技术领域technical field
本发明属于水利工程中混凝土振捣施工领域，具体涉及一种基于混合异构数据源的振捣质量评价方法。The invention belongs to the field of concrete vibrating construction in hydraulic engineering, and particularly relates to a vibrating quality evaluation method based on mixed heterogeneous data sources.
背景技术Background technique
在大坝混凝土浇筑过程中，振捣是最核心的工艺，然而，由于混凝土振捣密实机理复杂，目前仍没有统一的理论能够定量分析振捣作业是否满足要求。传统振捣质量控制手段是通过专业的施工人员监测振捣施工过程，控制振捣棒插入位置、插入深度、振捣时间等参数，然后钻芯取样测量混凝土抗压强度，以此控制混凝土振捣质量。但以上方法存在以下缺陷：(1)振捣参数的控制大多数依靠现场施工人员的监测，主观性较强，无法保证参数控制的准确性；(2)有限的测样点不能反映出整个仓面的振捣作业质量；(3)无法实时获得混凝土抗压强度，如果出现振捣质量不合格的情况无法及时补救。因此，有必要研究出一种实时、连续的混凝土振捣施工质量监控和动态评价方法，以实现全仓面混凝土振捣质量的精细化控制，这对于确保后期水利枢纽安全运行具有十分重要的意义。In the process of dam concrete pouring, vibrating is the core technology. However, due to the complicated mechanism of concrete vibrating and compacting, there is still no unified theory that can quantitatively analyze whether the vibrating operation meets the requirements. The traditional vibrating quality control method is to monitor the vibrating construction process by professional construction personnel, control parameters such as the insertion position, insertion depth, and vibrating time of the vibrating rod, and then drill core samples to measure the compressive strength of concrete, so as to control the concrete vibrating. quality. However, the above methods have the following defects: (1) The control of vibration parameters mostly relies on the monitoring of on-site construction personnel, which is highly subjective and cannot guarantee the accuracy of parameter control; (2) The limited sampling points cannot reflect the entire warehouse. (3) The compressive strength of concrete cannot be obtained in real time, and if the vibrating quality is unqualified, it cannot be remedied in time. Therefore, it is necessary to develop a real-time and continuous concrete vibrating construction quality monitoring and dynamic evaluation method to realize the fine control of the concrete vibrating quality of the whole warehouse surface, which is of great significance to ensure the safe operation of the later water conservancy project. .
目前，在混凝土振捣施工质量控制以及其表观质量分析方面已经拥有众多研究成果。在振捣施工实时监控方面，为实现混凝土振捣质量的精准量化评价，田正宏等研发出了混凝土振捣效果的3D可视化系统；Gong等利用UWB定位系统对振捣棒尖端进行精确定位，开发了一种可视化系统实时监控混凝土振捣情况。Tian等基于GPS等技术开发了混凝土振捣作业实时监控系统，通过实时监测的振捣轨迹、振捣时间、作用半径等数据对振捣情况进行可视化并反馈。上述研究以单个振捣棒为监控对象，无法满足大坝混凝土连续高强度浇筑的要求。钟桂良等为满足大坝连续高强度施工的要求，研发出振捣可视化监控系统，实时监控振捣台车多个振捣棒的插入位置、倾角、振捣时间以及插入深度等参数。上述研究成果均能实现振捣参数的实时监控，但振捣质量是否达标大多还是依靠专业人员主观判断，难免有失客观性。At present, there have been many research results in the quality control of concrete vibrating construction and the analysis of its apparent quality. In terms of real-time monitoring of vibrating construction, in order to achieve accurate quantitative evaluation of concrete vibrating quality, Tian Zhenghong et al. developed a 3D visualization system for concrete vibrating effect; Gong et al. A visualization system is developed to monitor the concrete vibration in real time. Tian et al. developed a real-time monitoring system for concrete vibrating operations based on technologies such as GPS, and visualized and fed back the vibrating conditions through real-time monitoring of the vibrating trajectory, vibrating time, and radius of action. The above research uses a single vibrator as the monitoring object, which cannot meet the requirements of continuous high-strength pouring of dam concrete. In order to meet the requirements of continuous high-strength construction of dams, Zhong Guiliang et al. developed a vibrating visualization monitoring system to monitor the insertion position, inclination angle, vibrating time and insertion depth of multiple vibrating rods of the vibrating trolley in real time. The above research results can realize real-time monitoring of vibrating parameters, but whether the vibrating quality meets the standard mostly depends on the subjective judgment of professionals, which inevitably loses objectivity.
