A prediction method for soil engineering parameters considering the influence of soil skeleton damage

Abstract

The present invention discloses a soil engineering parameter prediction method that can consider the influence of soil skeleton damage, including collecting data, the data including evaluation indicators and physical indicators; establishing a functional relationship and upper and lower limits between the evaluation indicators and the physical indicators; defining the value range of the physical indicators based on general engineering experience, generating a random sample set based on the functional relationship between the evaluation indicators and the physical indicators, the upper and lower limits, and the value range of the physical indicators; constructing a RBF neural network, training and validating the RBF neural network, and inputting the prediction set into the prediction network for prediction; selecting the evaluation indicator to be predicted, inputting the physical indicators in the measured data into the prediction network for prediction, obtaining the predicted value of the evaluation indicator and determining whether it meets the upper and lower limits. If so, the predicted value is used as the value of the evaluation indicator. Establishing the relationship between the physical indicator and the evaluation indicator, inputting the data based on the relationship into the RBF neural network for training and prediction, thereby improving the accuracy of prediction.

Description

Technical Field

[0001] The present invention relates to the field of soil bodies affected by soil skeleton damage, and in particular to a soil body engineering parameter prediction method that can take into account the influence of soil skeleton damage. Background Art

[0002] Soil skeleton degradation can be categorized into physical degradation and biochemical degradation based on different mechanisms. Physical degradation is the process by which soil undergoes changes in particle size distribution or loss of solid mass under the action of external forces, such as the crushing of aggregate or ballast in earth-rock dams or hydraulic erosion of soil. Biochemical degradation is the process by which the active components of the soil skeleton are converted into liquid and gas phases through chemical or biochemical reactions. Failure to accurately assess the impact of soil skeleton degradation on geotechnical structures can lead to serious geotechnical engineering problems, such as cracking and excessive settlement of earth-rock dam panels caused by aggregate crushing, and environmental geotechnical hazards caused by the biochemical degradation of municipal solid waste in landfills. For example, in 2008, a leachate leak occurred in Atlanta, USA, spilling 8,000 gallons of leachate into surrounding wetlands. In 2015, the face panels of the Campos Novos Dam in Brazil suffered severe cracking during the initial filling period, damaging the anti-seepage system. A major factor contributing to the Campos Novos Dam failure was the high loads imposed by the impoundment on the dam, which resulted in significant particle fragmentation of the aggregate at the base of the dam, leading to excessive settlement and deformation, and cracking of the face panels. In 2016, landfill sludge flow in Panyu, Guangdong, China, severely polluted the surrounding environment. When developing future-oriented geotechnical engineering, soil skeleton degradation is an unavoidable issue. Therefore, predictive assessment of geotechnical engineering impacts of soil skeleton degradation is crucial for preventing potential hazards.

[0003] The prediction of evaluation indicators of the three major characteristics of soil affected by soil skeleton damage is divided into three parts. In the permeability characteristics, the saturated permeability coefficient and unsaturated permeability coefficient are predicted; in the compression characteristics, the modified primary compression index and modified secondary compression index are predicted; in the strength characteristics, the strength parameters cohesion and internal friction angle are predicted.

[0004] After sorting and analyzing the relevant data, it was found that the parameters (saturated permeability coefficient, unsaturated permeability coefficient, modified primary compression index, modified secondary compression index, numerical cohesion and internal friction angle) are related to physical indicators (components, porosity, burial depth, age, cellulose to lignin ratio, volume moisture content, degradation status, etc.). The relationship between physical indicators and parameters was discussed and parameter prediction was carried out to provide a basis for engineering applications.

[0005] For example, in the prediction of landfill permeability characteristics, the saturated permeability coefficient is predicted within the specified physical indicator range and function range by establishing a functional relationship between the relevant physical indicators and the saturated permeability coefficient, and the predicted saturated permeability coefficient is judged: whether it meets the corresponding value range of the saturated permeability coefficient. If the predicted saturated permeability coefficient meets the saturated permeability coefficient requirements, it means that the project is reliable in terms of permeability characteristics. Otherwise, it means that there are problems with the project that need to be inspected and improved in a timely manner, providing guidance and basis for the next step of the project plan.

[0006] MATLAB is a mathematical software used in fields such as deep learning. It allows for algorithm implementation, data processing, and visualization, providing a robust presentation of prediction results. The RBF neural network prediction model is primarily used to predict parameters related to permeability, compression, and strength in soils affected by soil skeleton damage, providing a reliable basis for practical engineering applications.

[0007] However, in the field of soils affected by soil skeleton damage, no research has yet been conducted on the prediction of relevant parameters. Existing prediction research in the general field of soils has also failed to establish relationships between physical indicators and corresponding parameters. Other parameter prediction methods, however, mostly rely on experimentally obtained sample data for training and prediction, resulting in small sample sizes and lack of universal applicability. Furthermore, prediction methods are based on BP neural networks, which learn using a gradient descent error backpropagation algorithm. This results in slow learning speeds and the potential for local minima. RBF neural networks, on the other hand, avoid this local minima problem. As a feedforward neural network, they only require a small number of weights to adjust, resulting in stronger generalization and prediction capabilities. Their approximation, classification, and learning speed are superior to BP neural networks. Summary of the Invention

[0008] The present invention provides a soil engineering parameter prediction method that can take into account the influence of soil skeleton damage, so as to overcome the above technical problems.

[0009] A soil engineering parameter prediction method that takes into account the influence of soil skeleton damage includes:

[0010] Step 1: Collect field measured data, survey measured data, and indoor test data, wherein the data include evaluation indicators and physical indicators. The evaluation indicators include saturated permeability coefficient, modified primary compression index, modified secondary compression index, gas phase intrinsic permeability coefficient, liquid phase intrinsic permeability coefficient, gas phase relative permeability coefficient, liquid phase relative permeability coefficient, triaxial consolidation drained test strength parameter, and triaxial consolidation undrained test effective strength parameter. The physical indicators include burial depth, porosity, age, cellulose to lignin ratio, initial degradable component, volumetric water content, porosity, compressible component, dry weight, axial strain, and initial porosity.

[0011] Step 2: Establish the functional relationship between the evaluation indicators and the physical indicators as well as the upper and lower limits, including the relationship between the burial depth and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the porosity and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the porosity, age and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the porosity, age and initial degradable components and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the age and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the age and initial degradable components and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the ratio of cellulose to lignin and the saturated permeability coefficient as well as the upper and lower limits, and the relationship between the cellulose and lignin ratio and the saturated permeability coefficient as well as the upper and lower limits. The relationship between the ratio of cellulose to lignin and the initial degradable components and the saturated permeability coefficient, as well as the upper and lower limits, the relationship between porosity and the inherent permeability coefficient of the gas phase, as well as the upper and lower limits, the relationship between porosity and the inherent permeability coefficient of the liquid phase, as well as the upper and lower limits, the relationship between volumetric water content and the relative permeability coefficient of the gas phase, as well as the upper and lower limits, the relationship between volumetric water content, porosity and the relative permeability coefficient of the gas phase, as well as the upper and lower limits, the relationship between volumetric water content, porosity and burial depth and the relative permeability coefficient of the liquid phase, as well as the upper and lower limits, the relationship between volumetric water content and the relative permeability coefficient of the liquid phase, as well as the upper and lower limits, the relationship between volumetric water content, porosity and the relative permeability coefficient of the liquid phase The relationship between the permeability coefficient and the upper and lower limits, the relationship between the volumetric water content, porosity and burial depth and the liquid phase relative permeability coefficient and the upper and lower limits, the relationship between the initial porosity and the modified primary compression index and the upper and lower limits, the relationship between the compressible component and the modified primary compression index and the upper and lower limits, the relationship between the dry weight, the compressible component content and the modified primary compression index and the upper and lower limits, the relationship between the initial porosity and the modified secondary compression index and the upper and lower limits, the relationship between the compressible component content and the modified secondary compression index and the upper and lower limits, the relationship between the initial porosity and the compressible component content and the modified secondary compression index and the upper and lower limits, The relationship between axial strain and strength parameters of triaxial consolidation drained test and their upper and lower limits; the relationship between age and strength parameters of triaxial consolidation drained test and their upper and lower limits; the relationship between initial void ratio and strength parameters of triaxial consolidation drained test and their upper and lower limits; the relationship between initial void ratio, axial strain and strength parameters of triaxial consolidation drained test and their upper and lower limits; the relationship between axial strain and effective strength parameters of triaxial consolidation undrained test and their upper and lower limits; the relationship between age and effective strength parameters of triaxial consolidation undrained test and their upper and lower limits; the relationship between age and effective strength parameters of triaxial consolidation undrained test and their upper and lower limits;

