基于带有偏差单元的IRN模型的大体积混凝土温度预测与控制

Forecast and control on temperature of massive concrete based on the irn model with the error unit

分析了大体积混凝土温度的各种影响因素 ,结合神经网络的特点 ,提出运用神经网络中带有偏差单元的 IRN(Internally Recurrent Net)模型进行大体积混凝土温度预测与控制的新思路。通过实验 ,该模型可达到较好的效果 ,具有一定的可操作性和实用性。为今后大体积混凝土温度预测与控制提供了可借鉴的方法。

The factors that influence the temperature of the massive concrete is discussed in the paper. Based on the characteristics of artificial neural network model, a new model of artificial neural network, the Internally Recurrent Net(IRN) model with the error unit is proposed and applied to the forecast and control on temperature of massive concrete. The actual examples verify that the model is ideal and the new method is practicable and beneficial to the forecast and control on temperature of massive concrete.