A machine learning-based method, apparatus, and equipment for quality control of self-compacting concrete construction.

Abstract

This application relates to the field of big data technology and provides a method, apparatus, and equipment for quality control of self-compacting concrete construction based on machine learning. The method includes: acquiring environmental information corresponding to a target scenario; generating a pouring simulation model for self-compacting concrete based on the environmental information; importing preset construction behavior data into the pouring simulation model; and performing behavior control on the construction behavior data to obtain controlled behavior data. This method can improve the accuracy of performance prediction and, when performance data does not meet performance indicators, perform behavior control on the desired construction behavior data, thereby improving the accuracy of construction behavior and the pouring quality of self-compacting concrete.

Description

Technical Field

[0001] This application belongs to the field of big data technology, and in particular relates to a method, apparatus and equipment for quality control of self-compacting concrete construction based on machine learning. Background Technology

[0002] Concrete, as one of the main building materials in modern infrastructure and housing construction, directly impacts the quality of buildings by ensuring its performance meets construction requirements. For self-compacting concrete, its actual performance fluctuates significantly due to factors such as construction conditions and raw material price variations. Current self-compacting concrete application techniques generally require post-pouring performance testing to determine if it meets quality requirements, making it difficult to guarantee construction quality. When performance indicators fail to meet quality requirements, demolition, reinforcement, and re-inspection are often necessary, greatly extending the construction time.

[0003] It is evident that existing technologies for using self-compacting concrete cannot accurately predict its performance indicators in different pouring scenarios, making it difficult to guarantee construction quality and resulting in significant fluctuations in the building construction cycle. Summary of the Invention

[0004] This application provides a machine learning-based method for controlling the construction quality of self-compacting concrete and an electronic device. This method can solve the problem that existing self-compacting concrete technologies cannot accurately determine the performance indicators of the concrete in different pouring scenarios, making it difficult to guarantee construction quality and resulting in large fluctuations in the building construction cycle.

[0005] In a first aspect, embodiments of this application provide a machine learning-based method for controlling the construction quality of self-compacting concrete, including:

[0006] Obtain environmental information corresponding to the target scene; the environmental information includes: the dimensions of the pouring space and the scene type;

[0007] Based on the environmental information, a pouring simulation model for self-compacting concrete is generated; the pouring simulation model is obtained by machine learning training based on preset historical valid data; the historical valid data is obtained after data cleaning of all historical project data.

[0008] The preset construction behavior data is imported into the pouring simulation model, and the construction behavior data is subjected to behavior control to obtain control behavior data; the control behavior data is the behavior data that makes the performance data of self-compacting concrete meet the target scenario to the corresponding performance index.

[0009] In one possible implementation of the first aspect, the step of importing preset construction behavior data into the pouring simulation model and performing behavior control on the construction behavior data to obtain control behavior data includes:

[0010] Based on the construction behavior data, a simulated construction object is created in the pouring simulation model; the construction behavior data includes: at least one construction device and at least one construction worker;

[0011] Based on the construction behavior data, multiple construction steps are determined, and the simulated construction object is sequentially controlled to execute each of the construction steps in the pouring simulation model;

[0012] During the execution of the construction steps, the state data of the simulated concrete in the pouring simulation model is determined in real time; the state data includes: mechanical performance parameters, durability performance parameters, and carbon emission parameters.

[0013] If any state data at any time is detected to fail to meet the performance index, then behavior control data is generated based on the performance deviation corresponding to the state data and the associated construction steps.

[0014] The construction behavior data is adjusted based on the behavior control data, and the operation of creating a simulated construction object in the pouring simulation model based on the adjusted construction behavior data is returned to be executed until all the construction steps are completed and the state data meets the performance index.

[0015] In one possible implementation of the first aspect, determining the state data of the simulated concrete in the pouring simulation model in real time during the execution of the construction steps includes:

[0016] During the execution of the construction steps, real-time data of the simulated concrete in the pouring simulation model is collected.

[0017] Based on the scenario type, determine the scenario influencing factors corresponding to the state data; the scenario type includes: ambient temperature, ambient humidity, and location slope;

[0018] Based on the real-time data and the scene influencing factors, a state prediction curve corresponding to the simulated concrete is constructed.

[0019] Based on the predicted completion time corresponding to the completion of all the aforementioned construction steps, the state data is determined in the state prediction curve; the state data is the data corresponding to the predicted completion time in the state prediction curve.

[0020] Correspondingly, if any state data at any time is detected to fail to meet the performance index, then behavior control data is generated based on the performance deviation corresponding to the state data and the associated construction steps, including:

[0021] Calculate the area of ​​the curve deviation between the predicted state curve and the standard state curve;

[0022] The deviation index corresponding to the state data is calculated based on the weight value corresponding to each time zone in the deviation area of ​​the curve; the weight value is determined based on the time interval length between each time zone and the execution time of the current construction behavior.

[0023] If the deviation index is greater than the preset deviation threshold, then behavior control data is generated based on the performance deviation corresponding to any of the state data and the associated construction steps.

[0024] In one possible implementation of the first aspect, before generating the pouring simulation model corresponding to the self-compacting concrete based on the environmental information, the method further includes:

[0025] Obtain the historical valid data and the historical prediction data corresponding to each historical valid data; match the historical environment corresponding to each historical valid data with the environmental information;

[0026] Based on the data deviation corresponding to each of the historical valid data points and the historical predicted data points, determine the data confidence level of each of the historical valid data points;

[0027] Based on multiple preset data feature dimensions, determine the historical feature indicators corresponding to each of the historical valid data;

[0028] Based on the historical feature indicators and data confidence levels corresponding to each of the historical valid data, calculate the indicator deviation coefficient corresponding to each data feature dimension.

[0029] Based on the index deviation coefficient, the key feature dimension corresponding to the environmental information is determined from all the data feature dimensions;

[0030] Based on the historical valid data, the pouring simulation model is constructed according to the historical feature indicators corresponding to the key feature dimensions.

[0031] In one possible implementation of the first aspect, calculating the index deviation coefficient corresponding to each data feature dimension based on the historical feature index corresponding to each of the historical valid data and the data confidence level includes:

[0032] For any historical valid data point, calculate the index deviation factor between any historical valid data point and other historical valid data points based on the data confidence level; the index deviation factor is specifically:

[0033]

[0034] Where Devt(i) is the indicator deviation factor corresponding to the i-th historical valid data; ConLv(i) is the data confidence level corresponding to the i-th historical valid data; M is the total number of historical valid data. For the i-th historical valid data, the historical feature index is the k-th data feature dimension. For the j-th historical valid data, the historical feature index is the k-th data feature dimension.

[0035] The indicator deviation coefficient is calculated based on each of the indicator deviation factors and the collection time corresponding to the historical valid data; the indicator deviation coefficient is specifically:

[0036]

[0037] Where DevtLv is the deviation coefficient of the index; Histime(i) is the collection time corresponding to the i-th historical valid data; Curtime is the current time; and α is a preset adjustment constant.

