Bin lattice length determination method for mass concrete skip-bay construction

Abstract

The invention discloses a lattice length determination method for mass concrete skip method construction, which adopts a big data analysis mode, combines the actual condition of a construction site, and defines various indexes such as the raw material composition of mass concrete and the overall specification of the mass concrete to determine the lattice length of the mass concrete. The method comprises the following steps: establishing a database of trial mass concrete skip method construction, then performing analysis and calculation by utilizing a machine learning calculation model, finding out influence indexes, inputting mass concrete related indexes in a to-be-constructed project in an established graphical user interface, and adjusting the influence indexes according to an output result, so as to realize the determination of the optimal bin length. According to the length determination method, the actual property of the mass concrete is comprehensively analyzed, the calculation result is accurate, the calculation process is low in cost, and the method has important actual application value.

Description

Technical Field

[0001] The invention belongs to the field of concrete construction, and in particular to a method for determining the cell length of large-volume concrete in skip-cell construction. Background Art

[0002] Concrete is one of the most widely used materials in construction projects. In some projects, it is usually necessary to pour very large volumes of concrete, such as the foundation slabs of residential buildings, terminal slabs, and long-span bridge foundations. This type of concrete is called mass concrete. Concrete releases a lot of heat during the hydration process, which in turn affects the stress distribution inside the concrete and causes concrete cracking. This phenomenon is particularly significant for mass concrete.

[0003] At present, in order to improve the temperature rise cracking problem of large-volume concrete, post-casting strips are usually used to avoid cracking. However, the use of post-casting strips will affect the integrity of the construction, the process is relatively complicated and the effect is not ideal. Therefore, some scholars have proposed the use of the skipping method for large-volume concrete. The skipping method does not require the setting of post-casting strips, but divides the large-volume concrete into multiple bins. During construction, non-adjacent bins are poured simultaneously. After a certain period of curing, the concrete of adjacent bins is poured. The principle of the skipping method is to release the internal temperature of the large-volume concrete to form a time difference, thereby regulating the "resistance" and "release" effects of the internal stress of the concrete. In the skipping method construction, a good bin division length can ensure the smooth pouring of concrete in each bin and make the concrete performance superior. At present, when dividing the bins, the corresponding bin length calculation formula given in the "Technical Specifications for the Skipping Method of Super-Long Concrete Structures" is usually used, and the division is performed by calculating the temperature cracking index with reference to the actual plot characteristics.

[0004] However, there are still some problems when applying this method to divide the warehouse grid at the current stage: 1. Although the accuracy of the compartment division can be guaranteed by calculating the temperature cracking index, it requires multiple tests and analysis in the early stage, which requires corresponding test sites, the production of corresponding test blocks, and the acquisition of more technical parameters, which consumes corresponding manpower and material resources.

[0005] 2. The existing classification method does not take into account the characteristics of concrete raw materials that have the greatest impact on the cracking of mass concrete.

[0006] 3. When the actual construction length of the cell in an existing project exceeds the length given in the specification, there is a low-cost prediction that the large-volume concrete of the corresponding project will meet the requirements of the most suitable cell length under the specification.

[0007] Therefore, it is of great significance to seek a convenient, fast and highly feasible method to determine the length of the bin. Summary of the invention

[0008] The purpose of the present invention is to provide a method for determining the cell length of large-volume concrete in the skip-cell construction method, to solve the problem that the existing method for determining the cell length of the skip-cell method requires the calculation of temperature cracking indicators and multiple tests and analyses, which consumes a lot of manpower and material resources, and solves the problem of not considering the characteristics of concrete raw materials, and also solves the technical problem of predicting the available cell length at a lower cost.

