介绍了一种铸坯表面测温的方法，并依据铸坯表面温度，应用模式识别与人工神经网络相结合的方法，建立了板坯连铸的二冷水模型。该模型既可根据模式识别的分类图预报优化的二冷工艺参数，亦可根据钢种、中间包钢水温度、拉速及希望的二冷区不同部位的铸坯表面温度由PLS改善了的输入神经元网络，预报连铸不同回路的水量。该模型运用于生产实际后，铸坯表面裂纹大为减少，取得了较好的效果。
This paper introduces a method of measuring surface temperature of slab. Based on surface temperature of slab, the model of secondary cooling for slab continuous casting has been developed with PLSB-BPN. The model can provide optimal parameters by means of the classification map of pattern recognition. It can also predict the proper quantity of water for every secondary cooling water loops corresponding to desired surface temperature at different points on slab surface based on steel grade, temperature of molten steel in tundish and casting speed by using the artificial neural network with the input of PLS. Slab surface cracking has been decreased sharply with new model.
参考文献
| [1] | 蔡开科 .连铸坯裂纹[J].北京科技大学学报,1993,15(Z2):3-13. |
| [2] | 蔡开科.连续铸钢[M].北京:科学出版社,1990 |
| [3] | 徐荣军;姚忠卯;程大振 .热轧硅钢牌号PLS-ANN人工智能优化[J].钢铁,1997,32(01):36-39. |
| [4] | 黄世萍;刘洪霖;陈念贻 .初轧加热炉节能的模式识别优化[J].钢铁,1993,28(11):69. |
| [5] | 徐荣军;欧昌俗;姚忠卯 .模式识别及人工神经网络算法在16Mn钢生产中的应用[J].北京科技大学学报,1995,17(Z2):105-109. |
- 下载量()
- 访问量()
- 您的评分：
-
10%
-
20%
-
30%
-
40%
-
50%