Intelligent Processes in Automated Production Involving Industry 4.0 Technologies and Artificial Intelligence

Introduction

Industry 4.0 facilities are important about the upgrade of manufacturing processes (Rubmann et al., 2015). Automatism and image vision techniques can represent a good solution in order to improve quality control procedures (Massaro et al., 2018) according with ISO 9001:2015 standard, thus implementing Industry 4.0 control systems. Also traceability represent an important topic about process digitalization and process mapping. In this direction different technologies such as barcode, QR-code and RadioFrequency IDentification –RFID- can be adopted to trace and map production (Lotlikar et al., 2013). Traceability models should be integrated into management tools (Khabbazi et al., 2010) thus improving the knowedge base –KB-. Augmented reality – AR- can represent a further facility for monitoring manufacturing processes (Caudell et al., 1992), aided manufacturing (Novak-Marcincin et al., 2013), and quality check (Segovia et al., 2015). Specifically, by adopting marker-based approach in AR systems (Siltanen, 2012), it is possible to improve computer vision, image processing and computer graphics techniques. Image vision and image processing are suitable for automatic processes involving real time quality control of products. In this direction a study has been performed for welding monitoring of electronic boards by executing 2D and 3D image processing algorithms (Massaro & Vitti et al., 2018). In particular the 3D image processing could enhance and show hidden information (Cicala et al., 2014) thus contributing to view better possible defects and other construction details. Other sensors can be applied on production lines in order to achieve predictive maintenance (Massaro & Galiano et al., 2018). According with predictive calculus, data mining algorithms can be implemented into an information system in order to predict failures of single machines or of the whole production lines (Massaro & Maritati et al., 2018). In order to control the production during the time, can be adopted different charts mapping processes such as 4M charts (Favi et al., 2017), Plan Do Check Act -PDCA- (Chakraborty, 2016), Xm-R (Fouad et al., 2010), and p control charts able to check defect rate (Wang, 2009) and process quality (Acosta-Mejia, 1999). These charts can be joined with data mining and artificial intelligence –AI- algorithms to formulate a predictive maintenance scheduling model. In literature data mining has been implemented for multilevel alerting systems enabling predictive maintenance (Bastos et al., 2014), thus suggesting different industry research implementations and data flow architectures also in other application fields (Massaro, Maritati, Savino et al., 2018). Still remaining in the context of predictive maintenance, some authors applied artificial neural networks –ANNs- for the prediction of mechanical component failures and degradation prediction (Zhang, 2015), formulating innovative processes inherent in optimizing prediction and defining optimal training dataset (Krenek et al., 2016). These studies prove that different technologies can be applied in order to optimize production processes. According with the state of the art, the goal of the proposed chapter is to discuss preliminary research industry projects involving automated processes embedding artificial intelligence, data mining algorithms, and image processing tools together with Industry 4.0 enabling technologies. Specifically in this chapter are discussed some requirements of industry projects, following research and development –R&D- guidelines (Frascati, 2015): different cases of studies proves how it is possible to provide knowledge gain in production processes, supporting at the same time production traceability and quality assessment. The knowledge gain can be performed by innovative algorithms which can be structured in different ways depending on the specific requirements of the case of study. Knowledge gain can be achieved also by formulating association rules between the production processes or by applying scientific methodologies for data processing. In any case the first step to increase the knowledge in industry is to digitize the information as for Industry 4.0 logic. Some important topics of Industry 4.0 are Internet of Things -IoT-, Industrial Internet, Smart Manufacturing, and Cloud based Manufacturing (Vaidya et al. 2018). These topics (Oztemel et al., 2018) are implemented in different models enabling Industry 4.0 facilities (Basl et al., 2019). In this scenario artificial intelligence –AI- (Skobelev at al., 2017) could improve adaptive processes by predictive maintenance, collaborative robotics and rapid prototyping (OECD 2017). An upgraded Industry 4.0 architecture oriented on Industry 5.0 logic includes artificial intelligence, big data systems and IoT facilities (Özdemir et al., 2018). Following the state of the art the proposed chapter shows some cases of study involving Industry 4.0 facilities thus enhancing how the topics of Industry 4.0 can be applied in practical cases.