Background: In smart grid, a flexible demand response management mechanism is used to achieve the purpose of stabilizing the power grid, optimizing the power market, and rationally allocating resources. There are two types of demand response management in the demand response management mechanism: Price-based Demand Response (PDR) and Incentive-based Demand Response (IDR).
Methods: The paper studies the problem of privacy protection in IDR, and propose a method based on database digital watermark to protect user privacy. Segment the time, and then embed watermarks in the user’s consumption data of each time segment. At the end of each billing period, extract the watermarks from the data of each segment time, and then send the total consumption data of the user in this billing period to the power company. The power company could only know the total consumption data of the user, couldn’t know the user’s consumption data at any moment, which can prevent them from snooping the user privacy. The proposed digital watermarking algorithm is based on K-Means clustering and wavelet transform, the K-Means algorithm is used to cluster the database tuple data, and then wavelet transform is carried out on the available attribute values within the clusters, and the watermark is embedded in the transformed attribute values.
Result: The experimental results show that the proposed method is more robust when the database is under subset deletion attacks, subset substitution attacks and subset addition attacks. Besides, the computational cost is much more less.
Conclusion: The proposed digital watermark algorithm can embed the watermarks more decently and overcome the burden of watermark embedding caused by statistical feature control. Besides, the proposed method can protect the user privacy better than the other two methods.