Image Analysis of the Microstructure of Cement...My research looks at the microstructure of cement paste analysing the 3 phases: Pores, unhydrated cement (Clinker Grains) and hydrated

Post on 22-Sep-2020

1 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

Transcript

Image Analysis of the Microstructure of CementKaung Tin Khant

Supervisor: Dr. David Waite, Dr. Mark BlighResearch Theme: Resources and Infrastructure for the Future

Aims and Objectives

Method

Results and DiscussionRow 1 – SEM Image of Sample 1 Row 2 – X-Ray Tomography Image of Sample 2

• To obtain a model which could be used to compute the macroscale properties of cementthrough its microstructure.

• To identify the main cause of failure in cement by analysing its microstructure.• To improve the mechanical properties of cement by identifying an optimum structure of

cement at the microscale.

2. Modify the Image for use – Filter/Segment

Fig 1. Images of Cement

As can be seen from both images in Figure 1, the images obtained through both methods are ‘noisy’. Thus, filtering is required to be able to better identify the different materials found within the cement paste.The histograms are then used to obtain thresholds to segment at (ie. The 2 peaks). From Fig 2 and 3, it can be seen that both samples have varying quantities of clinker grains, pores, and gels. Despite this, Fig 4shows that the most likely cause for failure will be due to the pores (as indicated by the ‘red’ zones from the stress graphs). Properties of the cement samples were also obtained to help verify the model. Notethat these properties are just approximations from the code and have already been converted to the macroscale. No physical testing was completed to be able to compare these results to its actual values, butthe results do compare well to the general properties of concrete (Elastic Modulus of Concrete ranges from 17-30 GPa).

ConclusionThere is a lot of potential in this research and if a sufficient model was to be created,properties of cement paste could easily be computed by just obtaining an image ofits microstructure. This model could subsequently be used to identify the optimummicrostructure of cement. As a result, this could help in making a stronger cement, oralternatively make current day cement with less material; which would in turn saveon costs.

BackgroundCement is a commonly used material in the field of construction. Cement is a fine powderproduced by heating multiple materials in the kiln to form clinker, then grinding this clinker andgradually adding small amounts of other materials. Hydration is then used to form a cementpaste, which is used as a basic ingredient in making concrete. During hydration, some of thetricalcium silicate (C3S – Clinker Grains) is reacted to produce Calcium Hydroxide (CH, Depositedinto pores) and Calcium Silicate Hydrate (C-S-H – The gel that engulfs all the other materials).My research looks at the microstructure of cement paste analysing the 3 phases: Pores,unhydrated cement (Clinker Grains) and hydrated cement (the Gel).

Fig 2. Filtered Images - A. Black – Pores B. Gray – Gel/ C. White – Clinker Grains

Fig 3. Histogram of Filtered Images -Quantity of Materials at a specific Greyscale Fig 4. Stress Graphs

Table 1 : Properties of the Cement Displacement Reaction Force required to cause displacement (kN)

Bulk Elastic Modulus(GPa)

SEM Image 10% of Sample 1.81 21.61

X-Ray Tomography Image 10% of Sample 1.44 26.14

1. Obtain an image of the Microstructure of Cement through X-Ray Tomography (shown above) or Scanning Electron Microscopy (SEM).

3. Analyse the phases within the image (in this case using a Matlab code)

4. Results then repeat with different images

or parameters

Fig A. X-Ray Tomography Fig B. Filtering an Image Fig C. Different Phases of Cement Paste

References1. Huang , K. K. A. L., 2012. Statistical homogenization of elastic properties of cement paste basedon X-ray. pp. 1-11.

Acknowledgements I would like to thank Professor Chongmin Song from the Civil School of Engineering for providing mewith the tools to analyse the images. I would also like to thank Dr. David Waite and Dr. Mark Bligh fortheir invaluable training and guidance.

top related