This paper deals with the problem to decide how much flexibility should be given to the manufacturing system, when and how replacement in the facilities should be carried out to cope with the demand variation over the service life of the system. To make these decisions, the evaluation items, such as the cost for equipments, the operating cost, the cost for replacements and revenues, are considered as a cash-flow over the planning horizon. A method based on the optimality principle in dynamic programming is proposed for deciding the optimum reconstruction plan in which the net present worth of the cash-flow is adopted as the criterion for optimality. A numerical example is given to demonstrate the advantage of our method for reconstructing the system compared with the usual method in which the flexibility of the system is only considered to cope with the demand variation.
This paper gives an LQ optimal controller design method which is based on the iteration of experiment and controller re-design from the viewpoint of joint design of identification and control. This method provides us a systematic method of changing the weighting matrices to achieve better control performance based on the experimental data. Furthermore, the effectiveness of the proposed method is shown by experiments using an inverted pendulum system.
Automatic Gage Control (AGC) System is very important for cold tandem mill rolling of steel. We studied a new Gage and Tension Control System, based on Inverse Linear Quadratic (ILQ) Design Theory. At first, we designed the ILQ controller for the actual plant, the first stand of cold tandem mill. Next we analysed the essence of ILQ control law, and modified the original controller. Finally we installed the modified controller to the actual plant, and achieved very high accuracy of cold strip gage. The most important feature of this study is the analysis of the essence of ILQ controller based on the parametric representation of control law.
In this paper, we apply a turn over method for distributed parameter systems to a widely used heatexchanger system. The control law has both desirable properties of a pole assignment and an optimal control law, and the feedback gain is obtained without solving an infinite dimensional Riccati type differential equation. By this method, open loop poles existing in the right hand side of the turn over line are moved to the axial symmetry positions respecting to the turn over line drawn parallel to the imaginary axis, thus all closed loop poles are located in the specified region with desired degree of exponential stability. The distributed feedback law for a distributed heatexchanger system is proposed by using the technique without any approximation. We confirm by experiments that the constructed system has a short rise time and a short settling time. The state variables needed in the feedback law is estimated by a distributed observer. The distributed parameter observer is approximated by the method of weighted residuals and gives a good estimated value with only four sensors in the experiment.
In this paper, a method to control attitude of a space robot by its arm motion is considered. The proposed method is to generate large attitude change by the cyclic motion of the arm. Usually, only path planning of the arm has been treated for this kind of control. The method to modify the arm path by feedback of the attitude error is proposed in the paper. Numerical simulations have been executed to verify the validity of the feedback law. The results show that the method is effective to reduce the attitude error even when the discrepancy of the mass property exists between the space robot model for the path planning and that for the simulations.