A method for predicting tunnel slag properties based on parent rock mineral composition

Abstract

The present application relates to the technical field of geotechnical analysis, and in particular to a tunnel hole slag performance prediction method based on parent rock mineral composition. The method comprises: obtaining initial types and classification boundaries corresponding to each performance data; establishing initial prediction models corresponding to each performance data; obtaining a training set, a verification set and a test set of the initial prediction models; training the initial prediction models using the training set and the verification set to obtain a first prediction model; testing the prediction model using the test set to obtain a prediction accuracy; updating the classification boundaries or the initial prediction model according to the prediction accuracy, and training again to obtain a second prediction model; if the prediction accuracy corresponding to the second prediction model reaches a preset standard, stopping updating the classification boundaries or the initial prediction model, and if the preset standard is not reached, continuing to update until the preset standard is reached to obtain a final prediction model. The present application can accurately predict the performance of the hole slag.

Description

Technical Field

[0001] The present invention relates to the technical field of rock and soil analysis, and in particular to a method for predicting tunnel slag properties based on parent rock mineral composition. Background Art

[0002] In recent years, my country's railway construction has developed rapidly. During the tunnel excavation process, a large amount of tunnel slag was generated. The storage and treatment of tunnel slag has caused a large amount of land resources to be wasted and damaged the ecological environment along the railway. At the same time, due to the shortage of sand and gravel resources in my country and the difficulty in transporting raw materials, railway construction faces a severe shortage of concrete raw materials. If the tunnel slag is prepared into sand and gravel aggregates and admixtures and used locally for railway construction across the country, the problem of raw material shortage can be effectively alleviated. However, my country has a vast territory, and the geological conditions in different regions are different. The overall geology is relatively complex, and the properties of the tunnel slag parent rock are varied. It not only contains granite, gneiss, and diorite layers, but also contains weak strata such as slate and mudstone, as well as a large amount of fault crushed rock and sand.

[0003] Due to the wide variety of parent rock types, complex mineral compositions, and large property fluctuations, experimental methods are difficult to fully characterize the properties of railway tunnel slag. In contrast, parent rock property prediction models are crucial for clarifying the performance characteristics of railway tunnel parent rock. Based on information such as rock type and elemental composition, they can accurately predict the physical, chemical, and mechanical properties of the parent rock. However, traditional parent rock property prediction methods often suffer from low accuracy and limited applicability. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide a method for predicting the properties of tunnel slag based on the mineral composition of the parent rock. The technical solution adopted is as follows:

[0005] One embodiment of the present invention provides a method for predicting tunnel slag properties based on parent rock mineral composition, the method comprising:

[0006] Acquire a parent rock information data set, wherein the parent rock information data set includes a parent rock number, mineral composition, physical and chemical properties, and mechanical properties; and preprocess the parent rock information data set;

[0007] Classify each performance data of physical and chemical properties and mechanical properties, obtain the initial type and classification boundary corresponding to each performance data; establish the initial prediction model corresponding to each performance data;

[0008] Based on the parent rock information data set, a training set, a validation set, and a test set of an initial prediction model corresponding to each performance data are obtained; the initial prediction model is trained using the training set and the validation set to obtain a first prediction model; and the prediction model is tested using the test set to obtain a prediction accuracy rate;

[0009] According to the prediction accuracy, the classification boundary or the initial prediction model is updated, and a second prediction model is obtained by training again; if the prediction accuracy corresponding to the second prediction model reaches a preset standard, the updating of the classification boundary or the initial prediction model is stopped, and if the preset standard is not reached, the updating is continued until the preset standard is reached, and a final prediction model is obtained.

[0010] Preferably, the physicochemical properties and mechanical properties include:

[0011] The physicochemical properties and mechanical properties are obtained by experiments; the physicochemical properties include porosity, average pore size, density, harmful element content, pH, water absorption, thermal expansion coefficient and thermal conductivity; and the mechanical properties include hardness, compressive strength, shear strength, tensile strength, elastic modulus, Poisson's ratio, crushed stone value and machine-made sand crushed value.