[0038] In one possible implementation of the first aspect, before generating the pouring simulation model corresponding to the self-compacting concrete based on the environmental information, the method further includes:

[0039] Based on the existing feature values ​​corresponding to each of the historical engineering data in each data feature dimension, construct Gaussian distribution curves corresponding to each of the data feature dimensions respectively;

[0040] Based on the total amount of data in the data engineering project, determine the boundaries of abnormal distributions;

[0041] Historical engineering data within the abnormal distribution boundary of the Gaussian distribution curve are identified as abnormal historical data, while historical engineering data within the abnormal distribution boundary of the Gaussian distribution curve are identified as legitimate historical data.

[0042] According to the data calibration algorithm, the abnormal historical data is subjected to data calibration processing to obtain calibration data;

[0043] Based on the initial calibration data and the valid historical data, the historical valid data is obtained.

[0044] In one possible implementation of the first aspect, generating a pouring simulation model for self-compacting concrete based on the environmental information includes:

[0045] Based on the scenario type, obtain the standard casting model associated with the scenario type;

[0046] Based on the dimensions of the pouring space, a simulated pouring space is generated in the standard pouring model;

[0047] Based on the simulated pouring space, the pouring simulation model is generated.

[0048] Secondly, embodiments of this application provide a machine learning-based construction quality control device for self-compacting concrete, the device comprising:

[0049] An environmental information acquisition unit is used to acquire environmental information corresponding to the target scene; the environmental information includes: the size of the pouring space and the scene type.

[0050] The pouring simulation model generation unit is used to generate a pouring simulation model corresponding to self-compacting concrete based on the environmental information; the pouring simulation model is obtained by machine learning training based on preset historical valid data; the historical valid data is obtained after data cleaning of all historical project data.

[0051] The behavior control unit is used to import preset construction behavior data into the pouring simulation model, perform behavior control on the construction behavior data, and obtain control behavior data; the control behavior data is the behavior data that makes the performance data of self-compacting concrete meet the target scenario to the corresponding performance index.

[0052] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any of the first aspects above.

[0053] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in any of the first aspects above.

[0054] Fifthly, embodiments of this application provide a computer program product that, when run on a drone, causes the drone to perform the method described in any one of the first aspects above.

[0055] The beneficial effects of this application embodiment compared with the prior art are as follows: Electronic devices can acquire environmental information corresponding to the target pouring scenario when self-compacting concrete needs to be poured, and generate a matching pouring simulation model based on the environmental information. This allows subsequent pouring simulation operations of the pouring simulation model to adapt to the target pouring scenario, improving the accuracy of the pouring simulation. After constructing the corresponding pouring simulation model, various construction procedures during construction can be simulated in the pouring simulation model based on construction behavior data. This allows for determination of whether the self-compacting concrete poured using the aforementioned construction behavior data meets performance index requirements. If the performance index is not met, behavioral control is implemented to obtain corresponding control behavior data, thus achieving construction guidance for concrete pouring. Compared with existing concrete application technologies, this application embodiment does not require testing concrete performance indicators only after construction is completed. Instead, performance prediction can be performed through the pouring simulation model. Since the pouring simulation model is constructed based on the environmental information of the target scenario, the accuracy of performance prediction can be improved. Furthermore, if the performance data does not meet the performance index, behavioral control can be implemented on the expected construction behavior data, improving the accuracy of construction behavior and reducing the difficulty of using self-compacting concrete. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 This is a schematic diagram of the structure of a concrete construction management system provided in an embodiment of this application;

[0058] Figure 2 This is a schematic diagram illustrating the implementation of a machine learning-based method for controlling the construction quality of self-compacting concrete, as provided in an embodiment of this application.

[0059] Figure 3 This is a flowchart illustrating the specific implementation of a machine learning-based self-compacting concrete construction quality control method before step S203, as provided in the second embodiment of this application.

[0060] Figure 4 This is a flowchart illustrating the specific implementation of steps S2033 and S2034 in a machine learning-based construction quality control method for self-compacting concrete provided in the third embodiment of this application.

[0061] Figure 5 This is a schematic diagram illustrating the application of the state prediction curve provided in an embodiment of this application;

[0062] Figure 6 This is a flowchart illustrating the specific implementation of a machine learning-based self-compacting concrete construction quality control method provided in the fourth embodiment of this application before step S202.

[0063] Figure 7 This is a flowchart illustrating the specific implementation of a machine learning-based self-compacting concrete construction quality control method provided in the fifth embodiment of this application before step S202.

[0064] Figure 8 This is a flowchart illustrating the specific implementation of a machine learning-based self-compacting concrete construction quality control method provided in the sixth embodiment of this application in S202.

[0065] Figure 9 This is a schematic diagram of the structure of a machine learning-based self-compacting concrete construction quality control device provided in an embodiment of this application;

[0066] Figure 10 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0067] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0068] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0069] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0070] The machine learning-based construction quality control method for self-compacting concrete provided in this application can be applied to control equipment for machine learning-based construction quality control of self-compacting concrete. For example, Figure 1 A schematic diagram of the structure of a concrete construction management system according to an embodiment of this application is shown. See also Figure 1The concrete construction management system includes a control device 11 and an environmental information feedback terminal 12. The feedback terminal 12 can be set at the target scene 13 where concrete pouring is required. The target scene can have a groove, such as a groove 131, excavated for concrete pouring. The feedback terminal 12 can obtain the environmental information of the target scene 13 and feed it back to the control device 11. The control device 11 can determine control behavior data that matches the target scene 13 so that after pouring self-compacting concrete according to the control behavior data, the self-compacting concrete can meet the performance requirements, reduce the difficulty of using self-compacting concrete, and improve construction efficiency.

[0071] Compared to conventional concrete, self-compacting concrete is more susceptible to environmental factors during pouring, and its performance data can generally only be determined after the self-compacting concrete has solidified, which is quite difficult. This increases the difficulty of using self-compacting concrete. However, the embodiments of this application can construct a pouring simulation model that matches the target scenario before pouring self-compacting concrete, thereby calibrating the construction behavior data to improve the performance stability of self-compacting concrete pouring and increase the probability that the performance data of self-compacting concrete after pouring meets the performance indicators.

[0072] Please see Figure 2 , Figure 2 This illustration shows a schematic diagram of a machine learning-based method for controlling the construction quality of self-compacting concrete according to an embodiment of this application. This machine learning-based method is applied to the control device 11 described above; that is, the executing entity in this embodiment can be the control device 11. Specifically, the control device 11 is an electronic device, such as a computer, laptop, server, or smartphone. For ease of description, the executing entity will be described using an electronic device as an example. Specifically, the method includes the following steps:

[0073] In S201, environmental information corresponding to the target scene is obtained; the environmental information includes: the size of the pouring space and the scene type.

[0074] In this embodiment, the electronic device can first determine the target scenario where self-compacting concrete needs to be poured. This target scenario can be a specific location, or a building where concrete needs to be poured, etc., and can be determined according to the actual situation.

