All issues
Volume 9 (1973)
- Issue 6 Pages 625-
- Issue 5 Pages 499-
- Issue 4 Pages 373-
- Issue 3 Pages 251-
- Issue 2 Pages 123-
- Issue 1 Pages 1-
Volume 9, Issue 4
Displaying 1-17 of 17 articles from this issue
- |<
- <
- 1
- >
- >|
-
Yoshihiro IKEDA, Shiro HATA1973 Volume 9 Issue 4 Pages 373-377
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA closed loop system consisting of a thyristor amplifier frequently generates subharmonic self-oscillations, which cause deterioration of the accuracy of the system.
In this paper, the relations between the magnitude of 1/2-harmonic self-oscillation and the open loop gain of the control system are presented. From experimental results, it is shown that unbalance of elements and disturbance, such as fluctuation of the power source, promote the generation of the oscillation.View full abstractDownload PDF (581K) -
Hajime AKASHI, Hiromitsu KUMAMOTO1973 Volume 9 Issue 4 Pages 378-386
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper deals with the discrete least square approximation of multivariable functions whose values are given only at the lattice points.
First, the solution is obtained as a linear combination of the values of the function at the lattice points, and it is shown that the deterministic method requires evaluation of the sum of an extremely large number of terms.
Next, to overcome this difficulty, the problem is converted into that of evaluating the expected values of two probabilistic models. An algorithm based on simulating these models by means of choosing the lattice points at random is proposed for the evaluation of these expected values. The method of sampling the lattice points is also discussed.
Moreover, the following points are shown.
(i) The values of the parameters which are necessary to execute the algorithm can be obtained easily regardless of the number of independent variables.
(ii) In two typical cases the proposed algorithm requires less computational labour than the deterministic method.
(iii) The algorithm is effective in the following situations:
(a) The situation when the values (data) of the function to be approximated can be measured only at the lattice points
(b) The situation when the function is a composite function
A numerical example is also presented.View full abstractDownload PDF (1148K) -
Mikio KOBAYASHI, Yasujiro OSHIMA1973 Volume 9 Issue 4 Pages 387-392
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe purpose of this paper is to establish a new approximation method using the penalty function for solving constrained optimal control problems defined in Banach spaces.
The principal advantages of the present method are as follows: (a) Both upper and lower bounds of the cost functional are obtained, (b) The functional which is introduced to obtain the approximate solution does not contain parameters approaching zero or infinity and (c) An approximate solution which satisfies all the constraints is obtained.
In this paper, the constraints are given by equalities and inequalities, and all these constraints are manipulated by embedding them in the penalty function.View full abstractDownload PDF (838K) -
Suguru ARIMOTO, Brian PORTER1973 Volume 9 Issue 4 Pages 393-397
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe problem of the performance deterioration in an optimal regulator which is caused by the introduction of a Kalman filter as a state estimator is investigated. In this paper, time-invariant Kalman filters are used as state estimators for deterministic linear time-invariant systems like Luenberger observers, so that it is assumed that the gain matrix K of the filter can be chosen arbitrarily.
It is shown that there exist a positive constant ρ and a class of initial error-differencs, e(0), which are independent of the choice of filter gain matrices, such that performance deterioration, ΔJ, is bounded away from zero as shown by the form ΔJ≥ρ||e(0)||2unless the optimal feedback control law can be implemented by direct output feedback.View full abstractDownload PDF (603K) -
Hajime AKASHI, Hiroyuki IMAI1973 Volume 9 Issue 4 Pages 398-407
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSObservation subsystems must be optimized together with plant control strategy in general stochastic optimal control processes. However, even if the problem is limited to the optimization problem of plant control strategy only, it is very difficult to solve. Therefore, a practical observation subsystem must be optimized for a control process in which the feedback loop is closed by a feasible plant control strategy.
