[Colloquium] Peter Lin MS Presentation/Dec 17, 2021

[meganwoodward at uchicago.edu]

This is an announcement of Peter Lin's MS Presentation
===============================================
Candidate: Peter Lin

Date: Friday, December 17, 2021

Time:  2 pm CST

Remote Location: https://uchicago.zoom.us/s/2674223059?pwd=VUVzZmdrV3NXWndKbFBTRzlDT2wzQT09 Meeting ID: 267 422 3059 Passcode: 465902

Location: JCL 298

M.S. Paper Title: Exploring How to Couple Datacenters as Flexible loads to the Power Grid

Abstract: The rapid growth of datacenter (DC) loads can be leveraged to help meet renewable portfolio standard (RPS, renewable fraction) targets in power grids. The ability to manipulate DC loads over time (shifting) provides a mechanism to deal with temporal mismatch between non-dispatchable renewable generation (e.g. wind and solar) and overall grid loads, and this flexibility ultimately facilitates the absorption of renewables and grid decarbonization. To this end, we study DC-grid coupling models, exploring their impact on grid dispatch, renewable absorption, power prices, and carbon emissions, which cover the interests of the power grid, DCs, and non-DC customers. With a detailed model of grid dispatch, generation, topology, and loads, we consider coupling approaches categorized by who controls the temporal load shifting (DC, coordinator on top of DCs, DC and grid, grid) and what DC load information is shared with the grid (next hour/day's load, next day's load flexibility).

Basic results show that understanding the effects of dynamic DC load management requires studies that model the dynamics of both load and power grid.  Dynamic DC-grid coupling can produce large improvements: (1) reduce grid dispatch cost (-3%), (2) increase grid renewable fraction (+1.58%), and (3) reduce DC power cost (-16.9%).  However, it also has negative effects: (1) increase cost for both DCs and non-DC customers, (2) differentially increase prices for non-DC customers, and (3) create large power-level changes that may harm DC productivity. Improper datacenter-local control can produce more evident negative effects to both the grid and datacenters---higher dispatch cost and power prices, lower renewable fraction---by overshifting the load.

Addressing the negative effects, we further explore the impact of model variations---who controls dynamic DC loads, temporal resolution of price information, shared DC load information, and DC load continuity constraint. Among the various approaches, we find that by sharing 24-hour dynamic load schedule with the power grid in advance, utilizing hourly price information, and constraining load change, datacenter-local optimization can approximate the best performance of global grid-wide optimization, with DC capacity variation reduced and control autonomy preserved. These results suggest some directions on how to couple large-scale dynamic DC loads to the grid.

Advisors: Andrew Chien

Committee Members: Andrew Chien, John Birge, and Sanjay Krishnan