Acquisition de données de conductivité thermique par microfluidique et chémoinformatique
R. Jimenez · HAL (Le Centre pour la Communication Scientifique Directe) · 2023
Global warming-related climate change demands prompt actions to reduce greenhouse gas (GHG) emissions, particularly carbon dioxide. To reduce GHGs, biomass-based biofuels containing oxygenated compounds represent a promising alternative of energy source. To convert biomass into energy, several processes are operated at high pressure and high temperature conditions, which therefore required knowledge of thermophysical property data for their dimensioning, particularly thermal conductivity.Standard methods for measuring thermal conductivity involve the use of thermocouples and heat flow meters, but they can be time-consuming and error-prone when handling fluids. To overcome these challenges, smaller scale technologies like MEMS and Microfluidics have been proposed. By combining these two technologies, microfluidic devices incorporating MEMS have been created, known as Lab On Chip. This approach offers several advantages including a reduced reagent consumption, shorter operating times, and improved heat and mass transfers. Since then, research on microfluidic devices has advanced and the technology has improved, allowing for a wider range of operating conditions in temperature and pressure. However, existing microfluidic systems used for measuring thermal conductivity lack chemical inertness and adequate thermomechanical properties required for application in various fluid systems and conditions. Consequently, the creation of a microfluidic device that can withstand harsh conditions is still crucial.In this context, a microfluidic device has been developed to measure the thermal conductivity of liquids at different temperatures. The purpose of this is to generate new experimental data on thermal conductivity for oxygenated compounds that are not currently found in the existing literature. To ensure accuracy, the device was tested with liquids of known properties, before being used for generated new experimental data for which there is scarce information.Although microfluidics offers a great way for producing new data, it's important to consider other approaches such as modeling to supplement experimental data or to feed process simulators. The data available on the thermal conductivity of oxygenated compounds is limited and often varies depending on the source. This is especially true under extreme conditions. Machine learning can be used to develop predictive models that are powerful and can learn from existing data. For these models to be accurate and trustworthy, reference experimental data are required. These models can also be used to streamline the design of future experiments, saving time and money. By combining microfluidics and modeling, new thermal conductivity data can be generated, particularly at unexplored temperature conditionsOn this basis, a predictive model has been developed to determine thermal conductivity of liquids. In order to create accurate predictive models, it is necessary to have reliable and validated data. To ensure this, a database has been created, which includes both existing literature data and new experimental data obtain with the developed microfluidic device. After an extensive data collection, followed by data curation and experimental data generation, predictive models have been created and compared.Modeling and experiments complement each other. Together, they enhance the accuracy of predictions and the comprehension of the system under investigation.