[0075] In this embodiment, to enable scene reproduction in the casting simulation model, the electronic device can acquire environmental information corresponding to the target scene. This environmental information can include both the dimensions of the casting space and the scene type. Specifically, the dimensions of the casting space refer to the spatial information corresponding to the groove used for casting self-compacting concrete, such as the groove depth, width, and length. In scenarios where the groove is a complex cube, this information can include multiple cross-sectional diagrams, etc., and can be acquired according to the actual situation.

[0076] In this embodiment, the above-mentioned scene type can be used to determine the scene attribute information of the target scene. For example, the above-mentioned scene type can be determined according to the attributes of the horizontal plane where the groove is located, such as flat ground type and slope type, or it can be determined according to the location attributes of the groove, such as building foundation type, road surface type, pedestrian road surface type, etc. The above-mentioned scene type can include multiple types, that is, the scene type of the target scene is described by attributes of multiple dimensions, thereby improving the accuracy of subsequent scene simulation.

[0077] In one possible implementation, the aforementioned environmental information can be collected by a feedback terminal in the target scene. This collection terminal may include multiple sensors, which can acquire environmental information of the target scene and feed the environmental information back to electronic devices.

[0078] For example, the aforementioned feedback terminal may include temperature and humidity sensors, anemometers, and barometers to obtain meteorological-related environmental information such as air humidity, wind speed, and atmospheric pressure.

[0079] For example, the aforementioned feedback terminal may include a ground temperature probe, a soil moisture detector, and a groundwater level monitoring float to obtain geologically related environmental information such as the ground temperature range, soil information, and groundwater information in the target scenario.

[0080] Based on this, the aforementioned environmental information may also include: air humidity information, wind speed, atmospheric pressure, ground temperature range, soil information, and groundwater information, etc. Specific environmental information can be selected according to the actual situation and is not limited here.

[0081] In one possible implementation, the aforementioned environmental information can be obtained based on the design document corresponding to the target scenario. Since the relevant parameters for pouring corresponding to the target scenario can be determined during the design phase, such as the location and depth of the groove and the soil structure at the groove, this information can be recorded in the design document. The electronic device can extract key information related to the target scenario from the design document and perform feature data extraction on this key information, thereby obtaining the aforementioned environmental information.

[0082] In S202, a pouring simulation model for self-compacting concrete is generated based on the environmental information; the pouring simulation model is obtained by machine learning training based on preset historical valid data; the historical valid data is obtained by cleaning all historical project data.

[0083] In this embodiment, the electronic device can be equipped with a scene simulation model. After collecting the environmental information corresponding to the target scene, the corresponding scene object can be constructed in the scene simulation model. For example, a three-dimensional model matching the groove can be built in the scene simulation model, and the model attributes of the three-dimensional model can be adjusted according to the soil information corresponding to the target scene to simulate the material matching the soil information. For another example, if the environmental information includes groundwater information, a corresponding groundwater model can be built in the scene simulation model, and a permeability model corresponding to the groundwater model can be built according to the water level depth in the groundwater information to simulate the influence of groundwater on the groove. The specific type and number of three-dimensional models constructed can be determined according to the environmental information and are not limited here.

[0084] In this embodiment, the three-dimensional model created in the above-mentioned pouring simulation model is specifically an object related to self-compacting concrete, such as the groove model, groundwater model, and seepage model mentioned above. It may also include the solidification model and mechanical model of self-compacting concrete. The above model can be generated from a large amount of historical valid data collected during historical pouring processes. Through machine learning, the model data of each three-dimensional model is adjusted so that the three-dimensional model matches the measurement data of the historical usage process.

[0085] In this embodiment, since the amount of historical engineering data collected is large, including a large amount of abnormal data or missing data, the electronic device can clean the historical engineering data, such as removing abnormal data and filling in missing data, so as to obtain the corresponding valid historical data, thereby improving the effectiveness of the data for training and learning, and thus improving the accuracy of the simulation.

[0086] In some possible implementations, the data cleaning process for historical engineering data described above can be described as follows: First, noise and outliers are removed using data cleaning algorithms. For example, the Z-score method is used to standardize the historical engineering data, and a threshold is set to remove historical engineering data that exceeds the range. Next, the cleaned data is grouped using the K-means clustering algorithm, with a cluster size of 5. The similarity between data points and cluster centers is calculated using Euclidean distance, ultimately dividing the data into 5 clusters. Subsequently, principal component analysis is used to reduce the dimensionality of each grouped historical engineering data set, retaining 95% of the variance information, reducing the original data from 100 dimensions to 10 dimensions to reduce computational complexity and retain key features.

[0087] In S203, preset construction behavior data is imported into the pouring simulation model, and the construction behavior data is subjected to behavior control to obtain control behavior data; the control behavior data is the behavior data that makes the performance data of self-compacting concrete meet the target scenario to the corresponding performance index.

[0088] In this embodiment, the electronic device can store a standard construction process for self-compacting concrete and generate corresponding construction behavior data based on this standard construction process. Optionally, the electronic device can adjust the aforementioned standard construction process based on information such as the expected number of construction workers and the expected number of equipment corresponding to the target scenario, thereby obtaining construction behavior data corresponding to the target scenario.

[0089] In this embodiment, the above-mentioned construction behavior data may include multiple construction steps. Each construction step may be defined with corresponding construction intervals, pouring volume, pouring location and other information. The feature data contained in a specific construction step can be determined according to the construction type corresponding to that construction step.

[0090] In this embodiment, the electronic device can import the aforementioned construction behavior data into the constructed pouring simulation model, thereby determining the performance data of the self-compacting concrete under the operation based on the construction behavior data. If the performance data does not meet the performance index corresponding to the target scenario, it indicates that the aforementioned construction behavior data needs to be adjusted. Based on the adjusted construction behavior data, the operation of S203 is executed again. Through multiple iterations, until the performance data of the self-compacting concrete poured based on the adjusted construction behavior data meets the corresponding performance index, the adjusted construction behavior data, i.e., the aforementioned control behavior data, is output.

[0091] In this embodiment, the electronic device can generate a corresponding construction guidance document based on the control behavior data. Subsequently, the self-compacting concrete can be poured for the target scenario according to the construction guidance document, so as to increase the probability that the actual performance of the self-compacting concrete meets the performance indicators.

[0092] In one possible implementation, the electronic device can collect the actual performance information corresponding to the target scenario, calculate the predicted deviation between the actual performance information and the performance data obtained based on the control behavior data, and calibrate the casting simulation model based on the predicted deviation, thereby enabling model iteration and improving the accuracy of subsequent use.

