| The seawall engineering plays a positive role in maintaining the security of the coastal areas and ensuring the production and life of the people inside the dike areas and promoting the economic development of the inside seawall. With the advancement of seawall engineering management, strengthening the monitoring and studying the running state of seawall become an important foundation in the protecting of seawall security. The distribution of the seepage pressure is important to the safe operation of seawall, in this paper, the distribution seepage pressure have been studied under the change tide of seawall based on the actual monitoring data.The osmotic pressure of seawall was also rapidly changed by the affection of the changed tide level. In view of the effective of the rapidly changing seepage field, it is particularly significant to analyze the seawall’s unsteady seepage. The basic control differential equation has been studied which is suitable for seawall’s unsteady seepage from the current continuity principle and the Darcy’s law and analyzed the finite element method of seawall’s unsteady seepage field based on the variation principle. The change tidal acts as the external load in the study and with the similarity of the thermal analysis of the finite element software ANSYS and seepage analysis, the seawall’s pressure distribution of every moment can be obtained from the iterative calculation of ANSYS software, as a result, it laid a foundation for rational analyzing the seawall’s pressure condition.The main traditional methods for the permeability parameters include the field test and indoor test, however, with the progress of monitoring technique and calculation analysis method, the inversion method has been widely applied in the permeability parameters. In this paper, the numerical optimization inversion method and neural network parameter identification method were applied to the back analysis of seawall’s permeability parameter. The least squares objective function relative to the seawall’s permeability parameters is founded by the actual and simulation pressure time-series in the numerical optimization inversion method. Then the interval of parameter was gradually reduced by the scanning method and the optimal seawall’s permeability parameters were obtained by the conjugate gradient method. The mapping relation between the osmotic pressure sequence and the permeability parameters of seawall can be effectively established by the neural network. The sample of network can be obtained by numerical simulation technology and by the improved BP neural network for training, the seawall’s permeability parameters identification model can be founded with osmotic pressure sequence. Finally, the seawall’s permeability parameters are calculated from the trained neural network which with the measured pressure sequence putting in. With the two kind of methods for separately calculating and then extracting the simulation and actual of relevant osmotic pressure, it is found that the average relative errors of the simulation and actual osmotic pressure is small, meanwhile, the change rule of the curves and its imitative fitted very well, illustrated it can effectively obtain the seawall’s permeability parameters by the numerical optimization inversion method and neural network parameter identification method, as a result, it provided an important basis for revealing the osmotic pressure distribution of seawall.With the researching of the two inversion methods, it is found that the numerical optimization inversion method performs around the finite element analysis program, it has a clear thinker and it is concise and strong operability. Neural network parameter identification method built the links information of the pressure and permeability parameters in the network weights. It can be effectively inverted for the permeability parameters under a certain number of samples and the inversion result is ideal and stable, also has a better application prospect. |
PDF Full Text Request