[0009] To achieve the above object, the present invention adopts the following technical solution: A method for determining the length of a cell in a large-volume concrete skipping method construction, the steps of which are as follows: Step 1: Collect characteristic index data of existing large-volume concrete engineering plots, including raw material performance index of concrete in each plot, overall specification index of concrete in each plot, and actual construction cell length of each plot; Step 2, removing the data in the non-adaptive interval of the characteristic index in step 1, and then establishing a database of the characteristic index; Step 3: Based on the data in the database, the machine learning calculation model with optimized parameters is used for analysis to determine the final feasibility model; the raw material performance index of concrete in each block and the overall specification index of concrete in each block are input parameters, and the actual construction cell length of each block is output parameter; Step 4: Based on the final feasibility model, the SHAP method is used to analyze the influence of each input parameter on the output parameter, and the characteristic index that produces the influence is determined as the influence index; Step 5: Build a graphical user interface based on the final feasibility model, collect the characteristic index data of the large-volume concrete engineering plot to be constructed, input the data into the graphical user interface and output the predicted bin length L. out ; Step 6: Verify the predicted bin length L out Whether the maximum cell length L' required by the specification for the large-volume concrete project site to be constructed is met: Case 1: When L out ≤L', then the construction adopts L out As the cell length; Case 2: When L out >L', If the concrete mixing step has been completed, L' is used as the cell length during construction; If the concrete mixing step has not yet been carried out, the value of the influencing index is adjusted within the adaptability range and the cell length L is output again on the graphical user interface. new , adjust until the output L new ≤L', the concrete is mixed using the adjusted characteristic index data, and the construction is carried out using Lnew As the cell length.

[0010] The applicable concrete strength grade range in the database is 30MPa to 50MPa, the construction temperature range is 15℃ to 25℃, and the construction relative humidity range is 60% to 80%.

[0011] The performance indicators of raw materials for concrete include water-cement ratio, cement content, fly ash ratio, mineral powder ratio, fine aggregate ratio, coarse aggregate ratio and water reducer content. The overall specification indicators of concrete include the volume and pouring thickness of concrete.

[0012] The machine learning computational model uses the extreme gradient boosting model; The parameter optimization method of the extreme gradient boosting model adopts the particle swarm optimization algorithm to obtain the objective function of the final feasibility model; a graphical user interface is constructed based on the objective function, and the construction is established using the Tkinter module in Python software.

[0013] The influencing indicators are ranked from large to small in order of influence: fly ash content, concrete volume, water reducer content, water-cement ratio, mineral powder content, cement content and casting thickness.

[0014] In step six, a single influencing index, namely the fly ash content with the greatest impact, is adjusted. The fly ash content is positively correlated with the cell length, that is, the greater the fly ash content, the longer the cell length of the concrete.

[0015] In step six, two or more influencing indicators are adjusted in the order of influence from large to small.

[0016] The adaptability range of each characteristic indicator is as follows: The water-cement ratio is between 0.3 and 0.8, and the cement content is between 180 m 3 ~360m 3 The fly ash ratio is between 10% and 60%, the mineral powder ratio is between 0% and 40%, the fine aggregate to cement ratio is between 2 and 4.5, the coarse aggregate to cement ratio is between 3 and 6, the water reducer ratio is between 0 and 12 kg / m3, and the concrete volume is between 5000 m 3 ~15000m 3 The pouring thickness range is 0.4m~1.8m.

[0017] Compared with the prior art, the present invention has the following characteristics and beneficial effects: The present invention abandons the calculation idea of ​​using temperature cracking index to divide the cell in a single project, collects relevant data information of existing completed engineering examples of skip-cell construction, extracts nine material performance indicators of concrete itself and corresponding cell length in each large-volume concrete skip-cell construction plot as characteristic indicators, and establishes an information database of skip-cell construction; at the same time, a machine learning calculation model with optimized parameters is used to analyze the data in the database to obtain a final feasibility model, and the SHAP method is used based on the final feasibility model to analyze the influence of each characteristic indicator on the cell length, and the characteristic indicator that has an influence is determined as an influence indicator; then based on the final feasibility model, a graphical user interface is established using the Tkinter module in the Python algorithm. Finally, the characteristic indicators of the large-volume concrete in the project to be constructed are input into the graphical user interface, and the cell length can be output, and the most suitable cell length suitable for this engineering project can be obtained by adjusting each characteristic indicator.

