CRTS III slab ballastless track construction quality management method and system based on deep learning and BIM
By employing a construction quality management method based on deep learning and BIM, an evaluation system and model for CRTSⅢ type slab track was established. Predictions were made using the CRITIC-EWM combined weighting method and the SSA-LSTM model, and quality data was visualized in the BIM model. This achieved high accuracy and robustness in construction efficiency and informatization, solving existing technical problems related to construction site conditions, improving construction efficiency, and enhancing construction quality management.
Patent Information
- Application Number
- CN202511098458.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-06
- Publication Date
- 2025-12-09
AI Technical Summary
In the traditional CRTSⅢ type slab track construction quality management, the reliance on paper or verbal communication and the lack of real-time data support lead to poor decision-making timeliness and accuracy. Manual inspections are prone to overlooking problems, increasing communication costs and risks.
A construction quality management method based on deep learning and BIM is adopted. By identifying key sub-processes, an evaluation system is established, the evaluation value is calculated using the CRITIC-EWM combined weighting method, and the SSA-LSTM model is used for prediction. The quality data is then visualized in the BIM model.
It achieves highly accurate and robust construction quality prediction, improves construction efficiency and informatization, and helps to identify problems on-site in advance and avoid risks.
Smart Images
Figure CN121094618A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a construction quality management method and system for CRTSⅢ type slab track based on deep learning and BIM, belonging to the field of railway infrastructure construction quality management. Background Technology
[0002] With the rapid emergence of new technologies such as AI, big data, and IoT, informatization and intelligentization have become inevitable choices of the times. Building Information Modeling (BIM) serves as a digital tool for engineering design, and deep learning is the most promising technological approach to solving the major scientific and technological problem of strong artificial intelligence. How to apply it to traditional industries and realize the transformation of traditional industries towards informatization and intelligence has become a major research direction.
[0003] In the traditional CRTSⅢ type slab track construction quality management, the description of specific construction quality is generalized, and the transmission of quality information relies heavily on paper or oral communication, which cannot provide real-time data support and effective decision analysis. This affects the timeliness and accuracy of decision-making, resulting in opaque, easily lost, and difficult-to-trace information, increasing communication costs and risks. Furthermore, due to the reliance on manual visual inspection and experience judgment, problems and potential quality hazards are easily overlooked, affecting the quality of the project.
[0004] Therefore, this application proposes an improved construction quality management method and system for CRTSⅢ type slab track based on deep learning and BIM. Summary of the Invention
[0005] Purpose of the invention: In order to overcome the shortcomings of the existing technology, the present invention provides a construction quality management method and system for CRTSⅢ type slab track based on deep learning and BIM. The prediction model has high accuracy and good robustness. The accurate prediction values can help to discover problems in advance and avoid risks during on-site construction.
[0006] Technical Solution: To solve the above-mentioned technical problems, the present invention provides a construction quality management method for CRTSⅢ type slab track based on deep learning and BIM, comprising the following steps:
[0007] Step S1: Identify key sub-processes in construction, establish a construction quality evaluation system for CRTSⅢ type slab track, and obtain corresponding indicator construction data;
[0008] Step S2: Based on the CRTSⅢ type slab track construction quality evaluation system, the CRITIC-EWM combined weighting method is used to obtain the CRTSⅢ type slab track construction quality evaluation value.
[0009] Step S3: Based on the SSA-LSTM architecture and combined with the MPA-VMD enhancement algorithm, construct a quality prediction model for the construction process of CRTSⅢ type slab track, and predict subsequent evaluation values based on existing evaluation values.
[0010] Step S4: Convert the BIM model to GLTF format, display it on the web, and display the construction quality evaluation value and predicted value in the corresponding area of the BIM model.
[0011] Preferably, step S1 includes the following steps:
[0012] S11. The WBS method is used to decompose the construction process of CRTSⅢ type slab track. First, the entire construction process is divided into four key construction stages: base plate construction, track slab construction, self-compacting concrete pouring, and rail laying. These four construction stages are then further subdivided to complete the construction decomposition of CRTSⅢ type slab track.
[0013] S12. Based on the design specifications, the key factors affecting the construction quality of CRTSⅢ type slab track at each stage of construction were identified.
