All issues
Volume 6 (1970)
- Issue 6 Pages 483-
- Issue 5 Pages 385-
- Issue 4 Pages 289-
- Issue 3 Pages 195-
- Issue 2 Pages 97-
- Issue 1 Pages 1-
Volume 6, Issue 5
Displaying 1-14 of 14 articles from this issue
- |<
- <
- 1
- >
- >|
-
Zenzo MINOWA, Tadashi KAWASAKI, Teruko INAMATSU1970 Volume 6 Issue 5 Pages 385-394
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn order to establish a standard in the alcoholometry in Japan, a new alcoholometric specific gravity table is compiled at the National Research Laboratory of Metrology. The compiled tables give specific gravities of five places of decimals corresponding to every integral's percentage by weight in alcohol-water mixture at intervals of 5°C in temperature range from -20°C to 30°C.
The table has the following characteristics in the method of its compilation:
1) The specific gravity of pure ethyl-alcohol is determined by extrapolating the relation between specific gravities and concentrations of nearly pure ethyl-alcohol, whose water contents are measured by the Karl-Fischer method.
2) The relation between alcohol concentration and specific gravity of alcohol-water mixture is expressed by a polynomial of 11 degrees for 0, 5, 10, 15, 20, 25 and 30°C, by a polynomial of 7 degrees for -5 and -10°C and by a polynomial of 6 degrees for -15 and -20°C; the coefficients in the polynomials being determined statistically from many measurement values on concentration and specific gravity of alcohol-water mixtures.View full abstractDownload PDF (1343K) -
Sukita NAKAHARA1970 Volume 6 Issue 5 Pages 395-400
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSMirror contamination is one of the most difficult problems in measuring dew or frost-pointcontinuously by a dew-point instrument which detects dew or frost by specularly reflected light. In a recent paper, as a method of minimizing the effect of mirror contamination, the light diffusely scattered normally to mirror surface was adopted for reference of a balanced photocell bridge which was the detector of dew or frost. In order to obtain more effective results, near infrared rays are employed as specularly reflected light. The reason for it is that the intensity of specularly reflected light from a rough surface is related to σ/λ (Porteus formula), where λ is the wavelength of the light and σ is the r.m.s. roughness of the surface.
That is, in the case of near infrared rays, the specularly reflected light is strongly scattered and decreased in intensity by islands of frost (σ>λ), but scarcely decreased by monoethanolamine (M.E.A.) deposit (σ<λ). The r.m.s. roughness of M.E.A. deposit increased inproportion to t1/2 under the usual measuring condition.
By those experimental results and Porteus formula, the relation that the operating interval allowing a continuous frost-point measurement increases in proportion to λ2 is derived. The new type dew-point instrument is constructed using together the infrared specularly reflected light and the diffusely scattered one, and the interval of measuring frost-point is prolonged about 10 times or more.View full abstractDownload PDF (740K) -
Taro TAJIMA1970 Volume 6 Issue 5 Pages 401-407
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper we analyse the important technical problem when the first order temperature coefficient α of the standard voltage using zener diodes is required to be as small as 10ppm/degree, and then present a solution.
A semiconductor element named the multi-diode is made on trial, applying the temp. characteristics of the forward biased silicon diode. Two kinds of its designs are shown in this paper.
We can regulate the temp. slope of the forward voltage of the multi-diode step by step.We manufactured on trial the mufti-terminal type compensated zener diode by using the multi-diode. As a result of our experiment, we can control the first order temp. coeff. α of the standard voltage by the selection of the terminal of the multi-terminal type comp. zener diode, and then α become less than 5ppm/degree. This device is suitable to supply the standard voltage for the P.I.D. controllers and the digital meters etc.View full abstractDownload PDF (3087K) -
Shigeru ARAMAKI, Akimasa ISHIDA1970 Volume 6 Issue 5 Pages 408-414
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn a modern control system such as the numerical control system and the computer control system, servomechanism is often necessary to convert digital data to a real shaft angle. A synchro has been used as a D-A converter and a detector of an angle in such a servomechanism. But the synchro is not easy to use as a D-A converter when the given digital data are decimal or binary.
So the authors propose to use a 5-phase synchro and an 8-phase synchro.
The 5-phase synchro has five stator windings corresponding to the decimal number and the 8-phase synchro has eight stator windings corresponding to the binary number.
These digital synchros (a 5-phase synchro and an 8-phase synchro) have advantage in comparison with a usual synchro because D-A conversion is rather simple when digital data are decimal or binary.