[0018] The present invention can avoid most of the tests involved in calculating the risk of temperature cracking, significantly reducing the testing cost. When determining the timing of the cell length, the present invention fully considers the influence of factors such as cement, fly ash, mineral powder, aggregate, etc. on the cell length division, so as to adjust the proportion in a targeted manner according to the needs later. The present invention can predict the most suitable cell length for the construction of large-volume concrete skipping method according to the characteristics of different plots, assist in construction, and promote the engineering structure to achieve the best quality effect.

[0019] The present invention can comprehensively analyze the performance of relevant raw materials and concrete specifications involved in concrete, provide the most suitable cell length for the construction site, and ensure the best quality of large-volume concrete. The method is easy to implement and low in cost, and has important practical engineering promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0021] Figure 1 It is a heat map of the correlation between the characteristic indicators of the present invention.

[0022] Figure 2 It is a ranking diagram of the influence of each characteristic index on the prediction result of the present invention.

[0023] Figure 3 It is an influence diagram of a single characteristic index of the present invention on the prediction result.

[0024] Figure 4 It is the SHAP value development result of each characteristic index of the present invention.

[0025] Figure 5 It is a schematic diagram of a graphical user interface according to the first embodiment of the present invention.

[0026] Figure 6 It is a schematic diagram of a graphical user interface according to the second embodiment of the present invention.

[0027] Figure 7 This is a schematic diagram of a graphical user interface of the third embodiment of the present invention. Figure 1 .

[0028] Figure 8 This is a schematic diagram of a graphical user interface of the third embodiment of the present invention. Figure 2 . DETAILED DESCRIPTION

[0029] At the current stage, machine learning methods can quickly extract target parameters through big data analysis, thereby achieving the purpose of performance prediction and optimization. Therefore, using machine learning methods to study the division of compartments in skip-compartment construction is of great significance for improving work efficiency and guiding construction.

[0030] Construction of database: Before machine learning, a qualified database needs to be established. First, the selection and statistics of indicators in the database: There are many factors that affect the construction of large-volume concrete by skipping bin method. When dividing the bins for skipping bin method construction, the inherent performance of concrete materials has an important influence. The factors affecting the inherent performance of concrete materials are mainly cementitious materials, water-cement ratio and aggregates. Therefore, when the present invention sets and selects indicators, the first characteristic indicators to be considered are water-cement ratio (w_c) and cement content (ccon). In large-volume concrete, fly ash is usually added as a cementitious material to reduce the heat of hydration, and sometimes mineral powder is added as a compensating cementitious material for later strength. Therefore, fly ash content (fmh) and mineral powder content (kf) are also used as characteristic indicators. In concrete, according to the difference in particle size, aggregates can be divided into fine aggregates and coarse aggregates, both of which are also important components of concrete. Therefore, the present invention uses the mass ratio of fine aggregate to cement (X / c) and the mass ratio of coarse aggregate to cement (C / c) as characteristic indicators. Next is the admixture in concrete. This invention only considers water reducer as an admixture, so the water reducer content (jsj) is set as the characteristic index. Finally, for large-volume concrete, thickness and volume are also important indicators. Therefore, this study also sets thickness (h) and volume (tj) as characteristic indicators. At the same time, this invention also takes the actual cell length (L) that meets the specifications in the skip-cell method construction as a characteristic index.