在混凝土表观质量评价方面，Liu等基于图像处理技术对混凝土表观缺陷进行检测，进而评价混凝土表面质量；李琛等结合深度学习和数字图像处理技术，提出一种混凝土表观缺陷的检测方法；王佳俊等提出一种基于卷积神经网络的混凝土振捣表面质量识别方法，用于辅助振捣施工决策；王超等利用LFNet代替传统的卷积神经网络对混凝土表观裂隙进行识别，实现精准评价分析；Cui等将卷积神经网络和物联网技术相结合，提出一个实时监控框架，确保混凝土的振捣质量。以上研究大多集中于事后的混凝土裂缝识别、表观质量评估等方面，缺少振捣过程中实时评价的研究。In the evaluation of concrete apparent quality, Liu et al. detected the apparent defects of concrete based on image processing technology, and then evaluated the surface quality of concrete; Li Chen et al. combined deep learning and digital image processing technology, and proposed a detection method of concrete apparent defects ; Wang Jiajun et al. proposed a method for identifying surface quality of concrete vibrating based on convolutional neural network, which is used to assist vibrating construction decision-making; Wang Chao et al. used LFNet instead of traditional convolutional neural network to identify apparent cracks in concrete to achieve accurate Evaluation and analysis; Cui et al. combined convolutional neural network and Internet of Things technology to propose a real-time monitoring framework to ensure the vibrating quality of concrete. Most of the above studies focus on post-event concrete crack identification and apparent quality assessment, and lack of research on real-time evaluation during the vibrating process.
综上所述，目前针对水利工程中的混凝土振捣作业，主要集中在对振捣参数的监控；针对混凝土表面质量评价方面，大多通过图像处理技术对混凝土表观质量进行评价，很难对振捣施工过程进行实时控制，而且缺少将图像数据和数值数据二者相结合的异构数据的有关研究，数据的格式、形态差距愈大，相互之间的信息相关性、重复性、冗余性愈少，互补性愈强，如果不能对施工过程中异构数据源进行有效挖掘，不但会造成资源的浪费，也会影响施工评价结果。To sum up, for the concrete vibrating operation in water conservancy projects, the monitoring of the vibrating parameters is mainly concentrated; for the evaluation of the concrete surface quality, the image processing technology is mostly used to evaluate the apparent quality of the concrete, and it is difficult to evaluate the vibration parameters. The construction process is controlled in real time, and there is a lack of research on heterogeneous data combining image data and numerical data. , the stronger the complementarity, if the heterogeneous data sources in the construction process cannot be effectively excavated, it will not only cause a waste of resources, but also affect the construction evaluation results.
发明内容SUMMARY OF THE INVENTION
本发明为解决公知技术中存在的技术问题而提供一种基于混合异构数据源的振捣质量评价方法，该方法给出的振捣质量评价结果更加客观、准确。In order to solve the technical problems existing in the known technology, the present invention provides a vibrating quality evaluation method based on mixed heterogeneous data sources, and the vibrating quality evaluation result provided by the method is more objective and accurate.