[0014] S13. Comprehensively analyze the construction sub-processes and factors affecting construction quality, establish a construction quality evaluation index system for CRTSⅢ type slab track, and classify the evaluation levels of each index in conjunction with relevant standards.
[0015] Preferably, step S2 includes the following steps:
[0016] S21. Clean and standardize the indicator data in the CRTSⅢ type slab track construction quality evaluation system obtained in step S1 (there is no prediction indicator system in step S1).
[0017] S21. The EWM method is used to calculate the information entropy and weights. The CRITIC algorithm is used to calculate the information carrying capacity and weights. The combined weights are then calculated by combining the two weights.
[0018] S22. After obtaining the combined weights, the construction quality evaluation value of CRTSⅢ type slab track can be calculated. The evaluation value is calculated using a weighted summation method. The accurate construction quality score is obtained by classifying the evaluation levels based on the calculated weights.
[0019] Preferably, step S3 includes the following steps:
[0020] S31. Arrange the construction quality evaluation values of CRTSⅢ type slab track obtained in step S2 in chronological order and use them as input values.
[0021] S32. The model uses the MPA algorithm to find the optimal number of decomposition layers K and the penalty factor β. The optimal parameters are then substituted into the VMD algorithm to decompose the construction quality scoring dataset, and finally n modal components are obtained.
[0022] S33. The decomposed modal components are fed into the SSA-LSTM model for prediction. In the SSA-LSTM model, the SSA algorithm performs a comprehensive search on the key parameters of the LSTM model. Then, the optimal parameters obtained from the search are substituted into the LSTM model. After obtaining the predicted values of all modal components, the predicted values are superimposed to obtain the complete prediction sequence and output the prediction result.
[0023] Preferably, step S4 includes the following steps:
[0024] S41. Create a BIM model of CRTSⅢ type slab track and substructure in Revit;
[0025] S42. Obtain the evaluation and prediction values obtained from steps S2 and S3, and map them to the orbital structure;
[0026] S43. Perform secondary development on Revit to convert the Type III slab model from .rvt format to .gltf format;
[0027] S44. By using the material.color property in the Three.js framework to assign colors to the model corresponding to the quality level, the quality data can be displayed in a model-based manner, thereby enabling precise and efficient management of the construction quality of CRTSⅢ type slab track.
[0028] A construction quality management system for CRTSⅢ type slab track based on deep learning and BIM is characterized by including a project introduction module, a construction quality evaluation module, a construction quality prediction module, a non-conforming CRTSⅢ type slab management module, and a 3D display module. The project introduction module provides relevant project information, the construction quality evaluation module evaluates the project based on the project parameters provided by the project introduction module, the construction quality prediction module provides a construction quality prediction based on the evaluation information, and if the project is found to be non-conforming, the non-conforming CRTSⅢ type slab management module identifies the reasons for the non-conformity and issues a warning.
[0029] Beneficial effects: Compared with the prior art, the advantages of this application are:
[0030] 1) The prediction model based on VMD-LSTM has high accuracy and good robustness. The accurate predictions can help identify problems in advance and avoid risks during on-site construction.
[0031] 2) The MPA-VMD-SSA-LSTM coupled prediction model proposed in this invention optimizes the number of modes K and the penalty factor β of VMD through MPA, and adaptively optimizes the hyperparameters of the LSTM network using the SSA algorithm, thus constructing a multimodal time series analysis framework of "decomposition-reconstruction-prediction". This model demonstrates significant advantages in predicting the construction quality of CRTSⅢ type track: based on the measured dataset, the RMSE of its prediction results is consistently below 0.3, and the coefficient of determination R ≥ 0.99, showing a significant improvement in prediction accuracy compared to the traditional LSTM model, and exhibiting good robustness. Accurate predictions can help identify problems in advance during on-site construction and avoid risks.
[0032] 3) The BIM-based construction quality management system enables the visualization of data, which is more intuitive than traditional paper records and can help construction personnel make immediate adjustments, thereby improving construction efficiency.
[0033] 4) The CRTSⅢ type slab track construction quality management method and system based on deep learning and BIM can improve the communication efficiency of current construction and make construction more information-based and intelligent. Attached Figure Description
[0034] Figure 1 This is a flowchart of the present invention.
[0035] Figure 2 This is a flowchart of the construction quality evaluation process of the present invention.
