Bovic Kilundu Y'Ebondo

Dr. Bovic Kilundu Y'Ebondo

orcid

Research Focus & Metrics

Representative Publications

View All Publications
26 August 2025 in Mechanical Systems and Signal Processing
Verified
Journal Article
Closed access
SKS. Moloverya Kavugho
S. Moloverya Kavugho
GAUTHIER NGANDU KALALAGAUTHIER NGANDU KALALA
GAUTHIER NGANDU KALALA
LRL. Rasolofondraibe
L. Rasolofondraibe
Bovic Kilundu Y'EbondoBovic Kilundu Y'Ebondo
Bovic Kilundu Y'Ebondo
XCX. Chiementin
X. Chiementin

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.

8 May 2026 in Measurement Science and Technology
Verified
Journal Article
Closed access
TKTeddy Kayala
Teddy Kayala
Chiementin XavierChiementin Xavier
Chiementin Xavier
FBFabrice Bolaers
Fabrice Bolaers
HKHubert Kazadi
Hubert Kazadi
Bovic Kilundu Y'EbondoBovic Kilundu Y'Ebondo
Bovic Kilundu Y'Ebondo
YTYessine Toumi
Yessine Toumi
Thrust ball bearings are critical components in rotating machinery subjected to predominant axial loads; thus, fatigue-induced damage is a major operational concern. This paper presents a novel method to estimate the remaining useful life (RUL) of thrust ball bearings by integrating vibration acceleration signals with the spalling growth mechanism. A phenomenological degradation law, derived from damage and contact mechanics principles (Paris’s law and Hertzian contact theory), is employed to model the trend of the health indicator (HI) extracted from vibration signals. This degradation follows a power-law function of the applied axial load and the accumulated number of cycles, characterized by three time-variant coefficients: ε, η and ζ . These coefficients are dynamically updated using a specifically designed extended Kalman filter (EKF) to enable accurate projection of the HI trajectory. By applying an experimentally derived end-of-life threshold, the RUL is estimated with high accuracy across a prolonged prediction horizon. Validation using run-to-failure experimental data demonstrates that this hybrid, physics-informed EKF model significantly outperforms pure empirical models, providing robust and reliable RUL prognosis.

20 March 2026 in Measurement Science and Technology
Verified
Journal Article
Closed access
ENEphraïm Moshi Ngele
Ephraïm Moshi Ngele
GAUTHIER NGANDU KALALAGAUTHIER NGANDU KALALA
GAUTHIER NGANDU KALALA
ABAbir Boujelben
Abir Boujelben
Chiementin XavierChiementin Xavier
Chiementin Xavier
Bovic Kilundu Y'EbondoBovic Kilundu Y'Ebondo
Bovic Kilundu Y'Ebondo
Reliable fault diagnosis is essential for condition monitoring of rotating machinery, yet the scarcity of experimental data often limits the effectiveness of machine learning algorithms. A promising approach lies in constructing hybrid datasets that combine experimental and numerical data. However, if performed arbitrarily, such hybridization may degrade diagnostic reliability. This study proposes a systematic methodology for building hybrid databases for ball bearing fault diagnosis across multiple fault severities and varying operating conditions. The methodology is evaluated along three complementary axes: (i) single-condition performance, (ii) expandability to a wide range of speeds through the identification of a reliable variation range, and (iii) transferability of calibration parameters to neighboring regime operating conditions. Both homogeneous (quantitative enrichment) and heterogeneous (qualitative enrichment with unseen fault labels) hybridizations are investigated. Results demonstrate that even a small experimental fraction (≈10%) substantially enhances classification accuracy compared with purely numerical models, while stable performance (>90%) is achieved from about 30% of experimental data. In multi-speed scenarios, hybridization maintains generalization provided sufficient experimental anchoring, though the reliable range remains narrow and asymmetric. In heterogeneous configurations, high accuracy is preserved even with a single EC, as numerical enrichment effectively compensates for missing severities, provided the enrichment is balanced with the operating range.

5 November 2026 in Measurement Science and Technology
Verified
Journal Article
Closed access
TKTeddy Kayala
Teddy Kayala
Chiementin XavierChiementin Xavier
Chiementin Xavier
FBFabrice Bolaers
Fabrice Bolaers
HKHubert Kazadi
Hubert Kazadi
Bovic Kilundu Y'EbondoBovic Kilundu Y'Ebondo
Bovic Kilundu Y'Ebondo
YTYessine Toumi
Yessine Toumi
Reliable estimation of the remaining useful life (RUL) of rolling bearings remains challenging under time-varying operating conditions such as variable load and rotational speed. Variations in operating conditions induce disturbances in vibration-based health indicators (HIs), which can obscure intrinsic degradation dynamics and degrade prognostic accuracy. Most existing approaches either assume stationary conditions or rely on data-driven models requiring extensive labeled datasets and offering limited physical interpretability. This paper proposes a physics-informed prognostic framework for bearings operating under changing conditions, based on an Innovation-Corrected Extended Kalman Filter (ICEKF). The approach combines vibration-based HI extraction with a cumulative phenomenological degradation model, whose parameters are recursively updated using ICEKF. A key feature of the proposed filter is the correction of the innovation term using sensitivity Jacobians, which removes the direct influence of load and speed variations from the measurement update. This enables estimation of a regularized HI trajectory driven solely by degradation dynamics. Degradation onset is detected using an unsupervised two-stage slope-based method. The corrected HI trajectory is extrapolated using a locally weighted power-law regression to estimate the RUL, with uncertainty quantified analytically. The methodology is experimentally validated on the SURVIB test bench using five run-to-failure tests on 51 207 thrust ball bearings subjected to variable load-speed profiles. The results demonstrate improved RUL prediction accuracy compared to classical regression methods, standard EKF, and particle filtering approaches, while enabling physically consistent estimation of spalled surface evolution.