Institution profile

Oak Ridge National Laboratory

Academic institutionnorthamerica · us
Official website
Research library396linked papers
Opportunities0open roles
Selected work

Representative Papers

Using Image Processing Techniques to Identify and Quantify Spatiotemporal Carbon Cycle Extremes

Nov 01, 20222022 IEEE International Conference on Data Mining Workshops (ICDMW)

This study addresses the critical limitations—poor temporal continuity and cross-boundary detection—in identifying carbon-cycle extreme events (STEs). We propose a novel spatiotemporal feature detection method rooted in image processing: modeling carbon flux anomalies as a 3D spatiotemporal manifold defined by local neighborhood structures, and automatically identifying contiguous, cross-regional STEs via 3D connected-component analysis. This work pioneers the integration of neighborhood-based topology and 3D connectivity into STE detection, formally defining the “spatiotemporal extreme manifold” (STE) and overcoming the constraints of conventional pixel-wise thresholding approaches. Leveraging multi-source GPP remote sensing data and Earth system model outputs, we conduct attribution analysis: among 100 detected STEs, the top contributors account for over 75% of negative carbon anomalies; STE area and intensity exhibit strong concordance with carbon loss magnitude; and climate drivers are attributed in a structurally dependent manner—revealing non-linear, spatially heterogeneous forcing mechanisms.

2 citationsRead paper

Fast Relax-and-Round Unit Commitment with Sub-hourly Mechanical and Ramp Constraints

Feb 03, 2026

This work addresses the computational challenges of solving unit commitment (UC) problems under high-volatility load and distributed generation scenarios, where conventional approaches struggle with sub-hourly mechanical and ramping constraints and suffer from slow solution times. To overcome these limitations, the paper proposes a heuristic relaxation-and-rounding method that avoids linear approximations and instead leverages existing continuous optimization solvers enhanced with tailored heuristics to efficiently handle complex operational constraints. The approach preserves model fidelity while achieving speedups of several orders of magnitude. Experimental results demonstrate that the method successfully solves large-scale, sub-hourly UC instances that are intractable for current state-of-the-art tools, thereby substantially improving computational feasibility and practical applicability.

1 citationsRead paper

What Trace Powers Reveal About Log-Determinants: Closed-Form Estimators, Certificates, and Failure Modes

Jan 18, 2026

Log-determinant estimation is crucial in Gaussian processes and Bayesian model comparison, yet conventional methods fail under high condition numbers. This work proposes a closed-form estimator based on matrix trace powers: by leveraging derivatives of the moment-generating function of normalized eigenvalues, combined with the log-transform $K(t) = \log M(t)$, local integer-point interpolation, and spectral lower-bound constrained optimization, it enables efficient computation. Requiring only $m = 4$–$8$ trace power evaluations, the method achieves $O(m)$—effectively constant-time—complexity. It also establishes, for the first time, a fundamental limitation: finite positive moments cannot uniformly approximate arbitrary spectral distributions. The resulting verifiable upper and lower bounds, together with tail-sensitivity analysis, offer rigorous error control and failure diagnostics, ensuring both accuracy and reliability even in high-condition-number regimes.

1 citationsRead paper

Numerical Simulation Informed Rapid Cure Process Optimization of Composite Structures using Constrained Bayesian Optimization

May 30, 2025

To address structural deformation in composite material curing arising from thermo-chemo-mechanical coupling, this paper proposes an efficient process optimization method integrating Gaussian process surrogate modeling with constrained Bayesian optimization (cBO). The objective is to minimize curing-induced deformation while strictly satisfying critical process constraints—particularly complete cure. This work represents the first systematic application of cBO to such multi-physics coupled optimization problems. Compared with conventional genetic algorithms requiring over 1,000 iterations, cBO achieves convergence in fewer than 50 iterations for both flat-plate and L-shaped composite structures, yielding over 20× computational speedup (>96% reduction in function evaluations) and maintaining a prediction error below 4%. The approach successfully attains dual optimization goals: strict constraint satisfaction and global minimization of residual deformation.

1 citationsRead paper

A General Framework for Error-controlled Unstructured Scientific Data Compression

Sep 16, 2024IEEE International Conference on e-Science

To address insufficient neighbor information exploitation and uncontrollable error in lossy compression of unstructured mesh scientific data, this paper proposes an error-bounded multi-component compression framework. First, irregular mesh data are interpolated onto a regular grid; then, the interpolated field and residual are compressed separately. This is the first method to achieve generic, strictly error-bounded compression for arbitrary-topology unstructured meshes. By decoupling interpolation from residual encoding, the framework seamlessly integrates with mainstream floating-point compressors such as ZFP and FPZIP. Evaluated on four representative scientific datasets, it achieves 2.3–3.5× higher average compression ratios than state-of-the-art methods while rigorously satisfying absolute error bounds ranging from 1e−6 to 1e−2—demonstrating superior accuracy-compression trade-offs.

1 citationsRead paper
Recent publications

Latest Papers