Introduction

In the context of the MarPCM experiment (Porter et al., 2023), we develop a neural network-based algorithm to help process images from the melting bridge experiments (Varas et al., 2021). Both solid and liquid phases are present during melting and, since the liquid PCM has a higher refractive index than air, traversing light rays are concentrated, as with a converging lens, which visually enlarges the shape of the solid PCM. This distortion can be a major obstacle in the optical measurement of the time-dependent liquid fraction since the apparent amount of solid is magnified.

The network is trained using synthetic images generated with a ray-tracing algorithm. This technique is based on following light rays backwards from the camera to the source and accounts for reflection, refraction, and absorption in the formation of the images. The rendered scene replicates the MarPCM setup.

Further details can be found in Martinez et al. (2023).


Methods

We combine the following methods:
- Numerical simulation of the phase change with thermocapillary effects. PCM melting is described using an enthalpy-porosity-based formulation of the Navier-Stokes equations and resolved using the finite element method (Varas et al., 2021).
- Generation of synthetic images of the melting process via ray-tracing. Images are produced using an in-house code. The rendered melting bridge scene is created using simulations of the melting process in microgravity while the code is validated using images from ground experiments; see the left panels of the figure.
- Image classification using neural networks. Synthetic images are used as inputs to a multiple-layer neural network. The network classifies each image in accord with its liquid fraction and provides an estimate of the real liquid-to-PCM volume fraction. Images are pre-processed using singular-value decomposition.


Results

An example of experiment image processing is given in the right panel of the figure. The liquid fraction is a critical parameter determining how well the real phase boundary can be determined from the image.


Conclusions

The results indicate that the melting process can be tracked much more accurately during the latter stages of melting. The initial melting stage, associated with a shallow layer of liquid near the surface of the bridge, is largely obscured by optical effects. Judged by the present numerical simulations and synthetic images, the developed neural network is capable of classifying the images and providing a good estimate of the liquid fraction over time.

Introduction

Organic phase change materials (PCMs) are used for energy storage and thermal management in numerous applications including space missions. Their low thermal conductivity, however, has impeded even wider use. Besides improvements in heat transport from natural convection, recent studies have shown that incorporating a free surface can significantly enhance performance via thermocapillary convection (Salgado Sánchez et al., 2020a).

Here, we analyze the enhancement achieved by combining natural and thermocapillary convection (Borshchak Kachalov et al., 2021) and by using optimal container geometries adapted to the natural shape of the evolving phase boundary to maximize the melting rate (Huang et al., 2022).


Method

The melting of n-octadecane is modeled numerically using the enthalpy-porosity formulation of the Navier-Stokes equations, which considers the solid and liquid phases of the PCM as a single continuous phase whose physical properties depend on temperature.

The process is analyzed first in a rectangular container with purely conductive heat transport as a reference. Each enhancement strategy — natural convection, thermocapillary convection and container design — is then applied sequentially in the manner shown for the five optimization paths (P.1 – P.5) in the figure. With thermocapillary effects, we restrict consideration to containers having a flat thermocapillary interface, i.e., open containers with a pinned free surface (Salgado Sánchez et al., 2020b).


Results

The PCM volume V significantly affects the melting process and largely determines the heat transport enhancement, measured by the ratio G between the melting time for the reference case and that of the given scenario. This improvement further depends on the type of dominant convective flow. While gravitational convection provides an enhancement ratio G on the order of 2 for small V and 5-8 for large V, thermocapillary flow can increase this relative enhancement by factors of 8-22 and 4-7, respectively.


Conclusions

The results demonstrate the efficacy of all three strategies for improving PCM performance. Often the first implemented strategy provides the greatest improvement relative to the reference case while supplemental strategies yield more modest gains, although not negligible ones. On the whole, melting times can be reduced by factors between 5 and 32 compared to the purely conducting reference configuration.

Introduction

MarPCM will investigate the efficacy of thermocapillary convection for augmenting the heat transfer rate of passive phase change materials (PCMs) in microgravity that incorporate a free surface; so-called thermocapillary-enhanced PCMs (TePCMs). Compared to conduction, thermocapillary flows can increase heat transport by a factor of two or more. The experiment seeks to understand heat and mass transport mechanisms during melting and solidification and assess the practicality of using TePCMs as passive control devices for space missions.

Here, we briefly revise the design, performance, and ground results of the cuboidal MarPCM cell. For further details, the reader is referred to Porter et al. (2023) and Martinez et al. (2024).

Experiment

The experiment consists of two units: the Electronics and Computer Unit (ECU) and the Experiment Cells and Diagnosis Unit (ECDU). The ECU contains the computer and other electronic components for power conditioning. The ECDU is an exchangeable module that includes two experiment cells (ECs) and elements for diagnosis and data acquisition. Three different ECDUs, holding a total of six different ECs, will be launched to the ISS in 2026 (subject to the final operations schedule).

The cuboidal EC has interior dimensions (in mm) of 22.5 (length) x 15 (height) x 25 (depth). Two types of cuboidal cells are designed. The Marangoni cell has a 5 mm layer of air above the PCM to allow thermocapillary convection during the phase change while the Reference cell is completely filled with PCM so that melting dynamics is driven solely by conduction.

To ensure precise temperature measurements, we select fast-response NTC temperature sensors, which are coupled with high-performance thermoelectric modules and a precision temperature controller for the thermal control of the ECs. The optical setup includes two cameras equipped with both telecentric and fixed focal length objective lenses, and an LED panel for illumination; see the left panel of the figure.

