Dany Katamba Mpoyi

Mr. Dany Katamba Mpoyi

Politecnico di Bari, Bari, Puglia, Italy
orcid

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
23 December 2022 in Applied Sciences
Verified
Journal Article
Open access
PDF file
Claudia BarileClaudia Barile
Claudia Barile
KCKatia Casavola
Katia Casavola
GPGiovanni Pappalettera
Giovanni Pappalettera
Vimalathithan Paramsamy KannanVimalathithan Paramsamy Kannan
Vimalathithan Paramsamy Kannan
Dany Katamba MpoyiDany Katamba Mpoyi
Dany Katamba Mpoyi
In this research paper, the acoustic emission technique and a deep learning framework based on two types of pre-trained CNN models (alexNet and squeezeNet) and a new model are proposed to characterize and classify the mechanical behavior of AlSi10Mg components. Specimens are built in a Selective Laser Melting machine with different bed orientations along X, Y, Z, and 45 degrees. Tensile tests are performed, and AE signals are recorded from these tests. To characterize the elastic and plastic deformation stages, a time-frequency domain analysis was performed using CWT-based spectrograms. Three different categories of damage classification strategies were implemented, and CNN models were trained for each strategy. CNN models including AlexNet, SqueezeNet, and the new model were used. Several training modes were performed to determine the CNN model that can accurately classify AE data. Understanding the minimum set of AE signals needed to train the CNN while having 100% accuracy and understanding the parameters affecting the accuracy of a CNN and the training time for the efficient classification of AE signals are the main objectives of this work. The results obtained demonstrated that the new simplified CNN model proposed can accurately classify the AE signals in a short time compared to AlexNet and SqueezeNet.

8 August 2025 in Measurement Science and Technology
Verified
Journal Article
Open access
SCSimone Carone
Simone Carone
CCCaterina Casavola
Caterina Casavola
Dany Katamba MpoyiDany Katamba Mpoyi
Dany Katamba Mpoyi
GPGiovanni Pappalettera
Giovanni Pappalettera
Bolt loosening in industrial vehicles poses substantial safety and operational risks, highlighting the need for reliable monitoring systems. Despite progress in diagnostic techniques, research specifically focused on detecting bolt loosening in complex rotating joints remains limited. This study introduces a novel approach that integrates vibration analysis with deep learning to enhance fault detection in bolt loosening in a rotating joint of a custom sewer cleaning vehicle transmission. The method uses kurtogram-based frequency domain analysis to convert vibration signals into images, which are then processed using a customized convolutional neural network (CNN). This approach eliminates the need for manual feature extraction and enhances diagnostic reliability. The results indicate that a single accelerometer (sensor A1) outperforms multi-sensor configurations. Furthermore, measurements from the accelerometer showed the highest accuracy, demonstrating that longitudinal vibrations are the most reliable for detecting bolt loosening. This study contributes to predictive maintenance and intelligent fault detection, offering a scalable, high-performance solution for monitoring bolt loosening. The integration of kurtogram-based spectral analysis and deep learning paves the way for next-generation autonomous fault detection systems, capable of real-time deployment in dynamic and complex industrial settings.

24 January 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
Claudia BarileClaudia Barile
Claudia Barile
KCKatia Casavola
Katia Casavola
Dany Katamba MpoyiDany Katamba Mpoyi
Dany Katamba Mpoyi
GPGiovanni Pappalettera
Giovanni Pappalettera
This study investigates the classification of damage modes in adhesively bonded carbon fiber-reinforced plastic (CFRP) composites, a critical factor in advancing lightweight automotive design. Adhesive bonding, replacing traditional riveting, improves structural integrity while reducing weight and CO2 emissions. Mechanical testing on CFRP composites was performed, and acoustic emission (AE) signals were collected to evaluate damage mechanisms. A deep autoencoder (DAE) framework was developed to automate damage characterization by reducing AE signal dimensionality through singular value decomposition (SVD) and classifying features using the k-means algorithm. This approach effectively identified three primary damage modes: matrix cracking, interfacial debonding, and fiber breakage. Traditional AE features, such as entropy and amplitude were also classified and validated using spectral analysis. The DAE-based strategy demonstrated superior capability in real-time damage mode differentiation. Fractographic analysis confirmed crack growth in the adhesive layer, leading to interfacial debonding, fiber-matrix separation, and eventual fiber rupture. These findings highlight the DAE framework’s effectiveness in enhancing damage mode characterization, offering valuable insights for optimizing the structural performance of bonded CFRP composites in automotive applications.

