All issues
Volume 7 (1971)
- Issue 6 Pages 499-
- Issue 5 Pages 393-
- Issue 4 Pages 295-
- Issue 3 Pages 197-
- Issue 2 Pages 99-
- Issue 1 Pages 1-
Volume 7, Issue 5
Displaying 1-14 of 14 articles from this issue
- |<
- <
- 1
- >
- >|
-
Pyong Sik PAK, Yutaka SUZUKI, Katsuhiko FUJII1971 Volume 7 Issue 5 Pages 393-400
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSBy applying the modern control theory, we could design a state feedback control system of a multivariable process as well as a single-input/single-output process. However, in existing processes all the internal state variables are rarely accessible. Therefore in order to realize the optimal feedback control, it becomes necessary to estimate or reconstruct the state variables of the process from its input and output signals which are directly measurable. In order to meet this necessity, the following three methods have so far been proposed i.e., the Heq method, the method by a process model and the method by an observer. In this paper, the optimal feedback control systems by utilizing these three methods are introduced. The mutual relations of these methods are discussed along with the similarities between these control systems and classical ones. In addition, these control systems are compared in the aspect of the sensitivities to process parameter variations. It is verified that the observer method is superior to the Heq method in realizability and also to the model method in the following two aspects: (1) no uncontrollable mode is produced, and (2) the controlled variable can be made insensitive to the change of the process parameters.View full abstractDownload PDF (1032K) -
Nariyasu MINAMIDE, Kahei NAKAMURA1971 Volume 7 Issue 5 Pages 401-409
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn solving optimal control problems, there arise situations in which Pontryagin's “Maximum Principle” may provide no information for determining the optimal controls. Such situations are called “Singular” and the corresponding solutions “Singular controls”. The main objective of the present paper is to study the necessary and sufficient conditions for the existence (or non-existence) of singular controls to that class of control problems which, when suitable function spaces are introduced, can be regarded as special versions of the follwing Banach space minimization problem. Let X, Y and Z be real Banach spaces. Let T: X→Y and S: X→Z be bounded linear transformations. Let M be a closed convex body in Z.
[Problem F2] Given ξ∈Y and η∈Z, find an element u∈X statisfying (1) ||u||≤ρ and (2) η-Su∈M, while minimizing ||ξ-Tu||.
Throughout the paper, it is assumed that η∈Z is a regular element, i.e., there exists at least one control u satisfying ||u||≤ρ such that η-Su∈int (M) (the interior of M). Existence theorem, uniqueness theorem, and necessary and sufficient conditions for both optimality and nonexistence of singular controls are demonstrated in a concise manner by using techniques of functional analysis.
Much stress is placed upon application of the theory to final value problems, time optimal problems and minimum fuel problems. Main results obtained to these problems can generally be phrased as follows. In order for the j-th component (u0)j(t) of the optimal solution u0(t) to be a Bang-Bang control (whence free from being singular), it is necessary and sufficient that the value of a criterion functional under consideration be made strictly smaller if the j-th component uj(t) of the control input u(t) is released from the amplitude constraint, while the other components remaining constrained.View full abstractDownload PDF (1221K) -
Masakazu OSHIRO, Takayoshi NAKAMIZO1971 Volume 7 Issue 5 Pages 410-416
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSAn approximate method is presented in this paper for determining a linear optimal control, under noisy observation, for a linear dynamical system with state-dependent noises. Since the state-dependent noise renders the optimal filter nonlinear, the so-called separation theorem will not be valid. It is primarily assumed in this paper that the controller consists of a linear optimal filter and a linear feedback gain matrix to be determined in a certain optimal fashion.
The filter can be derived independently of the optimal control but the optimal control requires knowledge of the filter dynamics. Thus it is shown that the solution of two-point boundary value problem is required for determining the filter gain and the feedback gain matrices. An approximate procedure circumventing a difficulty in computation of two-point boundary value problem is also proposed. Finally, a simple example is numerically solved for the comparative study.View full abstractDownload PDF (877K) -
An Analytical Method for Nonlinear NetworksKiyotaka SHIMIZU, Yuichiro ANZAI, Yoshio HAYASHI1971 Volume 7 Issue 5 Pages 417-426
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn general, fluid supply systems are described as nonlinear, dynamic, multi-terminal network systems with distributed sources, sinks and holders, in which the sources supply fluid based on demands at the sinks. Accordingly, it is desired to establish some mathematical method for determining the optimal supply curves which satisfy the predicted demand curves and minimize some objective function by utilizing the holders effectively.
