基于PSO-BP算法的Ti-10.2Mo-4.9Zr-5.5Sn合金本构关系研究
|
摘要
采用Gleeble-3500型热模拟试验机,对Ti-10. 2Mo-4. 9Zr-5. 5Sn合金进行等温恒应变速率压缩实验,研究其在变形温度943～1093 K,应变速率0. 001～10 s-1范围内的热变形行为,并构建一个层数为3×15×10×1的PSO-BP神经网络结构形式的本构关系模型。结果表明,合金的流变应力对变形温度和应变速率较为敏感,变形温度升高和应变速率减小都会使流变应力降低;在高温和低应变速率条件下,流变曲线大多呈现稳态流动特征,但在应变速率为10 s-1时,流动应力随应变增加呈下降趋势,软化现象较为显著;采用PSO-BP神经网络建立Ti-10. 2Mo-4. 9Zr-5. 5Sn合金本构模型,经过误差计算得出,该模型的相关系数和平均相对误差分别为0. 9892和2. 48%,预测值偏差在10%以内的数据点占91. 59%,具有良好的精度。
The isothermal compression test with constant strain rate was carried out to research the hot deformation behavior of Ti-10. 2 Mo-4. 9 Zr-5. 5 Sn alloy at 943-1093 K and strain rate of 0. 001-10 s-1 on Gleeble-3500 experiment machine,and a constitutive relation model of PSO-BP neural network with 3 × 15 × 10 × 1 was constructed. The results show that the flow stress of alloy is sensitive to deformation temperature and strain rate,both the increase of deformation temperature and decrease of strain rate cause the decrease of flow stress; the flow curves present steady states at high temperature and low strain rate. However,when the strain rate is 10 s-1,the flow stress decreases with the increase of strain,and the softening phenomenon is significant. A constitutive model of Ti-10. 2 Mo-4. 9 Zr-5. 5 Sn alloy was built based on PSO-BP neural network. The calculation correlation coefficient and average relative error of the model are calculated to be 0. 9892 and 2. 48% respectively; and the data points whose predicted value deviation is less than 10% account for 91. 59%,which indicates the model has good accuracy.
The isothermal compression test with constant strain rate was carried out to research the hot deformation behavior of Ti-10. 2 Mo-4. 9 Zr-5. 5 Sn alloy at 943-1093 K and strain rate of 0. 001-10 s-1 on Gleeble-3500 experiment machine,and a constitutive relation model of PSO-BP neural network with 3 × 15 × 10 × 1 was constructed. The results show that the flow stress of alloy is sensitive to deformation temperature and strain rate,both the increase of deformation temperature and decrease of strain rate cause the decrease of flow stress; the flow curves present steady states at high temperature and low strain rate. However,when the strain rate is 10 s-1,the flow stress decreases with the increase of strain,and the softening phenomenon is significant. A constitutive model of Ti-10. 2 Mo-4. 9 Zr-5. 5 Sn alloy was built based on PSO-BP neural network. The calculation correlation coefficient and average relative error of the model are calculated to be 0. 9892 and 2. 48% respectively; and the data points whose predicted value deviation is less than 10% account for 91. 59%,which indicates the model has good accuracy.
引文
[1] CHANG M C,LUO C W,HUANG M S,et al. High-temperature microstructural characteristics of a novel biomedical titanium alloy[J]. Journal of Alloys and Compounds,2010,499(2):171-175.
[2] CUI W F,JIN Z,GUO A H,et al. High temperature deformation behaviors forα+βtype biomedical titanium alloy Ti-6Al-7Nb[J]. Materials Science and Engineering A,2009,499(1-2):252-256.
[3]吴义舟,郭爱红.生物医用钛合金发展和研究现状[J].材料开发与应用,2010,25(2):81-85.WU Yizhou,GUO Aihong. The development and research present status of biomedical titanium alloys[J]. Development and Application of Materials,2010,25(2):81-85.
[4]张小波,王克鲁,鲁世强,等.基于BP神经网络的Ti3Al基合金本构关系模型的建立[J].特种铸造及有色合金,2013,33(3):224-226.ZHANG Xiaobo,WANG Kelu,LU Shiqiang,et al. Modeling of constitutive relationship of Ti3Al-based alloy based on BP neural network[J]. Special-cast and Non-ferrous Alloys,2013,33(3):224-226.
[5]孙宇,曾卫东,赵永庆,等.基于BP神经网络Ti600合金本构关系模型的建立[J].稀有金属材料与工程,2011,40(2):220-224.SUN Yu,ZENG Weidong,ZHAO Yongqing,et al. Modeling of constitutive relationship of Ti600 alloy using BP artificial neural network[J]. Rare Metal Materials and Engineering,2011,40(2):220-224.
[6]张德丰. MATLAB神经网络仿真与应用[M].北京:电子工业出版社,2009.ZHANG Defeng. MATLAB neural network simulation and application[M]. Beijing:Electronic Industry Press,2009.
[7]赵振江.基于PSO-BP神经网络的网络流量预测与研究[J].计算机应用与软件,2009,26(1):218-221.ZHAO Zhenjiang. Prediction and research on network traffic based on PSO-BP neural network[J]. Computer Application and Software,2009,26(1):218-221.
[8]何霞,张彦敏,宋克兴,等. Cu-0. 23Be-0. 84Co合金热变形行为[J].塑性工程学报,2015,22(2):105-110.HE Xia,ZHANG Yanmin,SONG Kexing,et al. Study on hot deformation behaviors of Cu-0. 23Be-0. 84Co alloy[J]. Journal of Plasticity Engineering,2015,22(2):105-110.
