Flow field design for the distribution of reactants and products on the electrode surface plays an important role in the overall performance of the fuel cell. It acts as a crucial factor when the laboratory scale fuel cell is scaled up for commercial applications. In the present work, a novel flow field design is proposed and its usefulness for the fuel cell applications are evaluated in a high-temperature polymer electrolyte fuel cell. The proposed geometry retains some of the features of serpentine flow field such as multiple bends, while modifications are made in its inplane flow path to achieve comparatively uniform reactant and product distribution. A threedimensional CFD model is developed to analyze the effectiveness of the proposed flow field. An HT-PEFC is fabricated and experimented with the proposed flow field for experimental validation. Furthermore, a low-cost current distribution mapping device is developed to validate the current density distribution on the electrode obtained from the CFD model. It exhibits a mismatch of 4% in the spatial distribution of current density between the modelling and experimental results. The proposed design is capable of achieving higher uniformity in current distribution across the active area (0.998 for modified serpentine and 0.96 serpentine) compared to serpentine flow field. This aids in boosting the current density of the cell by 27% at 0.57 V operations.

No abstract

Deoxyribo Nucleic Acid (DNA) microarrays are widely used to monitor the expression levels of genes in parallel. It is possible to predict human cancer using the expression levels from a collection of DNA samples. Due to the vast number of genes expression level, it is challenging to analyze them manually. In this paper, data mining approach is used to extract the prevailing information from DNA microarray with the help of multiresolution analysis tool. Dual Tree M-Band Wavelet Transform (DTMBWT) is employed for the extraction of features from the given dataset at the 2nd level of decomposition. K-Nearest Neighbor (KNN) classifier is used for cancer classification. Results show that KNN classifier classifies five different cancer datasets; Breast, Colon, Ovarian, CNS, and Leukemia with over 90% accuracy.