Search

CN-122022175-A - Assembly type light wall construction system and method based on AI and BIM

CN122022175ACN 122022175 ACN122022175 ACN 122022175ACN-122022175-A

Abstract

The invention relates to the technical field of intelligent construction, and discloses an assembly type light wall construction system and method based on AI and BIM, wherein the system comprises a data acquisition unit, a CAC-GAN model unit, a DPT model unit, a reinforcement learning unit, a ICAS model unit, a DBMF model unit, a SHMPS model unit, a decision support model unit and an interaction and update unit. The data acquisition unit integrates BIM model data through a standardized interface, unifies time stamps and stores the data into the central database. The DPT model unit utilizes a transducer to optimize a construction plan, the SHMPS model unit predicts faults and evaluates structural health, and the decision support model unit fuses the multi-source data to recommend optimal decisions. According to the invention, the data acquisition unit is combined with the CAC-GAN model unit to generate an accurate construction model and a dynamically adjusted construction plan, so that the real-time consistency of the model and the site construction state is realized.

Inventors

  • YANG SHIBAO
  • LIU SHUAI
  • Zou hu
  • CHEN JIANPENG

Assignees

  • 中化学建筑工程有限公司

Dates

Publication Date
20260512
Application Date
20260202

Claims (10)

  1. 1. The assembled light wall construction system based on AI and BIM is characterized by comprising a data acquisition unit, a CAC-GAN model unit, a DPT model unit, a reinforcement learning unit, a ICAS model unit, a DBMF model unit, a SHMPS model unit, a decision support model unit and an interaction and update unit; the data acquisition unit integrates BIM model data, environment sensor data and real-time progress reports through a standardized interface, and stores the integrated BIM model data, environment sensor data and real-time progress reports into a central database after unifying time stamps; the CAC-GAN model unit predicts a construction model by combining environmental factors and time characteristics, and dynamically adjusts input through a loss function; The DPT model unit analyzes the output of the CAC-GAN model unit by using a transducer, integrates BIM static data and BIM dynamic data, and optimizes a construction plan by supervised learning; the reinforcement learning unit generates a decision vector by taking BIM model data and real-time monitoring data as states, and dynamically optimizes strategies based on the completion, cost and rewarding functions of safety indexes; the ICAS model unit integrates the output of the preamble module and automatically executes the control instruction through a rule engine and a stream processing technology; the DBMF model unit compares the construction state with the plan deviation in real time and triggers a feedback mechanism; The SHMPS model unit predicts faults through sensor data by an anomaly detection algorithm and evaluates the health level of the structure by combining BIM model data; the decision support model unit fuses the multi-source data and dynamically recommends an optimal decision through a utility function; the interaction and updating unit feeds back data in real time through a visual interface, and continuously optimizes parameters through online learning.
  2. 2. The AI-and BIM-based fabricated lightweight wall construction system according to claim 1, wherein the specific process of integrating BIM model data, environment sensor data, and real-time progress reports through the standardized interface and storing the integrated time stamps in the central database includes processing BIM static data and BIM dynamic data using rest pi, processing environment sensor data using OPCUA, cleaning, normalizing, and time stamping the collected data, and storing the collected data in the form of a structured table in the central database.
  3. 3. The AI-and BIM-based fabricated lightweight wall construction system of claim 1, wherein the CAC-GAN model unit predicts the construction model in combination with environmental factors and temporal characteristics, and dynamically adjusts the input specific process by a loss function includes a generator inputting noise vectors, environmental factors and temporal characteristics to output a predicted construction model, a discriminator inputting real data and generated data to output a discrimination probability, and alternately updating the generator and the discriminator by a min-max gaming framework.
  4. 4. The AI-and BIM-based fabricated lightweight wall construction system of claim 1, wherein the DPT model unit analyzes the output of the CAC-GAN model unit using a transducer and integrates BIM static data and BIM dynamic data, and wherein the specific process of optimizing the construction plan by supervised learning includes inputting the output of the CAC-GAN model unit, the BIM static data and the BIM dynamic data, extracting the feature sequence using a multi-head attention mechanism of the transducer, and optimizing using supervised learning based on a loss function of a mean square error.
  5. 5. The AI-and BIM-based fabricated lightweight wall construction system according to claim 1, wherein the reinforcement learning unit generates a decision vector with BIM model data and real-time monitoring data as states, and the specific process of dynamically optimizing a strategy based on a bonus function of a degree of completion, a cost and a safety index includes generating the decision vector as an action with BIM model data and real-time monitoring data as states, and updating a state action value function by combining a bellman equation with a multi-index weighted sum of the bonus function.
  6. 6. The AI and BIM based fabricated lightweight wall construction system of claim 1, wherein the ICAS model unit integrates the preamble module output, and the specific process of automatically executing the control instructions through the rule engine and stream processing technique includes inputting the outputs of the CAC-GAN model unit, the DPT model unit, and the reinforcement learning unit, generating the instruction stream using the rule engine matching condition, and triggering the execution of the control instructions when the deviation exceeds the threshold.
  7. 7. The AI-and BIM-based fabricated lightweight wall construction system of claim 1, wherein the DBMF model unit compares the construction state with the planned deviation in real time and triggers the feedback mechanism of the resource reallocation by inputting the current construction state and the planned state, calculating a deviation vector, and triggering the feedback mechanism of the resource reallocation when the deviation vector exceeds a feedback threshold.
  8. 8. The AI-and BIM-based fabricated lightweight wall construction system of claim 1, wherein the SHMPS model unit predicts faults via an anomaly detection algorithm from sensor data and evaluates structural health levels in conjunction with BIM model data comprises inputting a sequence of sensor data, calculating isolation scores using an isolation forest algorithm to predict faults, and evaluating structural health levels in conjunction with BIM model data.
  9. 9. The system for constructing the fabricated lightweight wall based on AI and BIM as claimed in claim 1, wherein the decision support model unit fuses multi-source data, the specific process of dynamically recommending optimal decisions through utility functions comprises inputting all preamble module outputs, calculating utility values through multi-factor weighted sums of utility functions, and selecting decision options with maximum utility values as optimal decisions, the specific process of feeding back data in real time through a visual interface and continuously optimizing parameters through online learning comprises displaying data indexes through a Web-based instrument panel, and continuously optimizing model parameters through gradient updating.
  10. 10. The construction method of the assembled light wall based on AI and BIM is characterized by being applied to the assembled light wall construction system based on AI and BIM as claimed in claim 1, and comprises the following steps: s1, integrating data through a data acquisition unit and storing the data into a central database; s2, predicting a construction model through a CAC-GAN model unit, and optimizing a construction plan through a DPT model unit; S3, dynamically optimizing a strategy through a reinforcement learning unit, executing a control instruction through a ICAS model unit, and triggering a feedback mechanism through a DBMF model unit; s4, evaluating the structural health level through a SHMPS model unit, and recommending an optimal decision through a decision support model unit; And S5, feeding back data and optimizing parameters through the interaction and updating unit.