The ordinary Hardy-Cross' method for network system simulation is extended to the fluid supply system with holders. The problem formulated is analyzed by discretizing the continuous process so that a nonlinear programming technique is applicable. An algorithm of calculating the optimal control is developed by combining the extended Hardy-Cross' method and the Zotendijk's method of feasible directions.View full abstractDownload PDF (1372K) -
Successively Heightening Order MethodMinoru HIROTA1971 Volume 7 Issue 5 Pages 427-433
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA non-iterative algorithm is presented for obtaining the time optimal control in a process with feedback system. The application of the results should be limited to the case where the dynamics of the process are represented by the equations,
Xn=Xn-1, Xn-1=Xn-2, ……, X1=u (|u|≤1).One of the state variable can be limited in a region, |X1|≤L.
Successively Heightening Order Method is introduced to find the sign or zero of the control function without iteration. The method employs the synthesis of (n-1) th order system to give the time optimal control of n th order system. At first the synthesis is applied to the mathematical model to estimate X1E=0, X2E=0, …, XnE≠0. And then the control of that moment, u=-sgn (XnE), is applied to the process to decrease |XnE| towards zero. After a sampling time the procedure is repeated from the first step. If XnE becomes zero, the system degenerates to (n-1) th order system.View full abstractDownload PDF (898K) -
Ryuzo YOKOYAMA1971 Volume 7 Issue 5 Pages 434-440
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSBased on the generalized definition of an equivalent system, two phase variable canonical forms are given for linear multi-inputs and multi-outputs systems, i.e., the canonical form for control and the canonical form for observation. It is shown that a characteristic quantity called the stage sequence can be defined for each given system to uniquely determine the structure of its canonical form. Having very compact structures, these canonical forms give excellent representations of the physical and mathematical composition of systems and are particularly convenient for analyses of optimal control problems.View full abstractDownload PDF (716K) -
Takeshi SAITO1971 Volume 7 Issue 5 Pages 441-449
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper presents a fundamental theory of parameter identifiability for stochastic dynamical systems and its application to the case of stochastic linear dynamical systems. The most important result established in the former part of this paper is the fundamental theorem of identifiability which combines system equivalence and unique identifiability. In the latter part of this paper, three conditions of the unique identifiability of the parameters of stochastic linear dynamical systems are derived by using the above result and auxiliary lemmas. One of the results shows that under a considerably general condition, the transition matrix of linear system without input cannot be uniquely identified if the dimension of the output vector is lower than that of the state vector.View full abstractDownload PDF (1169K) -
Setsuo SAGARA, Masanori SUGISAKA1971 Volume 7 Issue 5 Pages 450-457
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSAn algorithm for nonlinear system identification via quasilinearization method is presented in this paper. It is assumed that a plant to be controlled may be modeled by a system described by nonlinear ordinary differential equations, and the observations may be expressed as the nonlinear functions with additive noise, involving states and unknown parameters.
The unknown parameters of nonlinear plant are estimated through the least mean square error criterion and the quasilinearization method, by which the nonlinear ordinary differential equation is linearized. The algorithm presented is an extension of the method proposed by B.R. Eisenberg and A.P. Sage, which is applicable for only nonlinear system identification when the observation is given as the linear combination of states.
On the other hand, one of the difficulties of the quasilinearization method is the small range of convergence. In order to investigate the range of convergence, a relation between the initial approximations by quasilinearization and the convergence is briefly shown with a simple example.
In the case of time-varying parameter system, the estimation is performed with approximating the variable parameter by means of a polynomial of time.
The examples demonstrate that the algorithm is an effective method of system identification for a variety of problems.View full abstractDownload PDF (874K) -
1971 Volume 7 Issue 5 Pages 457
Published: 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSDownload PDF (74K) -
Tomokazu ARAI1971 Volume 7 Issue 5 Pages 458-466
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSValues of dielectric constants were measured with mixtures of two organic-compound liquids. Sixty-seven combinations of two out of thirty different organic-compound liquids whose dielectric constants were ranging from about 2 to 40 were subjected to the measurement. Results of the measurement show that the dependence of the dielectric constant of the mixture liquid on its volume percent of the component liquid can be divided into six groups that can be expressed by six known theoretical relations. This will provide a guide for selecting appropriate combinations that have an acceptable linearity from application point of view over the entire range of the volume percent of composition. The temperature dependence of the dielectric constant ranging 5 to 10 at the room temperature was measured with thirty-one mixture liquids. The results are presented in a chart. This chart can be conveniently used in selecting an appropriate mixture liquid when the values of the dielectric constant and its temperature coefficient are specified.
