An Effective Scheme to Solve Critical Data Missing Problems for IoT-Based Smart Energy Management

The accurate imputation of missing load data in building energy consumption is essential for optimizing energy management and scheduling in Internet of Things (IoT)-based smart energy management systems. However, in real-world applications, building load data often suffers from the issue of missing critical samples due to IoT device failures and maintenance. To address this problem, we propose an effective scheme by designing a load data augmentation model named the DAM based on deep neural networks. In the DAM, the partial missing data are generated in each round, followed by stacking with the semi-dataset to perform a new generation round. After several rounds, the missing critical load data are recovered with high precision. A building load dataset collected from a real IoT-based energy-efficiency management system is used for evaluation in this work. Experimental results demonstrate that the proposed scheme can effectively replenish the missing critical data and exhibit excellent stability. Additionally, we compare the prediction performance of the DAM approach with other comparison methods. The results show that our proposed approach outperforms the comparison methods, achieving the highest R2 score of 0.963. Hence, the DAM approach presents an effective solution for addressing the problem of missing critical data in IoT-based smart energy management systems, which is vital for optimizing energy dispatch.