This paper presents dc-bus voltage control with a three-phase bi-directional inverter for dc distribution systems. The bi-directional inverter can fulfill both grid connection and rectification modes with power factor correction. The proposed control includes two approaches, one line-cycle regulation approach (OLCRA) and one-sixth line-cycle regulation approach (OSLCRA), which take into account dc-bus capacitance and control dc-bus voltage to track a linear relationship between the dc-bus voltage and inverter inductor current. Since both of the approaches require the parameter of dc-bus capacitance, this paper first presents determination of dc-bus capacitor size and an online capacitance estimation method. With the OLCRA, the inverter tunes the dc-bus voltage every line cycle, which can reduce the frequency of operation-mode change and current distortion. The OSLCRA adjusts current command every one-sixth line cycle to adapt to abrupt dc-bus voltage variation. The two approaches together can prevent dc-bus voltage from wide variation and improve the availability of the dc distribution systems without increasing dc-bus capacitance. Experimental results measured from a three-phase bi-directional inverter have verified the feasibility of the discussed control approaches.

No abstract

No abstract

An H H H 1 1 1 method for propulsion controlled aircraft (PCA) design is introduced. The PCA system is adopted for emergency fly-by-throttle control of an airplane with multiple control surface failure. The H H H 1 1 1 methodology is applied to the problem of matching by dynamic compensation the throttle-actuated crippled aircraft and the nominal control-surface-actuated aircraft model. The design procedure is used to synthesize model matching H H H 1 1 1 controllers for an L-1011, a commuter airliner, and a thrust vectoring aircraft models in longitudinal motion and an L-1011 model in lateral motion. Simulation results demonstrate that the concept is workable.

This paper presents a quantum-based algorithm for evolving artificial neural networks (ANNs). The aim is to design an ANN with few connections and high classification performance by simultaneously optimizing the network structure and the connection weights. Unlike most previous studies, the proposed algorithm uses quantum bit representation to codify the network. As a result, the connectivity bits do not indicate the actual links but the probability of the existence of the connections, thus alleviating mapping problems and reducing the risk of throwing away a potential candidate. In addition, in the proposed model, each weight space is decomposed into subspaces in terms of quantum bits. Thus, the algorithm performs a region by region exploration, and evolves gradually to find promising subspaces for further exploitation. This is helpful to provide a set of appropriate weights when evolving the network structure and to alleviate the noisy fitness evaluation problem. The proposed model is tested on four benchmark problems, namely breast cancer and iris, heart, and diabetes problems. The experimental results show that the proposed algorithm can produce compact ANN structures with good generalization ability compared to other algorithms.