A learning-assisted dynamic security enabled operational planning with transferable load

Towards penetrating increasing renewable energies, demand side has been gradually incorporated into power system dispatch due to its immense potential. However, extensive load shifts only for accommodating clean energies may jeopardize system security. To address this issue, a secure demand response strategy for operational dispatch is proposed based on learning assisted solving scheme. Firstly, backpropagation neural network (BPNN) is used to construct the security rule quantified by total transfer capability. This rule serves as a surrogate, is then embedded into demand side involved dispatch for implication of dynamic security margin. An efficient solving scheme is applied to solve the surrogate-assisted dispatch at the end. The modified IEEE 39-bus system is carried out to testify the proposed scheme. The results indicate that, our method enables efficient decision making to allow demand side to help balance production and demand without compromising the dynamic security.