[0012] Step 3: Define the value range of physical indicators based on general engineering experience, generate a random sample set based on the functional relationship between the evaluation index and the physical indicator, as well as the upper and lower limits and the value range of the physical indicator, and divide the sample set into a training set, a validation set, and a prediction set;

[0013] Step 4: Construct an RBF neural network, input the training set into the RBF neural network for training, input the validation set into the trained RBF neural network, and judge whether the relative error and determination coefficient of the RBF neural network meet the preset conditions. When the preset conditions are met, obtain the corresponding RBF neural network, save it as a prediction network, input the prediction set into the prediction network for prediction, and obtain the error and relative error between the predicted value and the actual value;

[0014] Step 5: Select the evaluation indicators that need to be predicted, input the physical indicators related to the evaluation indicators in the field measured data and survey measured data into the prediction network for prediction, obtain the predicted value of the evaluation indicator, and judge whether the predicted value of the evaluation indicator meets the upper and lower limits. If it meets the upper and lower limits, the predicted value will be used as the value of the rating indicator.

[0015] Preferably, the step 2 includes:

[0016] According to formulas (1) and (2), the relationship between burial depth and saturated permeability coefficient and the upper and lower limits are established:

[0017] K S =exp((-10+6.2×(Z+1) -0.15 )×ln(10)) (1)

[0018] K S =exp((-8.72+12×(Z+5) -0.7 )×ln(10))~exp((-9+5.1×(Z+1) -0.1 )×ln(10)) (2)

[0019] Where K S is the saturated permeability coefficient, in m / s; Z is the burial depth, in m;

[0020] The relationship between porosity and saturated permeability coefficient and the upper and lower limits are established according to formulas (3) and (4):

[0021] K S =exp((11.03×n-13.30)×ln(10)) (3)

[0022] K S =exp((9×n-9.5)×ln(10))~exp((11×n-14.5)×ln(10)) (4)

[0023] Where n is the porosity;

[0024] According to formulas (5) and (6), the relationship between porosity, age and saturated permeability coefficient and the upper and lower limits are established:

[0025] K S =exp((11.88×n+0.27×t-14.13)×ln(10)) (5)

[0026] K S =exp((12.47×n+0.27×t-13.02)×ln(10))~exp((12.47×n+0.27×t-15.8)×ln(10)) (6)

[0027] Where, t is the age, in years;

[0028] According to formulas (7) and (8), the relationship between porosity, age, initial degradable components and saturated permeability coefficient and the upper and lower limits are established:

[0029] K S =exp((13.9×n+0.25×t+0.029×B0-17.94)×ln(10)) (7)

[0030] K S =exp((13.9×n+0.25×t+0.029×B0-18)×In(I0))~exp((13.2×n+0.25×t+0.029×B0-17.3)×ln(10)) (8)

[0031] Wherein, B0 is the initial degradable component;

[0032] The relationship between age and saturated permeability coefficient and the upper and lower limits are established according to formulas (9) and (10):

[0033] K S =exp((-4.57-0.17×t)×ln(10)) (9)

[0034] K S =exp((-6.5-0.24×t)×ln(10))~exp((-2-0.09×t)×ln(10)) (10)

[0035] According to formulas (11) and (12), the relationship between age, initial degradable components and saturated permeability coefficient as well as the upper and lower limits are established:

[0036] K S =exp((-0.0044×t-0.68×B0+50.35)×ln(10)) (11)

[0037] K S=exp((-0.004×t-0.68×B0+49.36)×ln(10))~exp((-0.004×t+0.68×B0+51.60)×ln(10)) (12)

[0038] According to formulas (13) and (14), the relationship between the ratio of cellulose to lignin and the saturated permeability coefficient and the upper and lower limits are established:

[0039] K S =exp((-5.34+0.23×(C / L))×ln(10)) (13)

[0040] K S =exp((-9.3+0.7×(C / L))×ln(10))~exp((-2.5+0.1×(C / L))×ln(10))(14)

[0041] Where C / L is the ratio of cellulose to lignin;

[0042] The relationship between the ratio of cellulose to lignin and the initial degradable components and the saturated permeability coefficient, as well as the upper and lower limits, are established according to formulas (15) and (16):

[0043] K S =exp(-0.63×B0+0.09×(C / L)+46.11) (15)

[0044] K S =exp((-0.58×B0+0.09×(C / L)+46.11)×ln(10))~exp((-0.70×B0+0.01×(C / L)+50.11)×ln(10)) (16)

[0045] The relationship between porosity and gas phase intrinsic permeability coefficient and the upper and lower limits are established according to formulas (17) and (18):

[0046] k i G =exp((-13.7+4×ln(2×n+1))×ln(10)) (17)

[0047] k i G =exp((-13+4×ln(0.5n+1))×ln(10))~exp((-12.6+8×ln(n+1))×ln(10))(18)

[0048] Where k i G is the gas phase intrinsic permeability coefficient, unit is m 2 ;

[0049] The relationship between porosity and liquid phase intrinsic permeability coefficient and the upper and lower limits are established according to formulas (19) and (20):

[0050] k i L =exp((-15.6+8×ln(b+1))×ln(10)) (19)

[0051] k i L =exp((-15.9+4×ln(2×n+1))×ln(10))~exp((-14.8+4×ln(3×n+1))×ln(10)) (20)

[0052] Where k i L is the intrinsic permeability coefficient of the liquid phase, in m 2 ;

[0053] The relationship between volumetric water content and gas phase relative permeability coefficient and the upper and lower limits are established according to formulas (21, (22):

[0054]

[0055]

[0056] Where, is the relative permeability coefficient of the gas phase, in m 2 , θ is the volumetric water content;

[0057] According to formulas (23) and (24), the relationship between volumetric water content, porosity and gas phase relative permeability coefficient and the upper and lower limits are established:

[0058]

[0059]

[0060] Where, e is the porosity ratio;

[0061] According to formulas (25) and (26), the relationship between volumetric water content, porosity, burial depth and liquid phase relative permeability as well as the upper and lower limits are established:

[0062]

[0063]

[0064] The relationship between volumetric water content and liquid phase relative permeability coefficient and the upper and lower limits are established according to formulas (27) and (28):

