Vishal V. R. Nandigana

This paper presents a data-driven approach to solve heat conduction problems, in particular 2D heat conduction problems. The physical laws which govern such problems are modeled by partial differential equations. We examine temperature distributions of conductors that have square geometry subjected to various boundary conditions, both Dirichlet and Neumann. The data consists of images of these distributions in a semi-continuous form. Conventionally, such problems may be solved analytically or using numerical methods which can be computationally expensive. We attempt to use Image-Based Deep Learning algorithms such as encoder-decoders and variational auto-encoders which do not involve the physical laws of the problem. We also study the efficacy of deterministic models against probabilistic models and the feasibility of using image-based deep-learning methods for engineering applications.

In this paper, we perform all-atom molecular dynamics (AA-MD) simulations to predict noise in solid-state nanopores. The simulation system consists of ∼70,000 to ∼350,000 atoms. The simulations are carried out for ∼1.3 µs over ∼6500 CPU hours in 128 processors (Intel® E5-2670 2.6 GHz Processor). We observe low and high frequency noise in solid-state nanopores. The low frequency noise is due to the surface charge density of the nanopore. The high frequency noise is due to the thermal motion of ions and dielectric material of the solid-state nanopore. We propose a generalised noise theory to match both the low and high frequency noise. The study may help ways to study noise in solid-state nanoporous membranes using MD simulations.

In this paper, we report for the first time overlimiting current near a nanochannel using all-atom molecular dynamics (MD) simulations. Here, the simulated system consists of a silicon nitride nanochannel integrated with two reservoirs. The reservoirs are filled with 0.1M potassium chloride (KCl) solution. A total of ∼ 1.1 million atoms are simulated with a total simulation time of ∼ 1 μs over ∼ 30000 CPU hours using 128 core processors (Intel(R) E5-2670 2.6 GHz Processor). The origin of overlimiting current is found to be due to an increase in chloride (Cl-) ion concentration inside the nanochannel leading to an increase in ionic conductivity. Such effects are seen due to charge redistribution and focusing of the electric field near the interface of the nanochannel and source reservoir. Also, from the MD simulations, we observe that the earlier theoretical and experimental postulations of strong convective vortices resulting in overlimiting current are not the true origin for overlimiting current. Our study may open up new theories for the mechanism of overlimiting current near the nanochannel interconnect devices.

The ion transport measurements using various ion-exchange membranes (IEMs) face several challenges, including controllability, reproducibility, reliability, and accuracy. This is due to the manual filling of the solutions in two different reservoirs in a typical diffusion cell experiment with a random flow rate, which results in the diffusion through the IEM even before turning on the data acquisition system as reported so far. Here, we report the design and development of an automated experimental setup for ion transport measurements using IEMs. The experimental setup has been calibrated and validated by performing ion transport measurements using a standard nanoporous polycarbonate membrane. We hope that the present work will provide a standard tool for realizing reliable ion transport measurements using ion-exchange membranes and can be extended to study other membranes of various pore densities, shapes, and sizes.

The two-dimensional (2D) heterojunctions of layered transition metal dichalcogenides (TMDs) with different bandgaps are the basis of modern electronic and optoelectronic devices such as high-speed transistors, light-emitting diodes, diode lasers and so on. Although, complex heterostructures (HSs) have been widely fabricated in the vertical direction via van der Waals (vdWs) stacking of different TMDs, but, atomic stitching of such 2D materials in the horizontal direction is proven to be so far most challenging. Here, we report a two-step sequential growth of monolayer n-type MoSe2 - p-type WSe2 lateral junction using chemical vapor deposition (CVD), which was confirmed from Raman and photoluminescence measurements. This work could be extended to other families of TMDs and provide a platform for the development of new device functionalities such as in-plane transistors and diodes to be integrated within a single atomically thin layer.

