人工神经网络技术在高强混凝土中的应用——配合比设计初探

Application of Artificial Neural Network in the Design of High Performance Concrete——Primary Study on Mix Proportion

探讨了在多因素影响下,人工神经网络技术在混凝土配合比设计方面的实现手段.采用以正交设计试验作为学习样本模拟真实系统的方法,来模拟完全试验;同时,以部分试验数据为研究对象,通过自组织神经网络分类计算,构成学习样本来模拟真实系统,也得到了较为满意的结果.此项研究除提供了人工智能在混凝土配合比设计中的应用方法外,还在具体研究问题的背景下,为神经网络理论在确立学习样本的方法上寻求了一个可行的途径.

Through experimental studies on high performance concrete, the method of artificial neural network in the design of mix proportion of concrete was discussed. A method that actual system was simulated with experimental data of orthogonal test as learning samples was adapted and the object that complete trials were simulated was achieved. Meantime, taking partial experimental data as studying object, learning samples were constructed with cluster analysis of self adapting network to simulate actual system. Apart from providing the method of application of artificial intelligence in mix proportion design of concrete, a feasible method about building learning samples with the theory of neural network is obtained under the basic background.