[0036] Figure 3 This is a flowchart of the construction quality prediction process of this invention.
[0037] Figure 4 This is a flowchart illustrating the visual implementation of the present invention.
[0038] Figure 5 This is a diagram of the quality management system architecture of this invention.
[0039] Figure 6 This is a construction quality scoring chart for CRTSⅢ type slab track.
[0040] Figure 7 This is a schematic diagram of the relative error of the prediction model.
[0041] Figure 8 A quality assessment chart for track slab positioning construction.
[0042] Figure 9 The logic diagram for the quality evaluation and prediction module is shown below.
[0043] Figure 10 This is a diagram illustrating the construction quality evaluation and prediction module. Detailed Implementation
[0044] The invention will now be further described with reference to the accompanying drawings.
[0045] A construction quality management method for CRTSⅢ type slab track based on deep learning and BIM includes the following steps:
[0046] Step S1: Identify key sub-processes in construction, establish a construction quality evaluation system for CRTSⅢ type slab track, and obtain corresponding indicator construction data;
[0047] S11. The WBS method is used to decompose the construction process of CRTSⅢ type slab track. First, the entire construction process is divided into 4 key construction stages, namely base plate construction, track slab construction, self-compacting concrete pouring, and rail laying. Then, the 4 construction stages are further subdivided to complete the construction decomposition of CRTSⅢ type slab track.
[0048] S12. Based on the design specifications, the key factors affecting the construction quality of CRTSⅢ type slab track at each stage of construction were identified.
[0049] S13. Comprehensively analyze the construction sub-processes and factors affecting construction quality, establish a construction quality evaluation index system for CRTSⅢ type slab track, and classify the evaluation levels of each index in conjunction with relevant standards, as shown in Table 1 below: Classification Table of Construction Quality Evaluation Index for CRTSⅢ Type Slab Track.
[0050] Step S2: Based on the CRTSⅢ type slab track construction quality evaluation system, the CRITIC-EWM combined weighting method is used to obtain the CRTSⅢ type slab track construction quality evaluation value.
[0051] S21. Clean and standardize the index data in the CRTSⅢ type slab track construction quality evaluation index system obtained in step S1.
[0052] S22. The information entropy and weights are calculated using the EWM method, and the information carrying capacity and weights are calculated using the CRITIC algorithm. The weights w calculated using the entropy-weight method are then... EWM With the weights w determined by the CRITIC method CRITIC The combination weight w is calculated using the formula: w = w EWM +(1-α)w CRITIC Based on expert opinions and existing literature, we take α = 0.5;
[0053] After data processing in steps S23 and S21, the scoring values for each indicator are obtained according to the evaluation level classification table. For example, in the rail laying stage, the scoring values for the seven indicators—gauge, alignment, elevation, levelness, torsion, deviation from design elevation, and deviation from design centerline—are first obtained. Then, the data is input into the EWM-CRITIC evaluation method to calculate the combined weights of the corresponding indicators. Combined with the scoring values, the construction quality score for rail laying can be obtained. Other stages are calculated similarly, starting from the lower-level indicators and working upwards, finally yielding the construction quality evaluation value for CRTSⅢ slab track. After obtaining the combined weights, the calculation of the construction quality evaluation value for CRTSⅢ slab track can begin. The evaluation value calculation uses a weighted summation method. The combined weights, combined with the evaluation level classification established above, yield an accurate construction quality score.
[0054] Taking the track slab positioning during the construction of CRTS III type slab track as an example. According to the evaluation index system mentioned above, the corresponding track slab positioning stage layer was found, and the original data of the index layer were collected as shown in Tables 1 and 2.
[0055] Table 1 Original data of track slab external dimensions
[0056]
[0057]
[0058] Table 2 Original Data of Track Slab Fine-tuning Index Layer
[0059]
[0060] Based on the established evaluation index system and the evaluation level classification table, the original construction data was converted into scoring data, and the results are shown in Tables 3 and 4.
[0061] Table 3 Scoring data for track slab external dimensions
[0062]
[0063] Table 4 Scoring data for track slab fine-tuning index layer
[0064]
[0065] The weight calculation results are shown in Tables 5 and 6.
