Intelligent prediction method and system for grouting amount of building lifting layer

Abstract

The invention relates to the technical field of building foundation reinforcement and lifting, and provides an intelligent prediction method and system for the grouting amount of a building lifting layer, and the method comprises the steps: collecting the historical engineering data of a completed building grouting reinforcement and lifting project; performing normalization preprocessing on the historical engineering data to form preprocessed data; dividing the preprocessed data into a training set and a test set, and training an initial neural network model through the training set; inputting the test set into the initial neural network model, and then optimizing the initial neural network model according to an output result to form an optimized neural network model; and inputting the current engineering data into the optimized neural network model, outputting a prediction result, and then carrying out reverse normalization processing on the prediction result to obtain a grouting amount prediction value. The method can accurately predict the grouting amount required by building lifting, and provides reference for subsequent engineering.

Description

Technical Field

[0001] The present invention relates to the technical field of building foundation reinforcement and lifting, and in particular to a method and system for intelligently predicting the grouting amount of a building lifting layer. Background Art

[0002] As the core driving force of the new generation of industrial transformation, artificial intelligence technology is an important method to comprehensively improve the digitization, automation, informationization and intelligence of the engineering field. At present, in the field of building grouting and lifting, the prediction of the grouting amount for building lifting is an unresolved problem. If the grouting amount is too little, the building cannot be lifted; if the grouting amount is too much, it will cause waste; or if the grouting amount is not appropriate, the building will be lifted and tilted, leaving safety hazards. Summary of the invention

[0003] The purpose of the present invention is to solve at least one technical problem in the background technology and to provide a method and system for intelligently predicting the grouting amount of a building lifting layer.

[0004] To achieve the above object, the present invention provides a method for intelligently predicting the grouting amount of a building lifting layer, comprising:

[0005] Collect historical engineering data of completed building grouting reinforcement and lifting projects;

[0006] Perform normalization preprocessing on historical engineering data to form preprocessed data;

[0007] Divide the preprocessed data into a training set and a test set, and use the training set to train the initial neural network model;

[0008] Input the test set into the initial neural network model, and then optimize the initial neural network model according to the output results to form an optimized neural network model;

[0009] The current engineering data is input into the optimized neural network model, the prediction results are output, and then the prediction results are denormalized to obtain the grouting volume prediction value.

[0010] According to one aspect of the present invention, the historical engineering data includes: the initial settlement displacement value of the building monitoring point, the porosity of the lifted soil layer, the thickness of the lifted soil layer, the load pressure value on the upper part of the lifted soil layer, the grouting influence coefficient of the surrounding grouting holes and the grouting amount required during lifting.

[0011] According to one aspect of the present invention, the historical engineering data is normalized and preprocessed to form preprocessed data:

[0012] The historical engineering data is normalized and preprocessed using the proportional compression method to convert the data to between 0 and 1. Specifically:

[0013]

[0014] Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

[0015] According to one aspect of the present invention, the initial neural network model is trained by the training set:

[0016] The historical engineering data are divided into M groups, and 4M / 5 groups are used as training sets. The results of the input layer and output layer of the neural network are determined as the feature information corresponding to the training set. Then the neural network model is trained to determine the corresponding parameters in the neural network and form an initial neural network model.

[0017] According to one aspect of the present invention, the step of inputting the test set into the initial neural network model and then optimizing the initial neural network model according to the output result to form an optimized neural network model comprises:

[0018] The remaining M / 5 groups are used as test sets and input into the initial neural network model. The generalization ability of the initial neural network model is evaluated by the mean square error and the number of convergence steps.

[0019] The initial neural network model is optimized according to the evaluation results. Different numbers of hidden layer neurons are substituted and the corresponding network mean square error and convergence steps of the output are calculated respectively. The optimal number of hidden layer neurons is determined according to the minimum mean square error and the minimum convergence step, thus forming an optimized neural network model.

