現今世界上由於放射性能源大量的使用,處置放射性廢料的處置便成為重要的研究課題。目前對於放射性廢料的處置方法中較普遍的方法是利用深層處置場來封存放射性廢料,並於廢料與母岩中加一緩衝材料,緩衝材料的選取依各國家資源不同而有些許之差異,目前各國研究其工程性質發現,膨潤土與砂-膨潤土顆粒混合材料,具有良好的阻絕性及優良的熱學性質。 本文主要針對純膨潤土與依照不同體積比混和之膨潤土加花崗岩碎石,製作不同乾密度,含水量之土樣。設計一模具壓製試體,利用暫態熱針法量測熱傳導係數,由試驗結果可以得知,熱傳導係數會隨著土體之乾密度與含水量之增加而增加;與花崗岩碎石混合後,由於花崗岩可視為一加強材料,故熱傳導係數亦會隨著顆粒之體積比之增加而增加。利用微觀力學成分體積比觀念,發展等含水量法與等密度法計算體積比,應用於微分模式與Self-consistent scheme預測熱傳導係數,與De Vries and Campbell所發展之經驗式(1985),兩者與實驗值比較,可得一良好之預測結果。 For the use of radioactive energy, it is important to develop techniques for the disposal of radioactive wastes in the world. Recently, many countries plan to construct the disposal facility underground deeply. Radioactive waste are sealed in canisters, and buffer materials are filled between canisters and host rock. Buffer materials must have good engineering properties. One of the most important factor is the thermal property. Among the candidates of buffer materials in many other researches, we know that mixture of bentonite and crushed granite has good mechanical properties and thermal characteristics. The study aims at measuring thermal conductivity of buffer materials with different densities, water contents and granite contents. We improve the heat probe method (ASTM D5334) by designing proper instruments and techniques to measure the thermal conductivity of bentonite and sand-bentonite mixture. In the results we find that the thermal conductivities of buffer material blocks rise with the increasing of density or water content. With mixing of crushed granite, which has higher thermal conductivities, the thermal conductivity increases with granite volumetric fraction. Appling the concept of micromechanics, we develop the equal-density and the equal-water-content method to calculate the volumetric fraction of two individual batches to form one. In this way, Differential scheme and Self-consistent scheme can be applied to predict the thermal conductivities. In comparing with experiential method: the De Vries and Campbell model, both methods can match with experimental data well.