[0012] Preferably, the preprocessing of the parent rock information dataset includes:

[0013] The preprocessing of the parent rock information dataset includes data cleaning, data screening and data transformation of the data in the parent rock information dataset.

[0014] Preferably, the mineral composition includes:

[0015] The mineral composition includes parent rock lithology and three major rock types to which the parent rock lithology belongs, and parent rock element content; the parent rock is classified according to the parent rock lithology to obtain the three major rock types to which the parent rock lithology belongs.

[0016] Preferably, the initial type corresponding to each performance data includes:

[0017] For the initial type of one performance data, the number of performance data in each initial type is the same as the number of performance data in other initial types.

[0018] Preferably, the initial prediction model includes:

[0019] The initial prediction model includes an input layer, a hidden layer and an output layer; the number of hidden layers is a first preset number, and the number of neurons in the hidden layer is a second preset number.

[0020] Preferably, the prediction accuracy is:

[0021]

[0022] Wherein, δ represents the prediction accuracy; TP represents true positive; FP represents false positive; TN represents true negative; and FN represents false negative.

[0023] Preferably, the updating of the classification boundary or the initial prediction model according to the prediction accuracy includes:

[0024] The initial prediction model is updated for the activation function, the hidden layer and the neuron; and only one of the classification boundary, the activation function, the hidden layer and the neuron is updated each time;

[0025] The updating of the classification boundary comprises: setting a first updating strategy, and the first updating strategy specifically comprises: increasing or decreasing the classification boundary under the condition that the data quantity of each category is ensured to be within a preset interval;

[0026] The updating of the activation function comprises: selecting one of preset activation functions as the activation function of the initial prediction model each time;

[0027] The updating of the hidden layer comprises: under the condition that the number of hidden layers cannot exceed a set number of layers, one layer of hidden layer is added each time;

[0028] The updating of the neuron comprises: under the condition that the number of neurons in each layer of hidden layer is ensured to be less than a set number of neurons, a set number of neurons are added or reduced in each layer of hidden layer each time.

[0029] The embodiments of the present application have at least the following beneficial effects: the present application obtains a parent rock information dataset, and pre-processes the parent rock information dataset to improve the quality of the parent rock information dataset, and then improve the training quality of a subsequent prediction model and the prediction accuracy; further, each performance data in the physicochemical performance and the mechanical performance is classified to obtain an initial type and a classification boundary, which is more in line with the application scenario and improves the training quality of the prediction model; further, a training set, a validation set and a test set are divided, and the training set and the validation set are used to train an initial training model to obtain a first prediction model, and the prediction accuracy of the first prediction model is obtained to guide the subsequent updating of the classification boundary and the initial prediction model to obtain an optimal model; the classification boundary or the initial prediction model is updated according to the prediction accuracy, and a second prediction model is obtained by re-training; if the prediction accuracy corresponding to the second prediction model reaches a preset standard, the updating of the classification boundary or the initial prediction model is stopped, and if the preset standard is not reached, the updating is continued until the preset standard is reached to obtain a final prediction model; through the updating of the classification boundary or the initial prediction model and the re-training, the robustness and accuracy of the prediction model can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, and the advantages thereof, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without any creative effort.

[0031] Figure 1 A method flowchart of a tunnel hole slag performance prediction method based on parent rock mineral composition is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0032] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined object, the specific embodiments, structures, features and effects of a tunnel hole slag performance prediction method based on parent rock mineral composition according to the present application are described in detail as follows in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0034] The specific scheme of the tunnel hole slag performance prediction method based on parent rock mineral composition provided by the present application is specifically described below in combination with the accompanying drawings.

[0035] Embodiment:

[0036] The main application scenario of the present application is: machine learning technology has shown significant advantages in rock performance prediction, especially in processing large amounts of complex data and mining the internal laws between multiple factors, the accuracy and stability of machine learning model are superior to traditional prediction methods, and the application prospect in parent rock performance prediction is broad. Therefore, it is urgent to develop a parent rock performance prediction model based on machine learning method.

