Deep transfer learning based assistant system for optimal investment decision of distribution networks

With the rapid development of clean energy and the deepening of the interaction between supply and demand, power grid investment upgrading measures involve many new elements, such as clean energy installation and distribution automation. Traditional investment decision-making models are difficult to establish and solve. In view of this, this paper analyzes the investment benefit mechanism directly from the perspective of investment input-output relationship, and designs an interactive auxiliary investment decision-making system based on correlation rule mining. The system constructs an investment benefit mapping model from power grid investment measures to benefit output by means of deep transfer learning, and provides three objective functions, which consider the optimal economy, performance improvement and comprehensive index optimization, thus assisting decision makers to formulate investment alternatives according to different investment needs. A case demonstrates the decision-making process based on an actual power grid, and verifies the practicability and effectiveness of the system. (C) 2021 The Author(s). Published by Elsevier Ltd.