[0093] As can be seen from the above, the machine learning-based construction quality control method for self-compacting concrete provided in this application allows electronic devices to acquire environmental information corresponding to the target scenario of pouring when self-compacting concrete needs to be poured. Based on this environmental information, a matching pouring simulation model is generated, enabling subsequent pouring simulation operations to adapt to the target scenario and improve the accuracy of the pouring simulation. After constructing the corresponding pouring simulation model, various construction procedures can be simulated within the model based on construction behavior data. This allows for the determination of whether the self-compacting concrete poured using the aforementioned construction behavior data meets performance requirements. If the performance requirements are not met, behavioral control is implemented to obtain corresponding control behavior data, thus providing construction guidance for concrete pouring. Compared with existing concrete application technologies, the embodiments of this application do not require testing concrete performance indicators only after construction is completed. Instead, performance can be predicted through a pouring simulation model. Since the pouring simulation model is constructed based on the environmental information of the target scenario, the accuracy of performance prediction can be improved. Furthermore, if the performance data does not meet the performance indicators, behavioral control can be performed on the expected construction behavior data, thereby improving the accuracy of construction behavior and reducing the difficulty of using self-compacting concrete.

[0094] Figure 3 This document illustrates a flowchart illustrating the specific implementation of step S203 in a machine learning-based construction quality control method for self-compacting concrete provided in the second embodiment of this application. (See also...) Figure 3 As shown, relative to Figure 2 The embodiment provided in this application provides a machine learning-based method for controlling the construction quality of self-compacting concrete. S203 includes S2031~S2035, which are specifically described below:

[0095] In S2031, a simulated construction object is created in the pouring simulation model based on the construction behavior data; the construction behavior data includes at least one construction device and at least one construction worker.

[0096] In this embodiment, the pouring simulation model can store virtual models of different object types, including a concrete model for simulating self-compacting concrete, an equipment model for simulating construction equipment, and a human body operation model for simulating construction workers. Electronic devices can create matching simulated construction objects in the aforementioned pouring simulation model by reading configuration data related to construction equipment and workers from the construction behavior data.

[0097] In this embodiment, the configuration data may record the equipment type and quantity of the construction equipment. The electronic device can extract the equipment model associated with the equipment type from the material library, and create multiple equipment models corresponding to the above-mentioned equipment quantity in the pouring simulation model according to the equipment quantity, thereby obtaining the simulated construction object corresponding to the construction equipment.

[0098] In this embodiment, the aforementioned construction personnel are specifically used to simulate the construction behavior of personnel during the construction process. For example, the number of devices that can be controlled in parallel and the accuracy of controlling the devices. The electronic equipment can determine the control range corresponding to the simulated construction object based on the type, level, and number of the construction personnel. Subsequently, when executing construction steps, a random value can be determined from the construction range using a random number algorithm. This random value can then be used as the operating characteristic value for subsequent construction steps, thereby improving the accuracy of subsequent simulations.

[0099] In S2032, based on the multiple construction steps determined by the construction behavior data, the simulated construction object is sequentially controlled to execute each of the construction steps in the pouring simulation model.

[0100] In this embodiment, the electronic device can determine the construction steps that need to be completed when pouring self-compacting concrete based on construction behavior data. Each construction step can be set with a corresponding construction sequence and the interdependence between each construction step. The electronic device can control the simulated construction objects related to the construction steps in the pouring simulation model according to each construction sequence, so as to execute the corresponding construction steps through the simulated construction objects.

[0101] For example, the construction steps include: a self-compacting concrete mixing step, a pouring step, and a groove pouring step. The self-compacting concrete mixing step and the groove pouring step are two parallel steps, and the pouring step can only be implemented after the above two steps are completed. Based on this, the electronic device can control the simulated construction object related to the self-compacting concrete mixing to perform the above mixing step, and simultaneously control the simulated construction object related to the groove pouring step to perform the above pouring step. Specifically, whether the above two steps are parallel or serial can be determined based on the number of construction personnel in the construction behavior data, so as to determine the number of devices that can be controlled in parallel at the same time.

[0102] In S2033, during the execution of the construction steps, the state data of the simulated concrete in the pouring simulation model is determined in real time; the state data includes: mechanical performance parameters, durability performance parameters, and carbon emission parameters.

[0103] In this embodiment, the electronic device continuously collects the real-time status of the simulated concrete in the pouring simulation model while controlling each simulated construction object to perform any construction step. The simulated concrete is a simulation model built based on self-compacting concrete to simulate the state of self-compacting concrete during use. Specifically, during the concrete mixing stage, the simulated concrete model can be a mixture of material models corresponding to each raw material, simulating the mixing process of each material model to obtain the corresponding simulated concrete.

[0104] In this embodiment, the electronic device can predict the state of the self-compacting concrete after pouring based on the real-time state of the simulated concrete in the pouring simulation model, thereby obtaining the aforementioned state data. That is, the state data is not determined based on the state of the model concrete during the construction steps, but rather predicted based on its final state. The pouring simulation model has a state prediction module. The real-time state corresponding to the execution of the aforementioned construction steps is imported into the state prediction module, which can then output the aforementioned state data. Specifically, based on the performance evaluation dimensions of self-compacting concrete, the collected state data may include: mechanical performance parameters, such as the stiffness, hardness, and stress of the self-compacting concrete; durability performance parameters, such as deformation under different pressures and wear resistance levels; and carbon emission parameters, specifically the total amount and rate of carbon dioxide generated during the pouring of self-compacting concrete. The specific data type of the aforementioned state data can be determined according to actual indicator requirements and is not limited here.

[0105] In S2034, if it is detected that any state data at any time does not meet the performance index, then behavior control data is generated based on the performance deviation corresponding to the state data and the associated construction steps.

[0106] In this embodiment, the electronic device can match the above-mentioned state data with the performance indicators corresponding to the target scenario. If the state data meets the performance indicators corresponding to the target scenario, the construction step can continue to be executed, or after the current construction step is completed, the next construction step can be executed, and the operation of S2033 can be repeated until all construction steps are completed. If the state data of the simulated concrete meets the above-mentioned performance indicators after all construction steps are completed, the construction behavior data can be directly output without calibrating the construction behavior data.

[0107] In this embodiment, if the electronic device detects that a certain status data does not meet the performance index at any time, it indicates that there may be abnormal behavior in the execution of the construction step, and the actual behavior executed does not match the target scenario. In this case, it is not necessary to execute subsequent steps, but to control the behavior of the already executed construction steps.

[0108] In this embodiment, the behavior control of the construction steps can be performed as follows: based on the data type of the state data that does not meet the performance requirements, determine multiple executed steps associated with that data type, and adjust the data related to the executed steps in the construction behavior data according to a preset adjustment step size to obtain the behavior control data.

[0109] In S2035, the construction behavior data is adjusted based on the behavior control data, and the operation of creating a simulated construction object in the pouring simulation model based on the construction behavior data is returned based on the adjusted construction behavior data until the state data meets the performance index after all the construction steps are completed.

[0110] In this embodiment, the electronic device can adjust the construction behavior data according to the behavior control data, such as modifying some parameters, and return to execute the operation of S2031 to simulate the construction process again. Through iterative methods, the state data of the simulated concrete can be made to meet the performance indicators.