In this paper, from the above point of view, we consider the optimization problem of an observation subsystem mainly for linear control processes in which the feedback loop is closed by two typical plant control strategies. One is a direct feedback strategy, and in this case it is shown that the optimal observation control strategy can be obtained independetly of previously observed information and also that the computation procedure is comparatively simple. The other is an indirect feedback strategy (i.e., the plant control strategy with filter). In this case it is difficult to obtain the exact solution, but we introduce a suboptimal strategy which can be considered useful in many cases. It is shown in the example that good observation does not always result in good performance, even if the observation cost is not considered, when plant control strategy is not optimal.View full abstractDownload PDF (1143K) -
Yasuhiko TAKAHARA1973 Volume 9 Issue 4 Pages 408-414
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper we considered a realization theory of general stationary linear dynamic systems and, based on the consideration, discussed a characterization of constant coefficient linear ordinary differential equation systems in the framework of general systems theory.
We showed that a necessary and sufficient condition for an input-output response family to be realizable is that the family satisfy two properties, namely, input response consistency and state response consistency, and that the input response consistency, stationarity, strong causality and finite dimensionality can essentially characterize the differential equation systems.
Since we treat general linear systems, the realization problem of this paper is different from the conventional one. However, we demonstrated that our general result implies the ordinary one if the systems in consideration are restricted to the class of differential equation systems.
In conclusion we pointed out and demonstrated that our general treatment of linear systems leads to the construction of an axiomatic linear system theory, which will be discussed in a subsequent paper.View full abstractDownload PDF (976K) -
Ryuzo YOKOYAMA1973 Volume 9 Issue 4 Pages 415-420
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSFor a given decoupling problem of control system, an equivalent problem is defined. An advantage of the equivalent problem is that a system is descriebed in a simple canonical form so that calculations for analyses can be greatly reduced. The general structure of decoupled control systems is investigated through the equivalent problem. The main results obtained in this paper are the following.
(1) The possibility of a system decoupling can be identified only by the output matrix C and the stage sequence.
(2) In a decoupled control system, zeros and the order of each subsystem are specified by the output matrix C and the stage sequence, while poles are specified by an attached state feedback control law.
(3) A necessary and sufficient condition is developed for system decoupling with preserving the complete observability.View full abstractDownload PDF (795K) -
Hisashi ISHITANI, Hisao KOSHIHARA, Masumi ISHIKAWA, Nobuharu MIYATAKE, ...1973 Volume 9 Issue 4 Pages 421-428
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSRecent economic growth has improved our material standard of living, but at the same time has broght about several serious problems, such as environmental pollution. It then forces us to investigate how all of society will change in the future and to find countermeasures for the future. A useful tool is a dynamic model which satisfies the following conditions.
1) It contains various factors of the society, such as economy, environment, population, patterns of consumption, export, import, water resources, housing, land, transportation, social overhead capital, and so forth.
2) It deals with the long-range future, at least 20 to 30 years from now.
The above two conditions, however, make building a model extremely difficult.
There are two methods which deal with modeling of society, i.e. econometrics and System Dynamics. The authors have combined these two, taking their merits and demerits into consideration, and so the model thus constructed may be named the “à la SD”model. The object area here is the Kanto area, which seems the most important area in Japan. The model is composed of several submodels, of which some are as yet incomplete and at present replaced by proper exogeneous assumptions. Simulation runs over 30 years from 1964 under various policy assumptions show fairly adequate results.View full abstractDownload PDF (1171K) -
Hidekatsu TOKUMARU, Norihiko ADACHI, Takashi AMEMIYA1973 Volume 9 Issue 4 Pages 429-437
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA sufficient condition is presented for the overall stability of large scale interconnected system with zero input. In this paper it is assumed that the overall behavior of interconnected system is determined only by the outputs of subsystems. The concept of macroscopic stability is introduced from this point of view.
So far, several results have been obtained on this problem making use of a Liapunov function for each subsystem. However, to construct Liapunov functions, differential equations must be presented for each subsystem. Moreover, even if differential equations are given, the construction of Liapunov functions is in general not easy.
To overcome these difficulties, a simple method is presented for describing the dynamic relation between the input and the output of a subsystem.