[0065]

[0066]

[0067] Where, is the relative permeability coefficient of the gas phase, in m 2 ;

[0068] According to formulas (29) and (30), the relationship between volumetric water content, porosity and liquid phase relative permeability coefficient and the upper and lower limits are established:

[0069]

[0070]

[0071] According to formulas (31) and (32), the relationship between volumetric water content, porosity, burial depth and liquid phase relative permeability as well as the upper and lower limits are established:

[0072]

[0073]

[0074] The relationship between the initial porosity and the modified principal compression index, as well as the upper and lower limits, are established according to formulas (33) and (34):

[0075] C′ C =0.05×e0+0.07 (33)

[0076] C′ C =0.02×e0+0.01~0.09×e0+0.09 (34)

[0077] Where C′ C is the modified main compression index, e0 is the initial porosity ratio;

[0078] The relationship between the compressible component and the modified principal compressibility index and the upper and lower limits are established according to formulas (35) and (36):

[0079] C′ C =2.52E-3×P C +0.09 (35)

[0080] C′ C =2.52E-3×P C +0.02~2.52E-3×P C +0.21 (36)

[0081] Where, P C is the compressible component content;

[0082] The relationship between dry weight, compressible component content and modified principal compression index, as well as the upper and lower limits, are established according to formulas (37) and (38):

[0083] C′ C =-0.00854×ρ d +0.00187×P C +0.18 (37)

[0084] C′ C =-0.00854×ρ d +0.00187×P C +0.1~-0.00854×ρ d +0.00187×P C +0.3(38)

[0085] Where, ρ d is the dry weight in kN / m 3 ;

[0086] The relationship between the initial porosity and the modified secondary compression index, as well as the upper and lower limits, are established according to formulas (39) and (40):

[0087] C″ C =0.0023×e0+0.038 (39)

[0088] C″ C =0.0013×e0+0.012~-0.0023×e0+0.061 (40)

[0089] Where C″ C is the modified subcompression index;

[0090] According to formulas (41) and (42), the relationship between the compressible component content and the modified secondary compression index and the upper and lower limits are established:

[0091] C″ C =6.64e-4×P C -0.0167 (41)

[0092] C″ C =3e-4×P C -0.012~9.5e-4×P C -0.0046 (42)

[0093] According to formulas (43) and (44), the relationship between the initial porosity, the compressible component content and the modified secondary compression index, as well as the upper and lower limits, are established:

[0094] C″ C =0.00163×e0+2.95e-4×PC +0.008 (43)

[0095] C″ C =0.00163×e0+2.95e-4×P C +0.0035~-0.00163×e0+2.95×P C +0.014(44)

[0096] According to formulas (45), (46), (47), and (48), the relationship between axial strain and strength parameters of triaxial consolidation drainage test and the upper and lower limits are established:

[0097] ε a =0.45×(c-6.25) (45)

[0098] ε a =2.8E-3×(c) 2 ~0.9×c+2 (46)

[0099]

[0100]

[0101] Where, ε a is the axial strain; c is the cohesion, the unit is kPa, is the internal friction angle;

[0102] According to formulas (49), (50), (51), and (52), the relationship between age and strength parameters of triaxial consolidation drainage test and the upper and lower limits are established:

[0103] c=1.42×t+19.69 (49)

[0104] c=1.42×t+1.01~1.42×t+40.02 (50)

[0105]

[0106]

[0107] According to formulas (53), (54), (55), and (56), the relationship between the initial void ratio and the strength parameters of the triaxial consolidation drainage test and the upper and lower limits are established:

[0108] c=-15.17×e0+62.02 (53)

[0109] c=-3×e0+10.02~-21.03×e0+114.98 (54)

[0110]

[0111]

[0112] According to formulas (57), (58), (59), and (60), the relationship between the initial void ratio, axial strain, and the strength parameters of the triaxial consolidation drainage test, as well as the upper and lower limits, are established:

[0113] c=-12.13×e0+1.97×ε a +33.20 (57)

[0114] c=-12.13×e0+1.97×ε a +12.13~-12.13×e0+1.97×ε a +66.22 (58)

[0115]

[0116]

[0117] According to formulas (61), (62), (63), and (64), the relationship between axial strain and effective strength parameters of triaxial consolidation undrained test and the upper and lower limits are established:

[0118] ε a =0.58×(c d -5.00) (61)

[0119] ε a =0.50×(c d -15.01)~4.01×(c d -2) (62)

[0120]

[0121]

[0122] Where c d is the cohesion, unit is kPa, is the effective internal friction angle;

[0123] According to formulas (65), (66), (67), and (68), the relationship between age and effective strength parameters of triaxial consolidation undrained test and the upper and lower limits are established:

[0124] c d =-0.82×t+25.02 (65)

[0125] c d =-0.2×t-0.50~-0.82×t+58.10 (66)

[0126]

[0127]

[0128] According to formulas (69), (70), (71), and (72), the relationship between age and effective strength parameters of triaxial consolidation undrained test and the upper and lower limits are established:

[0129] c d =-0.82×t+25.02 (69)

[0130] c d =-0.2×t-0.50~-0.82×t+58.10 (70)

[0131]

[0132]

[0133] Where c d For cohesion, is the effective internal friction angle.

[0134] Preferably, the step three further comprises processing the saturated permeability coefficient, gas phase intrinsic permeability coefficient, liquid phase intrinsic permeability coefficient, gas phase relative permeability coefficient, and liquid phase relative permeability coefficient of the random sample set according to formulas (73), (74), (75), (76), and (77).

[0135] S=log(K S ) (73)

[0136] S i G =log(k i G ) (74)

[0137]

[0138]

[0139]

[0140] Among them, K S is the saturated permeability coefficient, k i G is the gas phase intrinsic permeability coefficient, k i L is the intrinsic permeability coefficient of the liquid phase, is the gas phase relative permeability coefficient, is the gas phase relative permeability coefficient.

[0141] Preferably, generating a random sample set according to the functional relationship between the evaluation index and the physical index, the upper and lower limit ranges, and the value range of the physical index includes defining the range of the physical index in MATLAB, inputting the functional relationship and the upper and lower limits between the physical index and the evaluation index into MATLAB, and generating a random sample set in MATLAB according to the continuous uniform random number syntax unifrnd(a, b, sz).

[0142] Preferably, the activation function of the RBF neural network is a Gaussian function, and its specific form is:

[0143]

[0144] Among them, σ k is the width of the kth hidden layer neuron in the RBF neural network, c k is the normalization constant of the kth hidden node, ||xc k || is the vector xc k The Euclidean norm of the input sample x and the center c of the kth hidden layer neuron k The radial distance between them, K represents the hidden layer neurons in the RBF neural network.

[0145] Preferably, the preset condition includes a determination coefficient R 2 Greater than 0.70, the relative error range is 0 to 15%.

