Search

Blade is a crucial part in steam turbine that chronically exposes to an atmosphere of high temperature, humidity, and pressure, which prone to cause fatigues in the structure. The root of a turbine blade is in undulation shape giving rise to the difficulties for inspection with conventional probes. This work proposes a novel flexible array eddy current testing (ECT) probe for the turbine blade root inspection. The proposed probe employs an in-plane differential scheme to reduce the common mode noise and increase the sensitivity to small defects. It works with three modes to improve the reliability of detecting defects in different orientations. It keeps the lift-off almost constant when scans the undulation shape structure, as the shape of the sensor array is flexible and is bended according to the sample surface. A prototype probe has been developed and tested by inspecting of a blade sample with artificial and nature defects. The probe can identify defect with dimensions length×wldth×depth=z.u mmxu.z mm×0.2 mm with various orientations and locations. The experiment results demonstrate that the probe has high sensitivity for defects at the root of turbine.

The missing of measurement is a phenomenon often happens in power system. In this paper, a method based on low rank matrix completion theory is proposed to recover the missing load data. The low rank matrix is constructed based on the linear approximation principle. Then the low rank matrix completion algorithm is derived based on the linearized Bregman method. Based on the real data of power grid, the method is tested and the error is calculated. The comparison of the error of recovered load data is made between low rank matrix completion algorithm and spline interpolation and extrapolation. The result shows that the algorithm based on low rank matrix completion is more effective and has stronger adaptability.

This paper addresses the problem of adaptive detection of a signal of interest in presence of Gaussian disturbance with unknown covariance matrix. The covariance matrices of the primary and the secondary data share a common structure while having different power levels. A Bayesian approach is proposed here, where the structure are assumed to be random, with an appropriate distribution. Moreover, we assume that the cell under test (CUT) contains a fictitious signal orthogonal to the nominal steering vector under the null hypothesis. Under above assumptions, we devise a Bayesian detector based on the generalized likelihood ratio test (GLRT). Interestingly, it is shown that the proposed detector coincides with the knowledge-aided adaptive coherence estimator (KA-ACE) previously designed in a previous paper by Wang et al. The result provides an alternative explanation of the good selectivity properties exhibited by the KA-ACE.

The optimization of national air route network (ARN) becomes imperative for the increase of air traffic. The Crossing Waypoints Location Problem (CWLP) is a crucial problem in the optimal design of ARN. Although CWLP is in fact a typical multi-objective problem, most state-of-art work solve it from a single objective way, which only considers the flight efficiency and other important properties of ARN are neglected, such as the airspace capacity. In this paper, we propose a multi-objective model for CWLP, in which the airspace capacity can be enhanced with minimizing the standard deviation of betweenness. Then we introduce a Multi-Objective Particle Swarm Optimization (MOPSO) to optimize the flight efficiency and the airspace capacity simultaneously. The application of redesigning the Chinese ARN shows our approach can alleviate the airspace congestion efficiently compared with the conventional methods.