[0031] Secondly, for the data in the database, in addition to statistically analyzing the mathematical features, it is also necessary to statistically analyze the correlation between various parameters. If there is an obvious linear relationship between the parameter indicators, the prediction accuracy of the machine learning model will be reduced. The present invention uses the Pearson correlation coefficient as a basis for judgment to evaluate the relationship between parameter indicators. The Pearson correlation coefficient is abbreviated as PCC, and the calculation formula of PCC is shown in formula (1). In the formula, i represents the data sample; N represents the number of data samples; y represents the correlation coefficient between the parameters.i and i 'represent the true value and the predicted value respectively; and Represent the mean of the true value and the predicted value respectively. According to the PCC judgment principle, the P value of the two characteristic indicators can be between -1 and 1. If the P value is closer to 0, it means that the correlation between the two characteristic indicators is weaker.

[0032] By calculating and analyzing the PCC value, we can get the correlation heat map between the characteristic indicators. Figure 1 As shown in the figure, it can be seen that most of the P values ​​between the characteristic indicators are between -0.4 and 0.4, which indicates that the linear relationship between the characteristic indicators is weak, so the impact on the prediction effect of the machine learning model is small, ensuring the smooth construction of the model.

[0033] The present invention adopts the currently widely used extreme gradient boosting (XGB) model as the ensemble learning algorithm. The ensemble learning model uses a decision tree as the base learner. When using this model for calculation and analysis, the objective function expression is represents the objective function of the XGB model, L represents the loss function, y x represents the actual value of the xth sample, u x Represents the xth input variable, Ω represents the regularization term, and M represents a constant. When predicting the length of the bin, it is necessary to optimize the parameters of the machine learning model. For the XGB model, the optimization method used is the particle swarm optimization method (PSO). The main adjustment parameters corresponding to different databases using this method include four, and the initial control range needs to be set to: eta value (learning rate of the machine learning model) is 0.02-0.8; gamma value (the minimum loss function drop required for the leaf node in the machine learning model to continue splitting) is 0.03-0.9; max_depth value (the maximum depth of the decision tree in the machine learning model) is 2-36; n_estimators value (the number of training decision trees in the machine learning model) is 12-680.

[0034] Global feature index analysis: In order to explain the prediction results of the machine learning model, the present invention uses Shapley additive explanation (SHAP). The SHAP method is based on the cooperative game theory and the SHAP value is used as a quantitative evaluation index. The SHAP value of a feature is calculated by the sum of the effects of the feature combined with other features on the regression results, and its calculation method is shown in formula (2).

[0035] Among them, φj(f, x) represents the SHAP value corresponding to feature J in the machine learning model of type f, and x represents the sample explained by the SHAP method. J represents the feature set, and S is the feature set that does not contain feature j.

[0036] The SHAP analysis method calculates the contribution of each feature indicator to the model prediction results by permuting and combining the feature values, thereby explaining the model's prediction results.

[0037] Using the XGB model, the SHAP analysis is used to comprehensively determine the impact of each feature index on the prediction results. The first is the impact of the feature index on the bin length. The results are shown in Figure 2 and Figure 3 As shown. Figure 2 It can be seen that the fly ash content has the greatest impact on the prediction results, followed by the volume of mass concrete, the water-reducing agent content, the water-cement ratio, the mineral powder content, the cement content and the thickness. The influence of coarse aggregate and fine aggregate can be ignored. For mass concrete, the main performance to be considered is the temperature effect that affects cracking. Fly ash has been proven to have a low heat release during the hydration process and can well regulate the release rate of cement hydration heat. Therefore, when dividing the length of the cell of mass concrete, the first consideration is the fly ash content. The second influencing factor is the volume of mass concrete. This is because the thickness of mass concrete changes little during the base plate construction process. The volume index is mainly controlled by the length and width. Therefore, the overall length and width directly affect the length of the cell. Then there are parameters such as water-reducing agent and water-cement ratio. These indicators mainly affect the flow properties and strength of concrete, which in turn affect the construction and maintenance of concrete foundations. As for the proportion of aggregate, its influence is small because it does not participate in the hydration heat effect of cementitious materials in large-volume concrete, and therefore does not affect the length division.