[9]杨建辉,张鹏,赵升吨. TC4-DT钛合金热变形本构方程[J].塑性工程学报,2016,23(2):120-125.YANG Jianhun,ZHANG Peng,ZHAO Shengdun. Hot deformation constitutive equation of TC4-DT titanium alloy[J]. Journal of Plasticity Engineering,2016,23(2):120-125.
[10]罗锐,程晓农,徐桂芳,等.新型Fe-20Cr-30Ni-0. 6Nb-2Al-Mo合金的热变形行为及本构模型[J].稀有金属,2017,41(2):132-139.LUO Rui,CHENG Xiaonong,XU Guifang,et al. Constitutive modeling for elevated temperature flow behavior of Fe-20Cr-30Ni-0. 6Nb-2Al-Mo alloy[J]. Chinese Journal of Rare Metals,2017,41(2):132-139.
[11]毛健,赵红东,姚婧婧.人工神经网络的发展及应用[J].电子设计工程,2011,19(24):62-65.MAO Jian, ZHAO Hongdong, YAO Jingjing. Application and prospect of artificial neural network[J]. International Electronic Elements,2011,19(24):62-65.
[12]沈昌武,杨合,孙志超,等.基于BP神经网络的TA15钛合金本构关系建立[J].塑性工程学报,2007,14(4):101-104.SHEN Changwu,YANG He,SUN Zhichao,et al. Based on BP artificial neural network to building the constitutive relationship of TA15 alloy[J]. Journal of Plasticity Engineering,2007,14(4):101-104.
[2] CUI W F,JIN Z,GUO A H,et al. High temperature deformation behaviors forα+βtype biomedical titanium alloy Ti-6Al-7Nb[J]. Materials Science and Engineering A,2009,499(1-2):252-256.
[3]吴义舟,郭爱红.生物医用钛合金发展和研究现状[J].材料开发与应用,2010,25(2):81-85.WU Yizhou,GUO Aihong. The development and research present status of biomedical titanium alloys[J]. Development and Application of Materials,2010,25(2):81-85.
[4]张小波,王克鲁,鲁世强,等.基于BP神经网络的Ti3Al基合金本构关系模型的建立[J].特种铸造及有色合金,2013,33(3):224-226.ZHANG Xiaobo,WANG Kelu,LU Shiqiang,et al. Modeling of constitutive relationship of Ti3Al-based alloy based on BP neural network[J]. Special-cast and Non-ferrous Alloys,2013,33(3):224-226.
[5]孙宇,曾卫东,赵永庆,等.基于BP神经网络Ti600合金本构关系模型的建立[J].稀有金属材料与工程,2011,40(2):220-224.SUN Yu,ZENG Weidong,ZHAO Yongqing,et al. Modeling of constitutive relationship of Ti600 alloy using BP artificial neural network[J]. Rare Metal Materials and Engineering,2011,40(2):220-224.
[6]张德丰. MATLAB神经网络仿真与应用[M].北京:电子工业出版社,2009.ZHANG Defeng. MATLAB neural network simulation and application[M]. Beijing:Electronic Industry Press,2009.
[7]赵振江.基于PSO-BP神经网络的网络流量预测与研究[J].计算机应用与软件,2009,26(1):218-221.ZHAO Zhenjiang. Prediction and research on network traffic based on PSO-BP neural network[J]. Computer Application and Software,2009,26(1):218-221.
[8]何霞,张彦敏,宋克兴,等. Cu-0. 23Be-0. 84Co合金热变形行为[J].塑性工程学报,2015,22(2):105-110.HE Xia,ZHANG Yanmin,SONG Kexing,et al. Study on hot deformation behaviors of Cu-0. 23Be-0. 84Co alloy[J]. Journal of Plasticity Engineering,2015,22(2):105-110.
[9]杨建辉,张鹏,赵升吨. TC4-DT钛合金热变形本构方程[J].塑性工程学报,2016,23(2):120-125.YANG Jianhun,ZHANG Peng,ZHAO Shengdun. Hot deformation constitutive equation of TC4-DT titanium alloy[J]. Journal of Plasticity Engineering,2016,23(2):120-125.
[10]罗锐,程晓农,徐桂芳,等.新型Fe-20Cr-30Ni-0. 6Nb-2Al-Mo合金的热变形行为及本构模型[J].稀有金属,2017,41(2):132-139.LUO Rui,CHENG Xiaonong,XU Guifang,et al. Constitutive modeling for elevated temperature flow behavior of Fe-20Cr-30Ni-0. 6Nb-2Al-Mo alloy[J]. Chinese Journal of Rare Metals,2017,41(2):132-139.
[11]毛健,赵红东,姚婧婧.人工神经网络的发展及应用[J].电子设计工程,2011,19(24):62-65.MAO Jian, ZHAO Hongdong, YAO Jingjing. Application and prospect of artificial neural network[J]. International Electronic Elements,2011,19(24):62-65.
[12]沈昌武,杨合,孙志超,等.基于BP神经网络的TA15钛合金本构关系建立[J].塑性工程学报,2007,14(4):101-104.SHEN Changwu,YANG He,SUN Zhichao,et al. Based on BP artificial neural network to building the constitutive relationship of TA15 alloy[J]. Journal of Plasticity Engineering,2007,14(4):101-104.