Evaluation model of construction quality data uncertainty: a study based on industry park project

This study proposes an evaluation and adjustment model for construction quality data uncertainty based on the research project of the Sichuan Provincial Building Industry Park. The core objective is to address the variability in concrete strength data caused by multiple factors during the construction process through a data-driven approach, which enables the accurate prediction of concrete strength. By analyzing construction quality data across different strength grades and curing ages, this study establishes a dynamic model centered on “data fitting” and “data prediction”. The model defines a reliability range for construction quality specific to the project, which facilitates the cleansing of outlier data and the adjustment of biased data. The results demonstrate that the model significantly enhances the reliability of the construction quality data. Further analysis of data from the same construction site revealed that it is highly applicable, and the date adjustment become more concentrated and stable. The fitting coefficient of this model approaches 1, which significantly improving the representativeness and reliability of the strength data. The findings provide a scientific basis for the dynamic assessment and optimization of construction quality, and robust support for quality control and prediction during the construction process.