[0111] In this embodiment of the application, by using construction behavior data to simulate the pouring process of self-compacting concrete during construction, and continuously predicting the state data of the simulated concrete in the pouring simulation model, abnormal steps can be detected in real time, and then adjusted. This can improve the iteration speed of control behavior data and the convergence speed of results, thereby improving the user experience and avoiding long waiting times for users to obtain the corresponding control behavior data.

[0112] Figure 4 This document illustrates a flowchart illustrating the specific implementation of steps S2033 and S2034 in a machine learning-based construction quality control method for self-compacting concrete provided in the third embodiment of this application. (See also...) Figure 4 As shown, relative to Figure 3 In the embodiment provided in this application, S2033 of the self-compacting concrete construction quality control method based on machine learning may include S401~S404, and S2034 may include S405~S407, as described in detail below:

[0113] During the execution of the construction steps, the state data of the simulated concrete in the pouring simulation model are determined in real time, including:

[0114] In S401, during the execution of the construction steps, real-time data of the simulated concrete in the pouring simulation model is collected in real time.

[0115] In this embodiment, since the simulated concrete is a virtual object created in the pouring simulation model, relevant parameters of the simulated concrete can be obtained. However, during the actual pouring process, it is impossible to collect the state of the compacted concrete in real time; state prediction can only be made based on experience, resulting in significant errors in state judgment. This embodiment, however, can determine real-time data through the simulated concrete, and the data type of the real-time data can be data related to performance indicators. This allows for the prediction of subsequent state data based on real-time data, so that the predicted state data can be compared with performance indicators to determine whether the state data meets the performance indicators.

[0116] In S402, the scene influence factors corresponding to the state data are determined according to the scene type; the scene type includes: ambient temperature, ambient humidity, and location slope.

[0117] In this embodiment, the electronic device can determine the scene influence factors that affect the state data based on the scene type corresponding to the target scene. It should be noted that the data type of state data can include multiple dimensions, such as mechanical performance, durability, and concrete stability. Different dimensions are affected by different environmental factors. The electronic device can also establish a correspondence between different types of state data and scene influence factors, thereby determining the scene influence factors affecting different state data and improving the accuracy of subsequent state prediction curves.

[0118] In S403, based on the real-time data and the scene influence factors, a state prediction curve corresponding to the simulated concrete is constructed.

[0119] In S404, the state data is determined in the state prediction curve based on the predicted completion time corresponding to the completion of all the construction steps; the state data is the data corresponding to the predicted completion time in the state prediction curve.

[0120] In this embodiment, the electronic device can establish a corresponding real-time state curve based on all real-time data collected during the execution of this construction step and in the preceding steps. Based on the real-time state curve, the influence coefficient of each scenario's influencing factor is determined. Then, based on the influence coefficients of each scenario's influencing factor and the aforementioned real-time state curve, the aforementioned state prediction curve is established. It should be noted that the real-time data includes data from multiple different dimensions; therefore, the number of state prediction curves constructed can also be multiple, specifically determined based on the data items contained in the real-time data. Furthermore, when calculating the deviation index subsequently, the corresponding deviation index can be calculated separately for each different data item.

[0121] For example, Figure 5A schematic diagram illustrating the application of a state prediction curve provided in an embodiment of this application is shown. See also: Figure 5 As shown, curve 51 includes the solid line segment 511 and the curved segment 512. Segment 511 is constructed based on real-time data collected during the executed steps and the current execution step. Based on segment 511, the influence coefficient of each scenario's influencing factor can be determined, thereby establishing a corresponding state prediction function. This state prediction function can then be used to predict the state of the simulated concrete, thus yielding segment 512. The end time of segment 512 corresponds to the completion of all construction steps.

[0122] In this embodiment, the electronic device can determine the predicted completion time after all construction operations are completed based on the state prediction curve, simulate the real-time data corresponding to the concrete, use the real-time data corresponding to the predicted completion time as the aforementioned state data, and match the state data with performance indicators to determine whether the construction requirements are met.

[0123] Correspondingly, if any state data at any time is detected to fail to meet the performance index, then behavior control data is generated based on the performance deviation corresponding to the state data and the associated construction steps, including:

[0124] In S405, the area of ​​the curve deviation between the state prediction curve and the standard state curve is calculated.

[0125] In this embodiment, the electronic device can simultaneously generate a standard state curve corresponding to the target scenario within the coordinate system where the state prediction curve is located. This standard state curve is generated based on the performance indicators corresponding to the target scenario.

[0126] In this embodiment, the electronic device can calculate the area of ​​deviation between the two curves to determine the degree of deviation between the simulated concrete state and the desired state. Continuing with... Figure 5 Taking this as an example, curve 52 is the standard state curve corresponding to self-compacting concrete, and then the area of ​​the curve deviation mentioned above, i.e., area 53, can be determined.

[0127] In S406, the deviation index corresponding to the state data is calculated based on the weight value corresponding to each time zone in the curve deviation area; the weight value is determined based on the time interval length between each time zone and the execution time of the current construction behavior.

[0128] In this embodiment, the electronic device can divide the aforementioned curve deviation area into multiple time-series partitions based on temporal relationships. Each time-series partition can correspond to a weight value. The longer the time interval between the partition and the execution time of the current construction action, the lower the prediction confidence, and therefore the smaller the corresponding weight value. Conversely, the shorter the time interval between the partition and the execution time of the current construction action, the higher the prediction accuracy and confidence, and the larger the corresponding weight value. By dividing the curve deviation area into time-series partitions, the electronic device can improve the accuracy of deviation index calculation. The aforementioned deviation index can be expressed as:

[0129]

[0130] Where EorroLv is the aforementioned deviation index; Weight(p) is the weight value corresponding to the p-th time series partition; S(p) is the area of ​​the region corresponding to the p-th time series partition; and P is the total number of time series partitions contained in the curve deviation area.

[0131] In S407, if the deviation index is greater than the preset deviation threshold, then behavior control data is generated based on the performance deviation corresponding to any one of the state data and the associated construction steps.

[0132] In this embodiment, when the deviation index is found to be greater than the deviation threshold, it indicates that the current state data does not meet the above performance index. Based on the abnormal state data, the associated construction steps are determined, and the performance deviation is obtained based on the difference between the state data predicted based on real-time data and the performance index. The performance deviation is converted into the corresponding adjustment step size, and the behavior control data is obtained based on the adjustment step size corresponding to all state data with one or more abnormalities.

[0133] In this embodiment of the application, by constructing a corresponding state prediction curve, the state data corresponding to the completion time can be determined based on the currently executed construction steps. Then, abnormal operation identification can be performed based on the state data, which can improve the accuracy of abnormal operation identification and thus improve the accuracy of behavior control.

