Single-site and multi-site reference controller designs

[摘要] This report presents the reference controller designs for the coordination of air conditioner load, solar photovoltaic (PV) generation and battery storage in a virtual power station electric network. The approach is a model-based predictive control (MPC) strategy that optimises the balance between loads and generation to achieve control objectives defined in an optimisation cost function, which may be customised. The present report illustrates the application of the proposed MPC strategy by controlling air conditioners load to follow PV generation assuming a short-term PV generation forecast is available. The proposed MPC strategy manipulates the demand of air conditioners using control commands consistent with the demand response capabilities specified by the AS/NZS 4755 standard. Namely, the MPC commands can control air conditioners to reduce their energy demand to 75% (DRM3), 50% (DRM2), or to switch off their compressor (DRM1). The performance of the proposed MPC strategy is illustrated in balancing fluctuations in aggregate PV generation by controlling the demand of ACs. The proposed MPC strategy may be applied through a single DRM signal channel to control demand of ACs in a single site, or at multiple sites. When more DMR signal channels are available, the proposed MPC strategy can control the population of ACs subdivided in clusters, each of which is associated with one DRM signal channel. Simulation results illustrate substantial improvements in tracking performance achieved for a heterogeneous population of 60 ACs.

[发布机构] CSIRO

[效力级别] Control Systems, Robotics and Automation [学科分类] 地球科学(综合)