[0037] Please refer to Figure 1 which shows a tunnel hole slag performance prediction method flowchart based on parent rock mineral composition provided by an embodiment of the present application, and the method comprises the following steps:

[0038] Step S1, obtain a parent rock information data set, the parent rock information data set includes parent rock number, mineral composition, physicochemical properties and mechanical properties; and the parent rock information data set is preprocessed.

[0039] In constructing the performance prediction model of the prediction mother rock, the mother rock data needs to be obtained, specifically, the mother rock needs to be investigated and tested, and the mineral composition, physicochemical performance data, and mechanical performance data of the mother rock are obtained based on the investigation and testing. The mineral composition, physicochemical performance data, and mechanical performance data of the mother rock are obtained based on experiments, and the mother rock also needs to be numbered to obtain the mother rock number. Further, the mother rock needs to be classified according to the lithology of the mother rock to obtain three main rock types to which the lithology of the mother rock belongs, which are magmatic rock, sedimentary rock, and metamorphic rock. The rock types of the mother rock are represented by encoding, in which 1 represents magmatic rock, 2 represents sedimentary rock, and 3 represents metamorphic rock. These numbers belong to category variables and only represent categories without size distinction. The mineral composition includes the lithology of the mother rock and the three main rock types to which it belongs, and the element content of the mother rock.

[0040] The mother rock number, mineral composition, physicochemical performance, and mechanical performance constitute the mother rock information data set, in which the physicochemical performance includes porosity, average pore size, density, harmful element content, pH, water absorption, thermal expansion coefficient, and thermal conductivity; and the mechanical performance includes hardness, compressive strength, shear strength, tensile strength, elastic modulus, Poisson's ratio, crushed value of crushed stone, and crushed value of machine-made sand. The performance data directly uses the test result data, which are all numerical variables, i.e., there are infinitely many possible values on the definition domain, such as the value of porosity (unit %) can be any number within its definition domain (0, 100).

[0041] Further, the mother rock information data set is preprocessed, specifically, the data in the mother rock information data set is cleaned, filtered, and transformed, and the blank data samples are removed through data cleaning to improve the quality of the data set. Thus, the preprocessed mother rock information data set can be obtained.

[0042] Step S2, each performance data in the physicochemical performance and the mechanical performance is classified to obtain the initial type and the classification boundary corresponding to each performance data; and an initial prediction model corresponding to each performance data is established.

[0043] Further, since the mother rock performance prediction is mainly applied to engineering scenarios, and in order to prevent overfitting in the training process, each performance data in the physicochemical performance and the mechanical performance needs to be classified, i.e., the finally predicted performance data is not a specific performance data value, but an interval or category to which it belongs.

[0044] In the classification of each performance data, the classification boundary needs to be determined. If the classification boundary adopted makes the data amount in each category after classification too different, the subsequent model trained using these data may learn knowledge that is biased towards the category with larger data amount, which may result in the prediction bias of the model used for prediction, making the model not practical. Therefore, when classifying each performance data, the statistical distribution of each performance data should be adjusted to ensure that the data amount in each category is close. Taking density as an example, the performance data is divided into 2.49-2.99, a total of 123 performance data. The performance data is divided into three categories, each category has 41 density data. At this time, the classification boundary is set to 2.63 and 2.71, i.e. the data is divided into 2.49-2.63, 2.63-2.71 and 2.71-2.99 three interval ranges. When classifying each performance data, N classification boundaries can obtain N+1 interval categories, i.e. N+1 initial types, and the data ratio is the proportion of the number of samples in each interval to the effective sample amount of the performance.

[0045] Therefore, each performance data needs to be classified, and the data amount in each type is the same. The classification boundary needs to be determined by the implementer according to the actual data distribution of each performance data. Thus, each performance data in the physical and chemical properties and mechanical properties is initially classified to obtain the initial type and classification boundary corresponding to each performance data.

[0046] Further, for each performance data, a corresponding prediction model is needed, which is referred to as an initial prediction model. The initial prediction model includes an input layer, a hidden layer and an output layer. The number of hidden layers is a first preset number, and the number of neurons in the hidden layer is a second preset number. Preferably, in the embodiment, the first preset number is 1 and the second preset number is 20. It should be noted that the implementer can determine the values of the first preset number and the second preset number according to the actual situation.

