Imbibition experiments

In this post, you will understand how we conducted the measurement of imbibition phenomenon . This study focus on the behavior of hemp shives and try to explain how it interact with cement and additives. The aim is to define the best powder mix-design.

Imbibition experiment

We had to test the powder to determine how we could get a good precision on our products. That is why, porosity, size of particules as well the porosity of the medium were identified as the key parameters.

Before all, we chose to study ciment alone with different porosities : 60%, 65% and 70%. This was chosen by the capability of cement particles to be agenced in the test tubes. As you can see below, we chose to fix the volume and manage the mass of cement inside of it.

Cement powder at different porosities

The main problem we had to face up is that a porosity gradient appear among the porous medium heigh. For example, at 70% of porosity, cracks appear mostly on the top of the tube rather than on the bottom.

Working at different ratios of cement/hemp shives is also a challenge. First because the hemp has a way lower density compared to cement powder.The video below highlight the imbibition process with a mix-design having a big amount of hemp particles.

https://video.wixstatic.com/video/c4908e_dcb2b0c6fb1545319c0bb16fee18d277/720p/mp4/file.mp4

Mix Cement/shives : 30/70%, 70%porosity

When we had defined the best composition and porosity of our mixes, we could first study the evolution of imbibition front. Using a camera, we studied frame by frame the shape and the speed of the front. Comparing all our compositions and porosities allowed us to more understand the role of each parameter.

Influence of porosity and hemp shive

As we were learning more about hemp behavior in cement paste, we chose to use MEB microscopy to see how the hemp particles were affected by the water during imbibition.

We higlighted that water tends to quarter the flakes of hemp with IR spectroscopyn. This has a huge importance in the evolution of particle during the hydratation of the binder and will naturally modify the mechanical resistance and isolation properties of our block.

Finally, we decided to also characterize imbibition using evolution of sample's mass while water imbibibed. Using indirect methods as weighing the diminution of water mass contained in a bulk, most of the results oriented our reflections the same way even though repetability wasn't good.

IA has also been carried out to help the data processing. This study was carried out using Python n to help frames analysis as it could take a lot of time to do it manually. The strategy is here to define a gray level to see the difference of a wet part compared to the dry one. Even though reflexions on the tube brang mistakes, we chose to use artificial intelligence to get an adaptative gray balance that shed a good light on our results.

This work have finally been a first step on the project and allowed us to know the parameters that should be tailored in the printer to reach high perofrmance and accuracy of the final element produced.