All issues
Volume 3 (1967)
- Issue 4 Pages 245-
- Issue 3 Pages 161-
- Issue 2 Pages 83-
- Issue 1 Pages 1-
Volume 3, Issue 1
Displaying 1-10 of 10 articles from this issue
- |<
- <
- 1
- >
- >|
-
Sunao KAWAI, Tadahiro MACHIYAMA, Mutsuo KOIZUMI1967 Volume 3 Issue 1 Pages 1-10
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSFrom the view point of control technique, many studies on the transient performance of heat-exchanger were reported. In most of them temperature is regarded as the input and output variables. But the fluid temperature is seldom handled as manipulated variable. Generally in actual cases, flow-rate of fluid is manipulated to control the output temperature.
The authors tried to describe the transient performance in the actual situations by assuming the flow-rate as manipulated variable and the output temperature as controlled variable.
In this paper, some experimental results and theoretical considerations on the shell & tube type heat-exchangers are reported. The considerations to describe the transient performance are based on the actual construction factors of the heat-exchanger and static characteristics derived by experiments and theoretical calculation.View full abstractDownload PDF (1248K) -
Kunihiko ICHIKAWA, Katsutoshi TAMURA, Katsuhiko HIDA1967 Volume 3 Issue 1 Pages 11-20
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper the authors present two examples of the application of the standard method of solving the optimal control problem, which they have been developing, for the time-optimal control problem of the second order systems; x+2u3x+u2x=u1 and x+2u3x+x=u1. They consider first the relation between the dimensions of control vector and state vector, and show that it is unreasonable to consider three control variables in the second order controlled system in the course of applying the maximum principle, and how these unreasonable problems can be dealt with. It is also shown that a system in which parameters are used as control variables is not a very special kind of system but it is only a type of nonlinear control systems.
In case of the system x+2u3x+x=u1, reverse time trajectories, which are constructed through the application of the standard method, intersect with each other in the limited region of x1x2-plane, and this fact has been giving one an impression that this problem is a very special one. The authors concluded that this phenomenon has come from the fact that maximum principle is a necessary condition for optimality and it does not give a sufficient condition. Complete solution of this problem is also given. Feedback control scheme, which is the main goal of the standard method, is obtained for both of the above second order systems.View full abstractDownload PDF (1049K) -
Syogo MATSUNAGA1967 Volume 3 Issue 1 Pages 21-25
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe author studied the dynamic characteristics of slab heating in the furnace under the considerations of fuel consumption and combustion gas flow characteristics through furnace.
In the analysis of
1. dynamics of heat transfer from combustion gas to slab,
2. dynamics of heat transfer in furnace,
3. dynamics of heat transfer in furnace with recuperator,
the factors which must be considered in the dynamical characteristics of the furnace were so chosen that the existent furnace can be operated in these manners, and these assumptions were so simplified that the analytical procedures go on easily.View full abstractDownload PDF (559K) -
Kensuke HASEGAWA1967 Volume 3 Issue 1 Pages 26-34
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA discrete stationary Markov process can be described by constant coefficient difference equations and analyzed conveniently by the z-transformation in the form of a generating function which reprensents a probability on each time after transition.
The signal flow graph is known as a powerful tool in linear network analysis because it represents network behaviors by Laplace transformed linear differential equations, showing graphically at the same time the cause and effect relationships between signals in the network.
R.W. Sittler proposed to make use of the signal flow graph in showing the transition of states in a Markov process. In his proposal, the generating functions are given as node variables being calculated by the node reduction or by the matrix method.
This paper briefs Sittler's method and develops its application systematically. At first, the Markov process is represented in a signal flow graph. After that, the author presents a new concept in understanding the Markov process. The Markov processes may be viewed as an ensemble of small electrical impulses transferring from node to node with a known transition probability which is expressed as the branch transmittance with a common dead time in the flow graph. The author defines the probability flow graph for the Markov process distinguishing it from the conventional flow graphs.
Using such a concept, the well-known characteristics of the Markov process are explained. Formulas are given to calculate the stationary distribution, the average arrival time and the generating function under a given condition.
In the last section, the transposed probability flow graph is defined for the Markov process involving gains and losses.
As an example, it is shown how to calculate the average expected gain in such a process.View full abstractDownload PDF (1181K) -
Masayuki SUZUKI, Masabumi TODOROKI, Ichiro SUGIURA1967 Volume 3 Issue 1 Pages 35-44
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe performance index sensitivities to plant parameter variation (PS) in optimum-control systems of both open and closed loop types are compared with each other. First, comparison is made with regard to two specific systems, namely a first-order linear stationary plant with the quadratic form performance index and the time optimal control process of a self-adjoint system with bounded norm control. The coefficients of PS of the two types are evaluated numerically in the two processes and are found to be always equal in their magnitudes.
Next, treating the above comparision problem generally, the conclusions are obtained as follows: if the performance index variation is stationary with respect to the variation of control signal around the optimal one and the closed and open loop optimum-control laws are differentiable with respect to the state variables (closed loop case) and the controller parameters (open loop case), the PS of the two types always have same magnitude and therefore the closed loop configuration is siginificant merely when the plant has no parameter variation and is disturbed only by so-called outer disturbances.View full abstractDownload PDF (1301K) -
Akira OHTE1967 Volume 3 Issue 1 Pages 45-53
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSCapacitive type displacement transducer has a number of advantages over the other types of displacement transducers; it is free from contact failure, simple in construction, etc.