[0020] According to one aspect of the present invention, the denormalization formula is:

[0021]

[0022] Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

[0023] To achieve the above object, the present invention also provides a grouting amount intelligent prediction system for a building lifting layer, comprising:

[0024] The historical engineering data collection module collects historical engineering data of completed building grouting reinforcement and lifting projects;

[0025] The data preprocessing module performs normalization preprocessing on the historical engineering data to form preprocessed data;

[0026] The model training module divides the preprocessed data into a training set and a test set, and trains the initial neural network model through the training set;

[0027] The model optimization module inputs the test set into the initial neural network model, and then optimizes the initial neural network model according to the output results to form an optimized neural network model;

[0028] The grouting volume prediction module inputs the current engineering data into the optimized neural network model, outputs the prediction results, and then denormalizes the prediction results to obtain the grouting volume prediction value.

[0029] To achieve the above-mentioned purpose, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the intelligent prediction method for the grouting amount of the lifting layer of a building as described above.

[0030] To achieve the above objectives, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method for intelligently predicting the grouting amount of a building lifting layer as described above is implemented.

[0031] According to the solution of the present invention, the present invention collects previous engineering data, and adopts a neural network algorithm to automatically learn and extract data features, train the data and generate a model. According to the generated model, the grouting volume can be predicted, which provides a reference for subsequent projects and is crucial to the development of the grouting field.

[0032] The present invention performs normalization preprocessing on the data through a proportional compression method, converting the data to between 0 and 1, which can avoid the excessive influence of certain feature data on the model results, and can significantly improve the verification accuracy and convergence rate of the model, thereby enhancing the model performance.

[0033] The present invention clarifies the optimal number of hidden layer neuron nodes by trial calculation. The number of hidden layer neurons obtained by this method makes the network convergence speed fast, the error small, and improves the generalization and fault tolerance of the network, laying a solid foundation for establishing an accurate grouting volume prediction model. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A flowchart schematically showing a method for intelligently predicting the grouting amount of a building lifting layer according to an embodiment of the present invention;

[0035] Figure 2It is a neural network algorithm diagram of Example 1;

[0036] Figure 3 This is a flow chart of the grouting volume prediction method of Example 1. DETAILED DESCRIPTION

[0037] The present invention will now be discussed with reference to exemplary embodiments. It should be understood that the embodiments discussed are only to enable those skilled in the art to better understand and thus implement the present invention, rather than implying any limitation on the scope of the present invention.

[0038] As used herein, the term “including” and variations thereof are to be interpreted as open-ended terms meaning “including, but not limited to.” The term “based on” is to be interpreted as “based, at least in part, on.” The terms “one embodiment” and “an embodiment” are to be interpreted as “at least one embodiment.”

[0039] Figure 1 The flowchart schematically shows a method for intelligently predicting the grouting amount of a building lifting layer according to an embodiment of the present invention. Figure 1 As shown, in this embodiment, the intelligent prediction method of the grouting amount of the building lifting layer includes:

[0040] Collect historical engineering data of completed building grouting reinforcement and lifting projects;

[0041] Perform normalization preprocessing on historical engineering data to form preprocessed data;

[0042] Divide the preprocessed data into a training set and a test set, and use the training set to train the initial neural network model;

[0043] Input the test set into the initial neural network model, and then optimize the initial neural network model according to the output results to form an optimized neural network model;

[0044] The current engineering data is input into the optimized neural network model, the prediction results are output, and then the prediction results are denormalized to obtain the grouting volume prediction value.

[0045] Further, according to one embodiment of the present invention, the historical engineering data include: the initial settlement displacement value of the building monitoring point, the porosity of the lifted soil layer, the thickness of the lifted soil layer, the load pressure value of the upper part of the lifted soil layer, the grouting influence coefficient of the surrounding grouting holes and the grouting amount required during lifting.

[0046] In this implementation, the initial settlement displacement value of the monitoring point in the engineering project is obtained from the monitoring report of the monitoring unit, and the unit is mm.

[0047] The essence of the porosity of the lifting soil layer is the ratio of the pore volume to the total volume of the soil, which can directly reflect the compressibility of the soil. The smaller the porosity n value of the soil, the denser the soil and the lower the compressibility; the larger the n value, the looser the soil and the higher the compressibility. The porosity n of the soil layer has a great influence on the grouting volume. The calculation formula of the porosity is: Among them: e is the porosity ratio, which is the ratio of the pore volume in the soil to the volume of its solid particles, which can be obtained from the geological survey data of the survey unit.

