This paper takes the energy-saving benefits of intelligent construction technology in green buildings as the core, and systematically researches the application effects of building information modeling (BIM), IoT-driven building energy management system (BEMS), and energy consumption dynamic anomaly detection model. Threedimensional visual modeling of building structure is realized through BIM technology, which solves the problems of structural stress, material optimization and site adaptability, and improves design safety and stability. The BEMS system is constructed based on the three-layer architecture of the Internet of Things, realizing real-time monitoring of energy consumption data and driving refined energy management and control. For massive energy consumption data processing, an approximation sampling model is proposed to determine the optimal sample size through iterative calculations, and the K-center clustering algorithm is used to dynamically detect energy consumption anomalies. Taking an office building as a case study, the energy consumption data of 100 consecutive days is extracted based on the IoT BEMS platform, and the energy consumption anomaly detection model is constructed by analyzing the dimensionality reduction. The power consumption on July 2 is detected to surge to 9124.16 KW with an anomaly score of 2.509, which is significantly higher than the normal threshold, while the power consumption on the remaining days is stable at 2462-3844 KW (anomaly score ≤ 0.409). The verification of energy-saving effect shows that the system realizes energy consumption of 2423.29 KW on July 1 by dynamically regulating the endend equipments, saving 1420.97 KW of electricity on a single day compared with 3844.26 KW of the traditional long-light and long-cooling mode, and the energy-saving rate reaches 36.96%.
