Szczegóły publikacji
Opis bibliograficzny
Neural algorithm for detecting inclusions in aluminum castings during drilling based on tool acoustic emission / Paweł GUT, Dawid Piecuch // W: 27th International Carpathian Control Conference (ICCC) [Dokument elektroniczny] : 1-3 June 2026, Szilvásvárad, Hungary : proceedings. — Wersja do Windows. — Dane tekstowe. — Piscataway : IEEE, cop. 2026. — Print on Demand(PoD) ISBN: 979-8-3195-3321-0. — e-ISBN: 979-8-3195-3320-3. — S. 186–189. — Wymagania systemowe: Adobe Reader. — Bibliogr. s. 189, Abstr. — Publikacja dostępna online od: 2026-07-07
Autorzy (2)
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 169804 |
|---|---|
| Data dodania do BaDAP | 2026-10-06 |
| Tekst źródłowy | URL |
| DOI | 10.1109/ICCC71363.2026.11593222 |
| Rok publikacji | 2026 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Institute of Electrical and Electronics Engineers (IEEE) |
Abstract
The article presents a neural algorithm for detecting inclusions in aluminum castings during the drilling process, based on the analysis of acoustic emission signals recorded directly from the cutting tool. The input data for the system are raw sound signals recorded with a microphone with a bandwidth of up to 20 kHz during the actual machining process. The data was obtained experimentally in laboratory conditions, covering two process states: drilling in homogeneous material and drilling in an area with material inclusions. A convolutional neural network, adapted for direct processing of one-dimensional time signals, was used for classification. The network architecture consists of a series of convolutional layers, ReLU activation functions, max-pooling operations, and fully connected layers, enabling automatic extraction of features relevant to the detection of structural inhomogeneities in the material. The model was trained as a binary classifier distinguishing between normal drilling and drilling in the inclusion area. The results confirm the effectiveness of the proposed approach in identifying inclusions based on acoustic emission signals, indicating the potential for applying deep learning methods in cutting process monitoring systems. The presented solution can serve as a basis for the development of intelligent diagnostic systems supporting the quality control of aluminum castings in industrial conditions.
