The Capacity Outage Probability Table (COPT) is the most common analytical model for the generation adequacy evaluation. A recursive algorithm can be used to form the COPT which demonstrates arrays of capacity levels of generation with their probabilities of occurrence. The number of states in a COPT is a critical task since more states generally mean a higher modelling accuracy and also a longer computation time. In this study, the wind farm generation adequacy using hourly wind speed data of St. John's, Newfoundland and Labrador, Canada will be investigated. Fuzzy c-means method is used to determine appropriate number of states in the COPT. The study results are demonstrated using Roy Billinton Test System (RBTS). Unlike the existing methods, the computation complexity is declined by using Fuzzy c-means method, and the model of a wind farm is simplified. This analytical analysis is useful for financial investigators and planners of wind farm generation projects.
In this study, a novel hierarchical object-based Random Forest classification approach is proposed for discriminating between different wetland classes in a sub-region located in the northeastern portion of the Avalon Peninsula. Particularly, multi-polarization and multi-frequency SAR data, including X-band TerraSAR-X single polarized (HH), L-band ALOS-2 dual polarized (HH/HV), and C-band RADARSAT-2 fully polarized images, were applied at different classification levels. The overall accuracy and kappa coefficient were determined in each classification level for evaluating the classification results. The importance of input features was also determined using the variable importance obtained by Random Forest. Using this new hierarchical RF classification approach, an overall accuracy of up to 92% was obtained for classifying different land cover types in the study area.
Computer controlled characterization of X-ray diffraction (XRD) spectra of asphalt binder were studied with respect to X-ray background type, profile fit and dimensional analysis. Four different background types (linear, parabolic, 3rd and 4th order polynomial) were used in order to assess the precision of fit and residual error of fit. Mathematical functions (Pearson VII, pseudo-Voigt, generalized Fermi) are also employed in this analysis. When varying the Pearson VII exponent (1.6, 2.0) and pseudo-Voigt Lorentzian (1.0) constant the results show an effect on the X-ray peak position and calculated average size of crystallite parameters. The interlayer distance between the aromatic sheets (dM) was determined to be between 4 to 5 angstroms and the distance between the saturated portions (d?) at 5 to 7 angstroms. For all cases the lowest residual error of fit was the 4th order polynomial X-ray background type.
This paper describes a new feature extraction method for optical character recognition (OCR) system for recognizing handwritten documents in Malayalam and its pronunciation. Malayalam is one of the 22 officially recognized language of India, Malayalam character recognition has gained immense popularity in the past few years. The intrinsic challenges present in this domain along with the large character set of Malayalam and its cursive structure further complicate the recognition process. This paper proposes a new efficient feature extraction method for training the feed-forward back propagation neural network by using discrete wavelet transform (DWT) and at the same time recognized character is pronounced with the help of text-to-speech system (TTS).
The process of training and skill acquisition is essential for workers in high-risk environments. A particularly important aspect of safety and emergency response training in such environments is learning a new environment. This can be facilitated through the use of a virtual environment simulator along with adequate performance assessment metrics. Currently, user performance is evaluated solely using behavioral metrics. However, the processing efficiency theory suggests that neural signals could provide additional information about an individual's skill level that behavioural metrics cannot. Recent studies have identified neural metrics, as measured by electroencephalography (EEG), associated with the learning of tasks such as flying an aircraft and performing surgery. However, neural indicators of learning a spatial environment have not been investigated. The results of this study expand the applicability of previous findings to offer an improved evaluation of spatial knowledge.
Abstract-This paper presents a new detection method for broken rotor bar fault (BRB) in induction motors based on the estimation of signal parameters via YULE Walker Auto Regression (YUL-AR) power spectral density estimate method and MUSIC pseudo spectrum. The performance of the two aforementioned technique are tested with the simulated stator current signal of an induction motor with BRB. The results obtained from technique show consistent result with established techniques. The technique is capable of correctly identifying the frequencies of the BRB characteristic components but limited accuracy result for amplitudes and harmonic frequencies components. . Results obtained proves that the proposed method is a promising choice for BRB detection in induction motors operating under normal and transient load condition.