本发明为解决公知技术中存在的技术问题所采取的技术方案是：一种基于混合异构数据源的振捣质量评价方法，采用以下步骤：The technical scheme adopted by the present invention for solving the technical problems existing in the known technology is: a vibration quality evaluation method based on mixed heterogeneous data sources, adopting the following steps:
1)收集已有工程的混凝土振捣数据，包括振捣施工过程数据和其对应时刻的混凝土表面图像数据，根据收集数据、参照相关规范对混凝土振捣质量进行评价，划分为合格和不合格二个等级；1) Collect the concrete vibrating data of the existing project, including the vibrating construction process data and the concrete surface image data at the corresponding time, and evaluate the concrete vibrating quality according to the collected data and referring to the relevant specifications, and divide it into qualified and unqualified two. grades;
2)采用卷积神经网络对混凝土表面图像数据进行特征提取，获得图像特征；2) Convolutional neural network is used to extract features from concrete surface image data to obtain image features;
3)将步骤2)中的图像特征与步骤1)中的振捣施工过程数据进行特征组合，获得组合特征数据；3) combining the image features in step 2) with the vibrating construction process data in step 1) to obtain combined feature data;
4)将步骤3)中的组合特征数据划分为训练集和测试集，训练集和测试集中均应包括合格和不合格二个等级；4) Divide the combined feature data in step 3) into a training set and a test set, and both the training set and the test set should include two levels of qualified and unqualified;
5)利用网格搜索法确定SVM模型的设置参数；5) Utilize the grid search method to determine the setting parameters of the SVM model;
6)将训练集中的组合特征输入到SVM模型中对其进行训练，直至精度不再提高，然后采用测试集的数据进行测试；6) Input the combined features in the training set into the SVM model for training until the accuracy is no longer improved, and then use the data of the test set for testing;
7)如果测试精度满足要求，获得振捣质量评价模型SVM，如果测试精度不满足要求，返回步骤5)调整参数，重复步骤5)～6)，直至测试精度满足要求获得振捣质量评价模型SVM；7) If the test accuracy meets the requirements, obtain the vibration quality evaluation model SVM, if the test accuracy does not meet the requirements, return to step 5) to adjust the parameters, and repeat steps 5) to 6) until the test accuracy meets the requirements to obtain the vibration quality evaluation model SVM. ;
8)实时采集现场混凝土振捣数据，包括振捣施工过程数据和混凝土表面图像数据，并采用卷积神经网络对混凝土表面图像数据进行特征提取，获得图像特征，图像特征与振捣施工过程数据进行组合后，输入振捣质量评价模型SVM中，振捣质量评价模型SVM实时输出振捣质量评价结果。8) Collect on-site concrete vibrating data in real time, including vibrating construction process data and concrete surface image data, and use convolutional neural network to extract features from the concrete surface image data to obtain image features. Image features are compared with vibrating construction process data. After the combination, input the vibration quality evaluation model SVM, and the vibration quality evaluation model SVM outputs the vibration quality evaluation results in real time.
所述步骤4)，训练集与测试集的数据比例为4:1。In the step 4), the data ratio of the training set and the test set is 4:1.
本发明具有的优点和积极效果是：通过卷积神经网络提取混凝土振捣表面图像特征，将所提取出的特征与振捣施工数据组合，输入振捣质量评价模型SVM中进行计算和分析，进而实时评价混凝土振捣施工是否合格。将混凝土图像数据和施工数据结合起来进行分析评价，采用多源异构数据融合的方法，实现振捣施工质量的实时控制，可以提高施工水平。不需要将图像型数据和数值型数据分别分析，减小了评价模型构建的工作量，实现了异构数据之间的互补，使得振捣质量评价结果更加客观、准确。The advantages and positive effects of the present invention are: extracting the surface image features of concrete vibrating through the convolutional neural network, combining the extracted features with the vibrating construction data, inputting the vibrating quality evaluation model SVM for calculation and analysis, and then Real-time evaluation of whether the concrete vibrating construction is qualified. The concrete image data and construction data are combined for analysis and evaluation, and the method of multi-source heterogeneous data fusion is adopted to realize the real-time control of the vibrating construction quality, which can improve the construction level. There is no need to analyze image data and numerical data separately, which reduces the workload of evaluating model construction, realizes the complementarity between heterogeneous data, and makes vibrating quality evaluation results more objective and accurate.
具体实施方式Detailed ways
为能进一步了解本发明的发明内容、特点及功效，兹例举以下实施例进行详细说明如下：In order to further understand the content of the invention, features and effects of the present invention, the following examples are hereby described in detail as follows:
一种基于混合异构数据源的振捣质量评价方法，采用以下步骤：A vibrating quality evaluation method based on mixed heterogeneous data sources, adopts the following steps:
1)收集已有工程的混凝土振捣数据，包括振捣施工过程数据和其对应时刻的混凝土表面图像数据；根据收集数据、参照相关规范对混凝土振捣质量进行评价，划分为合格和不合格二个等级；1) Collect the concrete vibrating data of the existing project, including the vibrating construction process data and the concrete surface image data at the corresponding moment; evaluate the concrete vibrating quality according to the collected data and referring to the relevant specifications, and divide it into qualified and unqualified two. grades;
振捣施工数据与混凝土表面图像数据时间节点一一对应；振捣施工过程数据包括混凝土级配、塌落度、振捣时间和振捣深度、角度等。The vibrating construction data corresponds to the time nodes of the concrete surface image data one by one; the vibrating construction process data includes concrete gradation, slump, vibrating time, vibrating depth, angle, etc.
