We test through simulation the influence of lock-off time on the aggregated flexibility of thermostatically controllable loads (TCLs), which is modeled by virtual battery (VB) models.
The VB model provides a simple but effective way to quantify the aggregated flexibility of TCLs, but its efficacy is not clear when there is lock-off time for TCLs.
TCLs cannot be turned on during the lock-off period, and so lock-off time imposes inevitable restrictions on not only the flexibility of individual TCLs but also the aggregated flexibility of a group of TCLs.
We first test the influence of lock-off time on the efficacy of VB models with two artificial signals. One of the signal is of low frequency while the other is of high frequency. We then use real-world data to test the influence.
One set of data is from a duck curve from CAISO while the other set of data is from AGC signals from PJM. The duck curve changes slowly while the AGC signals change rapidly over time.
From both the artificial signals and real-word data, we find that lock-off time has a more significant impact on rapid signals than on slow signals.
Revised: March 16, 2020 |
Published: August 6, 2019
Citation
Wang P., D. Wu, and J. Wang. 2019.Impacts of Lock-off Time on Virtual Battery Model from Thermostatically Controlled Loads. In IEEE Power & Energy Society General Meeting (PESGM 2019), August 4-8, 2019. Piscataway, New Jersey:IEEE.PNNL-SA-139486.doi:10.1109/PESGM40551.2019.8973442