30 April 2024 in International Journal of Damage Mechanics
Verified
Journal Article
Closed access
CBClaudia Barile
Claudia Barile
CCCaterina Casavola
Caterina Casavola
Dany Katamba MpoyiDany Katamba Mpoyi
Dany Katamba Mpoyi
GPGiovanni Pappalettera
Giovanni Pappalettera
VKVimalathithan Paramsamy Kannan
Vimalathithan Paramsamy Kannan
This article presents a classification of Acoustic Emission (AE) signals from AlSi10Mg specimens produced via Selective Laser Melting (SLM). Tensile tests characterized the mechanical properties of specimens printed in different orientations (X, Y, Z, 45°). Initially, a study quantified damage modes based on the stress-strain curve and cumulative AE energy. AE signals for each specimen (X, Y, 45°, Z), across deformation stages (elastic and plastic), and damage modes were analyzed using continuous wavelet transform to extract time-frequency features. A novel convolutional neural network, based on artificial bee colonies and fuzzy C-means, was developed for scalogram classification. Data augmentation with Gaussian white noise enhanced the approach. Cross-validation ensured robustness against overfitting and suboptimal local maxima. Evaluation metrics, including the confusion matrix, precision-recall curve, and F1 score, demonstrated the algorithm's high accuracy of 92.6%, precision-recall curve of 92.5%, and F1 score of 92.5% for AE signals based on printing direction (X, Y, 45°, Z). The study highlighted the potential for improving AE signal classification related to elastic and plastic deformation stages with 100% accuracy. For damage modes, the algorithm achieved a confusion matrix accuracy of 90.6%, a precision-recall curve of 90.4%, and an F1 score of 90.5%. This approach demonstrates high accuracy in classifying AE signals across different printing orientations, deformation stages, and damage modes of AlSi10Mg specimens manufactured through SLM.

15 June 2026 in Applied Sciences
Verified
Journal Article
Open access
PDF file
CBCIKALA BAGALWA Bienvenu
CIKALA BAGALWA Bienvenu
Bovic Kilundu Y'EbondoBovic Kilundu Y'Ebondo
Bovic Kilundu Y'Ebondo
Dany Katamba MpoyiDany Katamba Mpoyi
Dany Katamba Mpoyi
KCKatia Casavola
Katia Casavola
GPGiovanni Pappalettera
Giovanni Pappalettera
Tool wear directly affects surface quality, dimensional accuracy, and manufacturing cost in milling operations, making reliable wear state classification essential for process control. This paper presents an offline deep learning framework for multiclass tool wear classification using the UC Berkeley milling dataset (NASA-Ames). Statistical features are extracted from vibration, acoustic emission, and spindle motor current signals, and dimensionality is reduced from 78 to 9 informative variables using LASSO regression. A four-layer Long Short-Term Memory (LSTM) network then models the temporal evolution of tool degradation across three wear states: healthy, degraded, and failed. Two model variants are compared: Model A uses sensor-derived features only, while Model B additionally incorporates feed rate and depth of cut as inputs. To prevent data leakage, partitioning is performed at the machining-case level rather than at the individual window level. Model A achieves 92% classification accuracy; Model B reaches 95%, demonstrating that cutting conditions provide contextual information that resolves ambiguity between wear states produced under different machining regimes. These results confirm that combining multisensor feature fusion, LASSO-based selection, and sequential deep learning constitutes an effective framework for tool wear classification in milling.