Advanced Materials Research Vols. 466-467

Abstract: Defect is one of the important factors resulting in gear fault, so it is significant to study the technology of defect diagnosis for gear. Class imbalance problem is encountered in the fault diagnosis, which causes seriously negative effect on the performance of classifiers that assume a balanced distribution of classes. Though it is critical, few previous works paid attention to this class imbalance problem in the fault diagnosis of gear. In imbalanced problems, some features are redundant and even irrelevant. These features will hurt the generalization performance of learning machines. Here we propose PREE (Prediction Risk based feature selectionfor EasyEnsemble) to solve the class imbalanced problem in the fault diagnosis of gear. Experimental results on UCI data sets and gear data set show that PREE improves the classification performance and prediction ability on the imbalanced dataset.

Abstract: The terminal display system is a basic component of digitized tank integrated simulator. The combination of VAPS platform which is installed on PCs with VC++6.0 development environment is the basis of simulation system. And work in practice that can not only reduce design costs but also shorten design cycles. First, the structure, design concepts and realization methods of this simulation system are introduced. The system architecture diagram and the development process are given later. The research of several key issues is focused on, including the display of simulation interface as well as external drivers, and finally the development of the terminal display system for the tank integrated simulator is achieved so that met the demands for training purpose.

Abstract: In this paper, a sliding mode control approach combined with the boundary saturation function approach is put forward and used in a pneumatic force servo system to achieve an exact force control. First, a comparatively accurate model of the system is obtained by using the system identification approach and an analysis is made on the time-varying nature of the model. Then, it is designed by use of the boundary saturation approach, thus overcoming the system instability caused by the non-linearity of the proportional pressure valve and the change of the temperature inside the air cylinder. Finally, the performance of the pneumatic force servo control system is simulated and a comparison is made with the PID control. Results show the feasibility and effectiveness of the approach.

Abstract: In the paper, the idle speed control problem for hybrid vehicles engine is investigated. The mathematics equations about dynamic characteristic of engine’s primary subsystems are presented in the view of the physical characters of engine. A new back stepping sliding mode controller was designed based on the simplified model of engine, and the simulation results show the great improvement of engine control performance, especially to overcome the disturbance.

Abstract: The characteristics of filling machine pipe orifice laminar flow, turbulent flow and the energy loss in fluid were studied in this paper. And, the influence of the characteristics on the forming droplet in pipe orifice and the setting-off for fluid were analyzed too.

Abstract: At present the design of the aircraft control surface simulator is all about how to eliminate the redundant moment and increase the tracking accuracy of load simulator. In this paper, for the abnormal phenomena appeared in the semi-physical simulation loop of the load simulator, a simulation analysis was made to illustrate that there is a strong coupling effect between the load simulator and the actuator simulation loop, and increasing the tracking accuracy of load simulator simply can not eliminate the effect on the semi-physical simulation system. Therefore, there are some limitations in the design of load simulator based on its performance indexes only.

Abstract: In this paper, a T-shaped defected microstrip structure (DMS) which can provide rejection band and slow-wave is studied in detail, then two low pass filters based on T-shaped DMS are proposed, designed and manufactured. Consistent simulation and measurement results have confirmed the design concept and indicated that the proposed filters have a series of advantages such as simple in design, little in electromagnetic leak, and also compact in size which facilitates the integration to some degree.

Abstract: This paper introduces design of a multimode controller for a Tactical Ballistic Missile autopilot that combined optimal control method with classical frequency domain analysis. In the paper, optimal control method is used to system structure design and to determine gain coefficient, frequency constraints in classical control theory are used to limit the weighting coefficient matrix of objective function in order to ensure system robustness. The simulation results show that the corresponding system responses have better quickness and some robustness, but the overshoot becomes a bit larger with contrast to traditional method.

Abstract: This paper presents modeling random drift by traditional time series method. In the modeling process, the real time average algorithm is proposed which can extract the constant drift of the MEMS gyroscope effectively to get random drift, and a modified recursive extended least squares (RELS) method for parameter estimation of the autoregressive moving average models (ARMA) is presented. The modified RELS algorithm consists of two-stage RLS algorithm which can on line be implemented, and has the fast convergence rate. After modeling, the corresponding Kalman filter is designed to make compensation for gyroscope random drift. The compensation results for the practical testing data of the MEMS gyroscope show that the model established can reflect the gyroscope random drift tendency with reasonable accuracy, and the gyroscope random drift can be effectively reduced by the filtering method presented.

Abstract: In order to obtain more output power of photovoltaic (PV) array, which depends on solar irradiation and ambient temperature, maximum power point tracking (MPPT) techniques are employed. Among all the MPPT strategies, the Perturb and Observe (P&O) algorithm is more attractive due to the simple control structure. Nevertheless, steady-state oscillations always appear due to the perturbation. In this paper, a new MPPT method based on BP Neural Networks and P&O is proposed for searching maximum power point (MPP) fast and exactly, and its effectiveness is validated by experimental results using hardware platform based on microcomputer.