With this description of the dynamics of each subsystem, a sufficient condition is given for the macroscopic stability of interconnected systems. This condition dose not neccessarily contain all the results obtained so far, but it can be applied not only to finite dimensional systems but also to time-delay systems.View full abstractDownload PDF (1126K) -
Yasuo TACHIBANA1973 Volume 9 Issue 4 Pages 438-445
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSTwo identification methods of linear systems are discussed which explicitly make use of digital filters,
For the realization of these methods, two types of filter-banks and derivative filters of a certain order are used. A method is proposed by which these filters are designed as finite duration impulse response digital filters.
Also some numerical examples of the filters designed by this method are presented with the figures of their frequency responses.
As examples of identification using the proposed methods, two different results are given, One is the frequency response estimation of first and second order lag using signals with special spectra. The other is the parameter estimation of a turbine-generator system installed in some actual thermal power plant (156MW) using data observed under the condition of “on operation”.View full abstractDownload PDF (890K) -
Yoshifumi SUNAHARA, Akira OHSUMI, Masaaki IMAMURA1973 Volume 9 Issue 4 Pages 446-455
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper, the filtering problem is solved for a general class of nonlinear distributed parameter systems with a Gaussian white noise disturbance under noisy observations.
The mathematical models for the system and the observation mechanism are respectively described by a stochastic nonlinear partial differential equation and by an integral equation.
The principal line of attack is to introduce Girsanov's theorem of the transformation of absolutely continuous measures to the filtering theory. Using the differential generator extended to the case of stochastic partial differential equations, a version of conditional expectation is derived in a form of integro-differential equations for which the optimal filter dynamics is easily determined.
The remainder of this paper is devoted to the method of physical realization of nonlinear filters. By using the Taylor series expansion, an approximated filter dynamics is derived. Sample path behaviors of the approximated filter are also examined by digital simulation experiments.View full abstractDownload PDF (1221K) -
Yasuyuki FUNAHASHI, Yoshiaki ITO, Kahei NAKAMURA1973 Volume 9 Issue 4 Pages 456-462
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper presents a method of identification for discrete-time linear dynamic systems which is an extension of the methods proposed by G.N. Saridis and G. Stein and by T. Saito. A linear regression is formed and parameters of the system are estimated through the least mean-square type stochastic approximation.
The proposed method has the advantage, like the former methods, that few properties of both system and observation noises need be known. The major improvement of this method is that any type of input sequence can be used for identification. In the methods of Saridis and Stein and of Saito, the input sequence is restricted to a mutually independent random sequence. The length of the sampled weighting sequence to be estimated is shown to depend not only on the dimension of the system's state but also on the character of the input sequence.View full abstractDownload PDF (742K) -
Kenji KUBOTA1973 Volume 9 Issue 4 Pages 463-471
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper a global maximum searching method is developed for one-dimensional multimodal unknown functions. Justifiable properties of one-dimensional functions are analyzed, and they are used in the method of making approximate functions and global maximum searching.
In maximum searching the domain is divided into a set of subdomains, and within the set, the sub-domain whose estimated maximum is the largest is chosen. At a properly selected point within the subdomain an evaluation of the function is made, and the sub-domain is dividbd into two smaller ones for more accurate estimation.
This process is continued until the estimated maximum become equal to the evaluated maximum.
For estimation of the maximum of a sub-domain a parameter of uncertainty is introduced. The larger the uncertainty is, the larger the estimated maximum is made. A large estimated maximum means either that the true local maximum is large or that a closer search of the subdomain is required to reduce the uncertainty.
This uncertainty also plays a role of changing searching modes from global to concentrated local search and to checking if the maximum obtained is the true maximum.
