Weld fault identification using X-ray pictures is a useful nondestructive testing technique. Traditionally, this job has relied on skilled human experts, while the extraction and classification of heterogeneity required their personal involvement. To overcome those challenges, various approaches are used, including machine learning (ML) and image processing technologies. Although the detection and categorization of low contrast and poor-quality images have been improved, the end result is still unsatisfactory. Unlike earlier ML-based research, this paper provides a new deep learning network-based classification approach. In this work base convolutional neural network architecture with optimization techniques was used to enhance the performance of the architecture to obtain a best performance result in simpler CNN architecture. In this work we have obtained overall classification accuracy of 89% using simple CNN with optimization technique.

Background/Objectives: Last decades have witnessed a rapid growth in the development of harder, difficult and complexity to machine metals and alloys. AWJM is one of the most freshly built up nontraditional machining processes in processing various types of hard-to-cut materials nowadays. It is an economical method for heat sensitive materials that cannot be machined by processes that produce heat while machining. Machining parameters play a lead role in determining the economics of machine and machining quality. This paper investigates the prediction of MRR and Surface roughness on Lead Tin Alloy using the Artificial Neural Network (ANN). Methods/Statistical analysis: In this work, the influence of five AWJM parameters of the process on SR and MRR of an American element referred as Lead Tin Alloy which is machined by AWJM was experimentally performed and analyzed. According to RSM design, different experiments have been performed with the combination of input parameters on this American element. Findings: Outcome depicts the minimum error attained for data belonging to test is 1.063814%% for MRR and 0.208967018% for SR. Also the maximum error obtained is about 9.475104% for MRR and 9.070886429% for SR. By training the network deviations may occur but error is reduced because this technique is heuristic.

The design, production, and characterization of coir fiber-reinforced composites employing epoxy resin modified with CNSL are investigated in this work. A 1:1.25 epoxy-to-hardener ratio, 15-20% CNSL, and vacuum bagging were used to create fiber arranged for sun-dried coir composites. Structural and thermal properties were assessed using mechanical tests (tensile, flexural, and impact) as well as FTIR, TGA, and SEM investigations. The findings show that CNSL improved impact resistance and thermal stability, while interlocking coir composites had greater mechanical strength. However, impact hardness was decreased by over-mercerization. The results show that coir fiber composites are environmentally friendly substitutes for construction, automotive, and aerospace applications. To improve performance, future studies can optimize fiber loading and biodegradability.