Method and device for predicting settlement trend value of building

Abstract

The invention relates to the technical field of settlement monitoring, in particular to a method for predicting a building settlement trend value, which comprises the following steps of: testing the settlement amount of a building in each period through settlement observation points so as to obtain a period test value in-time ordinal sequence of each settlement observation point; establishing a traditional GM (1, 1) model for the time sequence sequence to predict so as to obtain a time response function; constructing an equal-dimensional metabolism GM (1, 1) model through an iteration strategy; obtaining an optimal dimension according to the development coefficient of the time response function; and according to the isometric metabolism GM (1, 1) model under the optimal dimension, obtaining the ash action amount and the development coefficient, and obtaining the settlement trend value of the building. According to the method, the settlement trend value of the building can be effectively predicted, the defects of subjectivity and randomness of parameter selection in a traditional method are overcome, and reliable technical support is provided for engineering practice. The invention further provides a device for predicting the settlement trend value of the building.

Description

Technical Field

[0001] The present invention relates to the technical field of settlement monitoring, in particular to a method and device for predicting the settlement trend value of a building. Background Art

[0002] Building settlement refers to the overall sinking of a building due to compression and deformation of the foundation soil under the influence of its own weight and external loads. Excessive or uneven building settlement can cause structural damage and even lead to safety accidents. Therefore, accurately predicting building settlement trends is of great engineering significance.

[0003] Currently, building settlement prediction methods primarily fall into two categories: theoretical calculation and data fitting. Theoretical calculation methods, based on soil mechanics principles, utilize a mechanical model of the foundation soil for calculation. However, due to the complexity of geological conditions and the difficulty in obtaining parameters, their prediction accuracy is often unsatisfactory. Data fitting methods, including regression analysis, time series analysis, and grey system theory, rely on measured settlement data to create mathematical models for prediction.

[0004] Grey system theory has been widely used in building settlement prediction due to its adaptability to small sample sizes and incomplete information. The traditional GM(1,1) model is the most fundamental prediction model in grey system theory, but it has limitations when dealing with non-monotonic sequences. To address the nonlinear characteristics of building settlement, researchers have proposed various improved grey prediction models, such as the metabolic GM(1,1) model and the grey Verhulst model.

[0005] However, existing grey prediction models still have the following problems in their application: first, there is a lack of systematic methods for selecting model parameters, which often rely on empirical judgment; second, there is a lack of direct and effective methods for predicting the final settlement trend of buildings. Therefore, a method that can systematically determine model parameters and directly predict the settlement trend of buildings is urgently needed. Summary of the Invention

[0006] In order to solve the above problems, the present invention provides a method and device for predicting the settlement trend value of a building, which can effectively predict the settlement trend value of a building, solves the disadvantages of subjectivity and arbitrariness in parameter selection in traditional methods, and provides reliable technical support for engineering practice.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is:

[0008] A method for predicting a building settlement trend value comprises the following steps:

[0009] S1. Settle settlement observation points on the building, and test the settlement of the building in each cycle through the settlement observation points to obtain a periodic test value of each settlement observation point, and obtain a time series of the settlement of the building according to the periodic test value;

[0010] S2. Establishing a traditional GM(1,1) model for the time series to predict the time response function;

[0011] S3. Constructing an isodimensional metabolic GM(1,1) model through an iterative strategy;

[0012] S4. Obtaining the optimal dimension m* according to the development coefficient of the time response function;

[0013] S5. According to the equal-dimensional metabolism GM(1,1) model under the optimal dimension m*, the ash action amount and the development coefficient are obtained, and the settlement trend value of the building is predicted by the ash action amount and the development coefficient.

[0014] Furthermore, in step S1, the periodic test values ​​of each settlement observation point are x (1) (1), x (1) (2),…x (1) (k),…x (1) (n), where x (1) (k) is the test value in the kth cycle; x (1) (n) is the test value in the nth cycle;

[0015] The time series is X (0) ={x (0) (1),x (0) (2),…,x (0) (k),…,x (0) (n)}.

