CN-121980664-A - Intelligent pre-assembly method for shallow circular bin sliding mode template based on multi-source data fusion

CN121980664ACN 121980664 ACN121980664 ACN 121980664ACN-121980664-A

Abstract

The invention relates to the technical field of pre-assembly, and discloses an intelligent pre-assembly method of a shallow circular bin sliding mode template based on multi-source data fusion, which comprises the following steps: performing geometric registration on the point cloud data set and the BIM data set to generate a registered point cloud data set, correlating physical features and dynamic features with the registered point cloud data set to obtain a joint feature vector set, and weighting the registered point cloud data set to obtain a weighted point cloud data set. The method captures the association among construction actions through neighbor node aggregation to support policy generation, then generates the policy by using an efficiency, precision and safety multi-objective rewarding function driving policy, finally forms a progressive verification system through error bench quantity verification, stress verification and dynamic verification, comprehensively guarantees the feasibility of the scheme from static compliance to dynamic stability, breaks through the limitation of the traditional dependence on a single model, and realizes the high precision, high safety and strong adaptability of construction pre-assembly.

Inventors

GAO WEIJIAN
Yan Guobi
WU QILONG
GUO JING
CHEN XINGHUA
Shi Zongbu

Assignees

中铁城建集团第二工程有限公司

Dates

Publication Date: 20260505
Application Date: 20260408

Claims (10)

1. An intelligent pre-assembly method for a shallow circular bin sliding mode template based on multi-source data fusion is characterized by comprising the following steps: S101, acquiring multi-source data, analyzing and cleaning the multi-source data to obtain a standardized data set, and extracting geometric, physical and dynamic characteristics from the standardized data set, wherein the standardized data set comprises a point cloud data set, a BIM data set and a drawing geometric data set; S102, performing geometric registration on a point cloud data set and a BIM data set to generate a registered point cloud data set, associating physical features and dynamic features with the registered point cloud data set to obtain a combined feature vector set, and weighting the registered point cloud data set to obtain a weighted point cloud data set; S103, acquiring material parameters and combining physical characteristics to form mechanical constraint, constructing a process causal graph according to a combined characteristic vector set and a construction sequence, converting template parameters into legal symbol sets on the basis of the mechanical constraint, and designing a reward function to obtain an initial assembly strategy; s104, carrying out multidimensional verification on the initial assembly strategy by combining mechanical constraint to obtain the final assembly strategy.
2. The intelligent pre-assembly method of the shallow circle warehouse slip form template based on multi-source data fusion according to claim 1, wherein the geometric registration of the point cloud data set and the BIM data set is carried out, and the specific steps are as follows: extracting normal vectors from the point cloud data set, calculating curvature, screening out high-curvature points as point cloud characteristic points, and obtaining a point cloud characteristic point set; extracting BIM characteristic points by applying a template plane equation based on the BIM data set to obtain a BIM characteristic point set; randomly selecting point cloud features, searching corresponding nearest neighbor points in BIM feature points, and obtaining a rough registered point cloud data set through a transformation matrix; Gradually BIM data sets on the basis of the roughly registered point cloud data sets to obtain initial point pair sets, restraining and filtering the initial point pair sets to generate constraint and filter point pair sets, and obtaining registered point cloud data sets by minimizing an objective function: Wherein R is a rotation matrix, t is a translation vector, For updating the point pair set after each iteration, For points in the coarsely registered point cloud dataset, Is a point in the BIM dataset.
3. The intelligent pre-assembly method of the shallow circular warehouse slipform template based on multi-source data fusion according to claim 1, wherein the associating the physical features and the dynamic features with the registered point cloud data set comprises: Extracting splicing key points of the templates from the registered point cloud data sets to obtain a key point set, wherein the key point set comprises three-dimensional coordinates, normal vectors and curvature values of the key points; Calculating the wind load component according to the normal vector, and calculating the thermal deformation according to the thermal expansion coefficient and the temperature difference of the template material; And combining the key point set, the concrete lateral pressure, the wind load component and the thermal deformation quantity to form a combined characteristic vector set.
4. The intelligent pre-assembly method of the shallow circular warehouse slipform template based on multi-source data fusion according to claim 1, wherein the construction process causal graph is constructed according to a combined feature vector set and a construction sequence, and the method comprises the following steps: each construction action is regarded as a node, and physical attributes in the construction actions and key point features in the combined feature vector set are bound with the node to generate a node set; defining a precedence dependence relationship in construction work according to the construction sequence, defining the precedence dependence relationship in construction work as an edge, and carrying out weight calculation to obtain an edge set; And connecting the node set with the edge set to obtain the process causal graph.
5. The intelligent pre-assembly method of the shallow circular warehouse slipform template based on multi-source data fusion according to claim 4, wherein the construction process causal graph is constructed according to a combined feature vector set and a construction sequence, and the method further comprises: Initializing node characteristics in the process causal graph, carrying out neighbor node aggregation on each node, and collecting characteristic vectors of neighbor nodes: wherein: Is a set of neighbor nodes of a node, In order for the attention coefficient to be a factor of attention, As a matrix of parameters, To activate the function.
6. The intelligent pre-assembly method of a shallow circular warehouse slip form template based on multi-source data fusion according to claim 1, wherein the designing of the reward function comprises: Dividing the reward function into an efficiency term, an accuracy term and a safety term respectively, wherein the efficiency term is defined as the reciprocal of assembly time, the accuracy term is defined as the reciprocal of deviation, and the safety term is defined as the difference value between allowable stress and actual stress; and respectively carrying out weighted summation on the efficiency item, the precision item and the safety item to obtain the rewarding value.
7. The intelligent pre-assembly method for the shallow circular warehouse slip form template based on multi-source data fusion according to claim 6, wherein the initial assembly strategy comprises the following steps: splitting the splicing action to obtain discrete action and continuous action; Encoding the assembly sequence into integers for discrete actions, setting fine adjustment displacement for continuous actions, and combining the discrete actions and the continuous actions into action vectors; Generating an initial strategy based on the current splicing state, and measuring the relative income of a certain splicing action in the current splicing state: wherein: For the expected return of the selection action a in the current splice state s, The reference value of the current assembly state s; and updating the initial strategy by calculating the relative benefits and adopting a maximized objective function to obtain the initial assembly strategy.
8. The intelligent pre-assembly method of the shallow circular warehouse slipform template based on multi-source data fusion according to claim 7, wherein the updating the initial strategy by using the maximized objective function comprises: wherein: for the action probability of the new policy, Representing the probability of an action of the old policy, Is a clip parameter.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and running on the processor, wherein the processor implements the intelligent pre-assembly method of the squat silo slipform template based on multi-source data fusion of any of claims 1 to 8 when the computer program is executed by the processor.
10. A computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and the computer program is executed to implement the intelligent pre-assembly method of the shallow circular warehouse sliding mode template based on multi-source data fusion according to any one of claims 1 to 8.