[0038] See also Figure 3 As shown in the figure, for each characteristic index, its influence on the prediction result of the cell length is positively correlated in the red part and negatively correlated in the blue part as the SHAP value increases. As can be seen from the figure, except for individual difference points, as the SHAP value increases. For example, the relationship between the fly ash content (fmh) and the cell length is a positively correlated influence trend, that is, when the fly ash content is larger, the cell length of the concrete is longer.

[0039] Single feature indicator analysis: see Figure 4 The figure shows the SHAP value development results of each characteristic index. Each dot in the figure represents a sample data point corresponding to the characteristic index. Figure 4It can be seen that for each characteristic indicator, as the value changes, the distribution of SHAP values ​​will show a certain trend of change, which is related to the division of bin length. Figure 4 (a) is the relationship between the SHAP value of fly ash content (fmh) and the cell length. Figure 4 (b) is the relationship between the SHAP value of volume (tj) and the length of the bin. Figure 4 (c) is the relationship between the SHAP value of the water reducer content (jsj) and the cell length. Figure 4 (d) is the relationship between the SHAP value of water-cement ratio (w_c) and the length of the bin. Figure 4 (e) is the relationship between the SHAP value of the mineral powder dosage (kf) and the cell length. Figure 4 (f) is the relationship between the SHAP value of cement content (ccon) and the cell length. Figure 4 (j) is the relationship between the SHAP value of thickness (h) and the length of the cell. Figure 4 (h) is the relationship between the SHAP value of the mass ratio of coarse aggregate to cement (C / c) and the cell length. Figure 4 (i) is the relationship between the SHAP value of the mass ratio of fine aggregate to cement (X / c) and the cell length.

[0040] For example, for Figure 4 In the relationship diagram of the SHAP value of fly ash content (fmh) and the cell length in (a), it can be seen that except for the two difference points where the fly ash content is around 15%, the SHAP values ​​of the remaining data points all show an approximately increasing trend with the increase of fly ash content, which indicates that with the increase of fly ash content, the cell length tends to increase. The change trends of other characteristic parameters that can affect the cell length can be compared with fly ash analysis. When the change trend of characteristic indicators is not obvious or the law is not clear, the amount of data in the database can be increased to make the model analysis more comprehensive and the model operation more instructive.

[0041] The Tkinter module in Python was used to build the graphical user interface to establish a GUI for predicting the most suitable bin length in the skip bin construction. Figure 5-Figure 8 As shown, the GUI includes two modules, an input module and an output module. After inputting the characteristic index at the input module, click "calculate" to make a prediction, and the predicted cell length of the group of concrete can be obtained.

[0042] Embodiment 1: Based on the construction of a large concrete basement exterior wall in a housing construction project, the present invention is used to determine the most suitable cell length for the skipping cell construction of the project. The project site covers an area of ​​6575m 2 , the pouring thickness of concrete is 0.4m. The strength of concrete is 35MPa.

[0043] The specific construction steps are as follows: Step 1: Collect data on characteristic indicators of existing large-volume concrete engineering plots, generally collect relevant data information on engineering examples of skip-bin construction reported in existing literature. The characteristic indicators include the raw material performance indicators of concrete in each plot, the overall specification indicators of concrete in each plot, and the actual construction bin length of each plot that meets the specifications; the raw material performance indicators of concrete include water-cement ratio, cement content, fly ash ratio, mineral powder ratio, fine aggregate ratio, coarse aggregate ratio and water reducer content, and the overall specification indicators of concrete include the volume and pouring thickness of concrete.

[0044] Step 2, remove the data in the non-adaptive range of the characteristic indicators in step 1, and then establish a database of characteristic indicators; the applicable concrete strength grade range in the database is 30MPa~50MPa, the construction temperature condition range is 15℃~25℃, and the construction relative humidity condition range is 60%~80%.