[0146] The present invention provides a soil engineering parameter prediction method that can take into account the influence of soil skeleton damage. The prediction is performed through the RBF neural network, filling the gap in the prediction of parameters related to permeability characteristics, compression characteristics and strength characteristics in geotechnical engineering. The prediction is based on a basis by analyzing the parameter relationship between relevant physical indicators and permeability characteristics, compression characteristics and strength characteristics, rather than randomly inputting data for training and prediction, thereby improving the accuracy of the prediction. The RBF neural network is used to train and predict data, and parameters that meet the site conditions can be determined according to actual conditions. This method has guiding significance for the initial design of the project, improves the reliability of the engineering design, provides a reliable basis for the analysis of subsequent service performance, and also provides a certain foundation for subsequent practical work. BRIEF DESCRIPTION OF THE DRAWINGS

[0147] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0148] Figure 1 It is a flow chart of the method of the present invention;

[0149] Figure 2 It is the structural diagram of the RBF neural network of the present invention;

[0150] Figure 3 is the relationship between the physical indicators and parameters of the present invention;

[0151] Figure 4 The present invention generates random sample group data according to MATLAB;

[0152] Figure 5 It is the training result of the present invention;

[0153] Figure 6 It is a comparison chart of the predicted value and the measured value results of the present invention;

[0154] Figure 7 This is a graph showing the relative error between the age and initial degradable components of the present invention;

[0155] Figure 8 This is the law of R2 values of the present invention for predicting the next 50 ages and initial degradable components;

[0156] Figure 9 is the relative error value range of the present invention at different ages;

[0157] Figure 10 It is a framework diagram of the prediction system of the present invention;

[0158] Figure 11 It is a flow chart of the prediction system development of the present invention. DETAILED DESCRIPTION

[0159] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0160] Figure 1 It is a flow chart of the method of the present invention, as shown in Figure 1 As shown, the method of this embodiment may include:

[0161] Step 1: Collect field measured data, survey measured data, and indoor test data, wherein the data include evaluation indicators and physical indicators. The evaluation indicators include saturated permeability coefficient, modified primary compression index, modified secondary compression index, gas phase intrinsic permeability coefficient, liquid phase intrinsic permeability coefficient, gas phase relative permeability coefficient, liquid phase relative permeability coefficient, triaxial consolidation drained test strength parameter, and triaxial consolidation undrained test effective strength parameter. The physical indicators include burial depth, porosity, age, cellulose to lignin ratio, initial degradable component, volumetric water content, porosity, compressible component, dry weight, axial strain, and initial porosity.

[0162] Step 2: Establish the functional relationship between the evaluation indicators and the physical indicators as well as the upper and lower limits, including the relationship between the burial depth and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the porosity and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the porosity, age and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the porosity, age and initial degradable components and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the age and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the age and initial degradable components and the saturated permeability coefficient as well as the upper and lower limits, the relationship between the ratio of cellulose to lignin and the saturated permeability coefficient as well as the upper and lower limits, and the relationship between the cellulose and lignin ratio and the saturated permeability coefficient as well as the upper and lower limits. The relationship between the ratio of cellulose to lignin and the initial degradable components and the saturated permeability coefficient, as well as the upper and lower limits, the relationship between porosity and the inherent permeability coefficient of the gas phase, as well as the upper and lower limits, the relationship between porosity and the inherent permeability coefficient of the liquid phase, as well as the upper and lower limits, the relationship between volumetric water content and the relative permeability coefficient of the gas phase, as well as the upper and lower limits, the relationship between volumetric water content, porosity and the relative permeability coefficient of the gas phase, as well as the upper and lower limits, the relationship between volumetric water content, porosity and burial depth and the relative permeability coefficient of the liquid phase, as well as the upper and lower limits, the relationship between volumetric water content and the relative permeability coefficient of the liquid phase, as well as the upper and lower limits, the relationship between volumetric water content, porosity and the relative permeability coefficient of the liquid phase The relationship between the permeability coefficient and the upper and lower limits, the relationship between the volumetric water content, porosity and burial depth and the liquid phase relative permeability coefficient and the upper and lower limits, the relationship between the initial porosity and the modified primary compression index and the upper and lower limits, the relationship between the compressible component and the modified primary compression index and the upper and lower limits, the relationship between the dry weight, the compressible component content and the modified primary compression index and the upper and lower limits, the relationship between the initial porosity and the modified secondary compression index and the upper and lower limits, the relationship between the compressible component content and the modified secondary compression index and the upper and lower limits, the relationship between the initial porosity and the compressible component content and the modified secondary compression index and the upper and lower limits, The relationship between axial strain and strength parameters of triaxial consolidation drained test and their upper and lower limits; the relationship between age and strength parameters of triaxial consolidation drained test and their upper and lower limits; the relationship between initial void ratio and strength parameters of triaxial consolidation drained test and their upper and lower limits; the relationship between initial void ratio, axial strain and strength parameters of triaxial consolidation drained test and their upper and lower limits; the relationship between axial strain and effective strength parameters of triaxial consolidation undrained test and their upper and lower limits; the relationship between age and effective strength parameters of triaxial consolidation undrained test and their upper and lower limits; the relationship between age and effective strength parameters of triaxial consolidation undrained test and their upper and lower limits;

[0163] The second step includes:

[0164] According to formulas (1) and (2), the relationship between burial depth and saturated permeability coefficient and the upper and lower limits are established:

[0165] K S =exp((-10+6.2×(Z+1) -0.15 )×ln(10)) (1)

[0166] K S =exp((-8.72+12×(Z+5) -0.7 )×ln(10))~exp((-9+5.1×(Z+1) -0.1 )×ln(10))(2)

[0167] Where K S is the saturated permeability coefficient, in m / s; Z is the burial depth, in m;

[0168] The relationship between porosity and saturated permeability coefficient and the upper and lower limits are established according to formulas (3) and (4):

[0169] K S =exp((11.03×n-13.30)×ln(10)) (3)

[0170] K S =exp((9×n-9.5)×ln(10))~exp(11×n-14.5)×ln(10)) (4)

[0171] Where n is the porosity;

[0172] According to formulas (5) and (6), the relationship between porosity, age and saturated permeability coefficient and the upper and lower limits are established:

[0173] K S =exp((11.88×n+0.27×t-14.13)×ln(10)) (5)

[0174] K S =exp((12.47×n+0.27×t-13.02)×(10))~exp((12.47×n+0.27×t-15.8)×ln(10)) (6)

[0175] Where, t is the age, in years;

[0176] According to formulas (7) and (8), the relationship between porosity, age, initial degradable components and saturated permeability coefficient and the upper and lower limits are established:

[0177] K S =exp((13.9×n+0.25×t+0.029×B0-17.9)×ln(10)) (7)

[0178] K S=exp((13.9×n+0.25×t+0.029×B0-18)×In(I0))~exp((13.2×n+0.25×t+0.029×B0-17.3)×ln(10)) (8)

[0179] Wherein, B0 is the initial degradable component;

[0180] The relationship between age and saturated permeability coefficient and the upper and lower limits are established according to formulas (9) and (10):

[0181] K S =exp((-4.57-0.17×t)×ln(10)) (9)

[0182] K S =exp((-6.5-0.24×t)×ln(10))~exp((-2-0.09×t)×ln(10)) (10)

[0183] According to formulas (11) and (12), the relationship between age, initial degradable components and saturated permeability coefficient as well as the upper and lower limits are established:

[0184] K S =exp((-0.0044×t-0.68×B0+50.35)×ln(10)) (11)

[0185] K S =exp((-0.004×t-0.68×B0+49.36)×ln(10))~exp((-0.004×t-0.68×B0+51.60)×ln(10)) (12)

[0186] According to formulas (13) and (14), the relationship between the ratio of cellulose to lignin and the saturated permeability coefficient and the upper and lower limits are established:

[0187] K S =exp((-5.34+0.23×(C / L))×ln(10)) (13)

