Carbon Emission Prediction Study For The Prefabricated Construction During Materialization Stage

Carbon emission control has become the core concern of society,and energy-efficient assembled buildings have good prospects for emission reduction.Compared with traditional buildings,the process of measuring carbon emissions in the assembly building construction stage is complicated,and the scope of carbon emissions identification differs in different stages of production,transportation and construction,which makes it difficult to quantify and quickly assess carbon emissions.Based on the carbon emission coefficient method,this paper constructs a BP neural network-based carbon emission prediction model and system for the construction phase of assembled buildings based on the study of carbon emission measurement models and factors based on BIM technology,which can be used to quickly assess carbon emissions and their distribution during the engineering design and pre-construction phases,and provide quantitative data and decision-making basis for the low-carbon construction of assembled buildings.The main research elements are as follows:Firstly,a carbon emission measurement model is constructed for the assembly building construction phase.Based on the clarification of the sources of carbon emissions from assembly building construction,the physical phase is decomposed into five parts and the carbon emission coefficient method is applied to establish a carbon emission measurement model for assembly building construction.Based on the existing industry standards and BIM technology to account for the basic data of labour,material and mechanical engineering quantities,a BIM technology-based carbon emission measurement model for the materialisation stage of assembly building was established,providing methods and tools to support the measurement of carbon emission data in multiple cases.Secondly,the carbon emission measurement model based on BIM technology was applied to measure the carbon emission data of 26 assembled building projects and analyse the key factors affecting carbon emission to provide training variables and data for further construction of the BP prediction model.Based on the carbon emission measurement results of the assembled building cases and the literature from the perspective of carbon sources,the main factors affecting the carbon emission measurement were identified based on the literature analysis method;10 key factors such as building storey height and prefabrication rate were analyzed by applying Pearson analysis.Finally,26 cases of carbon emission measurement of assembled buildings were used as the training data set for the model,and a BP neural network-based carbon emission prediction model was constructed for the physical stage of assembled buildings.The results show that the prediction model based on BP neural network can predict the carbon emissions of assembled buildings in the materialization stage with a prediction error of less than 5%,and based on this model,the "Carbon Emission Prediction System for Assembled Buildings" is built to achieve a rapid estimation of carbon emissions of assembled buildings.This paper combines the characteristics of assembled buildings to build a carbon emission prediction model for the materialization phase,which can simplify the calculation index and streamline the calculation steps,and provide quantitative data and decision basis for the rapid assessment of carbon emissions of proposed assembled projects.The assessment results can provide a theoretical basis for subsequent research on carbon emission prediction of assembled buildings.