But in general, in the use of capacitive transducer, some, disadvantage arises; it's sensitivity depends on source voltage and source frequency, and output frequency is high. Following technique of this transducer these disadvantages are eliminable.
This transducer has built-in oscillator which works with DC power source, and employs two capacitors whose capacitances change differentially. The transducer obtains a DC voltage in proportion to the difference between the two capacitance through its circuit employing diodes. At the same time, it obtains a DC voltage in proportion to the sum of two capacitance and negative feed back is performed so as to keep the sum constant. Thus the transducer is free from the effect of source voltage variation, frequency change, temperature change, etc., and very stable operation is assured.
Employing cylindrical differential air capactior, linear output voltage of the order of ±5.5V (2kΩ output impedance) is obtained for a variation of ±80 deg. angular displacement.
In this paper, principle, construction, circuit and characteristics of the transducer are described. Also some analysis of capacitance to DC conversion circuits are given.View full abstractDownload PDF (1889K) -
Using Concentric Circle Type Grid ElectrodesTomokazu ARAI, Bunjiro ICHIJO1967 Volume 3 Issue 1 Pages 54-61
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe dielectric method for accurate and rapid measurement of moisture content in granular materials is described in this paper. The different points from conventional usual method are as follows:
1) Substitution method, especialy for small capacitance circuit, has no effect of the parallel resistance of specimens using large sized electrodes.
2) The highest sensitivity of the measuring circuit is about 17μA/10-3 pF on a microammeter with a full scale range of 100μA. The circuit is not influenced by the parallel resistance value of specimens by several tens kΩ and over.
A construction and a capacitance measuring method of the concentric circle type grid electrodes are investigated. The penetrated height of the electrical field in the electrode is defined experimentaly. The dielectric property of wheat and sugar are measured by various electrodes.
The most important problem for dielectric measurement determining moisture content in grain is the dispersion of measuring values of capacitance. The dispersion is disccused on insertion of specimens and the size of electrodes. The results of experiments came out as expected.
Finally, moisture content in wheat, granulated sugar and sugar (first-class, second-class) measured by this method are described.View full abstractDownload PDF (978K) -
Koichiro DOI1967 Volume 3 Issue 1 Pages 62-65
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe flowmeter of reaction type is a flowmeter by which the flow-rate of a liquid is obtained by measuring the balancing force acted on the movable tube through which the liquid flows. A device of flowmeter of this type with a cupshaped movable tube has been reported by author and others in this journal. The flowmeter of reaction type here reported is a bee-line flowmeter, in which flow-in and flow-out are of the same direction. In such a flowmeter, the areas of inlet and outlet ends must be different to each other.
Experiment are done with two flowmeters, one of them has two bellows of equal effective areas and the other of unequal effective areas.View full abstractDownload PDF (1008K) -
Akira SUGIYAMA1967 Volume 3 Issue 1 Pages 66-72
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSGaussian random noise, having a uniform power spectrum at very low frequencies, is very useful for simulation studies of random processes which occur in many branches of science.
Noise generation by noisy resistors, diodes, phototubes and thyratrons requires elaborate regulation, and also has disadvantages of nonuniform power spectra at low frequencies and of hum-pickup.
In our experiment to remove such disadvantages, some transformations to the original noise were applied. This transforming technique is the one that simulates the process generating shot effect noise in a diode, by a radioactive source and a low-pass filter. A monostable multivibrator triggered by output pulses of a radiation detector actuated by the radioactive source yields a random phase impulse train with Poissonian appearances. When this impulse train is fed to the low pass filter whose cut-off frequency is low enough in comparison with the rate of occurrence of impulse, a continuous random wave is obtained at the output terminal of the low pass filter.
This process is similar to that of generating shot effect noise in a vacuum tube. As the rate of occurrence of impulse increases, the probability distribution of the output wave approaches normal (Gaussian).
In our experiment, a radium luminous paint and a self quenched Geiger-Müller tube were used as noise source and radiation detector, respectively. This generator has a transistorized true RMS meter to indicate the RMS value of the output wave and an elapsed time meter for maintaining G-M tube.
A χ2 (chi-squared) test for a histogram of the output wave verifies that the hypothesis that the experimental distribution is Gaussian can not be denied.
This method is valuable for practical use in obtaining DC∼50c/s Gaussian noise because of its simplicity, thereby making the maintenance of the noise-generator quite easy.View full abstractDownload PDF (1380K) -
Ryozo KIMIHIRA1967 Volume 3 Issue 1 Pages 73-82
Published: March 30, 1967
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSMany Important problems in engineering and physics require the solution of partial differential equations, such as Laplace's and Poisson's equqtions. In this paper, the method to solve the boundary problems of these equations with high accuracy and high calculation speed is discussed.
To solve the above equations, digital computers and registance networks are generally used. In this paper, the use of the Sampled Signal Analogue Computer which uses discrete analogue signals and lies between the digital computer and the registance networks is also proposed.
The errors of the solution consist of the discretization error ES, the operation error EC, and convergence error EN. These three errors are evaluated, and according to the result, the three apparatus are compared and discussed.
The digital computer is suitable to obtain solution of high accuracy (less than 0.1% in error) for the two dimentional problem.
The registance network is suitable to three dimentional problem and the solution of 0.01% in error can be obtained with the registance network of 0.1∼0.5% in accuracy.
The Sampled Signal Analogue Computer has accuracy of 0.5∼1% error, but the calculation time can be made shorter than the digital computer.View full abstractDownload PDF (1163K)
- |<
- <
- 1
- >
- >|