2)采用卷积神经网络对混凝土表面图像数据进行特征提取，获得图像特征(X1，X2，…，Xn)；2) Using a convolutional neural network to perform feature extraction on the concrete surface image data to obtain image features (X1 , X2 , . . . , Xn );
3)将步骤2)中的图像特征(X1，X2，…，Xn)与步骤1)中的振捣施工过程数据(Y1，Y2，…，Ym)进行特征组合，获得组合特征(X1，X2，…，Xn，Y1，Y2，…，Ym)；3) Feature combination of the image features (X1 , X2 , . . . , Xn ) in step 2) and thevibrating construction process data (Y1 , Y2 , . combined features (X1 , X2 , ..., Xn , Y1 , Y2 , ..., Ym );
4)将步骤3)中的组合特征数据划分为训练集和测试集，训练集和测试集中均应包括合格和不合格二个等级，训练集与测试集的数据比例建议为4:1，其中训练集用于训练SVM模型，测试集不参与模型的训练，用于评估SVM模型性能；4) Divide the combined feature data in step 3) into a training set and a test set. Both the training set and the test set should include two levels of qualified and unqualified. The data ratio of the training set and the test set is recommended to be 4:1, where The training set is used to train the SVM model, and the test set does not participate in the training of the model and is used to evaluate the performance of the SVM model;
5)利用网格搜索法确定SVM模型的设置参数；5) Utilize the grid search method to determine the setting parameters of the SVM model;
6)将训练集中的组合特征(X1，X2，…，Xn，Y1，Y2，…，Ym)输入到SVM模型中对其进行训练，直至精度不再提高，然后采用测试集的数据进行测试；6) Input the combined features (X1 , X2 , ..., Xn , Y1 , Y2 , ..., Ym ) in the training set into the SVM model to train it until the accuracy no longer improves, and then use the test set data for testing;
7)如果测试精度满足要求，获得振捣质量评价模型SVM；如果测试精度不满足要求，返回步骤5)，重复步骤5)～7)，直至测试精度满足要求获得振捣质量评价模型SVM；7) If the test accuracy meets the requirements, obtain the vibration quality evaluation model SVM; if the test accuracy does not meet the requirements, return to step 5), and repeat steps 5) to 7) until the test accuracy meets the requirements to obtain the vibration quality evaluation model SVM;
8)在振捣施工开始后，实时采集现场混凝土振捣数据，包括振捣施工过程数据和混凝土表面图像数据，并采用卷积神经网络对混凝土表面图像数据进行特征提取，获得图像特征，将图像特征与振捣施工过程数据进行组合后，输入到振捣质量评价模型SVM中，振捣质量评价模型SVM实时输出振捣质量评价结果。8) After the vibrating construction starts, collect the concrete vibrating data on site in real time, including the vibrating construction process data and the concrete surface image data, and use the convolutional neural network to extract the features of the concrete surface image data to obtain the image features, and convert the image data into the image data. After the characteristics are combined with the vibration construction process data, they are input into the vibration quality evaluation model SVM, and the vibration quality evaluation model SVM outputs the vibration quality evaluation results in real time.