The osmotic energy from a salinity gradient (i. e. blue energy) is identified as a promising non-intermittent renewable energy source for a sustainable technology. However, this membrane-based technology is facing major limitations for large-scale viability, primarily due to the poor membrane performance. An atomically thin 2D nanoporous material with high surface charge density resolves the bottleneck and leads to a new class of membrane material the salinity gradient energy. Although 2D nanoporous membranes show extremely high performance in terms of energy generation through the single pore, the fabrication and technical challenges such as ion concentration polarization make the nanoporous membrane a non-viable solution. On the other hand, the mesoporous and micro porous structures in the 2D membrane result in improved energy generation with very low fabrication complexity. In the present work, we report femtosecond (fs) laser-assisted scalable fabrication of μm to mm size pores on Graphene membrane for blue energy generation for the first time. A remarkable osmotic power in the order of μW has been achieved using mm size pores, which is about six orders of magnitudes higher compared to nanoporous membranes, which is mainly due to the diffusion-osmosis driven large ionic flux. Our work paves the way towards fs laser-assisted scalable pore creation in the 2D membrane for large-scale osmotic power generation.

Artificial intelligence (AI) and more specific subsets of AI such as machine learning (ML) and deep learning (DL) have become widely available in recent years. Open source software packages and languages have made it possible to implement complex AI based data analysis and modeling techniques on a wide range of applications. The application of these techniques can expedite existing models or reduce the amount of physical testing required. Two data sets were utilized to examine the effectiveness of multiple ML techniques to estimate experimental outcomes and to serve as a substitute for additional testing. To achieve this complex multi-variant regressions and neural networks were utilized to create estimating models. The first data sets of interest consist of a pool fire experiment that measured the flame spread rate as a function of initial fuel temperature for 8 different fuels, including Jet-A, JP-5, JP-8, HEFA-50, and FT-PK. The second data set consists of hot surface ignition data for 9 fuels including 4 alternative piston engine fuels for which properties were not available. When properties were not available multiple imputation by chained equations (MICE) was utilized to estimate fluid properties. 10 different ML techniques were implemented to analyze the data and R-squared values as high as 92% were achieved.

Negative differential resistance (NDR) is one of the nonlinear transport phenomena in which ionic current decreases with the increase in electromotive potential. Electro-osmosis, diffusio-osmosis, and surface charge density of pores are the driving forces for observing NDR in nanoscale ion transport. Here, we report electrodiffusioosmosis induced NDR using micro to millimeter size pores in a two-dimensional (2D) graphene-coated copper (Gr/Cu) membrane. Along with NDR, we also observed ion current rectification (ICR), in which there is preferential one-directional ion flow for equal and opposite potentials. The experimentally observed NDR effect has been validated by performing ion transport simulations using Poisson-Nernst-Planck (PNP) equations and Navier-Stokes equations with the help of COMSOL Multiphysics considering salinity gradient across the membrane. Charge polarization induced electro-osmotic flow (EOF) dominates over diffusio-osmosis, causing the backflow of low concentration/conductivity solution into the pore, thereby causing NDR. This finding paves the way toward potential applications in ionic tunnel diodes as rectifiers, switches, amplifiers, and biosensors.

In modern GDI engines, in-cylinder flows play a crucial role in the performance and emission characteristics. Adding baffles to the cylinder head is one way to enhance these flows, as baffle effectiveness is linked to its height. This study aims to investigate the influence of baffle height on the in-cylinder flows and its impact on the performance and emission characteristics of a four-stroke, four-valve, spray-guided GDI engine using computational fluid dynamics (CFD) analysis. The maximum baffle height is determined based on the geometry of the engine. Throughout the analysis, the engine operates at a constant speed of 1000 rev/min., under part-load conditions. From the results, the baffle heights of 2 mm and 3 mm significantly improve mixture stratification compared to the base engine. Furthermore, the indicated mean effective pressure (IMEP) for the engine with 2 mm and 3 mm baffle height improves by about 3% and 4%, respectively, compared to the base engine. In addition, with the increase in the baffle height, there is a significant reduction in the HC emissions with a marginal increase in NOX emissions.