[0188] K S =exp((-9.3+0.7×(C / L))×ln(10))~exp((-2.5+0.1×(C / L))×ln(10))(14)

[0189] Where C / L is the ratio of cellulose to lignin;

[0190] The relationship between the ratio of cellulose to lignin and the initial degradable components and the saturated permeability coefficient, as well as the upper and lower limits, are established according to formulas (15) and (16):

[0191] K S =exp(-0.63×B0+0.09×(C / L)+46.11) (15)

[0192] K S =exp((-0.58×B0+0.09×(C / L)+46.11)×ln(10))~exp((-0.70×B0+0.01×(C / L)+50.11)ln(10)) (16)

[0193] The relationship between porosity and gas phase intrinsic permeability coefficient and the upper and lower limits are established according to formulas (17) and (18):

[0194] k i G =exp((-13.7+4×ln(2×n+1))×ln(10)) (17)

[0195] k i G =exp((-13+4×ln(0.5n+1))×ln(10))~exp((-12.6+8×ln(n+1))×ln(10))(18)

[0196] Where k i G is the gas phase intrinsic permeability coefficient, unit is m 2 ;

[0197] The relationship between porosity and liquid phase intrinsic permeability coefficient and the upper and lower limits are established according to formulas (19) and (20):

[0198] k i L =exp((-15.6+8×ln(n+1))×ln(10)) (19)

[0199] k i L =exp((-15.9+4×ln(2×n+1))×ln(10))~exp((-14.8+4×ln(3×n+1))×ln(10)) (20)

[0200] Where k i L is the intrinsic permeability coefficient of the liquid phase, in m 2 ;

[0201] The relationship between volumetric water content and gas phase relative permeability coefficient and the upper and lower limits are established according to formulas (21, (22):

[0202]

[0203]

[0204] Where, is the relative permeability coefficient of the gas phase, in m 2 , θ is the volumetric water content;

[0205] According to formulas (23) and (24), the relationship between volumetric water content, porosity and gas phase relative permeability coefficient and the upper and lower limits are established:

[0206]

[0207]

[0208] Where, e is the porosity ratio;

[0209] According to formulas (25) and (26), the relationship between volumetric water content, porosity, burial depth and liquid phase relative permeability as well as the upper and lower limits are established:

[0210]

[0211]

[0212] The relationship between volumetric water content and liquid phase relative permeability coefficient and the upper and lower limits are established according to formulas (27) and (28):

[0213]

[0214]

[0215] Where, is the relative permeability coefficient of the gas phase, in m 2 ;

[0216] According to formulas (29) and (30), the relationship between volumetric water content, porosity and liquid phase relative permeability coefficient and the upper and lower limits are established:

[0217]

[0218]

[0219] According to formulas (31) and (32), the relationship between volumetric water content, porosity, burial depth and liquid phase relative permeability as well as the upper and lower limits are established:

[0220]

[0221]

[0222] The relationship between the initial porosity and the modified principal compression index, as well as the upper and lower limits, are established according to formulas (33) and (34):

[0223] C′ C =0.05×e0+0.07 (33)

[0224] C′ C =0.02×e0+0.01~0.09×e0+0.09 (34)

[0225] Where C′ C is the modified main compression index, e0 is the initial porosity ratio;

[0226] The relationship between the compressible component and the modified principal compressibility index and the upper and lower limits are established according to formulas (35) and (36):

[0227] C′ C =2.52E-3×P C +0.09 (35)

[0228] C′ C =2.52E-3×P C +0.02~2.52E-3×P C +0.21 (36)

[0229] Where, P C is the compressible component content;

[0230] The relationship between dry weight, compressible component content and modified principal compression index, as well as the upper and lower limits, are established according to formulas (37) and (38):

[0231] C′ C =-0.00854×ρ d +0.00187×P C +0.18 (37)

[0232] C′ C =-0.00854×ρ d +0.00187×P C +0.1~-0.00854×ρ d +.0.00187×P C +0.3(38)

[0233] Where, ρ d is the dry weight in kN / m 3 ;

[0234] The relationship between the initial porosity and the modified secondary compression index, as well as the upper and lower limits, are established according to formulas (39) and (40):

[0235] C″ C =0.0023×e0+0.038 (39)

[0236] C″ C =-0.0013×e0+0.012~-0.0023×e0+0.061 (40)

[0237] Where C″ C is the modified subcompression index;

[0238] According to formulas (41) and (42), the relationship between the compressible component content and the modified secondary compression index and the upper and lower limits are established:

[0239] C″ C =6.64e-4×P C -0.0167 (41)

[0240] C″ C =3e-4×P C -0.012~9.5e-4×P C -0.0046 (42)

[0241] According to formulas (43) and (44), the relationship between the initial porosity, the compressible component content and the modified secondary compression index, as well as the upper and lower limits, are established:

[0242] C″ C =-0.00163×e0+2.95e-4×P C +0.008 (43)

[0243] C″ C =-0.00163×e0+2.95e-4×P C +0.0035~-0.00163×e0+2.95e-4×P C +0.014(44)

[0244] According to formulas (45), (46), (47), and (48), the relationship between axial strain and strength parameters of triaxial consolidation drainage test and the upper and lower limits are established:

[0245] ε a =0.45×(c-6.25) (45)

[0246] ε a =2.8E-3×(c) 2 ~0.9×c+2 (46)

[0247]

[0248]

[0249] Where, ε a is the axial strain; c is the cohesion, the unit is kPa, is the internal friction angle;

[0250] According to formulas (49), (50), (51), and (52), the relationship between age and strength parameters of triaxial consolidation drainage test and the upper and lower limits are established:

[0251] c=1.42×t+19.69 (49)

[0252] c=1.42×t+1.01~1.42×t+40.02 (50)

[0253]

[0254]

[0255] According to formulas (53), (54), (55), and (56), the relationship between the initial void ratio and the strength parameters of the triaxial consolidation drainage test and the upper and lower limits are established:

[0256] c=-15.17×e0+62.02 (53)

[0257] c=-3×e0+10.02~-21.03×e0+114.98 (54)

[0258]

[0259]

[0260] According to formulas (57), (58), (59), and (60), the relationship between the initial void ratio, axial strain, and the strength parameters of the triaxial consolidation drainage test, as well as the upper and lower limits, are established:

[0261] c=-12.13×e0+1.97×ε a +33.20 (57)

[0262] c=-12.13×e0+1.97×ε a +12.13~-12.13×e0+1.97×ε a +66.22 (58)

[0263]

[0264]

[0265] According to formulas (61), (62), (63), and (64), the relationship between axial strain and effective strength parameters of triaxial consolidation undrained test and the upper and lower limits are established:

[0266] ε a =0.58×(c d -5.00) (61)

[0267] ε a =0.50×(c d -15.01)~4.01×(c d -2) (62)

[0268]

[0269]

[0270] Where c d is the cohesion, unit is kPa, is the effective internal friction angle;

[0271] According to formulas (65), (66), (67), and (68), the relationship between age and effective strength parameters of triaxial consolidation undrained test and the upper and lower limits are established:

[0272] c d =-0.82×t+25.02 (65)

[0273] c d =-0.2×t-0.50~-0.82×t+58.10 (66)

[0274]

[0275]

[0276] According to formulas (69), (70), (71), and (72), the relationship between age and effective strength parameters of triaxial consolidation undrained test and the upper and lower limits are established:

[0277] c d =-0.82×t+25.02 (69)

[0278] c d =-0.2×t-0.50~-0.82×t+58.10 (70)

[0279]

[0280]

[0281] Where c dFor cohesion, is the effective internal friction angle.