[0134] Figure 6 This diagram illustrates the specific implementation flowchart of a machine learning-based self-compacting concrete construction quality control method provided in the fourth embodiment of this application before step S202. See also... Figure 6 In contrast Figure 2 In any of the embodiments described above, the self-compacting concrete construction quality control method based on machine learning provided in this embodiment further includes, before S202: S601~S606, which are detailed below:

[0135] In S601, the historical valid data and the historical prediction data corresponding to each historical valid data are obtained; the historical environment corresponding to each historical valid data is matched with the environmental information.

[0136] In this embodiment, during the pouring of self-compacting concrete, the project can be designed, specifying the performance requirements for the self-compacting concrete. However, the actual usage and performance during pouring may differ from the design predictions. Therefore, the electronic device can determine the impact of different construction procedures on concrete performance by obtaining a comparison of the two sets of data from historical construction processes. This allows for the foundational construction of scenario simulation. The concrete performance and related construction data during actual pouring are considered historical valid data, while the requirements and construction specifications recorded in the design documents during the pouring process are considered historical predicted data.

[0137] In this embodiment, in order to improve the matching degree between the constructed pouring simulation model and the scene, the historical environment corresponding to the above-collected historical valid data is matched with the environmental information corresponding to the target scene. For example, the scene type of the historical environment is the same as the scene type of the target scene, and / or the actual spatial size of the historical environment matches the pouring space size of the target scene.

[0138] In S602, the data confidence level of each historical valid data is determined based on the data deviation corresponding to each historical valid data and the historical predicted data.

[0139] In this embodiment, the electronic device can calculate the data deviation between each historical valid data point and its corresponding historical predicted data. Specifically, if the historical valid data contains multiple different data points, the deviation factors for each data point are calculated separately, and the deviation factors corresponding to all data points are superimposed or weighted to calculate the aforementioned data deviation.

[0140] In this embodiment, there are multiple historical pouring events, resulting in multiple sets of data for both the aforementioned valid historical data and the predicted historical data. The data deviation for each set is calculated, and the corresponding data confidence level is determined based on the data deviation. A larger data deviation indicates a significant discrepancy between the expected and actual performance of the self-compacting concrete during pouring, potentially due to design or operational errors. In this case, there is a possibility of anomalies, and the corresponding confidence level is lower. That is, there is an inverse relationship between data deviation and data confidence level. The electronic device can convert the data deviation into the corresponding data confidence level using a preset conversion function.

[0141] In S603, the historical feature index corresponding to each of the historical valid data is determined according to multiple preset data feature dimensions.

[0142] In S604, the index deviation coefficient corresponding to each data feature dimension is calculated based on the historical feature index corresponding to each of the historical valid data and the data confidence level.

[0143] In this embodiment, the historical valid data contains multiple data dimensions, such as mechanical performance dimension, durability performance dimension, and carbon emission dimension. The electronic device can extract the historical feature indicators corresponding to each data dimension from each historical valid data.

[0144] In this embodiment, the electronic device can calculate the mean value of the indicator for the data feature dimension based on the historical feature indicators corresponding to all historical valid data. Then, it can calculate the indicator deviation between the historical feature indicators of each historical valid data and the mean value of the indicator, and add the corresponding data confidence level on the basis of the indicator deviation, thereby calculating the discrete contribution corresponding to the historical valid data. Based on the discrete contribution of all historical valid data, the indicator deviation coefficient of the corresponding data feature dimension can be calculated.

[0145] In one possible implementation, the calculation of the above-mentioned index deviation coefficient may include the following steps:

[0146] In S604.1, for any historical valid data, an index deviation factor between any historical valid data and other historical valid data is calculated based on the data confidence level; the index deviation factor is specifically:

[0147]

[0148] Where Devt(i) is the indicator deviation factor corresponding to the i-th historical valid data; ConLv(i) is the data confidence level corresponding to the i-th historical valid data; M is the total number of historical valid data. For the i-th historical valid data, the historical feature index is the k-th data feature dimension. For the j-th historical valid data, the historical feature index is the k-th data feature dimension.

[0149] In S604.2, the index deviation coefficient is calculated based on each of the index deviation factors and the collection time corresponding to the historical valid data; the index deviation coefficient is specifically:

[0150]

[0151] Where DevtLv is the deviation coefficient of the index; Histime(i) is the collection time corresponding to the i-th historical valid data; Curtime is the current time; and α is a preset adjustment constant.

[0152] In this embodiment of the application, the index deviation factor corresponding to different historical valid data is calculated respectively. This allows for the calculation of the degree of deviation between each historical valid data and the mean, thereby determining the dispersion of historical valid data in all historical pouring events (i.e., reflected by the index deviation factor). Furthermore, the greater the distance between a historical pouring event and the current time, the greater the difference between their pouring techniques and the smaller the corresponding reference contribution. Therefore, by determining the time difference with the current time as its weight value, the accuracy of the index deviation coefficient can be improved.

[0153] In S605, based on the index deviation coefficient, the key feature dimension corresponding to the environmental information is determined from all the data feature dimensions.

[0154] In this embodiment, the electronic device can select data dimensions whose values ​​are greater than a preset deviation threshold based on the index deviation coefficient as key feature dimensions. That is, the above key feature dimensions have large deviations in different pouring scenarios, which may be due to differences in construction processes. They need to be used as key variables for monitoring when building the model later.

[0155] In S606, the pouring simulation model is constructed based on the historical feature indicators corresponding to the key feature dimensions of the historical valid data.

[0156] In this embodiment, each key feature dimension can correspond to one or more construction steps, and one or more construction simulation objects can be determined based on the corresponding construction steps. The electronic device can determine the actual construction object associated with the key feature dimension from historical valid data, as well as the process parameters of the actual construction object, such as working hours, work intensity, and personnel level of the workers, thereby establishing the association between the key feature dimension and the construction simulation object. Based on the association between each key feature dimension and the construction simulation object, the above-mentioned pouring simulation model is generated.

[0157] In this embodiment of the application, by acquiring historical valid data, the key feature dimensions that will be affected by the process can be determined based on the degree of dispersion of each data feature dimension in the historical valid data. Then, a corresponding casting simulation model is generated based on the key feature dimensions, thereby improving the accuracy of the model.

[0158] Figure 7 This diagram illustrates the specific implementation flowchart of a machine learning-based self-compacting concrete construction quality control method provided in the fifth embodiment of this application before step S202. See also... Figure 7 In contrast Figures 2 to 6 In any of the embodiments described above, the self-compacting concrete construction quality control method based on machine learning provided in this embodiment further includes, before S202: S701~S705, which are detailed below:

[0159] In S701, Gaussian distribution curves corresponding to each of the historical engineering data and each of the data feature dimensions are constructed based on the existing feature values ​​of each of the historical engineering data and each of the data feature dimensions.

[0160] In S702, the boundary of abnormal distribution is determined based on the total amount of data in the data engineering data.

[0161] In S703, historical engineering data in the Gaussian distribution curve within the abnormal distribution boundary are identified as abnormal historical data, and historical engineering data in the Gaussian distribution curve within the abnormal distribution boundary are identified as legitimate historical data.