This paper presents the principle and practice of these digital synchros and examines the problem of multi-speed system.View full abstractDownload PDF (950K) -
Kikuo SAKAMOTO1970 Volume 6 Issue 5 Pages 415-421
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper discusses the problem of estimating state variables and nonstationary parameters for adaptive control system in the presence of random disturbance and mesurement noise.In this paper, a parameter is assumed to be “Frequency Band Limited Process”, so that the parameter equation can be described approximately with m prior parameters and white gaussian noise by introducing the sampling theory. This approximate parameter equation is applied to the nonlinear filtering theory for estimating a parameter from indirect data with the presence of noise. The number of m is changed properly according to the nonstationary variation of the parameter.
Experimental computer simulation of 1st order system confirms the practicability of the method.View full abstractDownload PDF (851K) -
Seihaku HIGUCHI, Kokichi TANAKA1970 Volume 6 Issue 5 Pages 422-428
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSMany methods have deen developed in consideration of the importance of identification. In the recent years the high accuracy devices such as the operational amplifier and the multiplier have been developed. Though the identification methods using analog computation have been long reported, in this paper the identification method of the low order process is treated re-evaluating the merits of analog computation.
First, the impulse response function of the process is expanded by means of orthnormal function filter, and the order of this process is decided from the discriminant which is expressed by the relation among the above expanded coefficients. For example, the form of this process is designated as the second-order lag system. At the same time, the computation rules which determine the values of the unknown parameters are also given. Finally, the approximate deviations from the true values of expansion coefficients are estimated.View full abstractDownload PDF (830K) -
1970 Volume 6 Issue 5 Pages 428
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSDownload PDF (78K) -
Kiyotaka SHIMIZU1970 Volume 6 Issue 5 Pages 429-438
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSOur purpose is to obtain necessary conditions that the optimal pure strategies for a minimizer u and a maximizer υ must satisfy. By a static game is meant a so-called continuous game, and by a dynamic game an ordinary differential game.
We obtained previously the necessary conditions for the continuous game such that the pay-off function F(u, υ) and the inequaiity constraints g(u, υ)≤0 are functions of only u and υ. The obtained conditions were Kuhn-Tucker conditions for the game. They were derived by considering a pair of Lagrangian functions. In this paper we extend these conditions to the case when there exist equality constraints f(x, u, υ)=0 where x is a state vector. We prove also a duality theorem for the game.
Next, we consider a differential game with inequality constraints g(u, υ)≤0, and use a technique with a pair of Lagrangian functions as we did for the static game. Following Berkowitz, we base our analysis on the calculus of variation in obtaining the necessary conditions for the optimal minimizer u0(t) and maximizer υ0(t) in correspondence to Euler equations, Clebsch condition etc. for simple minimization.
In addition, the maximum principle for the game is derived in the case when no inequality constrains exist.
Finally we study relations between a static problem and a dynamic problem, namely, between Kuhn-Tucker conditions and Euler, Clebsch conditions. We conclude that the former conditions are indeed derived from the latter conditions.View full abstractDownload PDF (1232K) -
Hikaru INOOKA1970 Volume 6 Issue 5 Pages 439-447
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSLinear time-invariant systems have been studied by using both the frequency domain method and the state space method. However, there has been little effort devoted to comparing the characteristic features of these two methods.
This paper is concerned with one of such problems that have been left aside. The problem is to explain the difference between the effect of state variable feedback and the effect of output feedback using a compensating element. The compensating element is so chosen that the closed-loop transfer function of the output feedback system is identical with that of the optimal regulator with state variable feedback. It is shown that the output feedback system is quite different from the system with state variable feedback and therefore the effects of these two ways of feedback are distinct from each other.View full abstractDownload PDF (1043K) -
Shigemi TAGAMI1970 Volume 6 Issue 5 Pages 448-453
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper shows the practical procedure how to construct the nonlinear filter with nongaussian nonstationary random input by using the orthogonal function weighted by the probability distribution.
As for the optimizing criterion, the mean squared value of the difference between the actual and the desired output signals of the filter is taken for the convenience of mathematics. And the mean squared value of the error is evaluated quantitatively.
The technique described here is a means for circumventing difficulties in the mathematical treatment, and in the actual design of the optimum nonlinear filter, it becomes considerably simplified to decide which design is physically realizable.View full abstractDownload PDF (706K) -
Katsutoshi TAMURA, Kunihiko ICHIKAWA1970 Volume 6 Issue 5 Pages 454-458
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA new iterative method based on gradient function is proposed for solving bang-bang control problems. The proposed algorithm is outlined as follows. Assume any bang-bang function as an initial guess and calculate the gradient function for the control. The control function is improved succesively via the gradient function also in this method, but unlike the conventional gradient method, the time interval in which the control is improved is small and control is varied as large as admissible during that interval. Thus the improved control is again a bang-bann control.
