PROJECT TYPE:

• Scientific Research Project

PROJECT DURATION:

• 1.10.2025. – 30.09.2029.

PROJECT LEADER AND ASSOCIATES: 

• Project leader: doc. dr. sc. Karlo Griparić (UNIPU, Faculty of Engineering) 
• Associates: izv. prof. dr. sc. Diego Sušanj (UNIPU, Faculty of Engineering)
• prof. dr. sc. Marin Žagar (Rochester Institute of Technology (RIT) Croatia)
• Dalibor Fonović, dipl.ing. (UNIPU, Faculty of Engineering)

PROJECT DESCRIPTION:

In many manufacturing industries, one of the essential techniques for producing mechanical components is material removal, specifically through machining using computer numerically controlled (CNC) machine tools of various types. Although CNC machine production has been in use since the mid-20th century, the manufacture of complex products is unimaginable without efficient, precise toolpath generation. Optimizing machining parameters significantly affects machining quality and accuracy, as well as efficiency, profitability, and energy consumption. Given the influence of numerous factors on material processing, adjusting machining parameters often requires significant practical experience from the programmer, which can lead to inconsistent processing quality and reduced efficiency. Traditional approaches to parameter optimization rely on geometric and heuristic methods, whose application is limited to bodies with simpler geometry and constant working conditions. The main goal of the project is to investigate the use of machine learning algorithms to improve the selection of machining parameters and tool path corrections by learning from historical simulation data and real-world machining operations. The project proposes a hybrid approach that combines conventional geometric algorithms with machine learning models such as deep learning and neural networks. The expected outcome is to propose a strategy for advancing intelligent manufacturing by demonstrating that advanced data processing methods can improve production processes in existing industrial environments.

PROJECT GOALS:

The project aims to conduct a comprehensive analysis of existing procedures for determining machining parameters. Based on this analysis, a key step is implementing a database of CNC programs. Furthermore, the feasibility of applying Machine Learning (ML) algorithms will be investigated to determine machining operations and parameters from the collected data. Through learning from historical simulation data and real-world machining operations, a hybrid model will be developed that merges conventional geometric algorithms with advanced techniques, such as deep learning and neural networks. This approach is crucial because traditional optimization methods often require extensive practical experience from programmers, leading to inconsistent processing quality and reduced efficiency. The ultimate goal is to propose a strategy for intelligent manufacturing that improves efficiency and accuracy and reduces energy consumption per product unit in existing industrial environments. An additional objective of the project is to strengthen human potential, specifically by increasing researchers' competencies and consolidating knowledge in CNC machining, data analysis, and machine learning.

PROJECT HOLDER: Juraj Dobrila University of Pula