Deep Learning Neural Networks for Forecasting the Abrasive Wear in Machining Tools for Cryogenic Treatment by Process Parameter Optimization

Abstract

Every year, the industries are found to lose a billion of rupees due to abrasive wear and tear and subsequently wastage of material by surface loss. The widely used surface techniques are nitriding and chemical/physical vapor deposition. Nitriding requires high temperature and a long processing time, and chemical/physical vapor deposition is limited by adhesive strength of coating to the base metal. To overcome these, hard-facing and cryogenic treatment processes are suggested to improve the wear properties. Hard-facing is a coating of harder material on base metal by suitable metal working process. Cryogenic treatment is one of the heat treatment process by which the material is subjected cryogenic temperatures from −150 to −273 °C. An attempt has been made here to optimize the cryogenic treatment time on slurry abrasive wear behavior of chromium carbide hard-faced surfaces on mild steel base metal. The strategy includes the implementation of artificial neural networks (ANN) with feed forward architecture (FFA) trained with backpropagation algorithm (BPA) at the start. However, coherent findings state that hybrid neural networks (HNN) and deep learning-based hybrid neural networks (DHNN) have substantial effect on identification of parameters related to cryogenic treatment as compared with the conventional ANN trained with BPA. Investigations reveal that this technology is of great use and finds major application in monitoring the abrasive wear and tear of industrial cutting tools. Therefore, a set of actual variables like temperature, speed, abrasive content and cryogenic treatment time are taken as the input variables so as to estimate and forecast the wear rate of the hard faced surfaces using ANN. Qualitative observations revealed that this ANN-based forecasting and estimation technique suggested timely replacements and over hauling of the industrial tools used in cryogenic treatment. New developments and domination of smart phones with Android technology find suitable application for online monitoring and forecasting of abrasive wear in various cutting, milling and drilling tools using Industrial Internet of Things (IIoT) which is a wireless network. The server–client approach used to implement this technology combines SOAP web service with socket programming for Transmission Control Protocol (TCP) and Internet Protocol—version 4 (IPv4) protocols.