CPS: Small: Scalable and safe control synthesis for systems with symmetries
Professors Necmiye Ozay and Johanna Mathieu serve as the principle investigator and co-principle investigator, respectively, on a project funded by the National Science Foundation to ensure that the control of energy systems is safe.
Electricity grids are controlled via a process called “load balancing,” where “load” is the amount of electricity supplied by a generating system at any given time. “Load balancing” is where electrical power plants continuously ramp up and down to produce enough electricity to meet demand. However, it becomes more difficult to supply the load the as more wind and solar power are added to the grid, for wind and solar produce electricity intermittently.
Electric loads like air conditioners can be controlled to help with load balancing. Specifically, the timing of their electricity consumption can be shifted over timescales of seconds to minutes to match the production of varying solar and wind power production. However, when these loads are controlled, problems can arise in the distribution network. For instance, voltages can deviate from their normal ranges. If the voltage is too low, appliances might not turn on; if it’s too high, appliances could be damaged.
To address these issues, Ozay and Mathieu are developing theory and algorithms that allow better control of the grid (allowing for increased renewable energy use) while eliminating potential safety issues that can arise to guarantee a certain level of performance.
I-DREEM: Impact of Demand Response on short and long-term building Energy Efficiency Metrics
This project looks at interactions between energy efficiency and demand response. It’s possible that improving the energy efficiency of a building makes it less responsive to grid needs, and conversely controlling a building’s load to satisfy grid needs – such as responding to variations in renewable generation – may reduce efficiency. Energy efficiency and demand response are both fundamentally dependent upon the controls inherent in the building. This project specifically focuses on buildings’ air conditioning systems. This project will test the hypothesis that varying the load could force the controls to move away from their optimal setting, and hence from the most efficient operating point.
Prof. Johanna Mathieu is the U-M principle investigator for this project with Prof. Ian Hiskens serving as a Co-PI. To study the efficiency/response interaction, they will undertake detailed physical modeling and conduct experiments on campus buildings to learn whether shifting energy consumption in time by changing temperature set-points throughout a building makes the building less efficient.
Another aspect of the project involves analyzing a data set collected over three years that includes one million consumers in California. This investigation will explore long-term trends in energy efficiency, with particular interest in consumers that participated in demand response programs.
The project is funded by the Department of Energy Building Technologies Office with SLAC National Accelerator Laboratory leading the overall project and U-M as a subcontractor. North Carolina State University is also collaborating.
Mitigating Phase Unbalance for Distribution Systems with High Penetrations of Solar PV
Power systems should ideally operate as balanced, three-phase systems, where the A-Phase, B-Phase, and C-Phase voltages are equal. However, rooftop solar power installations are often connected to a single phase, causing voltage imbalance. This can have negative consequences, such as overheating motors and inefficiency. As such, Profs. Johanna Mathieu and Ian Hiskens are U-M co-principle investigators for this project that explores ways distribution companies can mitigate that imbalance.
In urban areas, feeders – which are a type of transmission line – are shorter, and the higher density of rooftop solar tends to provide natural balancing. However, voltage imbalance is becoming quite a problem in rural areas, where the distribution feeders are longer, consumers are sparse, and motor loads are prevalent.
This project specifically explores strategies for balancing the power produced by rooftop solar photovoltaic (PV) installations. The focus is on the use of power inverters, which are electronic devices that transform the direct current power produced by solar PV into alternating current for grid connection. Individual power inverters cannot provide balancing, but coordinated control of many distributed power inverters can do so, allowing greater penetration of rooftop solar. The researchers are developing control strategies that provide the required coordination with minimal communications.
The project is led by Argonne National Lab and includes the University of Wisconsin-Madison and the National Rural Electric Cooperative Association (NRECA). NRECA is an organization that works with most of the small rural electric cooperatives, and their member utilities cover more U.S. area than all other utilities combined. Cooperatives are experiencing significant growth in solar installations on their long rural distribution networks and are particularly interest in the project outcomes.
Modeling and Analysis of Load Ensembles
Balancing the variable power production of renewable generation can be achieved by controlling large number of small loads, such as residential air-conditioners. However, such aggregations of load can behave in unexpected and unusual ways. For example, when the on/off cycling of air-conditioner loads become synchronized, excessive voltage fluctuations can result.
Prof. Ian Hiskens, the principle investigator on this project, is developing new ways to analyze the dynamic response of such load aggregations. The aim is to establish models that better capture extreme forms of behavior. These improved load models will in turn enable control schemes to be refined to ensure actions are always predictable and acceptable.
Overcoming the Technical Challenges of Coordinating Distributed Load Resources at Scale
Professors Mathieu and Hiskens have obtained a $2.9 million grant from the Advanced Research Projects Agency-Energy (ARPA-E) through the U.S. Department of Energy to improve the integration of renewable energy into the grid. The team includes researchers from the University of California-Berkeley and Los Alamos National Laboratory, and Pecan Street Inc., a nonprofit based in Austin, TX.
The variability inherent in renewable generation will soon stretch, and possibly overwhelm, the ability of conventional generation to provide load balancing. To address this problem, this project will develop strategies for manipulating air conditioning systems to provide balancing support, but in a way that ensures negligible impact on consumer comfort. This will be achieved by adjusting the on/off cycling of air conditioners, thereby making energy consumption more controllable.
The project will explore tradeoffs between control performance and communications requirements, with investigations incorporating both simulation and hardware-in-the-loop experimentation. The project will also explore distribution network and grid stability issues resulting from load control, and develop control strategies to overcome these issues. The final stage of the project will see the proposed communications/control scheme implemented on a distribution utility feeder in Texas.
Price, Generation, Emissions, and Transmission Impacts of Energy Storage in PJM
In this interdisciplinary project, Professors Johanna Mathieu (EECS) and Catherine Hausman (Public Policy) will explore the impacts of battery energy storage in the mid-Atlantic states. PJM Interconnection is the electricity market that operates the transmission network in all or parts of 13 states. It has seen a significant growth in battery storage providing frequency regulation, which is a fast time-scale load balancing service.
The researchers will analyze the effects of batteries on the physical network (i.e., dispatch, transmission congestion), as well as on electricity prices and grid emissions. They will use electricity network data to develop a variety of economic and engineering models. They will also develop and explore the impact of new mechanisms designed to properly incentive storage for its benefits.
This research has the potential to help guide future battery energy policies and electricity market designs. The project is funded by a grant from the Sloan Foundation.
For more information, check out the Michigan Power and Energy Laboratory (MPEL) webpage.