神经元网络在VD温度预报模型中的应用

在对冶金生产过程中的现场实际情况进行详细分析的基础上,采用神经元网络技术建立和开发了一个真空脱气炉（VD）钢水温度的预报模型.该模型从基础自动化系统和L2过程控制系统中获得所需的数据,并通过设计的输入数据对神经网络进行了训练、检验和性能评价.该模型通过L2过程计算机,以网络方式向操作人员预报并指示VD开盖温度,经过实际运行,该模型可以准确预报VD终点温度,实现了对VD终点温度的精确控制,对于保证生产的顺利进行和节约能源有着重要的意义.

Application of Neuron Network to VD Furnace Temperature Prediction Model

To resolve the problem that there is no validate temperature predict model, and to fulfill the control of vacuum degasser (VD)molten steel temperature, a neuron network model is designed to predict the molten steel temperature of VD furnace. This model gets the data needed from automation system and L2 process control machine. The data are used to train the neuron network, to check the model and to estimate the model. This model gets the data from L2 by network, and give the prediction temperature to the operator by the same method. The operator could control the production process according to the result of the model. The temperature control is achieved accurately at the end. This model system is useful to improve product and save energy.