International Journal of Convergence Computing (5 papers in press)
Special Issue on: Smart Converging Technology
Development of an Adaptive non-parametric model for estimating maximum efficiency of Disc membrane
by Anirban Banik, Tarun Kanti Bandyopadhyay, Sushant Kumar Biswal, Mrinmoy Majumder
Abstract: Membrane separation and filtration process are a technique of removing the impurities from the feed stream based on the pore size of the membrane bed. Permeate stream produced by the membrane are good quality due to this membrane found wide application in the field of water purification, gas-gas separation etc. In the concerned study, the ability of the cellulose acetate disc membrane for improving the quality of the rubber industrial effluent of Tripura has been investigated in pilot scale. GMDHmultilayered feedback algorithm has been implemented to predict the maximum efficiency of the membrane. The efficiency of the membrane is maximized for the optimal value of pore size, inlet velocity, and operating pressure. It has been found that efficiency of the membrane is maximized when the pore size of the membrane is kept 2.060538 μm, inlet velocity is 0.201896 m/sec and operating pressure is 694.7201 kPa. The performance of the prepared GMDH model is evaluated by using model evaluation technique like NSE, PBIAS, slope and Y-intercept, RSR. It has been found that software predicted data can be used for trouble shooting and optimal design of the membrane bed.
Keywords: membrane; GMDH; neural network; membrane separation technique; optimization.
Designing and Modelling of Grid Connected Photovoltaic System (Case Study: EEU Building at Hawassa City)
by Yishak Kifle, Baseem Khan, Jitendra Singh
Abstract: Renewable energy sources are integral part of the present energy system because of the various problems, associated with conventional energy sources. Solar and wind are the major renewable energy sources. The renewable energy sources are sufficient enough to meet the world energy requirement. In this paper, a comprehensive analysis on energy output calculation of the photo-voltaic (PV) system is presented. The solar radiation effect is articulated based on monthly average sun shine hours. After local area solar potential is assessed, small plant is developed and designed for EEU building electric power supply, which is connected to the grid. Therefore, there is huge reduction in grid supply demand. The study uses different metrological data and simulates the grid connected PV technologies to end users, for decision making in PV energy investment. Performance ratio and specific energy is calculated as indicator for the plant feasibility in the area.
Keywords: solar potential assessment; grid connected PV; performance ratio; specific energy.
Wheat Yield Forecasting Using Fuzzy Logic
by Tanya Sah
Abstract: Forecasting in general and crop yield forecasting, in particular, is considered a compound problem. Food and Agriculture Organizations report Global Agriculture towards 2050 accentuates the challenges which agriculture sector is going to face in near future. The report draws the attention towards the disparity between the demand and supply. It also highlights the issues that automation has introduced into the food and agriculture sector. Where there is a need to increase the production to feed the rapidly increasing population there is an equal necessity for coping up with the dwindling numbers of farm workers. As such, there is a need for a reliable forecasting algorithm capable of handling time series data. In this paper, we have proposed a fuzzy time series forecasting algorithm to forecast wheat yield. The reason to choose fuzzy over other forecasting methods is its capability of dealing with the vague, imprecise data and it outperforms many statistical conventional models in such conditions. The neural network has been used for training and defuzzification of the forecasted values.To attests the efficacy and the performance of the proposed method, it has been tested against the wheat production dataset.
Keywords: Fuzzy Logic; Wheat yield forecasting; Neural Network.
A novel method for reduction of leakage current in MOSFET
by Debasis Mukherjee, B.V.Ramana Reddy
Abstract: In this paper, structural modification of conventional bulk MOSFET has been proposed for minimization of subthreshold leakage current. Key structural features of bulk MOSFET have been kept unaltered. Comparison of conventional and proposed structure has been presented for a 20 nm NMOS with 0.8 volt Vdd. The proposed structure is capable of reducing subthreshold leakage current even at very low drain voltage when the gate voltage is zero. Around 55 percent reduction of OFF current has been obtained when drain voltage is at Vdd and gate voltage is zero. The methodology proposed does not have any special requirement at the circuit level, and can be combined with all circuit level methodologies. The proposed structure is named as Defensive MOSFET as it looks like a defensive shield. Structural dimensions of 20 nm MOSFET generation have been taken from the 2011 edition of International Technology Roadmap for Semiconductors or ITRS. All simulation processes have been executed by Sentaurus G-2012.06 Technology Computer Aided Design or TCAD software.
Keywords: 20 nm; leakage current reduction; NMOS.
Development of Real Time Monitoring System for Intake Water of Surface Water Treatment Plant with the help of Segmented Polynomial Neural Network
by Paulami De, Mrinmoy Majumder
Abstract: The Surface Water Treatment Plant (WTP) has an important role in the maintenance of water quality supplied to the consumers It was found from various studies that inflow water quality to a WTP plays an important role in the operation of the treatment plants. The nature and characteristics of the inflow water control the duration of treatment, dozing pattern etc. The temporal variation only one of the parameters like Turbidity of the inflow water can induce major adjustment in the plant operation and may also affect the efficiency of the WTP. Identification of temporal pattern can help the engineers to adapt the WTP operations and can save unnecessary wastage of plant resources. That is why in the present study a new model was proposed for prediction of the temporal patterns of various chemical parameters by adopting a new type of Artificial Neural Network model. The model was applied to predict the Turbidity concentration at the intake point of Bardowali WTP of Tripura which have a peri-urban settlement where Turbidity varies due to the urban runoff which disposes its suspended matters into the at the intake point of the WTP. According to performance metrics utilized the model was able to predict the temporal pattern of the model at 99% accuracy at a lag of 1 hour.
Keywords: Soft-computation; Polynomial Neural Network; Segment Separation; Water Treatment Plant.