R. Balasundaram

The use of thermoplastics has increased in recent years in a variety of fields. Some machining work is necessary to maximize the benefits of the polymer. The thermoplastic material used in this analysis was Derlin (32 mm in diameter and 15 mm in thickness), which was drilled to investigate the drilling characteristics using a Computer Numerical Control vertical milling centre. The experiments were carried out using an L9 orthogonal array with speed, feed, and pecking as input parameters to study its machining characteristics. Hard carbide tools and HSS drill bits were used in the work. Surface roughness and pecking were the outputs investigated. Analysis of variance (ANOVA) and the decision tree algorithm were used to examine the output parameters. According to the Taguchi and Decision Tree algorithm results, the significant factors for surface roughness in hard carbide tools were speed and feed. For HSS drill bits, the main dominating factor for surface roughness was speed. For roundness errors, pecking was the dominant factor. The decision tree was developed from the output, and the accuracy is high.

The process of turning hardened material with a Rockwell hardness of 45–70 on the C scale is known as hard turning. The current work involved dry turning of heat-treated oil-hardened non-shrinking (OHNS) die steel. The experiment was designed using an L9 orthogonal array. The experiment’s input parameters include speed, feed and depth of cut (DoC), while its outputs include material removal rate (MRR), surface roughness and machining time. This work uses the C 4.5-decision tree algorithm for the L9 orthogonal array, transforming low-level data into high-level knowledge through if-then-else rules. For each performance attribute, rules are framed based on certain conditions. The experimental results match the attribute-based rules.

Use of machine learning and artificial intelligence (AI) to analyze the complex interdependencies of production dataset has gained momentum in recent years. Machine learning and predictive algorithms are now used by manufacturers to fine-tune the quality of their products. WEDM of SS304 with process parameters such as pulse-on-time (Ton), pulse-off-time (T off), current (I), and voltage (V) was varied to study the effect of machining parameters such as Material Removal Rate (MRR) and surface roughness. Experiments were planned and executed according to the L’9 orthogonal array. Scanning Electron Microscope (SEM) was utilized to study the machined surface. An analysis of variance (ANOVA) was performed to determine the input and output significance. ANOVA results revealed that V (81.85%) and T_off (77.75%) for surface roughness. Further to determine the relationship between variables, various regression models based on machine learning was tested. The effectiveness of the regression models were tested. From their output it was concluded that the multilayer perception model had the highest correlation coefficient (0.999) for MRR while for surface roughness it was (0.995).

The selection of materials with the required properties was developing rapidly. There is no single rule to choose suitable material for a particular product. The complexes interrelation among the properties of composites makes more difficult to suggest suitable material for material scientist. The forecasting of relationship among material properties with throughput screening methods can help for the discovery of novel materials. The conventional way of selection of material need more investigations and time consuming. Recently, machine learning approaches are widely used in mechanical domain. Hence in this work, machine learning-based method was proposed to discover the relationship between the properties of natural fibres composites. To apply machine learning approach, there are 26 different natural fibres composites of Coir, Jute, Banana etc. were taken in to consideration. These materials are reinforced with pure export with different manufacturing methods such as Hand layup, Spray layup, Hot press moulding, and Compression moulding to produce the samples as per ASTM standards. The properties like density, tensile strength, hardness etc. were measured and interrelation among them was recognised using proposed method.

In this study, a polymer composite is made using chemically treated jute fiber and waste floor tile powder as an alternative source for roof tile application. The wear qualities were examined at various ages, and the outcomes were optimized. In order to improve the wetting properties of the jute fiber, it was chemically treated. MINITAB software was used to develop Taguchi method parameters such as jute fiber percentage, waste tile powder percentage, and NaOH chemical treatment using the MINITAB software. It was determined that hardness was the most important characteristic in terms of wear properties after the specimens were subjected to ageing and abrasion wear testing and hardness tests were carried out as per normal protocols. As a result of the waste tile powder addition, the surface and core pore formation rates were reduced and the wear index rates were low. Jute fiber with 15%, 9% tile powder, and 5% NaOH treatment were found to have the lowest wear index of the other specimen compositions tested, according to the wear index. Specimen made with 5% jute fiber addition, 9% tile powder inclusion, and 10% NaOH treatment, on the other hand, had more hardness. Degradation of the fibers and delamination are side effects of the ageing process. The wear resistance of the surface was increased by the use of waste tile powder.