[0048] When grouting is lifted, the grouting pipe is lifted while grouting. The thickness of the lifted soil layer will also affect the amount of grouting, the unit is m.

[0049] The load pressure value on the upper part of the raised soil layer is the uniformly distributed load transmitted to the raised soil layer of the foundation by the upper building load, and the unit is KPa.

[0050] Each monitoring point corresponds to a grouting hole K, and the grouting conditions of the grouting holes around the grouting hole K will also affect the grouting amount during the grouting lifting of the grouting hole K. The number of other grouting holes within a radius of 6m with the grouting hole K as the center is 0, 1, 2, 3, and 4, which correspond to the influence coefficients γ = 1.0, 0.8, 0.7, 0.6, and 0.5 respectively.

[0051] The grouting volume required for the corresponding grouting hole K at the monitoring point during the settlement and uplift stage is in m3.

[0052] In this embodiment, the above historical engineering data is the situation of a single grouting hole corresponding to the lifting area, and the grouting amount is the grouting amount required when a single grouting hole is used to lift the area.

[0053] Further, according to an embodiment of the present invention, the historical engineering data is normalized and preprocessed to form preprocessed data as follows:

[0054] The historical engineering data is normalized and preprocessed using the proportional compression method to convert the data to between 0 and 1. Specifically:

[0055]

[0056] Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

[0057] Further, according to one embodiment of the present invention, the initial neural network model is trained by the training set as follows:

[0058] The historical engineering data are divided into M groups, and 4M / 5 groups are used as training sets. The results of the input layer and output layer of the neural network are determined as the feature information corresponding to the training set. Then the neural network model is trained to determine the corresponding parameters in the neural network and form an initial neural network model.

[0059] Further, according to an embodiment of the present invention, the test set is input into the initial neural network model, and then the initial neural network model is optimized according to the output result to form an optimized neural network model, including:

[0060] The remaining M / 5 groups are used as test sets and input into the initial neural network model. The generalization ability of the initial neural network model is evaluated by the mean square error and the number of convergence steps.

[0061] The initial neural network model is optimized according to the evaluation results. Different numbers of hidden layer neurons are substituted and the corresponding network mean square error and convergence steps of the output are calculated respectively. The optimal number of hidden layer neurons is determined according to the minimum mean square error and the minimum convergence step, thus forming an optimized neural network model.

[0062] Further, according to an embodiment of the present invention, the formula for denormalization is:

[0063]

[0064] Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

[0065] According to the above scheme of the present invention, the present invention collects previous engineering data, and adopts a neural network algorithm to automatically learn and extract data features, train the data and generate a model. According to the generated model, the grouting volume can be predicted, which provides a reference for subsequent projects and is crucial to the development of the grouting field.

[0066] The present invention performs normalization preprocessing on the data through a proportional compression method, converting the data to between 0 and 1, which can avoid the excessive influence of certain feature data on the model results, and can significantly improve the verification accuracy and convergence rate of the model, thereby enhancing the model performance.

[0067] The present invention clarifies the optimal number of hidden layer neuron nodes by trial calculation. The number of hidden layer neurons obtained by this method makes the network convergence speed fast, the error small, and improves the generalization and fault tolerance of the network, laying a solid foundation for establishing an accurate grouting volume prediction model.

[0068] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a grouting amount intelligent prediction system for a building lifting layer, comprising:

[0069] The historical engineering data collection module collects historical engineering data of completed building grouting reinforcement and lifting projects;

[0070] The data preprocessing module performs normalization preprocessing on the historical engineering data to form preprocessed data;

[0071] The model training module divides the preprocessed data into a training set and a test set, and trains the initial neural network model through the training set;

[0072] The model optimization module inputs the test set into the initial neural network model, and then optimizes the initial neural network model according to the output results to form an optimized neural network model;

[0073] The grouting volume prediction module inputs the current engineering data into the optimized neural network model, outputs the prediction results, and then denormalizes the prediction results to obtain the grouting volume prediction value.

[0074] Further, according to one embodiment of the present invention, the historical engineering data include: the initial settlement displacement value of the building monitoring point, the porosity of the lifted soil layer, the thickness of the lifted soil layer, the load pressure value of the upper part of the lifted soil layer, the grouting influence coefficient of the surrounding grouting holes and the grouting amount required during lifting.

[0075] In this implementation, the initial settlement displacement value of the monitoring point in the engineering project is obtained from the monitoring report of the monitoring unit, and the unit is mm.

[0076] The essence of the porosity of the lifting soil layer is the ratio of the pore volume to the total volume of the soil, which can directly reflect the compressibility of the soil. The smaller the porosity n value of the soil, the denser the soil and the lower the compressibility; the larger the n value, the looser the soil and the higher the compressibility. The porosity n of the soil layer has a great influence on the grouting volume. The calculation formula of the porosity is: Among them: e is the porosity ratio, which is the ratio of the pore volume in the soil to the volume of its solid particles, which can be obtained from the geological survey data of the survey unit.

[0077] When grouting is lifted, the grouting pipe is lifted while grouting. The thickness of the lifted soil layer will also affect the amount of grouting, the unit is m.

[0078] The load pressure value on the upper part of the raised soil layer is the uniformly distributed load transmitted to the raised soil layer of the foundation by the upper building load, and the unit is KPa.

[0079] Each monitoring point corresponds to a grouting hole K, and the grouting conditions of the grouting holes around the grouting hole K will also affect the grouting amount during the grouting lifting of the grouting hole K. The number of other grouting holes within a radius of 6m with the grouting hole K as the center is 0, 1, 2, 3, and 4, which correspond to the influence coefficients γ = 1.0, 0.8, 0.7, 0.6, and 0.5 respectively.

[0080] The grouting volume required for the corresponding grouting hole K at the monitoring point during the settlement and uplift stage is in m3.

[0081] In this embodiment, the above historical engineering data is the situation of a single grouting hole corresponding to the lifting area, and the grouting amount is the grouting amount required when a single grouting hole is used to lift the area.

[0082] Further, according to an embodiment of the present invention, the historical engineering data is normalized and preprocessed to form preprocessed data as follows:

[0083] The historical engineering data is normalized and preprocessed using the proportional compression method to convert the data to between 0 and 1. Specifically:

[0084]

[0085] Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

[0086] Further, according to one embodiment of the present invention, the initial neural network model is trained by the training set as follows:

[0087] The historical engineering data are divided into M groups, and 4M / 5 groups are used as training sets. The results of the input layer and output layer of the neural network are determined as the feature information corresponding to the training set. Then the neural network model is trained to determine the corresponding parameters in the neural network and form an initial neural network model.

[0088] Further, according to an embodiment of the present invention, the test set is input into the initial neural network model, and then the initial neural network model is optimized according to the output result to form an optimized neural network model, including:

[0089] The remaining M / 5 groups are used as test sets and input into the initial neural network model. The generalization ability of the initial neural network model is evaluated by the mean square error and the number of convergence steps.

[0090] The initial neural network model is optimized according to the evaluation results. Different numbers of hidden layer neurons are substituted and the corresponding network mean square error and convergence steps of the output are calculated respectively. The optimal number of hidden layer neurons is determined according to the minimum mean square error and the minimum convergence step, thus forming an optimized neural network model.

[0091] Further, according to an embodiment of the present invention, the formula for denormalization is:

[0092]

[0093] Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

[0094] According to the above scheme of the present invention, the present invention collects previous engineering data, and adopts a neural network algorithm to automatically learn and extract data features, train the data and generate a model. According to the generated model, the grouting volume can be predicted, which provides a reference for subsequent projects and is crucial to the development of the grouting field.

[0095] The present invention performs normalization preprocessing on the data through a proportional compression method, converting the data to between 0 and 1, which can avoid the excessive influence of certain feature data on the model results, and can significantly improve the verification accuracy and convergence rate of the model, thereby enhancing the model performance.

[0096] The present invention clarifies the optimal number of hidden layer neuron nodes by trial calculation. The number of hidden layer neurons obtained by this method makes the network convergence speed fast, the error small, and improves the generalization and fault tolerance of the network, laying a solid foundation for establishing an accurate grouting volume prediction model.

[0097] Furthermore, to achieve the above-mentioned purpose, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the intelligent prediction method for the grouting amount of the lifting layer of a building as described above is implemented.

[0098] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the intelligent prediction method for the grouting amount of the lifting layer of a building as described above is implemented.

[0099] Based on the above solution, the technical solution of the present invention is described in detail in the following in combination with the accompanying drawings in the form of a specific embodiment.

[0100] Example 1

[0101] Five completed engineering projects in the field of building lifting were selected. The five research projects included different foundation types such as raft foundation, pile cap foundation, and single foundation. The settlement amounts were also significantly different. The buildings included low-rise, high-rise, and factory buildings. The data analyzed in this way would be more typical.

[0102] In this embodiment, the intelligent prediction method of the grouting amount of the building lifting layer includes:

[0103] Step 1: Collect engineering data of existing projects;

[0104] Collect engineering data of building grouting reinforcement and lifting projects of multiple completed projects, including:

[0105] ① Initial settlement displacement value S (m);

[0106] ② Porosity of the lifted soil layer n;

[0107] ③Thickness of the lifted soil layer h (m);

[0108] ④ Load pressure value P (KPa) on the upper part of the lifting soil layer;

[0109] ⑤ Grouting influence coefficient γ of the grouting holes around the grouting holes;

[0110] ⑥ The grouting volume Q (m3) required when the grouting hole is lifted.

[0111] The data statistics are shown in Table 1 below:

[0112]

[0113]

[0114]

[0115] Table 1

[0116] Step 2: Preprocessing of historical engineering data;

[0117] The proportional compression method is used to normalize the data in step 1 so that the data is converted to between 0 and 1. The specific method is as follows:

[0118]

[0119] Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

[0120] The data statistics after normalization are shown in Table 2:

[0121]

[0122]

[0123]

[0124] Table 2

[0125] Step 3: Train the initial neural network model;

[0126] The total number of data groups in step 2 is M = 62. Groups 1 to 50 are used as training sets for model training. The neural network algorithm is used when training the model. The neural network receives the original feature information through the input layer, processes and extracts the feature information through the hidden layer, and finally outputs the result through the output layer. Figure 2 shown.

[0127] Input layer: X1 = Data (①), X2 = Data (②), X3 = Data (③), X4 = Data (④), X5 = Data (⑤);

[0128] Hidden layer: The number of neurons is initially set to 2N+1=11, where N is the number of neurons in the input layer.

[0129] Y1=f(W11X1+W21X2+W31X3+W41X4+W51X5+b)

[0130] Y2=f(W12X1+W22X2+W32X3+W42X4+W52X5+b)

[0131] …

[0132] Y11=f(W111X1+W211X2+W311X3+W411X4+W511X5+b);

[0133] Output layer: Z1 = f(V1Y1 + V2Y2 + … + V11Y11) = Data(⑥);

[0134] Activation function f(a): Sigmoid function is selected, f(a) = 1 / (1+e -a ).

[0135] Among them, Wij and Vj are weights, b is a bias term, and the optimal weight and bias term values ​​are found through training with 50 sets of data to obtain the initial neural network model.

[0136] Step 4: Evaluate the optimization model;

[0137] The total number of data sets in step 2 is M = 62. Groups 51 to 62 are used as test sets to input into the initial model to evaluate the performance of the initial model on unseen data. The program shows that the mean square error is 0.00102356, the number of convergence steps is > 2000, and the generalization ability of the model is not good.

[0138] The model is optimized according to the evaluation results. Different numbers of hidden layer neurons are substituted into the program. When the number of hidden layer neurons is 9 to 19, the network will output the corresponding network mean square error and convergence steps, as shown in Table 3 below.

[0139]

[0140] Table 3

[0141] When the number of neurons in the hidden layer is 15, the network converges after 560 trainings. The convergence speed is fast and the error is small, which improves the generalization and fault tolerance of the network and obtains an optimized neural network model.

[0142] Step 5: Use the optimized neural network model to predict the grouting volume of the new project;

[0143] After normalizing the above ①②③④⑤ data of the new project, input them into the optimized neural network model to obtain the predicted value of ⑥ above. Then, the predicted value is denormalized to obtain the final grouting volume prediction value, which can provide a basis for construction. The above process is as follows: Figure 3 shown.