[0016] Furthermore, in step S2, the time series are accumulated to generate a new series, a grey differential equation is constructed for the new series, and the differential equation is solved to obtain a time response function.

[0017] Furthermore, the new sequence is X (1) ={x (1) (1), x (1) (2), …x (1) (k),…,x (1) (n)};

[0018] The grey differential equation is:

[0019]

[0020] Among them, a is the development coefficient; u is the gray action;

[0021] The least squares method is used to estimate the parameters a and u, and we obtain:

[0022]

[0023] in,

[0024]

[0025]

[0026] Solve the grey differential equation to obtain the time response function:

[0027]

[0028] Among them, a is the development coefficient; u is the gray action; e is the natural constant.

[0029] Furthermore, in step S3, under the condition of keeping the modeling dimension m unchanged, a sliding time window is constructed by removing the oldest data while supplementing the latest observations to update the data set, and the updated data set is reconstructed according to the traditional GM(1,1) prediction model to construct an equal-dimensional metabolic GM(1,1) model.

[0030] Furthermore, in step S4, the method for determining the optimal dimension m* includes the following steps:

[0031] S4.1 Perform a global search in the dimensional space m∈[3,n];

[0032] S4.2 For each dimension m, calculate the corresponding n-(m-1) development coefficients a;

[0033] S4.3 Calculate the ratio of the number of development coefficients q that satisfy |a| ≤ 0.2 to the total number of development coefficients in each dimension k1 = q / [n-(m-1)];

[0034] S4.4 Select the dimension that maximizes k1, and when multiple dimensions have the same k1 value, select the dimension with the smallest average value of all |a| as the optimal dimension m*.

[0035] Furthermore, in step S5, according to the equal-dimensional metabolism GM(1,1) model under the optimal dimension m*, the average value u1 of n-(m*-1) ash action quantities and the average value a1 of the absolute value of the development coefficient are obtained, and the settlement trend value of the building is u1 / a1.

[0036] Furthermore, the average value a1 of the absolute values ​​of the development coefficients is less than 0.2.

[0037] A building settlement trend value prediction device comprises a memory storing an executable trend value prediction algorithm and a central processing unit in communication connection with the memory to execute the method for predicting the building settlement trend value in claim 1.

[0038] The present application has the following advantages:

[0039] The building settlement trend value is predicted based on the settlement amount of the building in each period, avoiding the subjectivity and randomness of parameter selection in the traditional method; the building settlement trend value is directly predicted by the ratio of the grey action amount to the development coefficient through the equal-dimension metabolism GM(1, 1) model, the method is simple and intuitive; the prediction error accumulation problem caused by the outdated data of the traditional grey model is effectively overcome. The method has simple modeling and efficient calculation, and can update the prediction result in real time relying on the periodic detection data, providing dynamic risk early warning for the building management unit, and ensuring the use safety of the building. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 The figure is a flow chart of the method for predicting the building settlement trend value of a preferred embodiment of the present application.

[0041] Figure 2 The figure is a measured settlement graph of the building CJ06 monitoring point.

[0042] Figure 3 The figure is a graph of all a values under 3-14 dimensions

[0043] Figure 4 The figure is a measured settlement graph of a surrounding building of a certain subway station.

[0044] Figure 5 The figure is a graph of a values under 8-19 dimensions. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0046] 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. The terminology used in the description of the present application herein only for the purpose of describing specific embodiments and is not intended to limit the present application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0047] Please also refer to Figures 1 to 5 The method for predicting the building subsidence trend value of a preferred embodiment of the present application comprises the following steps:

[0048] S1. Subsidence observation points are arranged on the building, and the subsidence amount of the building in each period is tested through the subsidence observation points to obtain the period test value of each subsidence observation point, and a time series of the subsidence amount of the building is obtained according to the period test value.

