Innovation Award Submission Poster
Thermal Energy Storage Dispatch Optimization
The University of Texas at Austin, The University of Texas CHP and District Energy System
Many large utilities rely upon various demand-response strategies in lieu of adding capacity to meet peak demands. National grid operators, ERCOT for example, also utilize demand response strategies, such as load-shedding, to ensure grid stability. For utilities, demand response is typically much less expensive than a peaking plant, with obvious environmental benefits as well. In a smaller scale utility like UT Austin’s, the most common means of demand response would compromise research.
UEM’s demand response solution is highly unique, both because of physical limitations, and in the means in which it is dispatched. The system is comprised of two TES tanks of differing dimensions and capacities, each sited at hydraulically diverse locations at substantially different elevations. These constraints required some extremely clever controls programming, developed completely in-house, with the original TES-1 tank utilizing more common PID push/pull flow control loops, with both pressure and level limits. The newer TES-2 uses PID flow control, but with several trims or biases. These ensure system response is gradual, but responsive enough to load transient changes such that safety limits are not approached, and nearby sensitive HDPE chilled water pipe used throughout the DSMC is not compromised. Further, while most TES solutions are implemented as on/off “step” functions to the chilled water system in concert with a large singular load – as was UT Austin’s first TES tank originally – UEM’s new controls strategy utilizes very gradual ramp rates to mirror UT Austin’s electrical load profile, and as a result, flatten electrical load as much as possible without the use of intelligent, predictive algorithms. This control scheme provides a better and more efficient power plant heat rate allowing the turbine-generator sets to run at baseload when shifting the MW load from an on/off peak conventional day/night to, on
average, a more flattened curve.
The unique controls programming requires a dedicated, master-planned IT architecture to marry both chiller power consumption and generation data from spinning generation reserves, along with the campus chilled water
requirements and chilled water production from buildings, chillers and TES resources. This architecture is protected on secure, isolated networks, and in the future will feature real-time predictive regression modeling
discussed later in greater detail.
Compared to the average of three years prior to the implementation of TES-2 in 2018, UT Austin’s electrical load pattern approximates the typical sinusoidal pattern of most grids. Normalized for both weather and building square footage added over the years, the new dispatch strategy has proven extremely effective in flattening these daily loads to within an incredible 10% bandwidth, as shown below.
Nicholas J. Schroeder, University of Texas at Austin#AnnualConference#2019#ConferenceProceeding#Poster