Kongu Engineering College

About: Kongu Engineering College is a based out in . It is known for research contribution in the topics: Cluster analysis & Control theory. The organization has 2001 authors who have published 1978 publications receiving 16923 citations.

Fuzzy Rule Selection Using Hybrid Artificial Bee Colony with 2-Opt Algorithm for MANET

Abstract: The Mobile Ad-hoc Networks (MANET) is an independent and self-governing hosts of wireless communication that communicate using wireless links thus forming a dynamic and temporary network without any centralized infrastructure. The MANET nodes will not be stationary and the sender and the receiver may not always take similar paths of routing. This way routing becomes quite complicated. A technique that has emerged recently is known as the Opportunistic Routing (OR) which chooses one set of candidates for the purpose of forwarding packets (being compared to that of conventional forwarding made to an approach with one node). It also takes into consideration the nature of the broadcast. This work proposes fuzzy logic with hybrid optimization approach for optimal route selection in MANET applications. The proposed hybrid optimization is based on 2-Opt algorithm and the Artificial Bee Colony (ABC). A fuzzy rule system depends on the end-to-end delay at a node time tends to leave the network there are several packets that are dropped and many different route requests that are generated. The results of the simulation demonstrated the proposed fuzzy rule selection and its efficiency by using the ABC-2 Opt algorithm on being compared with the selection of rule by using the ABC.

Computer aided innovation method for detection and classification of cervical cancer using ANFIS classifier

Abstract: Early detection of cervical tumour is very important to minimise deaths due to cervical cancer. Further it provides a deep insight into the anatomical information of the normal and abnormal cervix and helps in planning for a good treatment well in advance. Numerous techniques are used to detect malignancy through image segmentation. One such segmentation technique is discussed here. The proposed technique uses Artificial Neural Network Fuzzy Inference system (ANFIS) and watershed segmentation techniques for image classification and processing and compares the results with known techniques. A comprehensive set of fuzzy rules was used in the experiment to classify abnormal images to the corresponding malignancy. The experiment shows that the proposed technique is feasible and provides greater accuracy in detection of tumour types.

Hydrogen gas production from sago industry wastewater using electrochemical reactor: Simulation and validation

Abstract: This study describes the production of hydrogen gas from sago industry wastewater under varying operating conditions, such as time of electrolysis, electrode surface area, and current intensity. The process was investigated by employing a four-factor three-level Box–Behnken statistical design (BBD). The results were examined using analysis of variance (ANOVA). This study developed a second-order polynomial model and utilized three-dimensional (3D) response graphs to study the interactive effect of process variables on the production of hydrogen gas. The optimum conditions for maximum hydrogen gas production were determined and found to be 45 cm2 for electrode surface area, 22 min for time of electrolysis, and 13 A for current intensity. The predicted hydrogen gas output in optimum conditions was 1.12 mL/L. These results show that it is possible to achieve efficient hydrogen gas production, by electrochemically processing in sago industry wastewater.

Dynamic mechanical analysis of snake grass fiber-reinforced polyester composites:

Abstract: The dynamic properties of randomly oriented short snake grass fiber-reinforced polyester composites were analyzed with the functions of fiber weight fraction, length, frequency, and temperature. Th...

Digital switching scheme for cascaded multilevel inverters

Abstract: In this paper, a new digital switching technique for cascaded H bridge multilevel inverter (CMLI) for the purpose of power quality improvement is proposed in order to ensure an efficient voltage utilization and better harmonic spectrum. The method presented shows a topology to control cascaded multilevel inverter that is implemented with multiple DC sources to get 2n+1−1 levels. Whereas a standard cascade multilevel inverter requires n DC sources or n number of transformers for 2n+1 levels at the output, where n is the number of inverter stages. Moreover, the experimental results are carried out on a scaled down prototype to prove the effectiveness of the proposed topology. This topology of firing can be implemented without using PWM technique which considerably reduces the switching losses and also brings down the Total Harmonic Distortion (THD). Simulation and experimental results show the superiority of this technique over the conventional methods.