Description

Assembly type light wall construction system and method based on AI and BIM Technical Field The invention relates to the technical field of intelligent construction, in particular to an assembly type light wall construction system and method based on AI and BIM. Background In the field of modern building engineering, with the wide application of Building Information Model (BIM) technology, the level of digitization, visualization and informatization of project management is significantly improved. BIM technology can provide accurate three-dimensional digital representation for the design, construction and operation and maintenance stages of building projects, including multi-dimensional information such as structure, material properties, time and cost. The technology greatly optimizes the building design flow, reduces the time and cost required by design modification, and improves the construction efficiency and quality. However, while BIM technology brings many advantages, its application in practical construction management still faces some key challenges. First, existing BIM-based construction management systems often fail to adequately implement real-time updating of data and dynamic optimization of flow. In the traditional application, once the BIM model is built, synchronous updating of the BIM model and the actual construction progress is not timely enough, so that deviation between the model and the actual condition of the site is caused. The static information processing mechanism limits the quick response capability to the emergency in the construction process, and influences the overall progress and quality control of projects. Disclosure of Invention In order to make up for the defects, the invention provides a DPDK-based high-speed data isolation method and an application method thereof, and aims to improve the problem that in the prior art, the actual construction progress is updated in time synchronously, so that the deviation occurs between a model and the actual situation on site. The first aspect of the invention provides a technical scheme that the assembly type light wall construction system based on AI and BIM comprises a data acquisition unit, a CAC-GAN model unit, a DPT model unit, a reinforcement learning unit, a ICAS model unit, a DBMF model unit, a SHMPS model unit, a decision support model unit and an interaction and update unit; the data acquisition unit integrates BIM model data, environment sensor data and real-time progress reports through a standardized interface, and stores the integrated BIM model data, environment sensor data and real-time progress reports into a central database after unifying time stamps; the CAC-GAN model unit predicts a construction model by combining environmental factors and time characteristics, and dynamically adjusts input through a loss function; The DPT model unit analyzes the output of the CAC-GAN model unit by using a transducer, integrates BIM static data and BIM dynamic data, and optimizes a construction plan by supervised learning; the reinforcement learning unit generates a decision vector by taking BIM model data and real-time monitoring data as states, and dynamically optimizes strategies based on the completion, cost and rewarding functions of safety indexes; the ICAS model unit integrates the output of the preamble module and automatically executes the control instruction through a rule engine and a stream processing technology; the DBMF model unit compares the construction state with the plan deviation in real time and triggers a feedback mechanism; The SHMPS model unit predicts faults through sensor data by an anomaly detection algorithm and evaluates the health level of the structure by combining BIM model data; the decision support model unit fuses the multi-source data and dynamically recommends an optimal decision through a utility function; the interaction and updating unit feeds back data in real time through a visual interface, and continuously optimizes parameters through online learning. According to the technical scheme, the data acquisition unit integrates BIM model data, environment sensor data and real-time progress reports in real time through RESTAPI and OPCUA protocols, the data acquisition unit stores the integrated time stamps into a central database to ensure synchronism of models and actual conditions on site, the CAC-GAN model unit and the DPT model unit realize accurate prediction of a construction model and dynamic optimization of a construction plan through generation of an antagonism network and a transducer mechanism, the reinforcement learning unit, the ICAS model unit and the DBMF model unit work cooperatively, sudden events in the construction process are responded quickly through decision vector generation, rule engine execution and deviation feedback mechanisms, the SHMPS model unit is combined with an abnormality detection algorithm and BIM data to evaluate structural health, the decision support model