It is intended that the data presented in this paper will provide useful information for the liquid-immersion method of measuring the dielectric constant of powder and grain materials.View full abstractDownload PDF (1092K) -
Kamekichi SHIBA, Jin-ichi KITAMURA1971 Volume 7 Issue 5 Pages 467-472
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe gas flowmeter here reported is a movable L tube flowmeter of angular momentum type with a thermometer and a manometer.
Such a flowmeter enables us to know the mass flow rate of a gaseous substance by measuring the torque on the movable L tube, the temperature and pressure at the inlet (or outlet) end of the tube.
The theory of the flowmeter assumes the uniformity of velocity at the inlet (or outlet) end. When this assumption is invalid, the discharge coefficient becomes necessarily to correct the error. The experimental results show that the discharge coefficient for gas is equal to that for liquid, and the latter can be known experimentally with liquid.View full abstractDownload PDF (656K) -
Katsumi OHNIWA1971 Volume 7 Issue 5 Pages 473-479
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper treats a thyristor servo-amplifier applying a digital technique. The principle of this amplifier is as follows. At first, an input signal is transformed into a binary signal of four bits. By the transformed input signal, thyristors which are connected with four DC supplies of 20E0V, 21E0V, 22E0V and 23E0V respectively, are switched on or off. For example, when the input signal is three volts, i.e., (0011) volts in the binary form, the thyristors connected with the sources of 20E0V and 21E0V are switched on and the others off. By the addition of 20E0V and 21E0V, the output signal becomes 3E0V. In this manner, we can obtain the output signal of the transformed input signal times E0.
The merits of this amplifier are that the control circuit is constructed by combining simple logic circuits, and that the input vs. output relationship is nearly linear. The upper frequency depends on the frequency of the AC source which produces the DC supplies. Using the AC source of 50Hz, the upper frequency is about 10Hz. However, the step response time is very short, being approximately equal to the switching time of the thyristors.View full abstractDownload PDF (1308K) -
Yo IKEBE, Takeshi NAKADA1971 Volume 7 Issue 5 Pages 480-487
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSRecently, electro-hydraulic servovalves have come to be used widely in hydraulic control, but the complexity of their construction is unfavourable. Needless to say, servovalves without any moving parts are desirable from the viewpoint of fast-response, efficiency and reliability.
So, the purpose of this research, as the first step to the desirable servovalves, is to develope a new servovalve adopting a piezoelectric flapper directly actuated by an electric signal. A piezoelectric flapper, adopted instead of a conventional torque motor and a flapper, consists of piezoelectric slabs and a steel slab. As the result of the preliminary experiments, it was clarified that this flapper could be practically used if the adverse effect of its inherently strong hysteresis is removed.
In this research in order to give the flapper equivalently linear characteristics, the input signal to the flapper is given in the form of a pulse width modulated wave. However, in this case the output of the pulse width modulator consists of the various frequency components depending on the input and the carrier frequency. It is well-known that the signal cannot be accurately transmitted, when the ratio of the carrier frequency to the input frequency becomes small. So, one of the most important points of this paper is to establish the conditions of the accurate signal transmission in the P.W.M. mode operation.
As the results of the experiments on the new servovalve, it has become clear that in spite of the strong hysteresis, the new flapper has equivalently linear characteristics, when driven by the pulse width modulated wave from the modulator. Futhermore it has been clarified theoretically and experimentally that for accurate signal transmission through the modulator, the carrier frequency must be higher than seven times the highest frequency of the input signal components.View full abstractDownload PDF (1040K) -
Iwao MORISHITA, Ryuji TAKANUKI1971 Volume 7 Issue 5 Pages 488-496
Published: October 30, 1971
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA nonsupervised learning algorithm for multi-category pattern classification is presented, and its steady-state behavior is investigated by means of analysis and simulation. The algorithm assigns each pattern in an input sequence x(t) to one of M possible categories by compairing the M values wi(t)'x(t), i=1, 2, …, M, where wi(t) is a variable weight vector adjusted according to a rule of differential-equation type. The algorithm minimizes a certain form of criterion function. Analytical techniques are developed for finding the steady-state solutions of the differential equation and for determining the stability of each solution found. By applying the techniques, the steady-state behavior of the algorithm to three types of patterns is analyzed in detail. It is shown that although the algorithm does not necessarily yield a unique solution, every solution yielded is associated with a reasonable classification such that one cluster or some adjacent clusters in the pattern distribution correspond to one category. Even if the total number of clusters in the distribution is smaller than M, no cluster is divided between two categories. Converging to the origin, unnecessary weight vectors vanish in the steady-state. These results are verified by digital computer simulations.View full abstractDownload PDF (2014K)
- |<
- <
- 1
- >
- >|