[0066] Table 5 Calculation Results of Weighting of Track Slab Dimension Index Layer
[0067]
[0068] Table 6 Calculation Results of Weights for Track Slab Fine-tuning Index Layer
[0069]
[0070] According to the weighted calculation method, the score values are weighted according to the combined weight values to obtain the evaluation results of the next level, as shown in Table 7.
[0071] Table 7 Calculation Results of Track Slab Positioning Score
[0072]
[0073] Similarly, following the steps above, the track slab positioning score can be calculated as shown in Table 8. The results are as follows: Figure 8 As shown.
[0074] Table 8 Quality Scoring of Track Slab Positioning Construction
[0075]
[0076] Similarly, following the above calculation process, the construction quality scores for the base plate construction, self-compacting concrete pouring, and rail laying can be obtained respectively. Combining the construction quality scores of each sub-process above, the construction quality scores of 12 CRTSⅢ type slab track can be calculated. The construction quality of 12 CRTSⅢ type slab track is shown in Table 9.
[0077] Table 9. Construction Quality Scoring for CRTSⅢ Type Slab Track
[0078]
[0079] Step S3: Based on the SSA-LSTM architecture and combined with the MPA-VMD enhancement algorithm, construct a quality prediction model for the construction process of CRTSⅢ type slab track, and predict subsequent evaluation values based on existing evaluation values.
[0080] S31. Arrange the construction quality evaluation values of CRTSⅢ type slab track obtained in step S2 in chronological order and use them as input values.
[0081] S32. The model uses the MPA algorithm to find the optimal number of decomposition layers K and the penalty factor β. The optimal parameters are then substituted into the VMD algorithm to decompose the construction quality scoring dataset, and finally n modal components are obtained.
[0082] S33. The decomposed modal components are fed into the SSA-LSTM model for prediction. In the SSA-LSTM model, the SSA algorithm will conduct a comprehensive search for the key parameters of the LSTM model. Then, the optimal parameters obtained from the search are substituted into the LSTM model. After obtaining the predicted values of all modal components, the predicted values are superimposed to obtain the complete prediction sequence and output the prediction result.
[0083] Using the existing evaluation model proposed above, 300 sets of scoring data were calculated to form a construction quality scoring dataset. The prediction performance of the MPA-VMD-SSA-LSTM prediction model was then validated, and the results are as follows:
[0084] The MPA-VMD-SSA-LSTM prediction model has a maximum relative error of only 0.806%, and the MAE, MRE, MSE, RMSE, coefficient of determination, and Hill's inequality coefficient are 0.217983, 0.002803, 0.080905, 0.284438, 0.99768, and 0.003571, respectively. It can be seen that the prediction effect is excellent and can achieve accurate prediction of construction quality.
[0085] Step S4: Convert the BIM model to GLTF format and display it on the web, showing the construction quality evaluation value and predicted value in the corresponding area of the BIM model, such as... Figure 9 As shown:
[0086] S41. Create a BIM model of CRTSⅢ type slab track and substructure in Revit;
[0087] S42. Obtain the evaluation and prediction values obtained from steps S2 and S3, and map them to the orbital structure;
[0088] S43. Perform secondary development on Revit to convert the Type III slab model from .rvt format to .gltf format;
[0089] S44. By using the `material.color` property of the Three.js framework to assign colors to the model corresponding to the quality level, the quality data can be displayed in a model-based manner. This enables precise and efficient management of the construction quality of CRTSⅢ type slab track, such as... Figure 10 As shown.
[0090] A construction quality management system for CRTSⅢ type slab track based on deep learning and BIM includes a project introduction module, a construction quality evaluation module, a construction quality prediction module, a non-conforming CRTSⅢ type slab management module, and a 3D display module. The project introduction module provides relevant project information, the construction quality evaluation module evaluates the project based on the project parameters provided by the project introduction module, the construction quality prediction module provides a construction quality prediction based on the evaluation information, and if the project is found to be non-conforming, the non-conforming CRTSⅢ type slab management module identifies the reasons for the non-conformity and issues a warning.
