A cognition-attitude-behavior framework for low-carbon residential demand response

The transition towards a low-carbon power system is catalyzing the proliferation of residential demand response (RDR). Beyond the pursuit of economic efficiency, heightened environmental consciousness is increasingly motivating residents to engage in decarbonization initiatives. However, existing research primarily relies on economic-based or data-driven methods for RDR modeling, which suffer from limited accuracy or interpretability. Furthermore, the users’ low-carbon preferences are seldom considered, leading to insufficient exploitation of residential carbon reduction potential. To address this gap, this study proposes a hybrid data- and model-driven cognition-attitude-behavior framework for low-carbon RDR, aiming to increase demand response participation rates and reduce peak-load carbon emissions. Specifically, in the cognition module, users with heterogeneous cognitive features are identified via a graph neural network-based classification method. In the attitude module, users’ willingness to pay (WTP) for green electricity is quantified, and the WTP uncertainty is modeled as ambiguity sets to reflect fluctuations in individual preference. In the behavior module, with the participation of tradable renewable energy certificates, a distributionally robust optimization-based pricing model using a Stackelberg game is developed to simulate interactions among distribution system operators, user groups, and individual end-users. Numerical results demonstrate that the proposed framework can effectively increase RDR participation rate by approximately 12%, while promoting renewable energy consumption as well as reducing carbon emissions by around 3%.