[0045] Step 3: Based on the data in the database, the extreme gradient boosting model is selected as the model for calculation application, and then the particle swarm optimization method is used to optimize the parameters of the database. At the same time, the optimal parameter values ​​after particle swarm optimization are: eta value is 0.1; gamma value is 0.6; max_depth value is 12; n_estimators value is 240. On the basis of determining the optimal parameters, the extreme gradient model is verified and analyzed to determine the final feasibility model and obtain the objective function of the final feasibility model. Among them, the raw material performance indicators of concrete in each block and the overall specification indicators of concrete in each block are input parameters, and the actual construction cell length of each block that meets the specifications is the output parameter.

[0046] Step 4: Based on the final feasibility model, the SHAP method is used to analyze the influence of each input parameter on the output parameter, and the characteristic indicators that have an impact are determined as the influencing indicators; the influencing indicators are ranked from large to small in order of influence: fly ash content, concrete volume, water reducer content, water-cement ratio, mineral powder content, cement content and casting thickness.

[0047] Step 5: Build a graphical user interface based on the final feasibility model, collect the characteristic index data of the large-volume concrete engineering plot to be constructed, input the data into the graphical user interface and output the predicted bin length L. out .

[0048] The detailed information of the mass concrete during construction in this embodiment mainly includes environmental conditions and characteristic indicators. The on-site temperature is 22°C and the relative humidity is 76%. The detailed information of the raw material components of the concrete is as follows: the water-cement ratio of the concrete is 0.42, and the cement content is 224kg / m 3 The ratio of fly ash to cement is 41%, the ratio of mineral powder is 0%, the ratio of fine aggregate to cement is 3.3, the ratio of coarse aggregate to cement is 4.77, and the ratio of water reducer is 7.7kg / m 3 The calculated volume of the mass concrete is 3945m 3 .

[0049] Step 6: Verify the predicted bin length L out Whether the maximum cell length L' required by the specification for the large-volume concrete project site to be constructed is met: Case 1: When L out ≤L', then the construction adopts L out As the cell length; Case 2: When L out >L', If the concrete mixing step has been completed, L' is used as the cell length during construction; If the concrete mixing step has not yet been carried out, the value of the influencing index is adjusted within the adaptability range and the cell length L is output again on the graphical user interface. new , adjust until the output L new ≤L', the concrete is mixed using the adjusted characteristic index data, and the construction is carried out using L new As the cell length.

[0050] In step six, a single influencing index, namely the fly ash content with the greatest impact, is adjusted. The fly ash content is positively correlated with the cell length, that is, the greater the fly ash content, the longer the cell length of the concrete.

[0051] In step six, two or more influencing indicators are adjusted in the order of influence from large to small.

[0052] The adaptability ranges of each characteristic index are as follows: water-cement ratio range is 0.3-0.8, cement content range is 180m 3 ~360kg / m 3 The fly ash ratio is 10% to 60%, the mineral powder ratio is 0% to 40%, the ratio of fine aggregate to cement is 2 to 4.5, the ratio of coarse aggregate to cement is 3 to 6, and the water reducer ratio is 0 to 12 kg / m 3 , the volume interval of concrete is 5000m 3 ~15000m3 The pouring thickness range is 0.4m~1.8m.

[0053] This embodiment sees Figure 5 As shown, substitute the characteristic index into the input parameter part of the graphical user interface and click the prediction button to get the output bin length L out =46.55m, compared with the recommended value L'=50m for the maximum cell length of large-volume concrete in wall structures in the "DB11T 1200-2023 Technical Code for Skipping Method of Super-Long Volume Concrete Structures", it meets the maximum value limit specified in the code, that is, L out <L', so the most suitable cell length for this project is 46.55m.

[0054] Therefore, as long as the cell length is less than or equal to this value when dividing the cells, the concrete in each cell can perform well without cracking. In addition, this project monitored the crack condition of the mass concrete during the curing period and the later maintenance stage, and found that the concrete in each cell was in good condition during the above period, without cracks and other undesirable conditions.

[0055] Embodiment 2: Based on the construction of a large concrete foundation slab of a housing construction project, the present invention is used to determine the most suitable cell length for the skipping cell construction of the project. The project site covers an area of ​​8560m 2 , the pouring thickness of concrete is 0.6m. The strength of concrete is 40MPa.

