To investigate the mechanical properties of early-age cemented tailings backfill under multi-stage cyclic loading, a series of multi-stage cyclic loading-unloading uniaxial compression tests were conducted using a WHY-600 press. The tests considered curing age, loading time, and loading gradient as variables to analyze the uniaxial compressive strength and fracture morphology of the backfill. Additionally, the evolution of wave velocity was examined through ultrasonic wave velocity testing. The findings indicate that the uniaxial compressive strength of the backfill increases with extended loading time. Notably, the strength variation from 150 s to 300 s is more pronounced than that from 300 s to 450 s. Furthermore, as the loading gradient increases, the uniaxial compressive strength initially rises and subsequently declines, suggesting that moderate static load damage enhances the compressive performance of the backfill. At an early stage, the failure mode of the backfill body under the influence of multi-stage cyclic loading and unloading transitions from X shear failure to a mixed shear failure primarily characterized by X conjugate shear failure, and subsequently to a mixed shear failure predominantly governed by Y shear. The loading duration and loading gradient exhibit a significant synergistic effect on the variation in wave velocity of the backfill. Specifically, when the loading duration exceeds 300 seconds and the loading gradient surpasses σ0.5, the wave velocity initially decreases with increasing loading cycles and subsequently recovers to some extent. Utilizing a long short-term memory neural network in conjunction with a genetic algorithm, a predictive model for the uniaxial compressive strength of early tailings consolidated backfill, subjected to multi-stage cyclic loading and unloading, has been developed. Upon validation, the model demonstrates correlation coefficients of 0.9814 and 0.9466, respectively, with a calculation error range of 3.21% to 12.47%, thereby indicating its high applicability and reliability.
近年来,学者们采用力学实验和数值模拟等方法对循环加卸载条件下充填体的力学性能等方面进行了广泛研究,分析了充填体在循环加卸载条件下的宏观力学特性(邝泽良等,2017)、声发射演化(邱华富等,2025)和损伤特征(叶永飞等,2018),为提升井下充填体的安全性提供了依据。然而,充填体在井下反复受到不同程度的压力载荷作用,包括铲运机、凿岩台车等采装设备以及工作人员,其强度将发生改变。同时,受不同施压时间和施压载荷的影响,充填体的强度变化存在差异性,这给高效持续性采矿作业造成一定困扰。而人工智能算法通过收集大量充填体强度演化数据,能够为有效预测不同压力载荷作用下充填体的承载能力提供可靠途径(Dakshith et al,2016;Shi et al,2022)。
基于现有文献数据对GA各超参数进行范围确定,并以LSTM的训练集中的最小均方根误差作为选取超参数的依据进行检验(Nakhaei et al,2013;谭天乐,2021),如表3所示。由表3可知,输入层神经元节点数为3,隐藏层节点数为8,输出层节点数为1,染色体长度设定为22,种群数量为27,交叉率为0.75,变异率为0.01,算法迭代次数为100。
DakshithR W, GregY,2016.Optimization of the catch bench design using a genetic algorithm[J].International Journal of Mining Science and Technology,26(6):1011-1016.
[2]
GanD Q, SunH K, LiuZ Y,et al,2023.Late mechanical properties and energy evolution mechanism of cemented tailings backfill under early damage[J].Engineering Failure Analysis,149:107249.
[3]
HeC D, SadeghpourH, ShiY X,et al,2024.Mapping distribution of fractures and minerals in rock samples using Res-VGG-UNet and threshold segmentation methods[J].Computers and Geotechnics,175:106675.
[4]
LiuZ Y, GanD Q, SunH K,et al,2025.Dynamic impact performance of cemented tailings backfill in a water-bearing environment:coupling effects and damage characteristics[J].Soil Dynamics and Earthquake Engineering,191:109249.
[5]
NakhaeiF, MosaviM R, SamA,2013.Recovery and grade prediction of pilot plant flotation column concentrate by a hybrid neural genetic algorithm[J].International Journal of Mining Science and Technology,23(1):69-77.
[6]
ShiL L, ZhangJ, ZhuQ Z,et al,2022.Prediction of mechanical behavior of rocks with strong strain-softening effects by a deep-learning approach[J].Computers and Geotechnics,152:105040.
ChengHaiyong, WuAixiang, WuShunchuan,et al,2022.Research status and development trend of solid waste backfill in metal mines[J].Chinese Journal of Engineering,44(1):11-25.
GanDeqing, ZhangYajie, LiuZhiyi,et al,2023.Study on the mechanism of the influence of early damage of cemented backfill on later mechanical properties[J].Chinese Journal of Rock Mechanics and Engineering,42(4):821-832.
GuoLijie, LiuGuangsheng, MaQinghai,et al,2022.Research progress on mining with backfill technology of underground metalliferous mine[J].Journal of China Coal Society,47(12):4182-4200.
LongHai, BaoRongtao, TanShilin,et al,2023.Experimental study on the mechanical properties of layered tailing sand cemented backfill and optimization[J].Gold Science and Technology,31(5):763-772.
KuangZeliang, ZhengXiaolong, ChenJun,et al,2017.Experimental study on acoustic emission characteristics of consolidated filling body under cyclic loading[J].Nonferrous Metals Science and Engineering,8(4):91-97.
LiHongbao, GanDeqing, XinyuE,et al,2021.Experimental study on the effect of tailings particle size on early strength of cement paste backfill[J].Metal Mine,50(3):34-39.
LiuJie, BaiTengfei, LiGuangbo,et al,2024.Research on the quantitative relationship between the uniaxial compressive strength of filling body and amount of hydration products[J].Gold Science and Technology,32(6):1046-1055.
QianTao, FengJian, ZhangJi,et al,2023.Experimental study on proportion optimization and strength characteristics of mixed tailings slurry[J].Metal Mine,52(8):196-204.
QiuHuafu, LiYuhang, LiuLang,et al,2025.Acoustic emission characteristics of cemented tailings backfill under cyclic loading[J].Journal of Xi’an University of Science and Technology,45(1):74-85.
TanTianle,2021.Optimization of propellant consumption based on genetic algorithm for spacecraft rendezvous under model predictive and inversive control[J].Journal of Astronautics,42(12):1542-1549.
WangShushuai, YangRenshu, LiYongliang,et al,2025.Study on mechanical properties and failure characteristics of cemented aeolian sand backfill[J].Journal of Mining & Safety Engineering,42(2):418-429.
WangZhongchang, WangYanwen, XiaHongchun,2022.Critical nonsilting velocity of slurry in transportation pipe with different bend angles[J].Mining and Metallurgical Engineering,42(3):41-45.
WuQiong, GuoJinping, WangXiaolin,et al,2025.Effect of temperature on strength and energy evolution characteristics of cemented backfill mass under dynamic loading[J].Mining and Metallurgical Engineering,45(1):27-34.
YangZhichun, MinHuaidong, YangFan,et al,2025.Fault diagnosis technology of relay protection in microgrid based on convolutional bidi-rectional long short-term memory network[J].Acta Energiae Solaris Sinica,46(1):420-428.
YeYongfei, ZhangYanan, LiShichao,et al,2018.Acoustic emission characterization of damage of cemented filling under cyclic loading[J].Gold Science and Technology,26(6):819-825.
ZhangYoufeng, FuYuhua, YuYao,2022.Experimental study on mechanical properties of cemented filling body with bentonite under repeated impacts[J].Mining and Metallurgical Engineering,42(1):30-34,40.
ZhengShuangjin, JiangHoushun, XiongMengyuan,et al,2025.Data driven and mechanistic model based prediction of rate of penetration[J].Drilling & Production Technology,48(1):78-87.