[0049] In step S1, the period test value of each subsidence observation point is x (1) (1), x (1) (2), … x (1) (k), … x (1) (n), wherein x (1) (k) is the test value in the kth period; x (1) (n) is the test value in the nth period.

[0050] The time series is X (0) ={x (0) (1), x (0) (2), …, x (0) (k), …, x (0) (n)}.

[0051] In this embodiment, the subsidence observation points are arranged on the building according to the requirements of the “Code for Measurement of Building Deformation” (JGJ 8), and the subsidence amount of the building in each period is tested. It should be noted that the subsidence test value of each period is the subsidence increment of the period, not the cumulative amount of subsidence.

[0052] S2. A traditional GM (1, 1) model is established for the time series for prediction to obtain a time response function.

[0053] In step S2, the time series is accumulated to generate a new sequence, a grey differential equation is constructed for the new sequence, and the differential equation is solved to obtain the time response function.

[0054] The new sequence is X (1) ={x (1) (1), x (1) (2), …, x (1) (k), …, x (1) (n)}.

[0055] The grey differential equation is:

[0056]

[0057] wherein a is a development coefficient; u is a grey action amount.

[0058] The parameters a and u are estimated by the least square method to obtain:

[0059]

[0060] in,

[0061]

[0062] Solve the grey differential equation to obtain the time response function:

[0063]

[0064] Among them, a is the development coefficient; u is the gray action; e is the natural constant.

[0065] By cumulative reduction, the predicted value is obtained:

[0066]

[0067] S3. Construct an isodimensional metabolic GM(1,1) model through an iterative strategy.

[0068] In step S3, under the condition of keeping the modeling dimension m unchanged, a sliding time window is constructed by removing the oldest data while supplementing the latest observations to update the dataset, and the updated dataset is reconstructed according to the traditional GM(1,1) prediction model to construct an equal-dimensional metabolic GM(1,1) model.

[0069] In this embodiment, the modeling dimension m (3≤m≤n) is kept constant, and a sliding time window is constructed by an iterative strategy of removing the oldest data while supplementing the latest observations. The traditional GM (1,1) prediction model is rebuilt based on the updated data set. Specifically, for a model with a dimension of m, the first m data points {x (0) (1), x (0) (2),…,x (0) (m)} to build a GM(1,1) model, and then use {x (0) (2), x (0) (3),…,x (0) (m+1)} to build a new GM(1,1) model, and so on, until {x (0) (n-m+1), x (0) (n-m+2),…,x (0) (n)} to establish the last GM(1,1) model. In this way, for the equal-dimensional metabolic GM(1,1) model with dimension m, a total of n-(m-1) sub-models can be established.

[0070] S4. Obtain the optimal dimension m* according to the development coefficient of the time response function.

[0071] In step S4, the method for determining the optimal dimension m* includes the following steps:

[0072] S4.1 performing a global search in the dimension space m e [3, n];

[0073] S4.2 for each dimension m, calculating its corresponding n-(m-1) development coefficients a;

[0074] S4.3 calculating the proportion k1 of the number of development coefficients satisfying |a|≤0.2 in the total number under each dimension;

[0075] S4.4 selecting the dimension with the maximum k1, and when multiple dimensions have the same k1 value, selecting the dimension with the minimum average value of |a| as the optimal dimension m*.

[0076] In this embodiment, |a|≤0.2 is an empirical criterion for judging the applicability of the model in the grey system theory, because when |a|≤0.2, the model has high simulation accuracy and is suitable for medium and long term prediction.

[0077] S5. According to the equal-dimension metabolism GM(1, 1) model under the optimal dimension m*, the grey action amount and the development coefficient are obtained, and the settlement trend value of the building is obtained by prediction based on the grey action amount and the development coefficient.