[0282] Step 3: Based on general engineering experience, define the value range of physical indicators, organize the physical indicators and parameter values that affect the permeability, compression and strength characteristics, and establish the functional relationship between physical indicators and parameters by analyzing different physical indicators. The test data table including burial depth, porosity, age, C / L, initial degradable component content, compressible component content, etc. is shown below.

[0283] Table 1 Saturated permeability coefficient values under different physical indicators

[0284] Physical indicators and parameters Buried depth (m) Porosity Age (year) C / L Buried depth (m) 0～60 - - - Porosity - 0.42～0.80 - - Age (year) - 3～10 1～10 - C / L - - - 0.40～4.15 Initial degradable components (%) - 38.47～77.00 80.00～91.00 80.00～91.00 Saturated permeability coefficient (m / s) <![CDATA[1.0×10 -9 ~1.0×10 -5 ]]> <![CDATA[3.42×10 -7 ~5.94×10 -4 ]]> <![CDATA[1.35×10 -9 ~4.60×10 -5 ]]> <![CDATA[1.35×10 -9 ~4.60×10 -5 ]]>

[0285] Table 2 Unsaturated permeability coefficient under different physical indicators

[0286]

[0287] Table 3 Corrected primary compression index and corrected secondary compression index under different physical indicators

[0288] Physical indicators and parameters Initial porosity Initial porosity Dry weight Compressible component (%) Initial porosity 1.48～4.10 1.48～4.96 - - Dry weight - - 2.44～12.33 - Compressible component (%) - 30.07～88.88 17.95～84.45 17.95～84.45 Modified Primary Compression Index 0.097～0.438 - 0.119～0.381 0.119～0.381 Modified Subcompression Index - 0.0071～0.063 - 0.0071～0.063

[0289] Table 4 Strength parameters under different physical indicators

[0290]

[0291]

[0292] Generate a random sample set based on the functional relationship between the evaluation index and the physical index, as well as the upper and lower limits and the value range of the physical index. Define the range of the physical index in MATLAB, input the functional relationship and upper and lower limits between the physical index and the evaluation index into MATLAB, and generate a random sample set in MATLAB according to the continuous uniform random number syntax unifrnd(a,b,sz).

[0293] Different physical indicators have different relationships with parameters. By collecting and organizing data, the functional relationship between the corresponding physical indicators is provided for each parameter. At the same time, the magnitude of the permeability characteristic parameter value is very small. In order to prevent large error fluctuations from affecting the prediction accuracy during the training process, the sample data within the specified function range are processed. That is, according to formulas (73), (74), (75), (76), and (77), the saturated permeability coefficient, gas phase inherent permeability coefficient, liquid phase inherent permeability coefficient, gas phase relative permeability coefficient, and liquid phase relative permeability coefficient of the random sample set are processed.

[0294] S=log(K S ) (73)

[0295] S i G =log(k i G ) (74)

[0296]

[0297]

[0298]

[0299] Among them, K S is the saturated permeability coefficient, k i G is the gas phase intrinsic permeability coefficient, k i L is the intrinsic permeability coefficient of the liquid phase, is the gas phase relative permeability coefficient, is the gas phase relative permeability coefficient.

[0300] The compression and strength characteristics data levels are normal, and the predicted data are relatively stable, so no data processing is required.

[0301] The sample set is divided into a training set, a validation set, and a prediction set. The sample set is automatically divided into 70% of the sample set data as the training set, 15% as the validation set, and 15% as the prediction set. The training set is used to train the network, the validation set is used to prevent overfitting, and the prediction set is used to verify the prediction results of the neural network.

[0302] The relative error and coefficient of determination R of the model 2 To evaluate the stability of the sample and the predictive ability of the neural network, since the generated random sample data is relatively discrete and the data fluctuates greatly, the range of the relative error is slightly larger, and the determination coefficient R is selected. 2 When it is greater than 0.70, the sample is considered to have good stability and the neural network has good predictive ability. When the relative error value range is 0-15%, it is considered to be relatively reasonable.

[0303] Step 4: Construct RBF neural network. Figure 2 is the RBF network structure diagram, X=[X1,X2,X3,…,X n ] is the input vector, the activation function (basis function) of hidden node i; W is the input weight matrix; W ik(k = 1, 2, ..., K) is the synaptic weight between the i-th node in the hidden layer and the n-th node in the output layer; Y is the output. The algorithm's principle is to construct the hidden layer space using radial basis functions (RBFs) as the "basis" of neurons, eliminating the need to map the input vector to the hidden layer space through weighted connections. The mapping from the hidden layer to the output layer is the linear summation of the hidden layer outputs.

[0304] Theoretical analysis and simulation results show that the form of RBF has little effect on the performance of RBF neural network. The activation function of the RBF neural network established in this prediction system is a Gaussian function, and its specific form is:

[0305]

[0306] Among them, σ k is the width of the kth hidden layer neuron, c k is the normalization constant of the kth hidden node, ||xc k || is the vector xc k The Euclidean norm of the input sample x and the center c of the kth hidden layer neuron k The radial distance between them, K represents the number of hidden neurons in the RBF neural network. The center and width of all hidden neurons constitute the hidden layer parameters of the RBF neural network. At x = c k The only maximum value is obtained at ||xc k As || increases, its value decays rapidly, that is, the activation degree of the neuron decreases. Therefore, it will only be activated when the center of the hidden layer neuron and the sample are close, making the entire RBF neural network present a "local mapping" feature.

[0307] For a given input sample x p , the output of the nth output node of the network is:

[0308]

[0309] Where w 0n Indicates bias.

[0310] For all P inputs, the actual output matrix of the network is:

[0311] Y=ΦW (80)

[0312] In this embodiment, the sample X in the input training set p =(r1,r2,…,r n ),r n The nth value of the input physical indicator corresponds to the expected output d pFor the corresponding parameter values of permeability, compression and strength characteristics, a mapping function F(x) is established with input and output samples. The p connection weights between the hidden layer and the output layer are determined, and the samples in the training set are input one by one to obtain the unknown coefficients w p , (p=1,2,…,P) linear equations.

[0313] Step 5: Select the evaluation indicators that need to be predicted, input the physical indicators related to the evaluation indicators in the field measured data and survey measured data into the prediction network for prediction, obtain the predicted value of the evaluation indicator, and judge whether the predicted value of the evaluation indicator meets the upper and lower limits. If it meets the upper and lower limits, the predicted value will be used as the value of the rating indicator.

[0314] Implement the prediction function of the RBF neural network in MATLAB and write the relevant program for the RBF neural network system. To achieve the accuracy of the prediction system, the RBF neural network needs to be trained and the measured values of physical indicators are input into the RBF neural network for prediction. The RBF neural network prediction value is obtained and the error between the measured and predicted values is analyzed to determine the prediction ability of the neural network. The prediction is implemented by writing code in MATLAB.

[0315] Table 5 RBF related training parameter values

[0316]

[0317] Table 6 Data input

[0318]

[0319]

[0320] Table 7 Prediction implementation process

[0321]

[0322] After the above RBF neural network prediction system is run, the predicted results of the output measured physical indicators are correlated with the measured results, and unreasonable prediction results are re-predicted.