尽管上面对本发明的优选实施例进行了描述，但是本发明并不局限于上述的具体实施方式，上述的具体实施方式仅仅是示意性的，并不是限制性的，本领域的普通技术人员在本发明的启示下，在不脱离本发明宗旨和权利要求所保护的范围的情况下，还可以做出很多形式，这些均属于本发明的保护范围之内。Although the preferred embodiments of the present invention have been described above, the present invention is not limited to the above-mentioned specific embodiments. The above-mentioned specific embodiments are only illustrative and not restrictive. Under the inspiration of the invention, many forms can be made without departing from the scope of the invention and the protection scope of the claims, which all belong to the protection scope of the invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210527157.4ACN115170466A (en) | 2022-05-16 | 2022-05-16 | Vibration quality evaluation method based on mixed heterogeneous data source |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210527157.4ACN115170466A (en) | 2022-05-16 | 2022-05-16 | Vibration quality evaluation method based on mixed heterogeneous data source |
| Publication Number | Publication Date |
|---|---|
| CN115170466Atrue CN115170466A (en) | 2022-10-11 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210527157.4APendingCN115170466A (en) | 2022-05-16 | 2022-05-16 | Vibration quality evaluation method based on mixed heterogeneous data source |
| Country | Link |
|---|---|
| CN (1) | CN115170466A (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107255637A (en)* | 2017-06-06 | 2017-10-17 | 河海大学 | A kind of grinding coagulation soil compactness detection method based on laser image |
| CN107683469A (en)* | 2015-12-30 | 2018-02-09 | 中国科学院深圳先进技术研究院 | A kind of product classification method and device based on deep learning |
| CN110298830A (en)* | 2019-06-24 | 2019-10-01 | 天津大学 | Fresh concrete based on convolutional neural networks vibrates apparent mass recognition methods |
| CN112288221A (en)* | 2020-09-22 | 2021-01-29 | 广东博智林机器人有限公司 | Vibration quality detection method, system and device and electronic equipment |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107683469A (en)* | 2015-12-30 | 2018-02-09 | 中国科学院深圳先进技术研究院 | A kind of product classification method and device based on deep learning |
| CN107255637A (en)* | 2017-06-06 | 2017-10-17 | 河海大学 | A kind of grinding coagulation soil compactness detection method based on laser image |
| CN110298830A (en)* | 2019-06-24 | 2019-10-01 | 天津大学 | Fresh concrete based on convolutional neural networks vibrates apparent mass recognition methods |
| CN112288221A (en)* | 2020-09-22 | 2021-01-29 | 广东博智林机器人有限公司 | Vibration quality detection method, system and device and electronic equipment |
| Title |
|---|
| 赵青林;周明凯;王传美;郑劲;: "基于图像模式识别的混凝土二维显微结构研究进展及问题分析", 混凝土, no. 09, 27 September 2006 (2006-09-27)* |
| Publication | Publication Date | Title |
|---|---|---|
| Fan et al. | Machine learning-based classification of quality grades for concrete vibration behaviour | |
| CN106228536B (en) | The earth and rockfill dam dam material grain composition method of inspection based on Digital Image Processing | |
| CN107038524B (en) | Comprehensive evaluation method for construction quality of roller compacted concrete dam considering parameter uncertainty | |
| CN117309661A (en) | Concrete quality on-line measuring system | |
| CN113640380B (en) | Multi-level classification method and system for rail damage detection | |
| CN108153989B (en) | Rapid prediction method of compaction quality of concrete dam considering the influence of parameter uncertainty | |
| CN106990018A (en) | An Intelligent Identification Method of Grouting Density of Prestressed Concrete Beams | |
| CN111881495A (en) | Prestress steel structure safety assessment method based on digital twinning | |
| CN106600153A (en) | Tailing pond dam break risk evaluating method | |
| CN107255637A (en) | A kind of grinding coagulation soil compactness detection method based on laser image | |
| CN110619483A (en) | Tunnel surrounding rock grade dynamic change and decision-making method based on multi-source data fusion analysis | |
| CN108509696A (en) | Ancient building health monitoring method and device based on three-dimensional laser scanning technology | |
| CN119377765A (en) | Method and system for inspecting appearance quality of pipe pile production | |
| Lorenzi et al. | Investigation of the potential for evaluation of concrete flaws using nondestructive testing methods | |
| CN103837533B (en) | Based on concrete temperature monitoring and the emulating bakce analysis method of thermal imaging system | |
| CN118774969B (en) | Tunnel excavation support management method and system based on advanced geological prediction information | |
| CN115170466A (en) | Vibration quality evaluation method based on mixed heterogeneous data source | |
| CN118189834B (en) | A method and system for detecting the gap thickness of CFST arch bridge based on cooling excitation intensity | |
| WO2025152290A1 (en) | Stone surface defect detection method and system | |
| CN118571381A (en) | A lightning damage assessment method for composite materials based on neural network | |
| Li et al. | Prediction of internal defect area in wooden components by stress wave velocity analysis | |
| CN111160490A (en) | Deep learning dangerous rock deformation prediction method and device based on multiple time sequences | |
| CN112199889B (en) | Drilling data analysis and evaluation method | |
| CN117288587A (en) | Rock tensile strength while drilling test method and system | |
| CN117309999A (en) | Method and system for detecting compaction performance of concrete in arch bridge pipe based on complementary fusion of multisource acoustic parameter model |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |