Development of a model for management of the quality testing process in Lean industrial systems based on artificial intelligence methods

[N. Z. Pajić]

142 листа

Ova disertacija istražuje primenu metoda veštačke inteligencije u oblasti kontrole kvaliteta u okviru Industrijske revolucije 4.0. Ovaj rad se bavi predviđanjem ishoda operacije lemljenja pomoću mašinskog učenja, kao i primenom mašinskog vida za klasifikaciju i detekciju defekata pri nalivanju lepka u automobilskoj industriji kroz dve studije slučaja. Takođe je prikazan idejni koncept evaluacije kupčevih zahteva pomoću veštačke inteligencije, kao i obezbeđenje kvaliteta laserski zavarenih spojeva. Prikazom korišćenja različitih algoritama i metoda veštačke inteligencije, ova disertacija predstavlja prikaz primene naprednih pristupa kontrole kvaliteta u proizvodnim sistemima. U radu je dokazano da automatska klasifikacija i detekcija defekta doprinosi efikasnosti i smanjenju grešaka u proizvodnim procesima, kao i da smanjuje troškove testiranja. Isto tako, primena veštačke inteligencije u obezbeđenju kvaliteta različitih procesa omogućava bržu i efikasniju proceduru analize defekata i upravljanja kvalitetom. Kroz složen pristup istraživanju, ova disertacija ima za cilj da doprinese napretku u oblasti kontrole kvaliteta u industrijskim okruženjima, smanjenju troškova i vremena, kao i poboljšanju standarda kvaliteta u okviru industrije 4.0. Kao rezultat ovog istraživanja, predstavljeni su novi rezultati i savremene metode koje bi mogle značajno uticati na budući razvoj u oblasti kontrole kvaliteta u industrijskom okruženju.

This dissertation explores the application of artificial intelligence methods in the field of quality control within the framework of Industry 4.0. The study focuses on predicting the outcome of soldering operations using machine learning, as well as applying machine vision for defect classification and detection during adhesive pouring operations in the automotive industry through two case studies. Additionally, it presents the conceptual concept of evaluating customer requirements using artificial intelligence, as well as ensuring the quality of laser-welded joints. By employing various algorithms and methods of artificial intelligence, this dissertation demonstrates the application of advanced quality control approaches in production systems. The research proves that automatic defect classification and detection contribute to efficiency and error reduction in manufacturing processes, as well as cost savings in testing. Similarly, the application of artificial intelligence in ensuring the quality of various processes enables a faster and more efficient procedure for defect analysis and quality management. Through a comprehensive research approach, this dissertation aims to contribute to advancements in quality control within industrial environments, reducing costs and time, and enhancing quality standards within Industry 4.0. As a result of this research, new findings and contemporary methods are presented that could significantly impact future developments in the field of quality control in industrial settings.

Veštačka inteligencija, Kontrola kvaliteta, Industrija 4.0, Mašinski vid, Mašinsko učenje, Klasifikacija defekta, Detekcija defekata, Automobilska industrija

Artificial Intelligence, Quality Control, Industry 4.0, Machine Vision, Machine Learning, Defect Classification, Defect Detection, Automotive Industry

004.8:658.56(043)

Tekst