Review on Artificial Neural Network and its Application in Foundry
Review on Artificial Neural Network and its Application in Foundry
复制标题
DOI：
10.4028/www.scientific.net/amm.380-384.2129
复制
发表时间：
2013-08
期刊：
影响因子：
--
通讯作者：
D. Shi;G. Gao
中科院分区：
文献类型：
--
作者：
D. Shi;G. Gao
人工神经网络具有较强的非线性处理能力和较高的容错能力，在铸造行业得到了广泛的应用。首先，以最常用的BP网络为例，简要介绍了人工神经网络的模型及其实现。然后从湿型砂制备质量优化、熔炼设备控制、铸造缺陷诊断与预测、铁水化学成分预测、连铸漏钢预报、铸铁石墨形态识别等方面详细综述了人工神经网络在铸造中的应用。最后指出，随着人工神经网络理论的突破，人工神经网络在铸造领域的应用将更加深入和广泛。
Artificial neural network (ANN) has been widely used in the foundry because of its strong nonlinear processing capacity and high fault tolerance. First, the model and its realization of ANN are briefly described by taking the most commonly-used BP network as an example. Then the application of ANN in foundry are reviewed in detail from the aspects of the quality optimization of greensand preparation, control of the melting equipments, diagnosis and prediction of the casting defects, chemical composition prediction of the molten iron, breakout forecast of the continuous casting process, and graphite shape identification of cast iron. Finally its pointed that the application of ANN in foundry will be deeper and broader with the breakthrough of ANN theory.