CAJ | 학술논문

为了对磁性材料行业的材料配方和性能配方之间的知识进行有效管理,和准确预测材料配方和性能之间的变化趋势,提出了一个基于神经网络的磁性材料设计方法,并在此基础上实现了磁性材料设计专家系统.首先对配方和性能的基础数据进行了管理,然后通过生产实验数据训练神经网络,实现了性能和配方的相互推导,最后对配方或性能的变化趋势进行了准确预测.实验结果表明,专家设计系统的趋势预测具有极高的精确性.企业实施该系统后较大幅度的提高了新产品的开发速度,提高了企业的市场竞争力.
위료대자성재료행업적재료배방화성능배방지간적지식진행유효관리,화준학예측재료배방화성능지간적변화추세,제출료일개기우신경망락적자성재료설계방법,병재차기출상실현료자성재료설계전가계통.수선대배방화성능적기출수거진행료관리,연후통과생산실험수거훈련신경망락,실현료성능화배방적상호추도,최후대배방혹성능적변화추세진행료준학예측.실험결과표명,전가설계계통적추세예측구유겁고적정학성.기업실시해계통후교대폭도적제고료신산품적개발속도,제고료기업적시장경쟁력.
To manage knowledge effectively between magnetic properties and alloy composite, and predict trend precisely, a magnetic design method based ANN (artificial neural network) is presented. And on basis of this method a design system of expert is realized. Firstly, basic data from magnetic properties and alloy composite is managed effectively. Then ANN is trained completely by original product data, through which deduction from properties to alloy composite could be easily. Experimental results show trend predication from expert system is high accuracy. After the implementation, the system improves the development speed of new products and the market competitiveness of enterprises.