[0323] Take the prediction of saturated permeability coefficient under the dual factors of saturated permeability characteristics and degradable components as an example. First, the data is sorted and analyzed and the corresponding functional relationship is established with the saturated permeability coefficient. The corresponding relationship diagram is shown in the figure below. Figure 3 As shown. Input the relationship into MATLAB to generate random sample group data, as shown Figure 4 The random data set obtained above is input into the RBF neural network for training, and the training results are as follows. Figure 5The measured physical index values are input into the RBF neural network to obtain the prediction results and corresponding error analysis results as shown in Figure 6 、 7 , as shown in Figure 8. Figure 9 The relative error distributions under the effects of age alone and age combined with initial degradable components are shown. The relative errors are categorized and organized according to different ages. The figure shows that the relative errors under the effects of age and initial degradable components are small, with a narrow range of variation. The predicted values are close to the measured values, with an average relative error of 8.5%. When the relative error reaches a maximum of 14.4%, the age is 1.5. The relative error under the influence of a single factor shows a more dramatic trend, reaching a maximum of 18.6% at age. This indicates that under the influence of age alone, the relative error increases with age. Comparing the relative error ranges under the effects of age, age, and initial degradable component content, it is found that the extreme values of the relative errors under the influence of different physical indicators correspond to different ages.

[0324] This embodiment can be used as a prediction system, such as Figure 10 As shown, the reliability of the project is predicted, including determining the characteristics of the unsaturated soil that need to be analyzed, including permeability, compression and strength characteristics. The permeability characteristics include saturated permeability and unsaturated permeability. The compression characteristics include primary compression and secondary compression. The strength characteristics include cohesion and internal friction angle. The field measured data, survey measured data, and indoor test data are collected, and data analysis is performed to analyze the trend of the relevant factor data. The corresponding functional relationship is established and the envelope diagram is drawn. The parameter database is constructed, and the prediction and output results are learned through the RBF neural network. The results are verified to determine whether they meet the value range of the specified parameters. If the saturated permeability coefficient, the modified primary compression index, the modified secondary compression index, the gas phase inherent permeability coefficient, the liquid phase inherent permeability coefficient, the gas phase relative permeability coefficient, the liquid phase relative permeability coefficient, the triaxial consolidation drainage test strength parameter and the triaxial consolidation undrained test effective strength parameter are all within their respective value ranges, then the reliability of the project is determined overall in combination with expert experience. If there are parameter values that do not meet the value range, the project repair is guided according to the parameter values and expert experience. The open flow chart of the prediction system is as follows. Figure 11 As shown in the figure, a functional relationship is established according to the relevant physical indicators, random data of physical parameters are generated according to continuous uniform random numbers, and input into the corresponding function range to generate a random sample group. The data is processed and divided into data sets, 70% is divided into training set, 15% is divided into validation set, and 15% is divided into test set, which are input into RBF neural network. The error range is selected and the relative error is set between 0 and 15%. The coefficient of determination R 2 If it is greater than 0.70, input the measured value, output the result, and verify the result.

[0325] Overall beneficial effects:

[0326] The present invention provides a soil engineering parameter prediction method that can take into account the influence of soil skeleton damage. The prediction is performed through the RBF neural network, filling the gap in the prediction of parameters related to permeability characteristics, compression characteristics and strength characteristics in geotechnical engineering. The prediction is based on a basis by analyzing the parameter relationship between relevant physical indicators and permeability characteristics, compression characteristics and strength characteristics, rather than randomly inputting data for training and prediction, thereby improving the accuracy of the prediction. The RBF neural network is used to train and predict data, and parameters that meet the site conditions can be determined according to actual conditions. This method has guiding significance for the initial design of the project, improves the reliability of the engineering design, provides a reliable basis for the analysis of subsequent service performance, and also provides a certain foundation for subsequent practical work.

[0327] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A soil engineering parameter prediction method that can take into account the influence of soil skeleton damage, characterized in that: include, Step 1: Collect field measured data, survey measured data, and indoor test data, wherein the data include evaluation indicators and physical indicators. The evaluation indicators include saturated permeability coefficient, modified primary compression index, modified secondary compression index, gas phase intrinsic permeability coefficient, liquid phase intrinsic permeability coefficient, gas phase relative permeability coefficient, liquid phase relative permeability coefficient, triaxial consolidation drained test strength parameter, and triaxial consolidation undrained test effective strength parameter. The physical indicators include burial depth, porosity, age, cellulose to lignin ratio, initial degradable component, volumetric water content, porosity, compressible component, dry weight, axial strain, and initial porosity. Step 2: Establish the functional relationship between the evaluation index and the physical index, as well as the upper and lower limits, including the relationship between the burial depth and the saturated permeability coefficient, as well as the upper and lower limits; Step 3: Define the value range of physical indicators based on general engineering experience, generate a random sample set based on the functional relationship between the evaluation index and the physical indicator, as well as the upper and lower limits and the value range of the physical indicator, and divide the sample set into a training set, a validation set, and a prediction set; Step 4: Construct an RBF neural network, input the training set into the RBF neural network for training, input the validation set into the trained RBF neural network, and judge whether the relative error and determination coefficient of the RBF neural network meet the preset conditions. When the preset conditions are met, obtain the corresponding RBF neural network, save it as a prediction network, input the prediction set into the prediction network for prediction, and obtain the error and relative error between the predicted value and the actual value; Step 5: Select the evaluation indicators that need to be predicted, input the physical indicators related to the evaluation indicators in the field measured data and survey measured data into the prediction network for prediction, obtain the predicted value of the evaluation indicator, and judge whether the predicted value of the evaluation indicator meets the upper and lower limits. If it meets the upper and lower limits, the predicted value will be used as the value of the rating indicator.

2. A soil engineering parameter prediction method that can take into account the influence of soil skeleton damage according to claim 1, characterized in that: The second step includes: The relationship between porosity and saturated permeability coefficient and their upper and lower limits, the relationship between porosity, age and saturated permeability coefficient and their upper and lower limits, the relationship between porosity, age and initial degradable components and saturated permeability coefficient and their upper and lower limits, the relationship between age and saturated permeability coefficient and their upper and lower limits, the relationship between age and initial degradable components and saturated permeability coefficient and their upper and lower limits, the relationship between cellulose to lignin ratio and saturated permeability coefficient and their upper and lower limits, the relationship between cellulose to lignin ratio and initial degradable components and saturated permeability coefficient and their upper and lower limits, the relationship between porosity and gas The relationship between the intrinsic permeability coefficient of the phase and the upper and lower limits, the relationship between porosity and the intrinsic permeability coefficient of the liquid phase and the upper and lower limits, the relationship between volumetric water content and the relative permeability coefficient of the gas phase and the upper and lower limits, the relationship between volumetric water content, porosity and the relative permeability coefficient of the gas phase and the upper and lower limits, the relationship between volumetric water content, porosity and burial depth and the relative permeability coefficient of the liquid phase and the upper and lower limits, the relationship between volumetric water content and the relative permeability coefficient of the liquid phase and the upper and lower limits, the relationship between volumetric water content, porosity and burial depth and the relative permeability coefficient of the liquid phase and the upper and lower limits, the relationship between volumetric water content, porosity and burial depth and the relative permeability coefficient of the liquid phase and the upper and lower limits, The relationship between the initial porosity and the relative permeability of the liquid phase and the upper and lower limits, the relationship between the initial porosity and the modified principal compression index and the upper and lower limits, the relationship between the compressible component and the modified principal compression index and the upper and lower limits, the relationship between the dry weight, the compressible component content and the modified principal compression index and the upper and lower limits, the relationship between the initial porosity and the modified secondary compression index and the upper and lower limits, the relationship between the compressible component content and the modified secondary compression index and the upper and lower limits, the relationship between the initial porosity and the compressible component content and the modified secondary compression index and the upper and lower limits, the relationship between the axial strain and the strength parameters of the triaxial consolidation drained test and the upper and lower limits, the relationship between the age and the strength parameters of the triaxial consolidation drained test and the upper and lower limits, the relationship between the initial porosity and the strength parameters of the triaxial consolidation drained test and the upper and lower limits, the relationship between the initial porosity and the axial strain and the strength parameters of the triaxial consolidation drained test and the upper and lower limits, the relationship between the initial porosity and the axial strain and the strength parameters of the triaxial consolidation drained test and the upper and lower limits, the relationship between the axial strain and the effective strength parameters of the triaxial consolidation undrained test and the upper and lower limits, the relationship between the age and the effective strength parameters of the triaxial consolidation undrained test and the upper and lower limits.