[0047] At this point, the initial category and classification boundary corresponding to each performance data can be obtained, and the initial prediction model corresponding to each performance data can also be obtained.

[0048] Step S3, obtaining the training set, validation set and test set of the initial prediction model corresponding to each performance data based on the parent rock information data set; training the initial prediction model using the training set and the validation set to obtain a first prediction model; and testing the prediction model using the test set to obtain the prediction accuracy.

[0049] Further, the training set, the verification set and the test set of the initial prediction model corresponding to each performance data are obtained based on the parent rock information dataset; in the different types of performance data, the input of the initial prediction model corresponding to the performance data belonging to the physicochemical performance is the three major rock categories of the parent rock and the content of one or more parent rock elements, and the output is the initial type of each performance data in the physicochemical performance; the input of the initial prediction model corresponding to the performance data belonging to the mechanical performance is the three major rock categories of the parent rock, the content of one or more parent rock elements and the porosity, and the output is the initial type of each performance data in the mechanical performance.

[0050] The parent rock information dataset is divided into the training set, the verification set and the test set of the initial prediction model corresponding to each performance data according to a certain proportion, and each performance data corresponds to the training set, the verification set and the test set of the initial prediction model;

[0051] The training set is the data source for model training, the verification set is used for evaluating the performance of the model during the training process and serving as the basis for adjustment to form a prediction model with good training performance, and the test set is independent of the model training data and is used for testing the accuracy and robustness of the trained model.

[0052] In the present application, the amount of test set data is 20% of the total data amount, and the remaining 80% of the data amount is used as training data, wherein the training data includes the training set and the verification set; meanwhile, due to the problem of data amount, the number of the training set and the verification set is small, so k-fold cross-validation is used in the training process to deal with the situation of insufficient data. Specifically, the data set of the remaining 80% of the data is divided into k equal parts of sub-samples, and in each iteration, one of the sub-samples is used as the verification set, and the remaining k-1 sub-samples are used as the training set for model training. Subsequently, the prediction performance of the trained model on the verification set is evaluated. This process is repeated k times to ensure that each sub-sample has the opportunity to be used as the verification set. Finally, the results of the k iterations are averaged. The advantage of k-fold cross-validation is that it can make full use of limited data for training and verification in the case of insufficient parent rock data in the dataset, and at the same time, it can reduce the possibility that the evaluation result is affected by the specific data division method. In the present application, the value of k is 10, and the implementer can adjust the folding parameter according to the actual situation.

[0053] Taking an initial prediction model corresponding to a performance data as an example, the initial prediction model is trained using its corresponding training set and validation set. The mean absolute error or mean square error is selected as the loss function. The specific loss function to be used needs to be determined by the implementer based on the actual situation. At the same time, there will be a loss value during the training process, and the loss value decreases rapidly with the increase in the number of iterations, indicating that the initial prediction model has good learnability. Preferably, the number of iterations in the embodiment of the present invention is 400. The implementer can adjust the number of iterations based on the actual situation. In the present invention, after the number of iterations reaches 400, the loss value tends to be flat. The smaller the value of the loss function, the better the prediction performance of the model. After training is completed, the first prediction model is obtained.

[0054] After obtaining the first prediction model, the data in the test set is input into the first prediction model to obtain the prediction results. The confusion matrix is ​​used to evaluate the accuracy. The relationship between the prediction results of the first prediction model on the test set and the actual categories is displayed in matrix form. The rows of the confusion matrix represent the actual categories corresponding to the samples, and the columns represent the predicted categories of the samples predicted by the first prediction model. Each cell contains the number of samples matching the corresponding category. The typical composition includes true positives (TP), false positives (FP), true negatives (TN), and false negatives (FN). The prediction accuracy is then calculated. The prediction accuracy is:

[0055]

[0056] Among them, δ represents the prediction accuracy; TP represents true positive examples; FP represents false positive examples; TN represents true negative examples; and FN represents false negative examples.