[0078] In step S5, according to the equal-dimension metabolism GM(1, 1) model under the optimal dimension m*, the average value u1 of the n-(m*-1) grey action amounts and the average value a1 of the absolute values of the development coefficients are obtained, and the settlement trend value of the building is u1 / a1. In this embodiment, u1 / a1 represents the steady-state response value of the system, which corresponds to the final trend value in the settlement process of the building.

[0079] In this embodiment, the average value a1 of the absolute values of the development coefficients is less than 0.2.

[0080] This embodiment can be applied to the prediction of building settlement where the settlement amount tends to be stable over time.

[0081] A building settlement trend value prediction device includes a memory storing an executable trend value prediction algorithm and a central processing unit in communication connection with the memory to execute the method for predicting the building settlement trend value in claim 1.

[0082] The method for predicting the building settlement trend value of this embodiment is illustrated by a comprehensive transportation hub building in Yangzhou Yancheng City:

[0083] According to the literature (Jiang Dong, Luo Yayuan, Jin Haibo, et al. Prediction of building settlement based on ARIMA-LSTM-XGBoost combination model [J / OL]. Information Technology for Civil Engineering, 1-6.), the actual settlement monitoring value of monitoring point CJ06 of the comprehensive transportation hub building in Yizheng City, Yangzhou City is as follows: Figure 2 As shown in Figure 2, the settlement data of the measuring point are analyzed.

[0084] Depend on Figure 2 As shown in Table 1, the building settlement decreases gradually in each cycle, showing a trend of gradually stabilizing. According to the method of the present invention, taking 12 dimensions as an example, all a and u values ​​in 12 dimensions are calculated, and the results are shown in Table 1.

[0085] Table 112 a and u values ​​under dimension

[0086] Serial number Start time / d End Time a value u-value Predicted trend value u / a (mm) 1 8 173 0.1261 1.0538 8.36 2 23 188 0.1781 1.1208 6.29 3 38 203 0.1989 1.1039 5.55 Average of absolute values / / 0.1677 1.0928 6.52

[0087] From Table 1, we can find that k1=1, and the average value of the absolute value of all a is 0.1677. According to this calculation, the value of a in all dimensions can be obtained, such as Figure 3 shown.

[0088] according to Figure 3 The optimal dimension of the building settlement prediction model is 14. At this time, k1=1, a1=0.1344<0.2 (and the smallest in all dimensional spaces), u1=1.0794, so the settlement trend value of the building is u1 / a1=8.03mm.

[0089] The method for predicting the building settlement trend value of this embodiment is described by taking the buildings around a subway station as an example:

[0090] According to the literature (Lu Shuyao, Wen Like. Application of improved grey Markov model in building settlement prediction [J]. Surveying and Spatial Geographic Information, 2025, 48(04): 195-198.), 19 consecutive settlement monitoring data of the building settlement monitoring point were selected as the analysis object, and the measured settlement values ​​are shown in Figure 4 .

[0091] Depend on Figure 4 It can be seen that the building's settlement was large in the early stage, and gradually decreased in the later stage, showing a trend of stabilization. According to the method of the present invention, taking 14 dimensions as an example, all a and u values ​​under 14 dimensions are calculated, and the results are shown in Table 2.

[0092] Table 2 a and u values ​​under 14 dimensions

[0093] Serial number Starting period Ending period a value u-value Predicted trend value u / a (mm) 1 1 14 0.0225 0.7067 31.41 2 2 15 0.0277 0.7796 28.14 3 3 16 0.0150 0.6368 42.45 4 4 17 0.0065 0.5371 82.63 5 5 18 0.0438 0.7063 16.13 6 6 19 0.0298 0.6098 20.46 Average of absolute values / / 0.0242 0.6627 27.38

[0094] From Table 2, we can find that k1=1, and the average value of the absolute value of all a is 0.0242. According to this calculation, the value of a in all dimensions can be obtained. For example, Figure 5 The values ​​of a in dimensions 8 to 19 are given.

