摘要为了获得焊条原材料成分与焊条熔敷金属力学性能之间的映射关系，建立了焊条原材料成分预测遗传算法神经网络模型，通过用46组数据对模型进行训练和验证，研究分析得出以下结论:酸性焊条焊芯中的C含量、中碳锰铁中C含量、还原钛铁矿中C含量、焊芯中Mn含量、中碳锰铁中Mn含量、碱性焊条焊芯中的C含量、中碳锰铁中C含量、还原钛铁矿中C含量、焊芯中Mn含量、中碳锰铁中Mn含量、焊芯中的Si含量、中碳锰铁中的Si含量、低度硅铁中的Si含量的预测平均相对误差总体在15%以内，可以满足实际生产要求。表明所建立的遗传神经网络预测模型是有效的，能够直接根据熔敷金属的力学性能较准确地预测其焊条原材料成分。7069
关键词：碳钢焊条 遗传算法 神经网络预测 力学性能
毕业设计说明书（论文）外文摘要
Title Genetic neural network prediction of carbon steel electrode formulation component
Abstract
In order to obtain the mapping relationship between the electrode raw material composition and the mechanical properties of the electrode deposited metal,we have established the electrode raw material composition predict genetic algorithm neural network model, by using 46 sets of data to train and validate,we can research and analysis the following conclusions: Acid electrode welding the C content in the core, the C content in the medium carbon ferromanganese, the C content of ilmenite, the Mn content in the core, the Mn content in the medium carbon ferromanganese, the alkaline electrode welding the C content in the core,the C content in the ferromanganese, the C content of ilmenite, Mn content in the core, the Mn content in the medium carbon ferromanganese, the Si content in the core and the Si content in the medium carbon ferromanganese, low-grade ferrosilicon Si content of the forecast overall average relative error of less than 15 percent to meet the actual production requirements. That the genetic neural network prediction model is valid to directly according to the mechanical properties of the deposited metal to more accurately predict the electrode raw material components.
Keywords Carbon steel electrode Genetic Algorithms
Neural network prediction Mechanical properties 目 次
1. 绪论 1
1.1 遗传算法与神经网络在预测应用方面的研究现状 1
1.1.1 基于遗传算法的神经网络在预测应用方面的研究现状 1
1.1.2 遗传算法与神经网络在焊接方面的国内外研究现状 2
1.2 本课题的研究内容 3
2. 碳钢焊条配方成分特点分析 4
2.1 酸性碳钢焊条配方成分特点分析 4
2.2 碱性碳钢焊条配方成分特点分析 8
3. 基于遗传算法的碳钢焊条配方成分神经网络预测 9
3.1 试验数据处理 9
3.2 碳钢焊条配方成分遗传神经网络预测模型建立 10
3.2.1 遗传算法 10
3.2.2 BP神经网络 13
3.3 碳钢焊条配方成分遗传神经网络模型训练与预测 14
3.3.1 遗传神经网络的训练 14
3.3.2 遗传神经网络的预测 17
3.3.3 遗传算法优化的BP网络模型与简单BP网络模型预测结果比较 24
4. 结论 24
致 谢 26
参 考 文 献 27 碳钢焊条配方成分遗传神经网络预测:http://www.youerw.com/cailiao/lunwen_4831.html