Apportionment of control in virtual power stations

Distributed energy resource management involves matching the production of and demand for power in electrical grids. This leads to the question, where should responsibility for resource use be placed? A number of potential management schemes exist in current work on complex large-scale distributed computer systems. This paper works towards a comparison of their relative performance. A test scenario is designed, based on the concept of a virtual power station. Simulations are run, using input data from an actual system, to produce estimates of the overall cost of electricity in a virtual power station over a given period. Results are presented that examine the effect of improving or weakening management precision at various points. These indicate that fairly weak management techniques may be sufficient. While the presence of a number of forms of management, including central control, local autonomous control, and variation of group composition, is significant, improved management does not bring very large additional benefits. This can have important implications when designing virtual power station implementations. For instance it indicates low-grade communication networks and control devices may provide the best return on investment.