废胎橡胶用量对保温砂浆导热系数的影响与神经网络预测

Impact of waste tire rubber particles content on thermal conductivity of thermal insulting mortar and its prediction model based on BP neural network

在利用和有效处置废胎橡胶的背景下，研究了以废胎橡胶颗粒作为骨料的砂浆导热系数随废胎橡胶颗粒用量的变化规律。在砂浆配合比设计中，保持废胎橡胶颗粒和玻化微珠的堆积体积之和为1m^3，其中废胎橡胶颗粒用量分别为0、125、250、375和500kg/m^3。试验结果表明：以废胎橡胶颗粒和玻化微珠作为骨料的砂浆28d抗压强度在0．32-2．64MPa；该类砂浆的热导系数为0．080-0．152W/（m·K），并随着废胎橡胶颗粒和水泥用量的增加而增大。根据废胎橡胶粉、玻化微珠、水泥、乳胶粉、减水剂和水灰比等因素建立的BP神经网络导热系数预测模型，得出的预测值与试验的实际数值比较吻合，预测精度能满足要求。

Under the background of making use of and effective disposal of waste tire rubber, thermal conductivity coefficient of thermal insulting mortar varied with the content of waste tire rubber particles （WTRP） as a part of an aggregate was studied. Several mortar mixtures were ＇prepared by keeping the accumulated volume of both WTRP and glazed hollow beads equal to 1 m^3 and by using WTRP contents of 0,125,250,375 and 500 kg/m^3 respectively. Results indicate that the 28 d compressive strengths of the mor- tars were 0.32-2.64 MPa. And the thermal conductivity coefficients were 0.080-0.152 W/fan.K） increased with both the WTRP and cement content increasing. According to six factors such as contents of WTRP, glazed hollow beads, cement,emulsion powder and wa- ter reducer, and the ratio of water to cement,a prediction model of the thermal conductivity based on BP neural network was estab- lished. The obtained predicted value matches the actual test value, so the prediction accuracy can meet the requirements.