[0056] Different from the first embodiment, the optimal parameter values ​​in the model after particle swarm optimization are: eta value is 0.2; gamma value is 0.7; max_depth value is 6; n_estimators value is 148.

[0057] The detailed information of the mass concrete during construction in this embodiment mainly includes environmental conditions and characteristic indicators. The on-site temperature is 17°C and the relative humidity is 68%. The detailed information of the raw material components of the concrete is as follows: the water-cement ratio of the concrete is 0.46, and the cement content is 185kg / m 3 The ratio of fly ash to cement is 35%, the ratio of mineral powder is 10%, the ratio of fine aggregate to cement is 3.1, the ratio of coarse aggregate to cement is 4.5, and the ratio of water reducer is 8.4kg / m 3 The calculated volume of the mass concrete is 5260m 3 .

[0058] See also Figure 6As shown, substitute the above characteristic index into the input parameter part of the graphical user interface and click the prediction button to obtain the output bin length L out =42.74m, compared with the recommended value L'=40m for the maximum cell length of large-volume concrete of the foundation slab in the "DB11T 1200-2023 Technical Code for Skipping Method of Super-Long Concrete Structures", it exceeds the maximum length of 40m specified in the code. In this embodiment, the concrete mixing step of the construction organization process has been completed at this time, so the input index is not adjusted, and L'=40m can be used as the cell length for construction.

[0059] Embodiment three: Based on the large-volume concrete construction of a certain infrastructure construction project, the present invention is used to determine the most suitable cell length for the skip-cell method of construction of the project. The project site covers an area of ​​19,600 m 2 , the pouring thickness of concrete is 0.8m. The strength of concrete is 45MPa.

[0060] Different from the first and second embodiments, the optimal parameter values ​​in the model after particle swarm optimization are: eta value is 0.4; gamma value is 0.4; max_depth value is 22; n_estimators value is 425.

[0061] The detailed information of the mass concrete during construction in this embodiment mainly includes environmental conditions and characteristic indicators. The on-site temperature is 24°C and the relative humidity is 80%. The detailed information of the raw material components of the concrete is as follows: the water-cement ratio of the concrete is 0.36, and the cement content is 232kg / m 3 The ratio of fly ash to cement is 48%, the ratio of mineral powder is 15%, the ratio of fine aggregate to cement is 3.05, the ratio of coarse aggregate to cement is 4.8, and the ratio of water reducer is 10.4kg / m 3 The calculated volume of the mass concrete is 13460m 3 .

[0062] See also Figure 7 As shown, substitute the above characteristic index into the input parameter part of the graphical user interface and click the prediction button to obtain the output bin length L out =41.15m, compared with the recommended value L'=40m for the maximum cell length of mass concrete in the foundation in "DB11T 1200-2023 Technical Code for Skipping Method of Super-Long Mass Concrete Structures", it does not meet the maximum value limit specified in the code.

[0063] In this embodiment, the concrete mixing step of the construction organization process has not yet been carried out, so the value of the influencing index is adjusted within the adaptability range, and the cell length L is output again on the graphical user interface. new , adjust until the output L new ≤L', the concrete is mixed using the adjusted characteristic index data, and the construction is carried out using L new As the cell length.

[0064] In this embodiment, the concrete mix ratio can be adjusted according to the actual engineering needs in combination with the mixing station work, and the fly ash dosage with the greatest impact is adjusted. Figure 8 As shown in the figure, when the fly ash content is adjusted to 32%, the cell length L is re-output. new =38.46m<L'=40m, which meets the maximum length range in the specification. Therefore, the actual cell length value of this project is adjusted to 38.46m, which is the most suitable cell length. At the same time, concrete is deployed using the adjusted characteristic index data.

