สำนักราชบัณฑิตยสภา

83 ธั ชวี ร์ ลี ละวั ฒน์ วารสารราชบัณฑิตยสถาน ปีที่ ๓๗ ฉบับที่ ๔ ต.ค.-ธ.ค. ๒๕๕๕ tile were prepared by immersing them in water for 24 hours. Each wet specimen was placed on supports at right angle with the corrugations and then loaded at midspan until the breaking point. The load and deflection at midspan of the specimen at a given point were automatically recorded. The results of bending tests were expressed as the breaking load per meter width. The impermeability test for cement tiles was determined as specified within the ISO 9933 standard (7) . A total of 8 specimens for each type of tile were sealed with the acrylic frame with the dimensions of 450-mm width, 500-mm length and 40-mm height above the tops of the corrugation. Thereafter, water was filled in the frame until the level of water was about 20 mm above the tops of the corrugations and stored for 24 hours. Traces of moisture on the underface of the sheet for each tile were visually observed. The warm water test was also conducted as specified within the ISO 9933 standard (7) . A total of 20 specimens for each type of tiles were transversally cut to a length of 30 cm. The speci- mens for each type of tiles were divided randomly into two lots of 10 specimens. The first lot of 10 specimens being classified as the control set was kept for bending test. Whereas the second lot of 10 specimens was immersed in water controlled at 60±0.5˚ C . for 56 days. The bending test was then carried out for both lots under wet conditioning. Each specimen was placed on supports at half one pitch on each side of the central and then loaded along with the top of corrugation until the breaking point. Further details of the test are specified in the ISO 9383 standard (8) . The breaking load of each specimen was automatically recorded. The lower estimation of the mean breaking load at 95% confidence level for the second lot immersed in warm water (L s ) and the upper estimation of the mean breaking load at 95% confidence level for the first control lot (L 1 ) were calculated. The ratio (L) of L s over L 1 at the 95% lower confidence limit was determined. Moreover, the surface of each specimen was observed using digital microscopy to detect cracks, leachate stains, and defects of coating. The heat-rain test was conducted as specified within the ISO 9933 standard (7) . A total of 8 specimens for each type of tile were placed in the plastic tank on the frame inclined at 25±5˚ and then subjected for 25 cycles of heat and rain consecutively. Each cycle contains wetting for 2 hours and 50 minutes, interval for 10 minutes, heating for 2 hours and 50 minutes, and then interval for 10 minutes. The wetting process used in the study was controlled by delivering water at the speed of 2.5 l/min/m 2 and the controlled temperature of 15±2˚ C . The heating process was