[0095] according to Figure 5 The optimal dimension of the building settlement prediction model is 14. At this time, k1=1, a1=0.0242<0.2 (and the smallest in all dimensional spaces), u1=0.6627, so u1 / a1=27.38mm. Taking into account the settlement value of 30.50mm in the first period, the final settlement trend value of the building is 27.38mm+30.50mm=57.88mm.

Claims

1. A method for predicting building settlement trend value, characterized in that: The steps include: S1. Settle settlement observation points on the building, and test the settlement of the building in each cycle through the settlement observation points to obtain a periodic test value of each settlement observation point, and obtain a time series of the settlement of the building according to the periodic test value; S2. Establishing a traditional GM(1,1) model for the time series to predict the time response function; S3. Constructing an isodimensional metabolic GM(1,1) model through an iterative strategy; S4. Obtaining the optimal dimension m* according to the development coefficient of the time response function; S5. According to the equal-dimensional metabolism GM(1,1) model under the optimal dimension m*, the ash action amount and the development coefficient are obtained, and the settlement trend value of the building is predicted by the ash action amount and the development coefficient.

2. The method for predicting a building subsidence trend value according to claim 1, wherein: In step S1, the periodic test value of each settlement observation point is x (1) (1), x (1) (2),…x (1) (k),…x (1) (n), where x (1) (k) is the test value in the kth cycle; x (1) (n) is the test value in the nth cycle; The time series is X (0) ={x (0) (1), x(0) (2),…,x (0) (k),…,x (0) (n)}.

3. The method for predicting a building settlement trend value according to claim 2, wherein: In step S2, the time series are accumulated to generate a new series, a grey differential equation is constructed for the new series, and the differential equation is solved to obtain a time response function.

4. The method for predicting building settlement trend value according to claim 3, characterized in that: The new sequence is X (1) ={x (1) (1),x (1) (2),…,x (1) (k),…,x (1) (n)}; The grey differential equation is: Among them, a is the development coefficient; u is the gray action; The least squares method is used to estimate the parameters a and u, and we obtain: in, Solve the grey differential equation to obtain the time response function: Among them, a is the development coefficient; u is the gray action; e is the natural constant.

5. The method for predicting building settlement trend value according to claim 4, characterized in that: In step S3, under the condition of keeping the modeling dimension m unchanged, a sliding time window is constructed by removing the oldest data while supplementing the latest observations to update the data set, and the updated data set is reconstructed according to the traditional GM(1,1) prediction model to construct an equal-dimensional metabolic GM(1,1) model.

6. The method for predicting building settlement trend value according to claim 4, characterized in that: In step S4, the method for determining the optimal dimension m* includes the following steps: S4.1 Perform a global search in the dimensional space m∈[3,n]; S4.2 For each dimension m, calculate the corresponding n-(m-1) development coefficients a; S4.3 Calculate the ratio of the number of development coefficients q that satisfy |a| ≤ 0.2 to the total number of development coefficients in each dimension k1 = q / [n-(m-1)]; S4.4 Select the dimension that maximizes k1, and when multiple dimensions have the same k1 value, select the dimension with the smallest average value of all |a| as the optimal dimension m*.

7. The method for predicting building settlement trend value according to claim 1, characterized in that: In step S5, according to the equal-dimensional metabolism GM(1,1) model under the optimal dimension m*, the average value u1 of n-(m*-1) ash action quantities and the average value a1 of the absolute value of the development coefficient are obtained, and the settlement trend value of the building is u1 / a1.

8. A method for predicting building settlement trend value according to claim 7, characterized in that: The average value a1 of the absolute value of the development coefficient is less than 0.

2.

9. A device for predicting building settlement trend value, characterized in that: It includes a memory for storing an executable trend value prediction algorithm and a central processing unit communicating with the memory to execute the method for predicting the building settlement trend value in claim 1.

Images