Page 478 - Decision Making Applications in Modern Power Systems
P. 478
Multistage and decentralized operations of Chapter | 16 437
FIGURE 16.16 Dual residuals of sync-ADMM and async-ADMM.
FIGURE 16.17 Objective values of sync-ADMM and async-ADMM.
maximum time lag among the EV agents to be bounded by L max , which in our
case is set to 5. The minimal number of agents N min in this case is also set to
5. Figs. 16.15 16.17 show the convergence performance of the asynchronous
ADMM algorithm compared to the sync version with regard to primal residual,
dual residual, and the objective value, respectively. Though it takes more steps
to converge (74 vs 271), asynchronous ADMM can still achieve the system
optimality with disturbances from the communication and control systems.
16.6 Conclusion
In this chapter, we introduce the integration strategies of EVs under micro-
grid scenarios with the California energy, ancillary service, and DR markets.
