Research On Optimal Design And Platform Development Of Immersion Test Of Collapsible Loess

Collapsible evaluation is particularly important for the construction of loess areas.The current methods of collapsibility evaluation mainly include indoor test and field immersion test.Compared with the indoor test,the data of the on-site immersion test is more accurate and more in line with the actual working conditions.However,the on-site immersion test has many problems such as many constraints on the site environmental conditions and the lack of quantification of the test design scheme.Therefore,it is very necessary to carry out the optimal design research of the collapsible loess flooding test system.In order to explore the optimal design scheme of loess flooding experiments in different application scenarios,a systematic study was conducted on the infiltration and collapsing deformation laws of different seepage hole layout schemes and test pit sizes through a combination of field tests and numerical simulations.Established a quantitative relationship between the infiltration hole layout and water consumption and infiltration time,the size of the test pit and the amount of collapsible deformation,integrated the relevant quantitative laws through intelligent learning algorithms,and developed a relevant software platform to achieve water seepage in the immersion test Engineering application goals of optimal hole layout,"predict large pits from small pits" and intelligent design of test plans.The main achievements are:(1)Based on the field test pit flooding test and numerical simulation,the law of self-weight collapsibility and water penetration of loess foundation after flooding was revealed.The results show that the ground deformation of the immersion test is the most intense in the center of the test pit,and the ground of the test pit is finally funnel-shaped.The water immersion boundary has a significant inhibitory effect on the deformation of the collapsing;The range of influence is approximately symmetrical "bulb type".At the same time,the test monitoring results are compared with Geostudio numerical simulation calculation results to verify the feasibility and rationality of numerical simulation in analyzing the test pit flooding test.(2)Based on the numerical simulation of Geostudio,the law of water infiltration when arranging water seepage holes with different depths,numbers and spacings in the test pit immersion test was studied.The results show that:setting seepage holes at the bottom of the test pit can effectively shorten the test time and reduce water consumption;when the total depth of the seepage holes is the same,a small number of large-depth seepage holes with a relatively large number of shallow seepage holes have the short test time and low water consumption Characteristics;the test water consumption is directly proportional to the number of seepage holes and inversely proportional to the depth of the seepage holes.(3)Based on the infiltration range obtained by seepage analysis,using Phase2 software,a dense model correction method was used to construct a submerged test collapsible simulation model.For different soil quality,reduction factor,collapsible soil layer thickness and test pit size The development law of sag changes was studied,the restraint effect of the size of the test pits was discussed,the correction proposal for determining the reasonable size of the test pit was put forward,and the universal correction formula for the size of the test pit was established.(4)Through numerical simulation,sample data of infiltration rules under different soil qualities,collapsible soil layer thickness,and seepage hole layout schemes are formed,and a test model for infiltration time and water consumption prediction based on deep belief network is constructed.(5)Based on Matlab-GUI,an on-site immersion test optimization design platform was developed,which realized functions including immersion test pit size correction,collapsible deformation prediction,comparison of results before and after optimization,infiltration time and water consumption prediction,intelligent design of test schemes,etc.