Hierarchical control of aluminum smelter loads for primary frequency support considering control cost

Abstract

Providing ancillary service with controllable loads has drawn increasing attention from academia and industry. This paper proposes a hierarchical control strategy for aluminum smelter loads (ASLs), a typical controllable load, to provide primary support in a cost-efficient way. At the higher level, the supervisory computer in the aluminum smelting plant (ASP) estimates the available power of ASLs and allocates the contracted reserve to ASLs by optimization. At the lower level, within the allocated reserve from the higher level, ASLs respond to frequency deviations automatically by changing their power consumption, which is implemented by a local control scheme based on a novel DC current control (DCC) scheme and a dynamic frequency threshold (DFT) approach. Compared with the existing PI-based DCC, the novel DCC employs model predictive control (MPC) to improve its transient performance. The DFT approach can avoid ASLs responding to frequency deviations in the serious anode effects (AE) situations. Case studies are carried out based on the data of an actual ASP. The results show that the hierarchically controlled ASLs can reduce the frequency drop with lower control cost and reduce the regulation dependency on the generation side. Besides, the high performance of the MPC-based DCC and the effectiveness of the DFT approach are also demonstrated.