ANN MODELING AND OPTIMIZING THE DRILLING PROCESS PARAMETERS FOR AA5052–GLASS FIBER METAL LAMINATE USING GREY-FUZZY LOGIC APPROACH

Abstract

The exterior layer of the fuselage skin structure is made of Glass-Reinforced Fiber Metal Laminates (GLARE-FMLs). Hole quality is impacted by various force components during the drilling process. The process variables of drilling are optimized to obtain quality holes by controlling such force components. Because of the very limited works on optimizing the significant input process variables in getting quality drilled holes on FML plates, the present investigation is aimed to analyze the impact of varying the significant drilling process variables referred from the literature study and the wt.% of Layered Double Hydroxide (LDH) mixed with the epoxy, on the force components generated during the drilling process. The FML plates proposed in the study have been drilled based on the test plan prepared by Taguchi’s DOE. Grey Relational Analysis has been used for discovering the best combination of process variables leading to good quality holes. Further, grey-fuzzy modeling technique has been employed to check the results of GRA. From the results, it is identified that 1000 rpm spindle speed, 20 mm/min feed rate and 4 wt.% inclusion of LDH to the epoxy resin combination within the considered range of input process variables were found to be the best input parameter combinations resulting in comparatively better output responses. Artificial Neural Network (ANN) model has been generated to bridge the input variables with the outputs measured. The fitness of the generated ANN model has been checked and reported. The effect of varying the process variables on obtaining the quality drilled holes has been revealed using Main Effect Plots (MEPs).