A5023758393- Model Reduction and Neural Networks
- Machine Fault Diagnosis Techniques
- Heat Transfer Mechanisms
- Risk and Safety Analysis
- Fluid Dynamics and Vibration Analysis
- Anomaly Detection Techniques and Applications
- Heat transfer and supercritical fluids
- Adversarial Robustness in Machine Learning
- Aerodynamics and Fluid Dynamics Research
- Heat Transfer and Optimization
- Fatigue and fracture mechanics
- Fluid Dynamics and Turbulent Flows
- Advanced Decision-Making Techniques
- Reliability and Maintenance Optimization
- Gear and Bearing Dynamics Analysis
- Fault Detection and Control Systems
- Turbomachinery Performance and Optimization
- Gaussian Processes and Bayesian Inference
- Engineering Diagnostics and Reliability
- Sports Performance and Training
- Neural Networks and Applications
Purdue University West Lafayette
2023
Sharif University of Technology
2020
In this paper, we adopt conformal prediction, a distribution-free uncertainty quantification (UQ) framework, to obtain confidence prediction intervals with coverage guarantees for Deep Operator Network (DeepONet) regression. Initially, enhance the frameworks (B-DeepONet and Prob-DeepONet) previously proposed by authors using split prediction. By combining our Prob- B-DeepONets, effectively quantify generating rigorous DeepONet Additionally, design novel Quantile-DeepONet that allows more...
In the pursuit of accurate experimental and computational data while minimizing effort, there is a constant need for high-fidelity results. However, achieving such results often requires significant resources. To address this challenge, paper proposes deep operator learning-based framework that limited dataset training. We introduce novel physics-guided, bi-fidelity, Fourier-featured network (DeepONet) effectively combines low- datasets, leveraging strengths each. our methodology, we begin...
This paper proposes a new data-driven methodology for predicting intervals of post-fault voltage trajectories in power systems. We begin by introducing the Quantile Attention-Fourier Deep Operator Network (QAF-DeepONet), designed to capture complex dynamics and reliably estimate quantiles target trajectory without any distributional assumptions. The proposed operator regression model maps observed portion its unobserved trajectory. Our employs pre-training fine-tuning process address...
This paper designs surrogate models with uncertainty quantification capabilities to improve the thermal performance of rib-turbulated internal cooling channels effectively. To construct surrogate, we use deep operator network (DeepONet) framework, a novel class neural networks designed approximate mappings between infinite-dimensional spaces using relatively small datasets. The proposed DeepONet takes an arbitrary continuous rib geometry control points as input and outputs detailed...
This paper designs surrogate models with uncertainty quantification capabilities to improve the thermal performance of rib-turbulated internal cooling channels effectively. To construct surrogate, we use deep operator network (DeepONet) framework, a novel class neural networks designed approximate mappings between infinitedimensional spaces using relatively small datasets. The proposed DeepONet takes an arbitrary continuous rib geometry control points as input and outputs detailed...
In the pursuit of accurate experimental and computational data while minimizing effort, there is a constant need for high-fidelity results. However, achieving such results often requires significant resources. To address this challenge, paper proposes deep operator learning-based framework that limited dataset training. We introduce novel physics-guided, bi-fidelity, Fourier-featured Deep Operator Network (DeepONet) effectively combines low datasets, leveraging strengths each. our...
Deep Operator Network (DeepONet) is a neural network framework for learning nonlinear operators such as those from ordinary differential equations (ODEs) describing complex systems. Multiple-input deep (MIONet) extended DeepONet to allow multiple input functions in different Banach spaces. MIONet offers flexibility training dataset grid spacing, without constraints on output location. However, it requires offline inputs and cannot handle varying sequence lengths testing datasets, limiting...
Estimation of remaining useful life (RUL) rolling element bearings (REBs) has a major effect on improving the reliability in industrial plants. However, due to complex nature fault propagation these components, their prognosis is affected by various uncertainties. This intensified when recorded data offline, which very common for many machines lower cost rather than online monitoring strategy. In present paper, order overcome shortcoming feed-forward neural network (FFNN) REBs prognostics,...