[0057] In this way, the initial prediction model for each performance data is trained, the first prediction model corresponding to each performance data is obtained, and the prediction accuracy of the first prediction model is also obtained.

[0058] Step S4, update the classification boundary or the initial prediction model according to the prediction accuracy, and train again to obtain the second prediction model; if the prediction accuracy corresponding to the second prediction model reaches the preset standard, stop updating the classification boundary or the initial prediction model; if it does not reach the preset standard, continue updating until it reaches the preset standard to obtain the final prediction model.

[0059] Due to the difference in performance data, if the same prediction model is used for prediction, the prediction accuracy will not meet the requirements, and the classification of performance data will also affect the final prediction results. Therefore, it is necessary to carefully update the classification boundary and the initial prediction model according to the prediction accuracy corresponding to the prediction model after each training. The update of the initial prediction model includes updating the activation function, hidden layer and neurons.

[0060] If the classification interval of each performance data is too much, the complexity of the model will be increased, which is not conducive to model training. Therefore, when updating the classification boundary, it needs to be adjusted according to the statistical distribution of each output variable to ensure that the amount of data of each class is close to

[0061] The activation function plays a very important role in the model training process. The main function is to add a nonlinear operation to all hidden layers and output layers, so that the output of the neural network is more complex and has stronger expression ability. Common activation functions include Sigmoid, Tanh and ReLU. Different activation functions have different sensitivities to different data types.

[0062] The hidden layer structure includes the number of hidden layers and the number of neurons in each layer. Increasing the number of hidden layers can help the model capture more complex features, but it may also cause overfitting. Reducing the number of hidden layers can reduce the complexity of the model and prevent overfitting, but it may reduce the learning ability of the model. Increasing the number of neurons can improve the learning ability of the model, but too many neurons may cause overfitting and increase the computational cost.

[0063] Specifically, when updating the classification boundary, activation function, hidden layer and neuron, only one item is updated each time, and the others are not updated. For example, when updating the multi-classification boundary, the initial prediction model remains unchanged, that is, the activation function, hidden layer and neuron remain unchanged.

[0064] For classification boundary update, it includes: setting a first update strategy, the first update strategy is: increasing or decreasing the classification boundary under the condition that the amount of data of each class is within a preset interval. It should be noted that the preset interval is 15% to 50% of the amount of each performance data, that is, the amount of data in each classification interval is not less than 15% and not more than 50%. In addition, the step size of increasing or decreasing the classification boundary needs to be determined by the implementer according to the actual situation.

[0065] For activation function update, it includes: in the embodiment of the present application, the preset activation function is an activation function, which has three activation functions: Sigmoid, Tanh and ReLU. Each time of updating selects one of the three activation functions as the activation function of the initial prediction model, and then trains it.

[0066] For hidden layer update, it includes: when the hidden layer is updated, the number of hidden layers cannot exceed the set number of layers, and each time one layer of hidden layer is added, wherein the set number of layers is 3 layers.

[0067] For neuron updating, that is, updating the number of neurons in each hidden layer, including: under the condition that the number of neurons in each hidden layer does not exceed the set number, increasing or decreasing the set number of neurons in each hidden layer each time, wherein the set number is 10, for example, there are two layers of neurons in the model, then the number of neurons in the two layers of neurons is increased by 10.

[0068] After each update of the first prediction model, the training data needs to be used for training to obtain a second prediction model, and then the prediction accuracy of the second prediction model obtained after the update is calculated, and so on. After each update, training is needed, and then the prediction accuracy of the trained prediction model is calculated, until the prediction accuracy reaches the preset standard, and the final prediction model is obtained. Preferably, the preset standard in the embodiment of the application is 80%. Finally, the final prediction model corresponding to each performance data of the physicochemical performance and the mechanical performance can be obtained, and the best parameters and the accuracy of the final prediction model of part of the performance data are shown in Table 1.

[0069] Table 1

[0070]

[0071] Thus, the final prediction model corresponding to each performance data can be obtained, and the performance of the parent rock can be predicted.

