Prediction of water inflow into a tunnel is a crucial prerequisite for the waterproof and drainage design of mountain tunnels in water-rich areas.Based on the proposed Baiyun Mountain Tunnel project in Guangzhou,a num...Prediction of water inflow into a tunnel is a crucial prerequisite for the waterproof and drainage design of mountain tunnels in water-rich areas.Based on the proposed Baiyun Mountain Tunnel project in Guangzhou,a numerical percolation model of random fractured rock of a tunnel underpassing a water reservoir is established to study the seepage characteristics of surrounding rock,the law of water inflow,and the change of lining water pressure,considering the local artificial boundary conditions for seepage in large rock mass,.In addition,the influences of rock permeability,fracture aperture,grouting circle thickness,and penetration are analyzed.The results show that:(1)Only fractures with aperture wider than 0.1 mm can play a significant role in water conduction in rocks with the permeability lower than 10^(-11)m^(2);(2)The greater the permeability difference between the fractures and rocks,the more remarkable the effects of fractures on the surrounding rock seepage field and cavern water inflow;(3)The sensitivity of grouting waterproof function to grouting circle thickness,grouting ring penetration,and rock permeability is significantly higher than that of tunnel buried depth and fracture aperture;(4)The lining water head is much more sensitive to the grouting circle thickness and penetration than to the tunnel buried depth;(5)With the grouting range enlarging,the impact of grouting circle permeability on the precipitation pressure role of the grouting ring increases;(6)For the interesting tunnel designed to be built at the depth of 70 m,the grouting circle with the thickness of 0.5 m and permeability of 10-^(14)m^(2)is recommended.展开更多
针对火电机组SO_(2)排放质量浓度的影响因素众多,难以准确预测的问题,提出一种改进向量加权平均(weighted mean of vectors,INFO)算法与双向长短期记忆(bi-directional long short term memory,Bi-LSTM)神经网络相结合的预测模型(改进IN...针对火电机组SO_(2)排放质量浓度的影响因素众多,难以准确预测的问题,提出一种改进向量加权平均(weighted mean of vectors,INFO)算法与双向长短期记忆(bi-directional long short term memory,Bi-LSTM)神经网络相结合的预测模型(改进INFO-Bi-LSTM模型)。采用Circle混沌映射和反向学习产生高质量初始化种群,引入自适应t分布提升INFO算法跳出局部最优解和全局搜索的能力。选取改进INFO-Bi-LSTM模型和多种预测模型对炉内外联合脱硫过程中4种典型工况下的SO_(2)排放质量浓度进行预测,将预测结果进行验证对比。结果表明:改进INFO算法的寻优能力得到提升,并且改进INFO-Bi-LSTM模型精度更高,更加适用于SO_(2)排放质量浓度的预测,可为变工况下的脱硫控制提供控制理论支撑。展开更多
文摘Prediction of water inflow into a tunnel is a crucial prerequisite for the waterproof and drainage design of mountain tunnels in water-rich areas.Based on the proposed Baiyun Mountain Tunnel project in Guangzhou,a numerical percolation model of random fractured rock of a tunnel underpassing a water reservoir is established to study the seepage characteristics of surrounding rock,the law of water inflow,and the change of lining water pressure,considering the local artificial boundary conditions for seepage in large rock mass,.In addition,the influences of rock permeability,fracture aperture,grouting circle thickness,and penetration are analyzed.The results show that:(1)Only fractures with aperture wider than 0.1 mm can play a significant role in water conduction in rocks with the permeability lower than 10^(-11)m^(2);(2)The greater the permeability difference between the fractures and rocks,the more remarkable the effects of fractures on the surrounding rock seepage field and cavern water inflow;(3)The sensitivity of grouting waterproof function to grouting circle thickness,grouting ring penetration,and rock permeability is significantly higher than that of tunnel buried depth and fracture aperture;(4)The lining water head is much more sensitive to the grouting circle thickness and penetration than to the tunnel buried depth;(5)With the grouting range enlarging,the impact of grouting circle permeability on the precipitation pressure role of the grouting ring increases;(6)For the interesting tunnel designed to be built at the depth of 70 m,the grouting circle with the thickness of 0.5 m and permeability of 10-^(14)m^(2)is recommended.
文摘针对火电机组SO_(2)排放质量浓度的影响因素众多,难以准确预测的问题,提出一种改进向量加权平均(weighted mean of vectors,INFO)算法与双向长短期记忆(bi-directional long short term memory,Bi-LSTM)神经网络相结合的预测模型(改进INFO-Bi-LSTM模型)。采用Circle混沌映射和反向学习产生高质量初始化种群,引入自适应t分布提升INFO算法跳出局部最优解和全局搜索的能力。选取改进INFO-Bi-LSTM模型和多种预测模型对炉内外联合脱硫过程中4种典型工况下的SO_(2)排放质量浓度进行预测,将预测结果进行验证对比。结果表明:改进INFO算法的寻优能力得到提升,并且改进INFO-Bi-LSTM模型精度更高,更加适用于SO_(2)排放质量浓度的预测,可为变工况下的脱硫控制提供控制理论支撑。
