Distributed Resource Integration with Transactive Control
- Written by Ronald B. Melton and Donald J. Hammerstrom
One of sixteen smart grid demonstration projects funded under the U.S. stimulus bill of 2009, the Pacific Northwest Smart Grid Demonstration is focused squarely on transactive control technology. The control technology developed in the project can be thought of as a distributed decision-making technique. Decisions at nodal points are informed by both local and global information, as represented in transactive control signals shared by neighboring nodes.
The Pacific Northwest Smart Grid Demonstration is one of the sixteen regional smart grid demonstration projects funded by the U.S. Department of Energy under the American Reinvestment and Recovery Act (ARRA). The primary technology focus of the project has been the development and deployment of transactive control technology. A fundamental purpose of this technology is to coordinate distributed smart grid assets including demand response, distributed generation and storage – in other words, distributed energy resources.
Besides managing distributed assets, the technology is intended to reduce the need for balancing services; mitigate the cost impacts of renewable resources on grid operations in both the bulk power system and in distribution systems; reduce peak loads to maximize asset utilization and minimize the need for new capacity; and reduce wholesale energy production and purchase costs.
Transactive control achieves these objectives by providing a means to effectively coordinate the interaction of large numbers of small assets with grid operations while respecting customer boundaries and providing a basis for incentivizing their participation. This in turn is intended to provide a smooth, stable and predictable response for grid operators.
The project’s transactive control technology can be thought of as a distributed decision-making technique. Decision points, referred to as transactive control nodes, may be located at points in the electric power system where one can affect the flow of power or manage a constraint. The local decisions are informed by local information and global information represented in transactive control signals shared by neighboring nodes.
There are two transactive control signals: the transactive incentive signal (TIS) and transactive feedback signal (TFS). We have formulated the TIS as a dynamic, price-like signal representing the unit cost of energy needed to supply demand served by that specific transactive control node. Calculation of the value of the TIS can take into account factors such as fuel cost; capacity constraints; costs associated with rates, markets or demand charges; profits; and so forth. The incorporation of these factors into forecasts for future time intervals can be thought of as representing dynamic cost for achieving operational objectives.
The TFS, on the other hand, is a dynamic prediction over future time intervals of planned consumption of energy based on consideration of elastic and inelastic load components; weather impacts; occupancy impacts; energy storage system operations; customer preferences and similar factors. The project has implemented functions to create the TFS for bulk inelastic load, building thermostats, water heaters, dynamic voltage control and battery storage.
Understanding the nodal structure and the relationship between nodes is key to the transactive control approach. In the most general sense, transactive control technology calls for deployment of nodes based on the electric system topology. As previously mentioned, nodes are deployed at the points in the topology where power flow can be affected or constraints must be managed. The TIS and TFS are exchanged between neighbors. Through the exchange of these forward-looking signals, any given node can be thought of in the following way:
- The node is informed by its neighbors about future costs and future needs. The node knows its own state and costs.
- The node updates its plans based on the information shared by neighbors.
- The node shares its updated plans with all neighbors.
The neighbors follow the same process updating their plans as needed. This process iterates, in a form of “market closing,” to convergence.
Considering the case where the node represents a distributed energy resource, we have described a mechanism for integration of the resource into the larger system through its participation in a distributed market where it makes locally optimal decisions about its actions. We can explore this further by considering the example of a battery storage system.
In our implementation we have chosen to consider battery systems as loads. Both charge and discharge cycles are included with discharge being treated as negative load or “negawatts.” The TFS and control signals for the battery storage system are generated by a function that provides charge and discharge rate targets, based on the battery system’s power capacity; its state of charge; the TIS and TFS inputs from neighboring nodes; the historical and predicted TIS and TFS for the battery storage node; and preferences set by the storage system owner that determine the responsiveness (elasticity) of the system. In our implementation all load or supply for the battery storage system is considered to be elastic and battery system inefficiencies such as losses and auxiliary loads are ignored.
The basic approach for the battery storage transactive control function is to predict the power consumed or supplied during each future prediction time interval (a process known as elastic load prediction) and then determine charge / discharge actions to achieve maximum benefit from the predicted incentive values. The maximum benefit is determined using an augmented cost function that seeks to achieve the maximum benefit from the value of the predicted future incentives. If the incentive is lower than usual, one should charge; if higher than usual, one should discharge.
This transactive control function for battery storage DER integration is being tested with a 100kW / 100 kWh battery system located at Lower Valley Energy in Wyoming. This system responds to a TIS representing regional needs such as support for wind integration. In this experiment, the battery system vendor has limits on the number of charge / discharge cycles, which prevents it from responding frequently. Data on the transactive control performance of this system and the other transactive control test cases included in the project is currently being collected. The analysis results will be included in the project’s reporting to the Department of Energy in January 2015.
In conclusion, there is increasing diversity in the resources deployed in electric power systems worldwide. New approaches are needed for effective integration of these distributed energy resources. Transactive control is one such approach offering the advantages of: coordination with the broader power system through the exchange of transactive control signals with neighboring system elements; maintenance of control by the owner of the distributed asset; and alignment of values for the system operator(s) and the asset owners.
This article is based upon work supported by the U.S. Department of Energy under Award Number DE-OE0000190.