(北京有色金屬研究總院, 北京100088)
摘 要: 在Geeble−1500熱模擬機(jī)上對(duì)7055鋁合金進(jìn)行熱壓縮試驗(yàn)，基于熱壓縮試驗(yàn)數(shù)據(jù)，建立流變應(yīng)力的反向傳播(BP)神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)模型和加工圖。結(jié)果表明：用人工神經(jīng)網(wǎng)絡(luò)能更精確地預(yù)測(cè)熱壓縮過程中的流變應(yīng)力，預(yù)測(cè)精度明顯高于線性經(jīng)驗(yàn)公式的；通過預(yù)測(cè)模型可以獲得樣本數(shù)據(jù)值范圍內(nèi)的非樣本數(shù)據(jù)變形條件下的流變應(yīng)力，其預(yù)測(cè)結(jié)果充分反映該合金的高溫變形特征；在本實(shí)驗(yàn)條件下，7055鋁合金在高溫變形時(shí)存在一個(gè)失穩(wěn)區(qū)，即變形溫度在實(shí)驗(yàn)溫度范圍內(nèi)應(yīng)變速率為0.025 s−1以上的區(qū)域；在375～425 ℃的范圍內(nèi)，應(yīng)變速率小于0.001 s−1的區(qū)域，最大功率耗散系數(shù)為0.45；EBSD技術(shù)分析表明在安全區(qū)發(fā)生部分動(dòng)態(tài)再結(jié)晶。利用加工圖確定了熱變形時(shí)的流變失穩(wěn)區(qū), 并且獲得了試驗(yàn)參數(shù)范圍內(nèi)熱變形的最佳工藝參數(shù), 其熱加工溫度為350−430 ℃低應(yīng)變速率區(qū)。
關(guān)鍵字: 7055鋁合金；流變應(yīng)力；熱變形；神經(jīng)網(wǎng)絡(luò)；加工圖
7055 aluminum alloy based on artificial neural networks
（General Research Institute for Nonferrous Metals, Beijing 100088, China）
Abstract:The isothermal compression of 7055 alloy was carried out at a Geeble-1500D simulator. According to the experimental results, a back-propagation (BP) neural network model of flow stress and a processing map were developed. The results indicate that the neural network can correctly reproduce the flow stress in the sampled data and it can also predict well the non-sampled data. The predicted curves can accurately reflect the flow behavior of 7055 alloy during the hot deformation. The processing map shows unsteady zones of high temperature deformation of 7055 alloy, including the zone under experimental temperatures and above strain rate of 0.025 s−1. At 375−425 ℃ and 0.001 s−1, the peak efficiency of power dissipation is 0.45. Electron backscatter diffraction (EBSD) observations show that there is dynamic recrystallization in the steady zones. The temperature of 350−430 ℃ and low strain rate are recommended to set the industrial forming conditions.
Key words: 7055 Al alloy; flow stress; hot deformation; artificial neural network; processing map