Claims

1. A method for determining the length of a cell in a large-volume concrete skipping method construction, characterized in that: The steps to determine are as follows: Step 1: Collect characteristic index data of existing large-volume concrete engineering plots, including raw material performance index of concrete in each plot, overall specification index of concrete in each plot, and actual construction cell length of each plot; Step 2, removing the data in the non-adaptive interval of the characteristic index in step 1, and then establishing a database of the characteristic index; Step 3: Based on the data in the database, the machine learning calculation model with optimized parameters is used for analysis to determine the final feasibility model; the raw material performance index of concrete in each block and the overall specification index of concrete in each block are input parameters, and the actual construction cell length of each block is output parameter; Step 4: Based on the final feasibility model, the SHAP method is used to analyze the influence of each input parameter on the output parameter, and the characteristic index that produces the influence is determined as the influence index; Step 5: Build a graphical user interface based on the final feasibility model, collect the characteristic index data of the large-volume concrete engineering plot to be constructed, input the data into the graphical user interface and output the predicted bin length L. out ; Step 6: Verify the predicted bin length L out Whether the maximum cell length L' required by the specification for the large-volume concrete project site to be constructed is met: Case 1: When L out ≤L', then the construction adopts L out As the cell length; Case 2: When L out >L', If the concrete mixing step has been completed, L' is used as the cell length during construction; If the concrete mixing step has not yet been carried out, the value of the influencing index is adjusted within the adaptability range and the cell length L is output again on the graphical user interface. new , adjust until the output L new ≤L', the concrete is mixed using the adjusted characteristic index data, and the construction is carried out using L new As the cell length.

2. The method for determining the cell length of large-volume concrete skipping cell construction according to claim 1 is characterized in that: The applicable concrete strength grade range in the database is 30MPa to 50MPa, the construction temperature range is 15℃ to 25℃, and the construction relative humidity range is 60% to 80%.

3. The method for determining the cell length of the large-volume concrete skipping cell construction method according to claim 1 is characterized by: The performance indicators of raw materials for concrete include water-cement ratio, cement content, fly ash ratio, mineral powder ratio, fine aggregate ratio, coarse aggregate ratio and water reducer content. The overall specification indicators of concrete include the volume and pouring thickness of concrete.

4. The method for determining the cell length of large-volume concrete skipping cell construction according to claim 1 is characterized in that: The machine learning computational model uses the extreme gradient boosting model; The parameter optimization method of the extreme gradient boosting model uses the particle swarm optimization algorithm to obtain the objective function of the final feasibility model; A graphical user interface is constructed based on the objective function, and the construction is established using the Tkinter module in Python software.

5. The method for determining the cell length of large volume concrete skipping cell construction according to claim 3 or 4, characterized in that: The influencing indicators are ranked from large to small in order of influence: fly ash content, concrete volume, water reducer content, water-cement ratio, mineral powder content, cement content and casting thickness.

6. The method for determining the cell length of large-volume concrete skipping cell construction according to claim 5 is characterized in that: In step six, a single influencing index, namely the fly ash content with the greatest impact, is adjusted. The fly ash content is positively correlated with the cell length, that is, the greater the fly ash content, the longer the cell length of the concrete.

7. The method for determining the cell length of large-volume concrete skipping cell construction according to claim 5 is characterized in that: In step six, two or more influencing indicators are adjusted in the order of influence from large to small.

8. The method for determining the cell length of large-volume concrete skipping cell construction according to claim 6 or 7, characterized in that: The adaptability range of each characteristic index is as follows: The water-cement ratio is between 0.3 and 0.8, and the cement content is between 180 m 3 ~360m 3 The fly ash ratio is between 10% and 60%, the mineral powder ratio is between 0% and 40%, the fine aggregate to cement ratio is between 2 and 4.5, the coarse aggregate to cement ratio is between 3 and 6, the water reducer ratio is between 0 and 12 kg / m3, and the concrete volume is between 5000 m 3 ~15000m 3 The pouring thickness range is 0.4m~1.8m.

Images