Description

Intelligent pre-assembly method for shallow circular bin sliding mode template based on multi-source data fusion Technical Field The invention relates to the technical field of pre-assembly, in particular to an intelligent pre-assembly method of a shallow circular bin sliding mode template based on multi-source data fusion. Background The shallow circular bin is a building structure commonly used for storing bulk materials, and is characterized in that the bottom surface is circular or polygonal, the height is relatively low, and in the construction process, the intelligent pre-assembly technology for construction of building materials is widely applied. Meanwhile, the movable shuttering technology is also introduced into the construction of the shallow circular bin, so that the position of the die can be flexibly adjusted during concrete pouring, and the accuracy and stability of the structure are ensured. The modern construction methods effectively improve the construction quality and efficiency of the shallow circular warehouse. However, in the prior art, due to the fact that the multi-source data collaboration is insufficient, the traditional pre-assembly technology depends on a single data source, geometrical information and physical characteristics are disjointed, real-time disturbance of dynamic working conditions is difficult to fuse, and decision bias is generated, meanwhile, the traditional geometrical registration method such as the traditional ICP algorithm often ignores physical constraint, is difficult to deal with high-noise point cloud data, and causes registration error accumulation, and the reliability of subsequent analysis is affected, the traditional decision model is difficult to adapt to complex scene change based on fixed rules or a single physical model, and cannot dynamically balance multi-objective optimization requirements of efficiency, precision and safety, in addition, the traditional verification means mostly depends on a single mode, lacks full-link risk pre-judging capability, and is easy to ignore micro dynamics effects. Disclosure of Invention In order to overcome the defects in the prior art, the embodiment of the invention provides an intelligent pre-assembling method for a shallow circular bin sliding mode template based on multi-source data fusion. In order to achieve the above purpose, the present invention provides the following technical solutions: an intelligent pre-assembly method of a shallow circular bin sliding mode template based on multi-source data fusion, comprising the following steps: S101, acquiring multi-source data, analyzing and cleaning the multi-source data to obtain a standardized data set, and extracting geometric, physical and dynamic characteristics from the standardized data set, wherein the standardized data set comprises a point cloud data set, a BIM data set and a drawing geometric data set; S102, performing geometric registration on a point cloud data set and a BIM data set to generate a registered point cloud data set, associating physical features and dynamic features with the registered point cloud data set to obtain a combined feature vector set, and weighting the registered point cloud data set to obtain a weighted point cloud data set; S103, acquiring material parameters and combining physical characteristics to form mechanical constraint, constructing a process causal graph according to a combined characteristic vector set and a construction sequence, converting template parameters into legal symbol sets on the basis of the mechanical constraint, and designing a reward function to obtain an initial assembly strategy; s104, carrying out multidimensional verification on the initial assembly strategy by combining mechanical constraint to obtain the final assembly strategy. Further, the geometric registration of the point cloud data set and the BIM data set comprises the following specific steps: extracting normal vectors from the point cloud data set, calculating curvature, screening out high-curvature points as point cloud characteristic points, and obtaining a point cloud characteristic point set; extracting BIM characteristic points by applying a template plane equation based on the BIM data set to obtain a BIM characteristic point set; randomly selecting point cloud features, searching corresponding nearest neighbor points in BIM feature points, and obtaining a rough registered point cloud data set through a transformation matrix; Gradually BIM data sets on the basis of the roughly registered point cloud data sets to obtain initial point pair sets, restraining and filtering the initial point pair sets to generate constraint and filter point pair sets, and obtaining registered point cloud data sets by minimizing an objective function: Wherein R is a rotation matrix, t is a translation vector, For updating the point pair set after each iteration,For points in the coarsely registered point cloud dataset,Is a point in the BIM dataset.