[0144] The formula for denormalizing the data is:

[0145]

[0146] Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

[0147] Those skilled in the art will appreciate that the modules and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0148] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and equipment can refer to the corresponding processes in the aforementioned method implementation methods, and will not be repeated here.

[0149] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0150] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of the present invention.

[0151] In addition, each functional module in the embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0152] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the energy-saving signal sending / receiving method of each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks.

[0153] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other.

[0154] It should be understood that the size of the serial numbers of each step in the content of the invention and the implementation methods of the present invention does not absolutely mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the implementation methods of the present invention.

Claims

1. An intelligent prediction method for the grouting amount of a building lifting layer, characterized in that: include: Collect historical engineering data of completed building grouting reinforcement and lifting projects; Perform normalization preprocessing on historical engineering data to form preprocessed data; Divide the preprocessed data into a training set and a test set, and use the training set to train the initial neural network model; Input the test set into the initial neural network model, and then optimize the initial neural network model according to the output results to form an optimized neural network model; The current engineering data is input into the optimized neural network model, the prediction results are output, and then the prediction results are denormalized to obtain the grouting volume prediction value.

2. The intelligent prediction method for grouting amount of a building lifting layer according to claim 1 is characterized in that: The historical engineering data include: initial settlement displacement value of building monitoring point, porosity of lifted soil layer, thickness of lifted soil layer, load pressure value of upper part of lifted soil layer, grouting influence coefficient of surrounding grouting holes and grouting amount required during lifting.

3. The intelligent prediction method for grouting amount of a building lifting layer according to claim 1 is characterized in that: The historical engineering data is normalized and preprocessed to form preprocessed data: The proportional compression method is used to normalize the historical engineering data and convert the data to between 0 and 1. Specifically: Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

4. The intelligent prediction method for grouting amount of a building lifting layer according to claim 1 is characterized in that: The initial neural network model trained by the training set is: The historical engineering data are divided into M groups, and 4M / 5 groups are used as training sets. The results of the input layer and output layer of the neural network are determined as the feature information corresponding to the training set. Then the neural network model is trained to determine the corresponding parameters in the neural network and form an initial neural network model.

5. The intelligent prediction method for grouting amount of a building lifting layer according to claim 4 is characterized in that: The test set is input into the initial neural network model, and then the initial neural network model is optimized according to the output result to form an optimized neural network model, including: The remaining M / 5 groups are used as test sets and input into the initial neural network model. The generalization ability of the initial neural network model is evaluated by the mean square error and the number of convergence steps. The initial neural network model is optimized according to the evaluation results. Different numbers of hidden layer neurons are substituted and the corresponding network mean square error and convergence steps of the output are calculated respectively. The optimal number of hidden layer neurons is determined according to the minimum mean square error and the minimum convergence step, thus forming an optimized neural network model.

6. The intelligent prediction method for grouting amount of a building lifting layer according to any one of claims 1 to 5, characterized in that: The formula for the denormalization is: Where: X is the historical engineering data; X max and X min are the maximum and minimum values ​​of historical engineering data respectively; T is the normalized data; T max and T min They are the maximum and minimum values ​​of the normalized data, which are 1 and 0 respectively.

7. Intelligent prediction system for grouting volume of building lifting layer, characterized in that: include: The historical engineering data collection module collects historical engineering data of completed building grouting reinforcement and lifting projects; The data preprocessing module performs normalization preprocessing on the historical engineering data to form preprocessed data; The model training module divides the preprocessed data into a training set and a test set, and trains the initial neural network model through the training set; The model optimization module inputs the test set into the initial neural network model, and then optimizes the initial neural network model according to the output results to form an optimized neural network model; The grouting volume prediction module inputs the current engineering data into the optimized neural network model, outputs the prediction results, and then denormalizes the prediction results to obtain the grouting volume prediction value.

8. An electronic device, characterized in that The method comprises a processor, a memory and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the method for intelligently predicting the grouting amount of the lifting layer of a building as described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for intelligently predicting the grouting amount of a building lifting layer according to any one of claims 1 to 6 is implemented.

Images