Performance of the method is evaluated for six test functions ranging from a unimodal up to a 13-modal one. The results show that the maximum points are obtained with high efficiency and with high reliability.View full abstractDownload PDF (1071K) -
Masato KIMURA1973 Volume 9 Issue 4 Pages 472-477
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSTaylor's equation for tool life has been examined at first from the view point of optimum and adaptive control of machine tools, and then general expressions of differential equations of tool wear have been derived from the mathematical view point. The tool wear ratet is a function of wear, cutting speed, feed rate and so on, which are functions of time. Applications of the method of variational calculus to problems of maximizing or minimizing functionals such as the volume of the metal removed for a certain time interval have been done by use of the differential equations mentioned above, and thus the optimum cutting speed, feed rate etc. can be determined. The meanings of some conventional criterion functions in control operations have been clarified from the standpoint of the differential equations.View full abstractDownload PDF (768K) -
Kamekichi SHIBA, Mikiya KAWAMURA1973 Volume 9 Issue 4 Pages 478-482
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe temperature distribution in a gas can not be correctly measured without compensating for the error due to radiation. The experimental gas thermometers reported in the preceeding paper enable us to know the correct temperature distribution in a gas by eliminating the errors due to radiation.
As examples, the temperature distributions of gas in a boundary layer and in a constant temperature box are experimentally investigated.
In the boundary layer adjacent to a heated body, the heat transfers through the gas by conduction. When the surface of the heated body is plane, the temperature gradient in the boundary layer is constant, and the heat flow can be derived from the measured temperature gradient and from the thermal conductivity of air, which is a known quantity.
The experimental results with the experimental thermoelectric gas thermometer agree very well with the heat flow measured by a Schmidt's belt equipped within the heated body. It is thus experimentally verified that the temperature of gas measured with the experimental thermometer is correct. On the contrary the heat flow obtained with the conventional thermistor thermometer shows an error of about 30%.
The experimental resistance gas thermometer has been used to measure the temperature distribution in a constant temperature box. The correct temperature distribution has been obtained without any effect of radiation from an inner wall higher in temperature than the air.
It is recognized that the experimental gas thermometer can be used successfully to measure the temperature distribution in gas.View full abstractDownload PDF (707K) -
Yoshiharu MASUDA, Bunjiro ICHIJO1973 Volume 9 Issue 4 Pages 483-489
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSRecently direct reading reactance meter have come into general use in various industries. This paper presents a circuit functioning as a direct reading capacitance meter suitable for measuring capacitances of comparatively high-loss materials. This circuit may find a wide range of application in industry. The most important qualities required of this type of equipment are sensitivity to capacitance change, stability during continuous operation for long hours and insensitivity to any resistance change in the materials. Using this circuit, the minimum equivalent parallel resistance of a specimen which has no influence on the capacitance measurement reaches as low as 3kΩ at 2MHz.
The measuring range of capacitance is about 10-2∼103pF. Stability was observed to be within less than 0.5μA in terms of zero drift during several hours' continuous operation. And the required sensitivity is obtained in the detecting circuit. This circuit has the advantage that the measuring part can be used at a distance of some meters from the detecting part.View full abstractDownload PDF (710K) -
Hisashi ITO, Seiji TAKATA1973 Volume 9 Issue 4 Pages 490-495
Published: August 30, 1973
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSAs a substitute for the common method of calibration of radiation pyrometers, which requires either a blackbody furnace provided with a standard thermocouple or a blackbody cavity immersed in a freezing metal of high purity, a simple method having a practically sufficient accuracy is reported.
A small sheet of pure metal, the melting point of which is known, is set in a refractory cavity in a furnace, and the pyrometer to be calibrated, connected with a recorder, is aimed at the metal sheet. The furnace is heated up and, in due course of time, an abrupt increase of the pyrometer indication is observed on the record. The terminal point of this increase corresponds to the temperature of a blackbody kept at the melting point of the metal, as, at the moment of the melting doum of the metal sheet, the condition of radiation in the cavity turns from an incomplete equilibrium to the complete one, because of the disappearance of the specularly reflecting surface of the metal sheet and the consequent disclosure of the roughly reflecting surface of the refractory cavity wall. The departure from the expected correspondence gives the amount of error of the pyrometer indication at the melting point of the metal.
The reproducibility of calibration by this method, estimated from the results of application to three different models of infrared radiation pyrometer, was within ±2K at the melting points of gold and of silver. This method is also applicable to the melting point of copper and to that of palladium, if proper care is taken.View full abstractDownload PDF (2618K)
- |<
- <
- 1
- >
- >|