Controlling distributed energy storage via the cloud

/Controlling distributed energy storage via the cloud
  • Day:

    April 11
  • Start Time:

    11:00 AM - 12:30 PM
  • Stream:

    Smart Energy Solutions

The uptake of behind-the-meter energy storage is projected to significantly increase in the short term.  For most applications, however, the full value of energy storage can only be realised if multiple (sometimes competing) value streams can be realised by the same energy storage system.  This requires advanced forecasting — of demand, generation, and price.  Day-ahead forecasts are often not sufficient and do not take into account updated information about the system and its environment;  ideally the target charging schedule for the storage system should be frequently updated.  This makes operation of distributed energy storage a natural candidate for cloud-based solutions, where weather forecasts and complex optimisation algorithms can be run on resources that are provisioned as required, while fast local controllers ensure that battery charge/discharge follows a globally optimal strategy.  However, implementation is not always straightforward, and fallbacks must be put into place in the event of unexpected failures or communication drop-outs.

 

In January 2017, project partners IBM Research Australia (forecasting and optimisation), Selectronic (inverters), and Relectrify (energy storage) were successful in obtaining a Victoria New Energy Jobs Fund grant to demonstrate interactivity of new energy technologies and services.  This talk will discuss progress to date of this project.  It will describe the increased value that can be attained across whole-of-system via cloud-based optimisation of distributed energy resources, while also discussing implementation challenges and lessons learned.

 

3 key takeaways from this presentation:

1.  There is tremendous value in cloud-based orchestration of distributed energy resources

2. Complex forecasts and optimisations are best conducted in the cloud, where resources can be provisioned as necessary

3.  However, implementation is not straightforward and fallbacks must be put into place to protect against unexpected failures and communication dropouts.