This study intends to improve the accuracy of the concentration and flow measurement system for binary gas mixtures that was developed by two of the present authors1). This system is composed of a venturi tube and a laminar flowmeter connected in series. A basic examination on the viscosity estimate for binary gas mixtures and a preliminary measurement of the flow coefficient of the venturi tube are carried out. The procedure of concentration calculation based on the basic study is also shown. The estimate accuracy of the mole fractions of component gases and the flow rate of binary gas mixtures are experimentally examined using three kinds of mixed gas: CO2-Air, CO2-Ar and Ar-He. The results indicate that the mole fractions can be estimated with accuracy of 5% and the flow rate can be estimated with accuracy of 2%.
In this paper, we consider an optimal control problem for a linear discrete time system with stochastic parameters. Whereas traditional stochastic optimal control theory only treats systems with deterministic parameters with stochastic noises, this paper focuses on systems with stochastic uncertain parameters. We derive an optimal control law for a novel cost function by which the designer can adjust the trade-off between the average and the variance of the states. Finally, a numerical simulation shows the effectiveness of the proposed method.
This paper considers the system identification problem for hysteresis systems. This problem plays an important role in achieving better control performance, because many actuators have hysteresis property. This paper proposes a method to identify linear dynamical systems having input hysteresis property of backlash type. The method is based on particle filter, which is known for its applicability to a wide class of nonlinear systems. Numerical examples are given to demonstrate the effectiveness of the proposed method in detail. Furthermore, experimental validation is performed for a DC servo motor system.
This paper proposes a design method of strongly stable generalized predictive control (GPC) focused on closed-loop characteristics. GPC can be extended by coprime factorization, and the extended controller can be designed to be stable by selecting newly introduced parameter. That is, strongly stable system, which means both the closed-loop system and its controller are stable, can be obtained. Although the authors have considered the design method of strongly stable system using coprime factorization, the steady state of output has not been considered when feedback loop was cut. In the case that the controlled plant is stable, the steady state output of strongly stable system is stable even if feedback loop is cut. But for safety, it is desirable that the steady state of output becomes as close to the steady state of closed-loop output as possible even if feedback loop was cut. Therefore this paper explores a design method of strongly stable GPC focused on closed-loop characteristics by algebraic calculation of newly introduced parameter in the extended GPC. The proposed method has the feature that the steady state of output becomes the same as the steady state of closed-loop output even if feedback loop is cut.