[0091] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A construction quality management method for CRTSⅢ type slab track based on deep learning and BIM, characterized in that, Includes the following steps: Step S1: Identify key construction sub-processes, establish a CRTSⅢ type slab track construction quality evaluation system, and obtain corresponding indicator construction data. Step S2: Based on the CRTSⅢ type slab track construction quality evaluation system, the CRITIC-EWM combined weighting method is used to obtain the CRTSⅢ type slab track construction quality evaluation value. Step S3: Based on the SSA-LSTM architecture and combined with the MPA-VMD enhancement algorithm, construct a quality prediction model for the construction process of CRTSⅢ type slab track, and predict subsequent evaluation values based on existing evaluation values. Step S4: Convert the BIM model to GLTF format, display it on the web, and display the construction quality evaluation value and predicted value in the corresponding area of the BIM model.
2. The CRTSⅢ type slab track construction quality management method based on deep learning and BIM according to claim 1, characterized in that, Step S1 includes the following steps: S11. The WBS method is used to decompose the construction process of CRTSⅢ type slab track. First, the entire construction process is divided into 4 key construction stages, namely base plate construction, track slab construction, self-compacting concrete pouring, and rail laying. Then, the 4 construction stages are further subdivided to complete the construction decomposition of CRTSⅢ type slab track. S12. Based on the design specifications, the key factors affecting the construction quality of CRTSⅢ type slab track at each stage of construction were identified. S13. Conduct a comprehensive analysis of the construction sub-processes and factors affecting construction quality, establish a construction quality evaluation index system for CRTSⅢ type slab track, and establish an evaluation level classification table for each index in conjunction with relevant standards.
3. The CRTSⅢ type slab track construction quality management method based on deep learning and BIM according to claim 1, characterized in that, Step S2 includes the following steps: S21. Clean and standardize the index data in the CRTSⅢ type slab track construction quality evaluation index system obtained in step S1. S22. The information entropy and weights are calculated using the EWM method, and the information carrying capacity and weights are calculated using the CRITIC algorithm. The weights w calculated using the entropy-weighting method are then... EWM With the weights w determined by the CRITIC method CRITIC The combination weight w is calculated using the formula: w = w EWM +(1-α)w CRITIC α = 0.5; S23. Obtain the data scores of each indicator according to the evaluation level classification table, then substitute the data into the EWM-CRITIC evaluation method to calculate the combined weight of the corresponding indicators. Combine the scores to obtain the construction quality score of rail laying. Calculate the construction quality score of each stage in this way. Combine the combined weights with the evaluation level classification established above to obtain the accurate construction quality score.
4. The construction quality management method for CRTSⅢ type slab track based on deep learning and BIM according to claim 1, characterized in that, Step S3 includes the following steps: S31. Arrange the construction quality evaluation values of CRTSⅢ type slab track obtained in step S2 in chronological order and use them as input values. S32. The model uses the MPA algorithm to find the optimal number of decomposition layers K and the penalty factor β. The optimal parameters are then substituted into the VMD algorithm to decompose the construction quality scoring dataset, and finally n modal components are obtained. S33. The decomposed modal components are fed into the SSA-LSTM model for prediction. In the SSA-LSTM model, the SSA algorithm performs a comprehensive search on the key parameters of the LSTM model. Then, the optimal parameters obtained from the search are substituted into the LSTM model. After obtaining the predicted values of all modal components, the predicted values are superimposed to obtain the complete prediction sequence and output the prediction result.
5. The CRTSⅢ type slab track construction quality management method based on deep learning and BIM according to claim 1, characterized in that, Step S4 includes the following steps: S41. Create a BIM model of CRTSⅢ type slab track and substructure in Revit; S42. Obtain the evaluation and prediction values obtained from steps S2 and S3, and map them to the orbital structure; S43. Perform secondary development on Revit to convert the Type III slab model from .rvt format to .gltf format; S44. By using the material.color property in the Three.js framework to assign colors to the model corresponding to the quality level, the quality data can be displayed in a model-based manner, thereby enabling precise and efficient management of the construction quality of CRTSⅢ type slab track.
6. A CRTSⅢ type slab track construction quality management system based on deep learning and BIM as described in any one of claims 1 to 5, characterized in that: It includes a project introduction module, a construction quality evaluation module, a construction quality prediction module, a non-conforming CRTSⅢ slab management module, and a 3D display module. The project introduction module provides relevant project information, the construction quality evaluation module evaluates the project based on the project parameters provided by the project introduction module, the construction quality prediction module provides a construction quality prediction based on the evaluation information, and if the project is found to be non-conforming, the non-conforming CRTSⅢ slab management module identifies the reasons for the non-conformity and issues a warning.