基于膜算法进化极限学习机的氧气转炉炼钢终点预报模型

Endpoint prediction model for basic oxygen furnace steelmaking based on membrane algorithm evolving extreme learning machine

氧气转炉炼钢的控制目标是终点温度和碳含量,但由于不能对其进行在线连续测量,直接影响了出钢的质量.针对该问题,提出一种基于膜算法进化极限学习机(ELM)的抗干扰终点预报模型.利用进化膜算法的全局寻优能力调整ELM网络参数,不仅避免了ELM网络受异常点影响出现过拟合现象,还可以寻找最优复杂度的ELM模型.将找到的ELM模型应用到转炉炼钢领域并建立终点碳含量和温度的预报模型.在仿真实验中,分别使用含有高斯噪声的标准sin C函数和氧气转炉炼钢实际生产数据进行仿真,结果表明所提模型在含噪声的数据中具有较好的预报精度和鲁棒性.

The goal of basic oxygen furnace （BOF） steelmaking is the endpoint of the temperature and carbon content .But it does not work to online continuous measurement ,which directly affects the quality of steel .For solving the above problem ,an anti-jamming endpoint prediction model of extreme learning machine （ELM ） based on evolving membrane algorithm is proposed .The parameters of ELM are adjusted by the global optimization ability of evolving membrane algorithm ,w hich not only avoids the overfitting of ELM affected by outliers ,but also finds the optimal ELM model .The ELM model is applied to the field of BOF steelmaking ,and the endpoint prediction model of carbon content and temperature is created .Simulations are implemented by the sin C function with the Gaussian noise and the production data of BOF steelmaking .The experimental results indicate that the proposed model has good prediction accuracy and robustness in the processing of data with noise .