[0162] In this embodiment, the electronic device can perform data cleaning on historical engineering data, filtering out abnormal data that clearly shows abnormal construction behavior. The filtering method involves constructing a corresponding Gaussian distribution curve based on all historical engineering data and determining the abnormal distribution boundary. Data outside the abnormal distribution boundary is considered to be far from the mean and can be identified as abnormal data; that is, data outside the abnormal distribution boundary of the aforementioned Gaussian distribution curve is regarded as abnormal historical data.

[0163] In S704, the abnormal historical data is calibrated according to the data calibration algorithm to obtain calibration data.

[0164] In S705, based on the first calibration data and the valid historical data, the historical valid data is obtained.

[0165] In this embodiment, since the historical data of the pouring records is limited, in order to increase the amount of training data that can be referenced, when abnormal historical data is identified, the abnormal historical data will not be discarded directly. Instead, the abnormal historical data can be calibrated to obtain calibration data, thereby increasing the training data samples that can be used when building the model and thus improving the accuracy of the simulation.

[0166] Figure 8 The flowchart illustrating the specific implementation of a machine learning-based self-compacting concrete construction quality control method according to the sixth embodiment of this application in step S202 is shown. See also Figure 8 In contrast Figures 2 to 6In any of the embodiments described above, the self-compacting concrete construction quality control method based on machine learning provided in this embodiment includes S2021 to S2023 in step S202, which are detailed below:

[0167] In S2021, a standard casting model associated with the scene type is obtained according to the scene type.

[0168] In S2022, a simulated pouring space is generated in the standard pouring model according to the pouring space dimensions.

[0169] In S2023, the pouring simulation model is generated based on the simulated pouring space.

[0170] In this embodiment, the electronic device can acquire different standard casting models according to different scenarios. For example, for a road surface scenario, it acquires a standard casting model related to the road surface scenario; for a slope scenario, it acquires a standard casting model related to the slope scenario. The electronic device can construct different standard casting models for different scenario types and generate corresponding simulated casting spaces in the above standard casting models according to the casting space dimensions. If the environmental information includes groundwater information, a groundwater environment model can also be created in the above standard casting models, thereby improving the accuracy of the simulation.

[0171] In this embodiment, Figure 9 This illustration shows a structural block diagram of a machine learning-based self-compacting concrete construction quality control device according to an embodiment of this application. The device includes units for performing various functions. Figure 2 The steps implemented by the first device in the corresponding embodiment are described in detail. Figure 2 and Figure 2 The relevant descriptions in the corresponding embodiments are shown below. For ease of explanation, only the parts relevant to this embodiment are shown.

[0172] See Figure 9 A machine learning-based self-compacting concrete construction quality control device includes:

[0173] The environmental information acquisition unit 91 is used to acquire environmental information corresponding to the target scene; the environmental information includes: the size of the pouring space and the scene type.

[0174] The pouring simulation model generation unit 92 is used to generate a pouring simulation model corresponding to self-compacting concrete based on the environmental information; the pouring simulation model is obtained by machine learning training based on preset historical valid data; the historical valid data is obtained after data cleaning of all historical project data.

[0175] The behavior control unit 93 is used to import preset construction behavior data into the pouring simulation model, perform behavior control on the construction behavior data, and obtain control behavior data; the control behavior data is the behavior data that makes the performance data of self-compacting concrete meet the target scenario to the corresponding performance index.

[0176] It should be understood that, Figure 9 In the structural block diagram of the device shown, each module is used to perform... Figures 2 to 8 The steps in the corresponding embodiments, and for Figures 2 to 8 The steps in the corresponding embodiments have been explained in detail in the above embodiments. Please refer to them for details. Figures 2 to 8 as well as Figures 2 to 8 The relevant descriptions in the corresponding embodiments will not be repeated here.

[0177] Figure 10 This is a structural block diagram of an electronic device provided in another embodiment of this application. For example... Figure 10 The electronic device 1000 of this embodiment includes a processor 1010, a memory 1020, and a computer program 1030 stored in the memory 1020 and executable on the processor 1010, such as a program for a machine learning-based self-compacting concrete construction quality control method. When the processor 1010 executes the computer program 1030, it implements the steps in the various embodiments of the machine learning-based self-compacting concrete construction quality control method, for example... Figure 2 S201 to S203 are described above. Alternatively, the processor 1010 may implement the above when executing the computer program 1030. Figure 9 The functions of each module in the corresponding embodiments, for example, Figure 9 For details regarding the functions of units 91 to 93, please refer to [link / reference needed]. Figure 9 The relevant descriptions in the corresponding embodiments.

[0178] For example, computer program 1030 may be divided into one or more modules, one or more of which are stored in memory 1020 and executed by processor 1010 to complete this application. One or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of computer program 1030 in electronic device 1000. For example, computer program 1030 may be divided into various unit modules, each with the specific functions described above.

[0179] Electronic device 1000 may include, but is not limited to, processor 1010 and memory 1020. Those skilled in the art will understand that... Figure 10This is merely an example of electronic device 1000 and does not constitute a limitation on electronic device 1000. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device may also include input / output devices, network access devices, buses, etc.

[0180] The processor 1010 may be a central processing unit, or it may be other general-purpose processors, digital signal processors, application-specific integrated circuits, off-the-shelf programmable gate arrays or other programmable logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0181] The memory 1020 can be an internal storage unit of the electronic device 1000, such as a hard disk or memory of the electronic device 1000. The memory 1020 can also be an external storage device of the electronic device 1000, such as a plug-in hard disk, smart memory card, flash memory card, etc. equipped on the electronic device 1000. Furthermore, the memory 1020 can include both internal storage units and external storage devices of the electronic device 1000.

[0182] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for controlling the construction quality of self-compacting concrete based on machine learning, characterized in that, include: Obtain environmental information corresponding to the target scene; The environmental information includes: the dimensions of the pouring space and the scene type; Based on the environmental information, a pouring simulation model for self-compacting concrete is generated; the pouring simulation model is obtained by machine learning training based on preset historical valid data; the historical valid data is obtained after data cleaning of all historical project data. The preset construction behavior data is imported into the pouring simulation model, and the construction behavior data is subjected to behavior control to obtain control behavior data; the control behavior data is the behavior data that makes the performance data of self-compacting concrete meet the target scenario to the corresponding performance index. Before generating the pouring simulation model corresponding to self-compacting concrete based on the environmental information, the method further includes: Obtain the historical valid data and the historical prediction data corresponding to each historical valid data; match the historical environment corresponding to each historical valid data with the environmental information; Based on the data deviation corresponding to each of the historical valid data points and the historical predicted data points, determine the data confidence level of each of the historical valid data points; Based on multiple preset data feature dimensions, determine the historical feature indicators corresponding to each of the historical valid data; Based on the historical feature indicators and data confidence levels corresponding to each of the historical valid data, calculate the indicator deviation coefficient corresponding to each data feature dimension. Based on the index deviation coefficient, the key feature dimension corresponding to the environmental information is determined from all the data feature dimensions; Based on the historical valid data, the pouring simulation model is constructed according to the historical feature indicators corresponding to the key feature dimensions. The step of calculating the index deviation coefficient corresponding to each data feature dimension based on the historical feature index and the data confidence level corresponding to each of the historical valid data includes: For any historical valid data point, calculate the index deviation factor between any historical valid data point and other historical valid data points based on the data confidence level; the index deviation factor is specifically: Where Devt(i) is the indicator deviation factor corresponding to the i-th historical valid data; ConLv(i) is the data confidence level corresponding to the i-th historical valid data; M is the total number of historical valid data. For the i-th historical valid data, the historical feature index is the k-th data feature dimension. For the j-th historical valid data, the historical feature index is the k-th data feature dimension. The index deviation coefficient is calculated based on the collection time corresponding to each of the historical valid data points and the index deviation factor; the index deviation coefficient is specifically: Where DevtLv is the deviation coefficient of the index; Histime(i) is the collection time corresponding to the i-th historical valid data; Curtime is the current time; and α is a preset adjustment constant.