Results

Ground tests on the cuboidal Marangoni setup confirm its satisfactory operation; see the right panel of the figure. Results indicate sufficient image quality for processing, precise and stable temperature measurements, accurate thermal control, coherent melting times with the applied thermal gradients, and repeatability in design, manufacturing, and integration of the experiment units.

Conclusions

The developed hardware provides acceptable performance while the ground results obtained are coherent and compliant with the experiment requirements; the feasibility of the proposed design is thus demonstrated. Furthermore, the design illustrates a comprehensive approach towards developing space experiments.

Absorption technology can play a vital role in solving global warming problems. Among all the components of absorption machines, the absorber stands out as the main factor contributing to the reduction in heat and mass transfer. Since the 1990s, the concept of using Marangoni convection to enhance heat and mass transfer on an absorbing interface has become widespread. Marangoni convection is a flow driven by surface-tension gradients, which result from interfacial inhomogeneities in either temperature or solute concentration. To highlight the influence of Marangoni convection in absorption problems, various soluble as well as dissolved in gas-phase surfactants were used in experiments and numerical simulations. The addition of a surfactant to the absorbing interface locally induces solutal convection, further accompanied by thermal and buoyant convection. The overall conclusion of previous research was that the existing research recognized the critical role played by the Marangoni convection, but there is still uncertainty about the physical background.
Solutal Marangoni convection is an intricate phenomenon exhibiting complex and unsteady flow patterns. The interplay of solutal, thermal, and buoyant convection in the presence of absorption adds further complexity.
To clarify the role of Marangoni convection we present the results of comprehensive numerical simulations in presence [1] and absence of gravity. The conventional solution used in an absorber is the binary mixture of LiBr–H2O. Thus, we investigate the convective instability in a LiBr–water binary mixture, triggered by a local perturbation of uniform absorption. The decrease in mass fraction initiates solutal convection, leading to a local temperature change that, in turn, induces thermal Marangoni convection. We explore fluid dynamics, heat and mass transfer, revealing different regimes. The level of absorption and pattern formation and strongly influenced by gravity

Phase change materials (PCMs) possess the unique ability to absorb and release heat during phase transitions, making them invaluable for passive thermal control mechanisms. Since the 1970s, solid Phase Change Material (PCM) systems have played an important role in various space missions. In terrestrial applications, the presence of convective flows in the liquid phase helps to solve the problem of PCM low thermal conductivity. However, this approach is not applicable in microgravity. As an alternative strategy, the use of the thermocapillary effect, where a non-uniform temperature induces surface tension gradients driving convective motion, has been proposed as a source of convective heat transport in microgravity PCM devices. The upcoming MarPCM/ISS project aims to explore alternative methods for optimizing heat management [1]. The numerous numerical simulations of the melting phase highlighted enhancing heat transfer efficiency due to Marangoni convection.

In practice, thermal operations typically involve complete solid-liquid-solid conversion cycles. To optimize PCM performance, we explored three different scenarios based on the temperature configuration between sidewalls before and after melting. We identified the most promising case, which involves the timely switching of temperatures between cold and hot walls at a certain moment, preceding full melting or attaining steady state. This specific moment corresponds to the beginning of a decrease in heat extraction efficiency. To implement this concept in practice, we propose rotating the PCM material inside the package while maintaining wall temperatures fixed. The suggested rotation approach has the potential to simplify technological design and facilitate manipulation of the system. Future microgravity experiments could validate the efficacy of this innovative solution.

[1] Porter at al. The “Effect of Marangoni Convection on Heat Transfer in Phase Change Materials” experiment, Acta Astronautica, Volume 210, 2023, Pages 212-223

Introduction

Fluids, with all their complex behavior, are an essential part of space exploration. The reduced gravity environment of typical space missions alters how fluids interact with their surroundings and respond to external forces. Fluid manipulation and control in such conditions is both crucial and challenging.

The ‘‘Thermocapillary-based control of a free surface in microgravity’’ (ThermoSlosh) experiment aims to study the effectiveness of thermal and inertial forcing for fluid control in microgravity applications. ThermoSlosh was presented to the ISSSP competition and awarded 2nd place.

Experiment and method

The heart of the experiment is a cylindrical cell, half filled with silicone oil and subjected to controlled temperatures and vibrations. The thermal system of the experiment is composed of two pairs of thermocouples and Peltier elements while the optical system includes a high-resolution camera, a fixed-focus lens, and an LED panel. The experiment cell and optical system are mounted on a base plate which is attached to a stepper motor that can impose controlled vibrations.

To understand and predict the experimental measurements, we numerically analyze the 2D dynamics of the fluid; see the left panel of the figure. The influence of the applied temperature difference, vibrational amplitude, fluid viscosity, and contact angle are explored, and the response of the free surface is characterized using the rise and stabilization times, and the overshoot of its motion.

Results

The right panel of the figure shows examples of the temporal response of the (average) surface orientation under different scenarios. The applied ΔT and the thermocapillary flow it generates initially induce counterclockwise rotation of the interface. The rise time characterizes when the free surface first reaches an average orientation perpendicular to ΔT. The surface, however, continues to rotate because of inertia and overshoots. Subsequent cycles of clockwise and counterclockwise rotation with decreasing amplitude define the stabilization time for its approach to equilibrium.

Conclusions

Simulations indicate that the thermocapillary effect can be used to control the orientation of a free surface in microgravity while supplemental vibrations can be added to increase responsiveness. This control can be further improved by using PID feedback to adjust the applied temperatures in real time. Controller gains can be selected to optimize the stabilization time, energy cost, and reduce steady-state errors. Overall, the results demonstrate the potential of thermal forcing and vibrations for fluid management in reduced gravity and support the types of experiments proposed in the frame of the ThermoSlosh project.