钢—聚丙烯纤维混凝土碳化深度预测模型算法研究

Study on prediction model algorithm of carbonation depth of steel-polypropylene fiber reinforced concrete

为进一步对混凝土碳化问题进行研究,优化目前碳化深度预测模型中准确度不足的缺陷,本文以钢—聚丙烯纤维混凝土为例,对其不同周期、不同纤维比例掺量影响下的碳化深度进行预测。基于FICK第一定律开展目标对象的碳化深度预测研究,引入二阶多项式、BP神经网络等拟合算法,对其碳化深度预测模型进行优化改进。改进后的预测结果表明相比于传统算法,二阶多项式拟合算法的误差精度大体上控制在5%以内,满足工程精度要求。同时,采用BP神经网络算法也可验证其结果的可靠性,为后续钢—聚丙烯纤维混凝土碳化性能研究及其耐久性寿命预测提供重要参考。

In order to further study the problem of concrete carbonization and optimize the defect of inaccuracy in current carbonization depth prediction models,this paper takes steel-polypropylene fiber concrete as example to predict the carbonization depth under the influence of different periods and different fiber proportions.Based on FICK’s first law,the research on the carbonation depth prediction of target objects was carried out.The second-order polynomials and BP neural networks were introduced to optimize the carbonization depth prediction model.The improved prediction results show that compared with the traditional algorithm,the error accuracy of the second order polynomial fitting method is generally controlled within 5%,which meets the engineering accuracy requirements.Meanwhile,BP neural network algorithm can be used to verify the reliability of the results,which provides an important reference for the follow-up study on carbonization performance and durability life prediction of steel-polypropylene fiber concrete.