P.H. Wanten and S.W.J. den Brok

Institut für Geowissenschaften, Universität Mainz, Becherweg 21, 55099 Mainz, Germany.
wanten @mail.uni-mainz.de, denbrok@mail.uni-mainz.de

Underground mining may locally lead to the development of a subsidence trough in areas where this is not desirable. Filling up this trough with "sand", in order to bring the ground level back to its original height, is not always possible where e.g. buildings or roads are present. We presently study the possibility of raising the surface artificially by the injection of swelling cement in undeep seated sand layers within the through.

Experiments were performed (i) to test the swelling capacity against a pressure (up to 1 MPa) of commercially available fine-grained (OPC-42.5 based) swelling cements, and (ii) to improve the injectability by mixing in additives that delay the hardening process and decrease the viscosity to make injection over long distances possible.

Preliminary results showed that expansions of up to 45% can be achieved when the cement used was allowed to harden without constrictions under atmospheric conditions. The expansion is caused by the formation of ettringite. When the cement was allowed to harden under an applied hydrostatic pressure of up to 1 MPa, then the expansion was reduced to around 10%. This appears to be due to the fact that now first the pore space within the cement itself, called the capillary porosity, (up to 30%) has to be filled before the expansion can take place.

Normally the hardening of these cements starts after 5 minutes. With the help of organic additives (mostly on a glucose basis) we managed to delay the start of the hardening by around 1 hour. This made it possible to inject the cement in sand (1 - 2 mm) in a 10 cm diameter translucent tube over a distance of 35 cm in 5 minutes. The injection pressure was between 0.2 and 0.5 MPa. The pore volume fraction in this sand was around 0.35 so the expected expansion was of the order of around 15%. The observed expansion under atmospheric conditions, however, was in the range of 2 - 10%.

In conclusion we are able to say that the effects with the swelling cement currently used are not good enough for our purpose. Its viscosity is not low enough and its expansion in various experiments is lower than expected. By reducing the W/C-ratio we will lower the capillary porosity of the cement and the expansion against an applied stress might be higher. Current research is directed towards finding the right additives in order to decrease the viscosity and increase the expansivity of the swelling cement, or to change the composition of the cement itself.