[0072] It should be noted that the above-mentioned sequence of the embodiments of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above describes a specific embodiment of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0073] Each embodiment in the present application is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment focuses on the difference from other embodiments.

[0074] The above only describes the preferred embodiments of the application, and does not limit the application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the application shall be included in the protection scope of the application.

Claims

1. A method for predicting tunnel slag properties based on parent rock mineral composition, characterized in that: The method includes: Acquire a parent rock information data set, wherein the parent rock information data set includes a parent rock number, mineral composition, physical and chemical properties, and mechanical properties; and preprocess the parent rock information data set; Classify each performance data of physical and chemical properties and mechanical properties, obtain the initial type and classification boundary corresponding to each performance data; establish the initial prediction model corresponding to each performance data; Based on the parent rock information data set, a training set, a validation set, and a test set of an initial prediction model corresponding to each performance data are obtained; the initial prediction model is trained using the training set and the validation set to obtain a first prediction model; and the prediction model is tested using the test set to obtain a prediction accuracy rate; The classification boundary or the initial prediction model is updated according to the prediction accuracy, and the second prediction model is trained again; if the prediction accuracy corresponding to the second prediction model reaches the preset standard, the updating of the classification boundary or the initial prediction model is stopped; if it does not reach the preset standard, the updating is continued until the preset standard is reached to obtain the final prediction model; The updating of the classification boundary or the initial prediction model according to the prediction accuracy includes: The initial prediction model includes updating the activation function, hidden layer, and neurons. When updating the classification boundary, activation function, hidden layer, and neurons, only one item is updated at a time. Updating the classification boundary includes: setting a first update strategy, wherein the first update strategy is specifically to increase or decrease the classification boundary under the condition that the amount of data of each category is within a preset range; The activation function update includes: selecting one of the preset activation functions as the activation function of the initial prediction model in each update; The update of hidden layers includes: adding one hidden layer each time under the condition that the number of hidden layers cannot exceed the set number of layers; The neuron update includes: ensuring that the number of neurons in each hidden layer does not exceed the set number, and each update increases or decreases the set number of neurons in each hidden layer.

2. The method for predicting tunnel slag properties based on parent rock mineral composition according to claim 1, characterized in that: The physical and chemical properties and mechanical properties include: The physical and chemical properties and mechanical properties are obtained through experiments; the physical and chemical properties include porosity, average pore size, density, harmful element content, pH, water absorption, thermal expansion coefficient and thermal conductivity; the mechanical properties include hardness, compressive strength, shear strength, tensile strength, elastic modulus, Poisson's ratio, crushing value of gravel and crushing value of machine-made sand.

3. The method for predicting tunnel slag properties based on parent rock mineral composition according to claim 1, characterized in that: The preprocessing of the parent rock information data set includes: Preprocessing of the parent rock information dataset includes data cleaning, data screening and data conversion of the data in the parent rock information dataset.

4. The method for predicting tunnel slag properties based on parent rock mineral composition according to claim 1, characterized in that: The mineral composition includes: The mineral composition includes the lithology of the parent rock and its three major rock types, and the element content of the parent rock; the parent rock is classified according to the lithology of the parent rock to obtain the three major rock types to which the lithology of the parent rock belongs.

5. The method for predicting tunnel slag properties based on parent rock mineral composition according to claim 1, characterized in that: The initial type corresponding to each performance data includes: For an initial type of performance data, the number of performance data in each initial type is the same as the number of performance data in other initial types.

6. The method for predicting tunnel slag properties based on parent rock mineral composition according to claim 1, characterized in that: The initial prediction model includes: The initial prediction model includes an input layer, a hidden layer and an output layer; the number of hidden layers is a first preset number, and the number of neurons in the hidden layers is a second preset number.

7. The method for predicting tunnel slag properties based on parent rock mineral composition according to claim 1, characterized in that: The prediction accuracy is: ， in, Indicates prediction accuracy; TP indicates true positive examples; FP indicates false positive examples; TN indicates true negative examples; and FN indicates false negative examples.

Images