Abstract: There are many obstacles in the process of green renovation of existing buildings in China, which not only slows down the development of green renovation, but also increases the risk of renovation projects. In this study, the risk evaluation index system for green renovation of existing buildings is established for the retrofitting party. According to the different degree of impact of each risk factor, the weight of each index is given by the AHP method. The Matlab software is used to train the established GA-BP neural network, which is applied to the post-risk evaluation of the green transformation of the building and the risk assessment of the proposed green transformation project. The results show that for the reformer, risk management in the decision-making and design stage should be increased.

Key words: green renovation, risk assessment, risk management, neural network, genetic algorithm