The control after improvement is shown to give surely a less value of cost functional than before improvement. The convergent control obtained from this algorithm satisfies the local optimal condition. It is examined whether the characteristics of the gradient function with continuous control function will still hold for this case. With regard to this point, the derivation and calculation of the gradient function are also mentioned.
The usefulness of the proposed algorithm is illustrated by numerical examples. It should be remarked that the algorithm needs no assumption of switching times of optimal control function.View full abstractDownload PDF (567K) -
Ryuichi OKA1970 Volume 6 Issue 5 Pages 459-464
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSControl problems defined over fixed finite time interval are discussed and some relations between stability, reachability and optimality are described. A typical problem in the deterministic dynamical systems is the “regulator problem” discussed by R.E. Kalman, but he solved it only in the case of linear systems without the constraints on control forces. Also in Kalman's solution, the time interval is free. Here, we discussed a regulator problem under the consideration of fixed finite time interval (α, β, T)-BIBO-stability (|ui(t)|≤α(i=1, 2, …, n) implies |yi(t)|≤β(i=1, 2, …, m) for all t∈[0, T]) in both the linear and nonlinear cases. In the stationary systems with some suitable conditions, it is concluded that a certain bounded condition on the norm of f(x(t)) for all t∈[0, T] is required as a sufficient condition to construct a regulator in the sense of (α, β, T)-BIBO-stability.
In the stochastic systems, it is very natural to desire that the control satisfies the given stochastic stability condition (defined by H.J. Kushner) and, at the same time, the optimality condition for the prescribed performance index.
However, a control satisfying the stability fails to satisfy the optimality, and vice versa. We discussed this problem specially in a linear, constant coefficients stochastic system written in terms of Ito's stochastic differential equation for a quadratic performance index. Let the given stochastic stability be (ρ, m, V(x(t)), T). We assume that there exists a matrix L such that 0≤V (x(t))≤x'Lx for t∈[0, T], then the non-positive condition of the matrix, composed of matrices in the system, the performance index and the matrix L, combines the stochastic stability and the optimality.View full abstractDownload PDF (662K) -
Nariyasu MINAMIDE, Kahei NAKAMURA1970 Volume 6 Issue 5 Pages 465-472
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn the present paper, the minimum effort control problem under phase constraint is considered by functional analysis approach. The abstract version of the problem can be stated as follows; Let X, Y and Z be real Banach spaces. Let S, from X into Y, and T, from X onto Z, be bounded linear transformations. With ξ∈Y and η∈Z, find an element (if one exists) u∈X satisfying η=Tu and ||ξ-Su||≤ε, while minimizing ||u||. Throughout the paper, the pair (ξ, η) is assumed to be regular in the sense that there exists at least one element u∈X which satisfies the constraints η=Tu and ||ξ-Su||<ε. Initially X, Y and Z are taken to be arbitrary Banach spaces and the necessary and sufficient conditions for the existence of the solution to a regular pair are discussed.
It is then required that either X be a reflexive Banach space, or each of the three Banach spaces be the conjugate of another normed space. After having established the existence of the solution, the primal problem is transformed into the dual one in the conjugate space by virtue of the Kahn-Banach produced hyperplane of support to a convex body. As a result, the optimal solution is completely characterized in terms of the hyperplane. Incidentally, the conjugate equation and the conjugate problem are derived. It is shown that this conjugate problem provides a useful computational procedure forr obtaining the explicit solution through an example.View full abstractDownload PDF (1098K) -
Akira MOHRI1970 Volume 6 Issue 5 Pages 473-481
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper, a computational method for the optimal control of a linear sampled-data control system is described.
In the linear stationary dynamical system x=Ax+Bu, the problem of minimizing the integral type criterion function which has the quadratic form as its integrand J=∫tf0(x'Qx+u'Ru)dt, isconsidered. In this case, the differential equation of the system, the value at each sampling point of variable x0 introduced by the equation dx0/dt=x'Qx+u'Ru and the so-called adjoint equation etc. are reduced to recurrence equations with constant coefficients. And if these coefficients are determined beforehand, the iterative computations to determine the optimal value of u are carried out not by numerical integration. And these coefficients, generally, are computed by numerical integration. But if the method described in this paper is used, computations of numerical integrations can be considerably reduced. Therefore, the time for computations can be shortened.
In this paper this computational method is described.View full abstractDownload PDF (929K)
- |<
- <
- 1
- >
- >|