Key points are not available for this paper at this time.
As renovating existing residential buildings shifts towards more detailed methodologies, conducting comprehensive diagnostic assessments before renovation is crucial for achieving successful outcomes. This research introduces an innovative large-scale diagnostic assessment method for existing residential buildings, addressing the inefficiencies, redundancies, and subjective biases present in traditional diagnostic processes through intelligent assistance modeling. The proposed method focuses on five key elements: construction year, exterior walls, windows, balconies, and shading devices, categorizing assessment levels into four grades (A, B, C, D) based on varying renovation intensities. Evaluation criteria are established for service life, thermal performance, degradation, and aesthetic quality. An intelligent assistance model, constructed using training datasets, enables rapid large-scale assessments, significantly reducing the evaluation time while maintaining an accuracy rate of over 95%. Empirical testing on residential buildings in Shenzhen confirmed the model’s effectiveness, demonstrating its superior accuracy and efficiency compared to traditional methods. A weighted analysis revealed that the impact of each factor on the building’s condition was as follows: exterior wall thickness (0.38), exterior finish deterioration (0.35), window deterioration (0.29), balcony deterioration (0.28), and construction year (0.23). Additionally, an interactive software application integrating the intelligent assessment model was developed. This study employed an interdisciplinary approach, combining machine learning and big data, offering a new perspective on building assessment and providing significant reference value for future residential renovation and urban planning.
Liu et al. (Wed,) studied this question.