Ashraf Kassem
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Ashraf Kassem has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Kursat Kara
School of Mechanical and Aerospace Engineering, Oklahoma State University, Stillwater, USA
8 shared publications - Omer San
Department of Mechanical and Aerospace Engineering, University of Tennessee, Knoxville, USA
8 shared publications - Shafi Al Salman Romeo
School of Mechanical and Aerospace Engineering, Oklahoma State University, Stillwater, OK 74078, USA
8 shared publications - Bipin Tiwari
Department of Engineering Mathematics and Computing, Madhav Institute of Technology and Science, Gwalior, India
7 shared publications - Lamisa Musharrat2 shared publications
Cole Kazemba
1 shared publication
Latest Publications
- Open Access
- 25 August 2026
- Published by Elsevier in SSRN Electronic Journal
Toward Real-Time Nonlinear Parameter Identification of Blunt-BodyEntry Capsules
Characterizing the nonlinear pitch damping of blunt-body reentry capsules is a mission-critical inverse problem. The pitch-damping sum is a strongly nonlinear function of the angle of attack whose cycle- averaged zero crossing sets the limit-cycle amplitude, and established Markov chain Monte Carlo (MCMC) frameworks recover it offline, in batches, from the complete trajectory record. This work adapts square-root joint unscented Kalman filtering to this problem, assimilating free-flight CFD trajectories of the Genesis Sample Return Capsule at Mach 1.44 one observation at a time, jointly estimating the dynamic states, the static-stability derivative, an empirical cubic restoring-moment correction for amplitude-dependent phase error, and a five-point spline damping curve. A Fisher- information analysis first selects the control-point grid whose parameters the data can individually identify, a well-posedness requirement specific to sequential covariance propagation; the square-root formulation then guarantees a valid covariance at every step, and a physics-aware process noise derives the parameter variances from interpretable drift tolerances. From a neutral initialization, a single 0.765 s pass recovers the angle-resolved damping curve, whose effective damping places the limit-cycle amplitude between 9.8◦–11.2◦ across interpolation bases, a range that brackets the independently established ≈11◦ value previously reported for this configuration and read from the oscillation envelopes. The learned model predicts withheld trajectories from their initial conditions alone, synchronizes its state from thirty sparse observations, tolerates an order-of-magnitude error in the assumed sensor noise, and incorporates a new flight in seconds: training costs 1.9±0.4 s compared with roughly 252 minutes for direct MCMC, laying the groundwork for a real-time digital twin of reentry-capsule dynamic stability.
- 25 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessToward Real-Time Nonlinear Parameter Identification of Blunt-BodyEntry Capsules
Characterizing the nonlinear pitch damping of blunt-body reentry capsules is a mission-critical inverse problem. The pitch-damping sum is a strongly nonlinear function of the angle of attack whose cycle- averaged zero crossing sets the limit-cycle amplitude, and established Markov chain Monte Carlo (MCMC) frameworks recover it offline, in batches, from the complete trajectory record. This work adapts square-root joint unscented Kalman filtering to this problem, assimilating free-flight CFD trajectories of the Genesis Sample Return Capsule at Mach 1.44 one observation at a time, jointly estimating the dynamic states, the static-stability derivative, an empirical cubic restoring-moment correction for amplitude-dependent phase error, and a five-point spline damping curve. A Fisher- information analysis first selects the control-point grid whose parameters the data can individually identify, a well-posedness requirement specific to sequential covariance propagation; the square-root formulation then guarantees a valid covariance at every step, and a physics-aware process noise derives the parameter variances from interpretable drift tolerances. From a neutral initialization, a single 0.765 s pass recovers the angle-resolved damping curve, whose effective damping places the limit-cycle amplitude between 9.8◦–11.2◦ across interpolation bases, a range that brackets the independently established ≈11◦ value previously reported for this configuration and read from the oscillation envelopes. The learned model predicts withheld trajectories from their initial conditions alone, synchronizes its state from thirty sparse observations, tolerates an order-of-magnitude error in the assumed sensor noise, and incorporates a new flight in seconds: training costs 1.9±0.4 s compared with roughly 252 minutes for direct MCMC, laying the groundwork for a real-time digital twin of reentry-capsule dynamic stability.
- 27 April 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 178
Discovery of governing equations for reduced-order aerodynamic modes of atmospheric entry capsules
- 27 April 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 178
Discovery of governing equations for reduced-order aerodynamic modes of atmospheric entry capsules
- 13 April 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 177
Data-efficient neural operator surrogates for nonlinear pitch dynamics of blunt-body re-entry vehicles
- 13 April 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 177
Data-efficient neural operator surrogates for nonlinear pitch dynamics of blunt-body re-entry vehicles
- 12 April 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 177
Emulator-accelerated Bayesian framework for nonlinear parameter estimation of entry capsule dynamic stability
- 12 April 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 177
Emulator-accelerated Bayesian framework for nonlinear parameter estimation of entry capsule dynamic stability
- 3 January 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 171
Data-driven reduced-order modeling framework for blunt-body aerodynamic characterization
- 3 January 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 171
Data-driven reduced-order modeling framework for blunt-body aerodynamic characterization
- 11 October 2025
- Published by Elsevier in Aerospace Science and Technology
- Vol. 168
Multifidelity data fusion for analyzing dynamic stability of atmospheric entry vehicles
- 11 October 2025
- Published by Elsevier in Aerospace Science and Technology
- Vol. 168
Multifidelity data fusion for analyzing dynamic stability of atmospheric entry vehicles
- 22 July 2025
- Published by Elsevier in Aerospace Science and Technology
- Vol. 167
Bayesian framework for estimating dynamic stability derivatives in 6-DoF blunt-body entry vehicles
- 22 July 2025
- Published by Elsevier in Aerospace Science and Technology
- Vol. 167
Bayesian framework for estimating dynamic stability derivatives in 6-DoF blunt-body entry vehicles
- Open Access
- 1 January 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (1)
An augmented physics informed neural network approach for blunt-body dynamics
This paper presents an ansatz-informed approach to modeling the dynamics of blunt-body entry vehicles by combining physics-based modeling with machine-learning techniques. The main focus is developing an augmented physics-informed neural network (PINN) to simulate the vehicle's behavior during atmospheric entry. The proposed PINN architecture is carefully crafted to incorporate existing knowledge about the fundamental physical principles governing the system. This integration enhances the accuracy and applicability of the model compared to conventional data-driven approaches. The proposed methodology is evaluated across various scenarios and demonstrates precise predictions of the vehicle's trajectory and essential parameters.
- 1 January 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (1)
Open AccessAn augmented physics informed neural network approach for blunt-body dynamics
This paper presents an ansatz-informed approach to modeling the dynamics of blunt-body entry vehicles by combining physics-based modeling with machine-learning techniques. The main focus is developing an augmented physics-informed neural network (PINN) to simulate the vehicle's behavior during atmospheric entry. The proposed PINN architecture is carefully crafted to incorporate existing knowledge about the fundamental physical principles governing the system. This integration enhances the accuracy and applicability of the model compared to conventional data-driven approaches. The proposed methodology is evaluated across various scenarios and demonstrates precise predictions of the vehicle's trajectory and essential parameters.