Diagnosis of transformer faults using multi-class AdaBoost algorithm
Low fault diagnosis accuracy is caused by the ineffectiveness of traditional shallow machine learning methods un exploring the connection between the oil-immersed transformer fault data. In response, this study suggests a method for diagnosing transformer faults based on multi-class adaBoost algorithms solves this issue. First, the SVM and the adaBoost algorithm are linked. The SVM is improved by the adaBoost approach, and the transformer defect data is thoroughly investigated. The IPSO is then used to optimize the SVM's parameters when the dynamic weight is added to the PSO algorithm. This is accomplished by updating the particle inertia weight in real-time. Lastly, by examining the relationship between the type of fault and the dissolved gas in the transformer oil, the uncoded ratio technique develops a novel gas set collaboration. The feature vector used as the input is produced using the enhanced ratio approach. The diagnosis method suggested in this paper has a significant increase in diagnostic accuracy when compared to conventional methods, according to simulations using 419 collection of transformer fault data and 117 groups of IECTC10 standard data that were gathered in China. Additionally, it has a fast confluence speed and a powerful search capability.
Published by: Chilla Kaveri, Chagam Reddy Bhargavi, Gandra Neeraja, Burandin Sayyad Dada Umar Hussain, Shaik Tabassum
Author: Chilla Kaveri
Paper ID: V9I2-1156
Paper Status: published
Published: April 3, 2023
Detection and prediction of air pollution using Machine Learning
The regulation of air pollutant levels is rapidly increasing and it's one of the most important tasks for the governments of developing countries, especially India. It is important that people know what the level of pollution in their surroundings is and takes a step towards fighting against it. The meteorological and traffic factors. burning of fossil fuels, industrial parameters such as powerplant emissions play significant roles in air pollution. Among all the particulate matter (PM) that determine the quality of the air. When its level is high in the air, it causes serious issues on people's health. Hence, controlling it by constantly keeping a check on its level in the air is important.
Published by: Patan Masthan Vali, D. P. Neeha Kousar, T. Sai Pranathi, K. Nandini, M. Saikanth, A. Ramesh Babu
Author: Patan Masthan Vali
Paper ID: V9I2-1158
Paper Status: published
Published: April 3, 2023
Implementation of ai based protective mask detector
The global impact of the corona virus disease is significant. Firmly stop the corona virus from spreading. A single-shot detector (SSD)-based object identification technique that focuses on accurate, real-time face mask detection in densely populated settings such as communities and workplaces where there are a lot of people is described. On the basis of two methodologies, we suggest a system in this project. Single-shot multi-box recognition, often known as SSD, is a technique for identifying people wearing face masks in an image in a single attempt. By removing the area recommendation network, which causes an accuracy loss, SSD is employed to accelerate the cycle. Implementing our application in closed-circuit television (CCTV) surveillance systems. It will identify who is wearing the mask and who is not by using mobilenetV2 and machine learning techniques. With the aid of the single shot detection technique, it can filter photographs on the spot and distinguish between them. The data collected during this process, such as image capture, is kept in the cloud to ensure that the application functions properly.
Published by: D. Sarika, C. Amrutha Sai, M. Ganesh Kumar, M. Arun Kumar, A. Bhargavi, B. Jyoshna
Author: D. Sarika
Paper ID: V9I2-1159
Paper Status: published
Published: April 3, 2023
Credit card fraud detection: An evaluation of Machine Learning methods performance using SMOTE and AdaBoost
Online card transactions have increased daily as a result of the development of technologies like e-commerce and financial technology (FinTech) apps. As a result, there has been an increase in credit card fraud that impacts banks, merchants, and card issuers. Thus, it is critical to creating systems that guarantee the confidentiality and accuracy of credit card transactions. In this study, we use imbalanced real-world datasets produced by European credit cardholders to create a machine learning (ML) based framework for detecting credit card fraud. In order to address the class imbalance problem, we resampled the dataset using the Synthetic Minority over-sampling Technique (SMOTE).
Published by: Kethe Meghana, Nidimamidi Thahseen, Duragadda Dhana Lakshmi, Vamsharajula Seenu, Vattam Veda Prakash, Pola Nikhila
Author: Kethe Meghana
Paper ID: V9I2-1155
Paper Status: published
Published: April 1, 2023
Proposed survey questionnaire for the diagnosis of irritable bowel syndrome: Based on patient complaints
Functional bowel disorders (FBD) are extremely common all over the globe. Enhancing knowledge of FBD is essential because they have a detrimental effect on the healthcare system. Irritable bowel syndrome (IBS) is one such FBD where recurrent abdominal pain is associated with a change in bowel habits. Many conditions mimic IBS which include celiac disease, microscopic colitis, inflammatory bowel disease (IBD), lactose and fructose intolerance, etc. In order to correctly differentiate these disorders from IBS, limited testing may be necessary. Considering these overlapping conditions we thought of developing a questionnaire to help the physicians diagnose IBS based on a scoring system. A set of questions were drafted based on patient complaints and each question had a scoring system. Basis the patient's response to each question the total score would help the physicians understand if the patient is suffering from IBS. This is a brand-new grading system that has not yet undergone testing. The medical community is urged to evaluate the scoring system's usefulness and provide input so that it can be improved.
Published by: Dr. Saurabh Srivastava, Dr. Ashish Kumar, Dr. Ànuj Maheshwari, Dr. Gayatri Kapse
Author: Dr. Saurabh Srivastava
Paper ID: V9I2-1150
Paper Status: published
Published: March 29, 2023
Investigation on thermal conductivity of polymer (epoxy) based composites
The current research investigates the influence of fiber volume fraction on effective thermal conductivity (keff) in polymeric materials. This study identifies a method to improve the insulating property of a traditional fiber-reinforced polymer composite. A quantitative relationship for the heat transfer coefficients of polymer composites reinforced with fiber is created utilizing the law of minimal thermal performance and the equal law of particular similar thermal conductivity. To validate this statistical equation, two sets of polymer composites with fiber concentrations ranging from 0 to 15.7 vol percent were hand-built. Natural fibers such as banana fibers are integrated into an epoxy matrix in one set of composites, whilst glass fiber is employed as a filler material in another set, although the matrix material remains unchanged. Thermal conductivities of these composite materials are tested in accordance with ASTM standard E-1530 using the Unit herm TM Model 2022 tester, which operates on the double shielded heat flow concept. Furthermore, using the commercially accessible finite element tool ANSYS, the finite element technique (FEM) is employed to quantitatively measure the k eff of such composites. The numerical values generated by the proposed statistical model are then compared to empirically measured values. The analytical and simulation results reveal that the appropriate heat conductivity value for both sets of composites steadily declines as fiber concentration increases. Because none of the models developed properly anticipated the rate of heat transfer of the composites, the results generated from the proposed system closely match the experimental data. This study shows that as the fiber loading in the composite increases, so does the heat transmission rate. The use of 15.7 vol percent glass fiber in epoxy resin reduces heat conductivity by around 8%, whereas a 12 percent decrease is observed when the banana fiber is used as a filler. This research backs up the conceptual approach while indicating that finite element analysis is an effective tool for such investigations. This thermal insulating, fiber-reinforced polymer composites have potential applications in insulating boards, food containers, thermo flasks, construction materials, and so on due to their low thermal conductivity and lightweight.
Published by: Suhas B. R., Manjunath S. B.
Author: Suhas B. R.
Paper ID: V9I2-1151
Paper Status: published
Published: March 24, 2023