3. The soil engineering parameter prediction method considering the influence of soil skeleton damage according to claim 1 is characterized in that: The second step includes: According to formulas (1) and (2), the relationship between burial depth and saturated permeability coefficient and the upper and lower limits are established: Where, is the saturated permeability coefficient, in m / s; is the burial depth, in m; According to formulas (3) and (4), the relationship between porosity and saturated permeability coefficient and the upper and lower limits are established: Where, is the porosity; According to formulas (5) and (6), the relationship between porosity, age and saturated permeability coefficient and the upper and lower limits are established: Where, is the age period, in years; According to formulas (7) and (8), the relationship between porosity, age, initial degradable components and saturated permeability coefficient and the upper and lower limits are established: Where, is the initial degradable component; The relationship between age and saturated permeability coefficient and the upper and lower limits are established according to formulas (9) and (10): According to formulas (11) and (12), the relationship between age, initial degradable components and saturated permeability coefficient as well as the upper and lower limits are established: According to formulas (13) and (14), the relationship between the ratio of cellulose to lignin and the saturated permeability coefficient and the upper and lower limits are established: Where, C / L is the ratio of cellulose to lignin; The relationship between the ratio of cellulose to lignin and the initial degradable components and the saturated permeability coefficient, as well as the upper and lower limits, are established according to formulas (15) and (16): The relationship between porosity and gas phase intrinsic permeability coefficient and the upper and lower limits are established according to formulas (17) and (18): Where, is the gas phase intrinsic permeability coefficient, in units of The relationship between porosity and liquid phase intrinsic permeability coefficient and the upper and lower limits are established according to formulas (19) and (20): Where, is the intrinsic permeability coefficient of the liquid phase, in units of The relationship between volumetric water content and gas phase relative permeability coefficient and the upper and lower limits are established according to formulas (21, (22): Where, is the gas phase relative permeability coefficient, in units of , is the volumetric water content; According to formulas (23) and (24), the relationship between volumetric water content, porosity and gas phase relative permeability coefficient and the upper and lower limits are established: Where, is the void ratio; According to formulas (25) and (26), the relationship between volumetric water content, porosity, burial depth and liquid phase relative permeability as well as the upper and lower limits are established: Where, is the gas phase relative permeability coefficient, in units of ; According to formulas (29) and (30), the relationship between volumetric water content, porosity and liquid phase relative permeability coefficient and the upper and lower limits are established: According to formulas (31) and (32), the relationship between volumetric water content, porosity, burial depth and liquid phase relative permeability as well as the upper and lower limits are established: The relationship between the initial porosity and the modified principal compression index, as well as the upper and lower limits, are established according to formulas (33) and (34): Where, To correct the main compression index, is the initial porosity ratio; The relationship between the compressible component and the modified principal compressibility index and the upper and lower limits are established according to formulas (35) and (36): Where, is the compressible component content; The relationship between dry weight, compressible component content and modified principal compression index, as well as the upper and lower limits, are established according to formulas (37) and (38): Where, is the dry weight, in units of ; The relationship between the initial porosity and the modified secondary compression index, as well as the upper and lower limits, are established according to formulas (39) and (40): Where, is the modified subcompression index; According to formulas (41) and (42), the relationship between the compressible component content and the modified secondary compression index and the upper and lower limits are established: According to formulas (43) and (44), the relationship between the initial porosity, the compressible component content and the modified secondary compression index, as well as the upper and lower limits, are established: According to formulas (45), (46), (47), and (48), the relationship between axial strain and strength parameters of triaxial consolidation drainage test and the upper and lower limits are established: Where, is the axial strain; is the cohesion, in units of , is the internal friction angle; The relationship between age and strength parameters of triaxial consolidation drainage test and the upper and lower limits are established according to formulas (49), (50), (51), and (52): According to formulas (53), (54), (55), and (56), the relationship between the initial void ratio and the strength parameters of the triaxial consolidation drainage test and the upper and lower limits are established: According to formulas (57), (58), (59), and (60), the relationship between the initial void ratio, axial strain, and the strength parameters of the triaxial consolidation drainage test, as well as the upper and lower limits, are established: According to formulas (61), (62), (63), and (64), the relationship between axial strain and effective strength parameters of triaxial consolidation undrained test and the upper and lower limits are established: Where, is the cohesion, in units of , is the effective internal friction angle; According to formulas (65), (66), (67), and (68), the relationship between age and effective strength parameters of triaxial consolidation undrained test and the upper and lower limits are established: According to formulas (69), (70), (71), and (72), the relationship between age and effective strength parameters of triaxial consolidation undrained test and the upper and lower limits are established: Where, For cohesion, is the effective internal friction angle.

4. The soil engineering parameter prediction method according to claim 1, wherein the method is characterized in that: The step three also includes processing the saturated permeability coefficient, gas phase intrinsic permeability coefficient, liquid phase intrinsic permeability coefficient, gas phase relative permeability coefficient, and liquid phase relative permeability coefficient of the random sample set according to formulas (73), (74), (75), (76), and (77). in, is the saturated permeability coefficient, is the gas phase intrinsic permeability coefficient, is the intrinsic permeability coefficient of the liquid phase, is the gas phase relative permeability coefficient, is the gas phase relative permeability coefficient.

5. The soil engineering parameter prediction method considering the influence of soil skeleton damage according to claim 1 is characterized in that: The method of generating a random sample set based on the functional relationship between the evaluation index and the physical index, the upper and lower limit ranges, and the value range of the physical index includes defining the range of the physical index in MATLAB, inputting the functional relationship and the upper and lower limits between the physical index and the evaluation index into MATLAB, and generating a random sample set in MATLAB according to the continuous uniform random number syntax unifrnd(a, b, sz).

6. The soil engineering parameter prediction method considering the influence of soil skeleton damage according to claim 1 is characterized in that: The activation function of the RBF neural network is a Gaussian function, and its specific form is: in, It is the first k The width of the hidden layer neurons, For the k The normalization constant of hidden nodes, is a vector The Euclidean norm of the input sample Hedi k The center of the hidden layer neurons The radial distance between K Indicates the number of hidden neurons in the RBF neural network.

7. The soil engineering parameter prediction method considering the influence of soil skeleton damage according to claim 1 is characterized in that: The preset conditions include the determination coefficient R 2 Greater than 0.70, the relative error range is 0~15%.

Images