads to the possibility of degradation judgement to a certain extent by the nondestructive test. The authors proposed the following guidelines for judging insulation detrioration in the partial discharge characteristics when the partial discharge measuring instruments are applied to an insulation test of relatively low voltage electric motor. (1) Partial discharge inception voltage (Vi): More than 1kV. Only for armature winding, more than 1.5kV. The windings which exhibit these test data have necessary insulation strength for operation. (2) Maximum partial discharge magnitude (Qmax): Below 10×10-9 coulomb. Only for armature winding, below 2×10-9 coulomb. The windings which exhibit these test data have necessary insulation strength for operation. (3) Partial discharge current(I): Below 3×104× 10-9A. Only for commutating pole winding, below 2×104×10-9A. The windings which exhibit these test data have necessary insulation strength for operation.抄録全体を表示PDF形式でダウンロード (861K) -
栗原 伸夫, 末定 泰彦, 飯岡 康弘1981 年 17 巻 1 号 p. 125-132
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーThis paper discusses a new system using a microcomputer for diagnosing a large-scale analog control system in an electric power plant or an industrial plant. When a trouble occurs in the control system, this diagnosis system can instruct the operators correct actions for the load operation and can indicate fault locations for the repair work.
The features of the diagnosis system are the following: (1) to detect failures in the control devices (such as sensors, actuators and operational modules) by using signal processing methods such as reference checks, static correlation checks, and reasonableness checks; and (2) to locate faulty control loops for operators to act against a plant shutdown by using a decision table method.
The diagnosis system has been applied to the automatic boiler control system of an actual thermal power plant. In this paper, we describe the results of diagnosis of an artificial trouble, and the results of a statistical analysis for effects of the signal processing.抄録全体を表示PDF形式でダウンロード (1155K) -
消化管の輸送機能のメカニズム梅谷 陽二, 伊能 教夫1981 年 17 巻 1 号 p. 133-138
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーPeristalsis can generally be seen in the digestive organs of animals, and serves various purposes such as transporting or mixing the contents of the digestive tracts. Although this peristaltic movement has been studied by many physiologists, its neural controlling mechanisms have not yet been made clear.
This paper describes a biomechanical study on the peristaltic transport mechanism of gastrointestines. First of all, referring to the physiological knowledge established so far, the authors made an assumption about the structure of a neural network which should exist in the intestine itself to control a peristaltic motion (as shown in Fig. 2). Secondly, in order to examine the functional property of receptors, a simulation was carried out on a digital computer, and it was found that the receptors had to hold such a property as are responding to the inside pressure caused by the stuffed contents to be transported. Finally, referring to the above mentioned results a mechanical model for peristalsis was made, and its performances of simulation was examined. It resulted that this model expressed physiological phenomena on the intestinal movement very well, and was able to transport solid stuffs in the various shape.
Thus the authors believe that the neural network model proposed by them is valid and represents a principal a neural controlling mechanism to transport contents in the intestine.抄録全体を表示PDF形式でダウンロード (3264K) -
神藤 久, 岩住 哲朗1981 年 17 巻 1 号 p. 139-146
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーRecently, the model simplification methods of the large scale linear dynamical systems have been proposed by many authors.
In this paper, using the aggregation method, we investigate the representation problem of the reduced-order models of singular perturbation theory and Routh-approximation method.
By virtue of this result, we obtain a simplified representation of the different reduced-order models, which enables us to apply the optimal control system design techniques via the aggregation method already proposed.抄録全体を表示PDF形式でダウンロード (959K) -
丹羽 昌平, 寺下 尚孝1981 年 17 巻 1 号 p. 147-149
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーPDF形式でダウンロード (402K) -
市川 惇信, 赤間 清, 小林 重信1981 年 17 巻 1 号 p. 150-152
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーPDF形式でダウンロード (423K) -
岡田 芳隆1981 年 17 巻 1 号 p. 153-155
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーPDF形式でダウンロード (363K) -
鈴木 嘉彦, 井沢 義明, 箱守 京次郎1981 年 17 巻 1 号 p. 156-157
発行日: 1981/02/28
公開日: 2009/03/27
ジャーナル フリーPDF形式でダウンロード (261K)
- |<
- <
- 1
- >
- >|