arXiv subjects

Rohit Saini

Publications and source records attributed to Rohit Saini.

4 recordsLinked to original sources

arXiv2026

PLSQLBench: Benchmarking LLM Systems for Executable Procedural Database Programming

We present PLSQLBench, to our knowledge the first benchmark for evaluating whether LLMs can write executable PL/SQL programs, with correctness measured through execution-based tests. Existing LLM evaluations largely target general-purpose code generation or declarative text-to-SQL, leaving procedural database programming underexplored. PLSQLBench contains 2,865 instances: 2,594 single-turn tasks and 271 multi-turn conversations spanning 978 turns. The benchmark combines complex schema-grounded tasks over enterprise-style Spider 2 databases, simpler schema-grounded tasks derived from Spider, and MBPP-derived procedural problems, covering varying levels of database grounding and procedural complexity. Experiments with eight LLMs reveal recurring difficulties in schema grounding, PL/SQL dialect fidelity, procedural control flow, exception handling, and cross-turn consistency. Tool-augmented LLM agents improve performance on several schema-grounded evaluations, although substantial gaps remain. These results highlight procedural database programming capabilities not directly assessed by conventional code generation or text-to-SQL benchmarks. Our code is available at https://github.com/oracle-samples/plsqlbench.

cs.CL↗

arXiv2025

Enhanced Performance and Stability of Perovskite Solar Cells with Ag-Cu-Zn Alloy Electrodes

Though the common metal electrode-based perovskite solar cells have achieved a power conversion efficiency of >25%, they also play a crucial role in accelerating the degradation of the cells. In this study, we investigated phase transition engineering in Ag electrodes via Cu and Zn alloying, transforming from a cubic to a tetragonal phase. These alloyed electrodes are then thermally deposited as back electrodes in perovskite solar cells. We conducted a comprehensive analysis of the pure Ag and Ag-Cu-Zn alloys deposited atop a hole-transport layer for use in Cs0.05(FA0.83MA0.17)0.95Pb(I0.83Br0.17)3-based solar cells. Our findings reveal that solar cells developed with pure Ag electrodes demonstrate a power conversion efficiency (PCE) of 18.71%, characterized by a fill factor (FF) of 74.8%, an open-circuit voltage (VOC) of 1.08 V, and a short-circuit current density (JSC) of 23.17 mA/cm2. Conversely, solar cells fabricated with optimized Ag0.875Cu0.120Zn0.005 electrodes exhibit enhanced performance metrics, with an FF of 72.5%, VOC of 1.12 V, and JSC of 23.39 mA/cm2, culminating in an elevated PCE of 19.02%. Moreover, this electrode demonstrates remarkable durability, sustaining operational integrity for 460 hours for the PSCs stored in the N2 glove box, in contrast to the 320 hours for cells with Ag electrodes. The Ag-Cu-Zn alloys exhibited high resistance to corrosion and good adhesion on the hole-transport material layer compared to a layer of Ag. These advancements may lead to the realization of cost-effective, durable, and efficient solar energy conversion systems.

cond-mat.mtrl-sci↗

arXiv2021

Assessment of Soot Formation Models in Lifted Ethylene/Air Turbulent Diffusion Flame

In the present study, soot formation in the turbulent lifted diffusion flame, consisting of ethylene-air is numerically investigated using three different soot modeling approaches and is comprehensively reported. For turbulence-chemistry interaction, Flamelet generated manifold (FGM) model is used. A detailed kinetics is used which is represented through POLIMI mechanism (Ranzi et al. 2012). Soot formation is modeled using two different approaches, semi-empirical two-equation approach and Quadrature methods of moments approach, where both the approaches consider various sub-processes such as nucleation, coagulation, surface growth and oxidation. The radiation heat transfer is taken into account considering four fictitious gasses in conjunction with the weighted-sum-of-gray gas (WSSGM) approach for modeling absorption coefficient. The experimental and earlier published numerical data from Köhler et al. (2012) and Blacha et al. (2011) are used for assessment of different soot modeling approaches. The discrepancies between numerical and experimental data are observed due to under-prediction of OH radicals concentration and poor fuel-air mixing ratios in the vicinity of the fuel jet region leading to early soot formation and the trends are unaffected after invoking radiation.

physics.flu-dyn↗

arXiv2021

Investigation of NOx in piloted stabilized methane-air diffusion flames using Finite-rate and Infinitely-fast chemistry based combustion models

The present work reports on the numerical investigation of NOx in three turbulent piloted diffusion flames of different levels of extinction. The study involves two-dimensional axisymmetric modeling of combustion in these flames with fairly detailed chemistry, i.e. GRI 3.0 mechanism. The main focus of the study is to analyze the effects of the two different combustion model approaches, such as infinitely fast chemistry based unsteady flamelet and finite rate chemistry based EDC, in predicting the NOx formation in three piloted methane jet flames (Sandia D, E, and F). The EDC approach is able to predict the passive scalar quantities but shows over-prediction in the reactive scalar quantities and NO prediction, while the unsteady flamelet modeling is found to be essential in predicting the accurate formation of slow kinetic species like NOx. The inability of flamelet and EDC approach in capturing localized flame extinction is observed, which lead to an over-prediction of NOx at larger downstream locations. Further, the dominance of NOx formation pathways is investigated in all three flames.

physics.flu-dyn↗