2. The method according to claim 1, characterized in that, The step of importing preset construction behavior data into the pouring simulation model and performing behavior control on the construction behavior data to obtain control behavior data includes: Based on the construction behavior data, a simulated construction object is created in the pouring simulation model; the construction behavior data includes: at least one construction device and at least one construction worker; Based on the construction behavior data, multiple construction steps are determined, and the simulated construction object is sequentially controlled to execute each of the construction steps in the pouring simulation model; During the execution of the construction steps, the state data of the simulated concrete in the pouring simulation model is determined in real time; the state data includes: mechanical performance parameters, durability performance parameters, and carbon emission parameters. If any state data at any time is detected to fail to meet the performance index, then behavior control data is generated based on the performance deviation corresponding to the state data and the associated construction steps. The construction behavior data is adjusted based on the behavior control data, and the operation of creating a simulated construction object in the pouring simulation model based on the adjusted construction behavior data is returned to be executed until all the construction steps are completed and the state data meets the performance index.

3. The method according to claim 2, characterized in that, During the execution of the construction steps, the state data of the simulated concrete in the pouring simulation model are determined in real time, including: During the execution of the construction steps, real-time data of the simulated concrete in the pouring simulation model is collected. Based on the scenario type, determine the scenario influencing factors corresponding to the state data; the scenario type includes: ambient temperature, ambient humidity, and location slope; Based on the real-time data and the scene influencing factors, a state prediction curve corresponding to the simulated concrete is constructed. Based on the predicted completion time corresponding to the completion of all the aforementioned construction steps, the state data is determined in the state prediction curve; the state data is the data corresponding to the predicted completion time in the state prediction curve. Correspondingly, if any state data at any time is detected to fail to meet the performance index, then behavior control data is generated based on the performance deviation corresponding to the state data and the associated construction steps, including: Calculate the area of ​​the curve deviation between the predicted state curve and the standard state curve; The deviation index corresponding to the state data is calculated based on the weight value corresponding to each time zone in the deviation area of ​​the curve; the weight value is determined based on the time interval length between each time zone and the execution time of the current construction behavior. If the deviation index is greater than the preset deviation threshold, then behavior control data is generated based on the performance deviation corresponding to any of the state data and the associated construction steps.

4. The method according to any one of claims 1-3, characterized in that, Before generating the pouring simulation model corresponding to self-compacting concrete based on the environmental information, the method further includes: Based on the existing feature values ​​corresponding to each of the historical engineering data in each data feature dimension, construct Gaussian distribution curves corresponding to each of the data feature dimensions respectively; Based on the total amount of historical engineering data, the boundaries of abnormal distributions are determined; Historical engineering data within the abnormal distribution boundary of the Gaussian distribution curve are identified as abnormal historical data, while historical engineering data within the abnormal distribution boundary of the Gaussian distribution curve are identified as legitimate historical data. According to the data calibration algorithm, the abnormal historical data is subjected to data calibration processing to obtain calibration data; Based on the initial calibration data and the valid historical data, the historical valid data is obtained.

5. The method according to any one of claims 1-3, characterized in that, The step of generating a pouring simulation model for self-compacting concrete based on the environmental information includes: Based on the scenario type, obtain the standard casting model associated with the scenario type; Based on the dimensions of the pouring space, a simulated pouring space is generated in the standard pouring model; Based on the simulated pouring space, the pouring simulation model is generated.

6. A machine learning-based construction quality control device for self-compacting concrete, characterized in that, include: The environmental information acquisition unit is used to acquire environmental information corresponding to the target scene. The environmental information includes: the dimensions of the pouring space and the scene type; The pouring simulation model generation unit is used to generate a pouring simulation model corresponding to self-compacting concrete based on the environmental information; the pouring simulation model is obtained by machine learning training based on preset historical valid data; the historical valid data is obtained after data cleaning of all historical project data. The behavior control unit is used to import preset construction behavior data into the pouring simulation model, perform behavior control on the construction behavior data, and obtain control behavior data; the control behavior data is the behavior data that makes the performance data of self-compacting concrete meet the target scenario to the corresponding performance index. The casting simulation model generation unit is also used for: Obtain the historical valid data and the historical prediction data corresponding to each historical valid data; match the historical environment corresponding to each historical valid data with the environmental information; Based on the data deviation corresponding to each of the historical valid data points and the historical predicted data points, determine the data confidence level of each of the historical valid data points; Based on multiple preset data feature dimensions, determine the historical feature indicators corresponding to each of the historical valid data; Based on the historical feature indicators and data confidence levels corresponding to each of the historical valid data, calculate the indicator deviation coefficient corresponding to each data feature dimension. Based on the index deviation coefficient, the key feature dimension corresponding to the environmental information is determined from all the data feature dimensions; Based on the historical valid data, the pouring simulation model is constructed according to the historical feature indicators corresponding to the key feature dimensions. The step of calculating the index deviation coefficient corresponding to each data feature dimension based on the historical feature index and the data confidence level corresponding to each of the historical valid data includes: For any historical valid data point, calculate the index deviation factor between any historical valid data point and other historical valid data points based on the data confidence level; the index deviation factor is specifically: Where Devt(i) is the indicator deviation factor corresponding to the i-th historical valid data; ConLv(i) is the data confidence level corresponding to the i-th historical valid data; M is the total number of historical valid data. For the i-th historical valid data, the historical feature index is the k-th data feature dimension. For the j-th historical valid data, the historical feature index is the k-th data feature dimension. The index deviation coefficient is calculated based on the collection time corresponding to each of the historical valid data points and the index deviation factor; the index deviation coefficient is specifically: Where DevtLv is the deviation coefficient of the index; Histime(i) is the collection time corresponding to the i-th historical valid data; Curtime is the current time; and α is a preset adjustment constant.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 5.

Images