SMARTER -Smart Manager for Adaptive and Real-Time decisions in building clustERs

Abstract:

Traditionally, buildings have been viewed as mere energy consumers; however, with the new power grid infrastructure and distributed energy resources, buildings can not only consume energy, but they can also output energy. As a result, this project removes traditional boundaries between buildings in the same cluster or between the cluster and power grids, transforming individual smart buildings into NetZero building clusters enabled by cyber-support tools.

To achieve this goal, a collaborative research team among three universities: Arizona State University (ASU), Drexel University (DU) and the University of Buffalo – SUNY (UB) with Siemens Cooperate Research being the industry partner has been formed. A synergistic decision framework is to be established for temporally, spatially distributed building clusters to work as an adaptive and robust system within a smart grid. Four inter-related research tasks are proposed: (1) develop a validated emulator as the benchmarking to evaluate the decision strategies; (2) develop a robust building energy consumption model with the capability of online calibration; (3) develop an adaptive decision algorithms based on the energy consumption models developed from task 2; (4) work closely with industry (Siemens) to validate the WHOLE decision framework. In Year 1 (2012-2013), the team has successfully developed and validated a transactive building cluster Hardware-in-Loop (HIL) testbed on real commercial and mixed-used buildings in collaboration with Siemens. In Year 2 (2013-2014), the team has taken two different paths in developing a suite of high fidelity building energy consumption models. Specifically, the DU team employs grey box modeling approach using a system identification method, and the ASU team has developed a metamodel recommender system to identify appropriate black box model for the buildings with unique characteristics. Together, these models could be fused and calibrated for more accurate energy consumption prediction (task for year 3). In parallel, the UB team has been leading the efforts, in collaboration with ASU team to develop bio-inspired algorithms (Genetic algorithms, swarm intelligence algorithms) for building cluster operation decisions. Adaptive strategies (dynamic pricing, dynamic non-cooling load, etc.) have been identified which are to be incorporated in the decision algorithms for adaptive decisions (task for year 3).

Tags:
License: CC-2.5
Submitted by Teresa Wu on