Analysis Of Pore Water Pressure In Gravel Soil Core Wall And Permeability Coefficients Back Analysis Of High Earth-Rock Dam

Seepage is one of the main causes of earth-rockfill dam failure.The seepage monitoring data of earth-rockfill dam can not only be used to directly analyze the dam seepage safety,but also be used to reverse the dam seepage coefficients,which can improve the reliability of seepage field analysis.However,for high earth-rockfill dam,the excess pore water pressure will be formed during the filling of the core wall.It can be seen from the monitoring data that the piezometric water level in the core wall of high dam may exceed the upstream water level or even higher than the dam crest,and the excess pore water pressure in the core wall is difficult to dissipate,which often exists for many years.It will directly affect the seepage safety in the early stage of high core wall dam,even cause the core wall hydraulic fracturing,and seriously interferes with the calculation,analysis and safety evaluation of seepage field.Therefore,the study on excess pore water pressure in the core wall of high earth-rockfill dam has important theoretical value and engineering significance.There is no effective method so far to analyze the excess pore water pressure in high core wall dam.Selecting Nuozhadu earth-rock dam with gravel soil core wall as the research object,it deeply analyzed the prototype monitoring data of pore water pressure in the core wall dam,adopted finite element method to calculate and analyze the consolidation of the dam,proposed a simplified calculation method for the dissipation of pore water pressure in the core wall,which was used to reject the excess pore water pressure in the core wall,and reversed the seepage coefficients of the dam by using a back analysis method of radial basis function(hereinafter referred to as RBF)neural network optimized by particle swarm optimization(hereinafter referred to as PSO).The main contents and achievements are summarized as follows:1.It analyzed the monitoring data of pore water pressure in core wall of Nuozhadu dam.It showed that there was a good positive correlation between the growth rate of pore water pressure and the core wall filling speed;The lower the elevation of monitoring points in core wall,the larger the pore water pressure;While the core wall filling was finished,the maximum pore water pressure accounted for 55.9%of the weight of the overburden,and the corresponding piezometric level was higher than the crest elevation;While the reservoir was impounded,the core wall pore water pressure rised.As the seepage coefficient of the core wall was small,the formation of seepage was lag and the influence of seepage on pore water pressure in core wall was limited.The increase of pore water pressure during impounding was mainly caused by impounding force and wetting deformation of upstream dam shell.More specifically,the impoundment water pressure was applied to the upper surface of the core wall,the vertical component had compression effect on the upper part of the core wall,the core wall was bent to the upstream due to the immersion and wetting of upstream dam rockfill,and the horizontal component of water pressure had a bending effect pointing downstream.The combination of the three effects increased the compressive stress of the core wall,which leaded to the increase of pore water pressure of the core wall.Compared with the core wall filling,the increase of pore water pressure caused by the rise of reservoir water level was small.2.Based on Biot's consolidation theory,the finite element method was used to calculate the consolidation of Nuozhadu dam.The results showed that the calculated pore water pressure during filling period was less than the measured value,and the calculated dissipation rate of pore pressure was also higher than that measured.This was because the precision of calculating volume strain in finite element calculation was not high and determining the consolidation parameters of the core wall with different stress conditions was very difficult.In order to reduce the arch effect,the pressure transfer between the dam shell and the core wall was reduced by loosing the filter material appropriately.In addition,during the dam filling,the outline of the dam was basically symmetrical with the core wall as the center,which resulted that the contour lines of major principal stress and minor principal stress in core wall basically keepping horizontal at the time of completion.Thus,the core wall consolidation could be regarded as the consolidation problem of layered soil.Because the horizontal seepage coefficient of the rolled core wall was several times larger than that of the vertical direction,the seepage water flowed horizontally and discharged through the upstream and downstream filter material during the consolidation process.On this basis,it derived and solved the soil consolidation and pore water pressure dissipation control equation of vertical bearing pressure and horizontal drainage of the core wall.But the stress conditions of different parts of the core wall were quite different,and the consolidation coefficients of different parts were different.The variation of consolidation coefficient was considered and the dissipation of excess pore water pressure in the dam was calculated in this research.The calculated water head was in good agreement with the measured value.This method that was simple in calculation without complex finite element analysis,involved few parameters,was low computational cost,and could meet the engineering accuracy requirement.Therefor,it had important engineering practical significance to analyze the pore water pressure dissipation of the core wall of high earth-rock dam.3.It proposed a back analysis method of RBF neural network optimized by PSO,which was used to reverse seepage coefficients of earth-rockfill dam.This method used PSO to optimize the structural parameters of RBF neural network,including the center and width of hidden layer and connection weight of output layer,and in order to approximatively replace the implicit expression of solving seepage coefficients through water head,it used RBF neural network to realize the mapping from measured water head to seepage coefficients through multiple Gaussian radial basis functions of hidden layer and linear functions of output layer.The excess pore water pressure was rejected from the measured value(It considered the pore pressure dissipation),and it reversed the seepage coefficients of the core walk filter material? and coarse rockfill material ? of Nuozhadu dam.The seepage coefficients obtained from inversion was used to calculate the steady seepage field and the calculated water head was in good agreement with the observed value,which showed that the methods of inversion of seepage coefficients and rejecting the excess pore water pressure were reasonable.