Energy | Andy Berres

What do I like about energy?

Energy is such a broad topic! After spending many years on the energy end use side of things (traffic and building energy), working at NREL has allowed me to move into the energy generation side of things.

Coming from an end use perspective, the interaction between supply and demand is particularly interesting to me. But there are many other things I’m also curious about.

When there are outages, what (or who) caused them? How do the outages affect people?
Where are the weak points of our grid, and how can we make it more resilient?
When there’s a surplus, what do we do with the energy? (Storage)
How do we best integrate different (and new) energy technologies into a hybrid system?
How do we prepare the electric grid for the energy transition? Can we meet the demand of large loads without resorting to fossil energy?
As we update infrastructure within the energy system, what challenges should we expect? What are the best ways to address them?

I haven’t had a chance to work on all of these questions yet, but I’m slowly working my way through them.

Power Outages and Resilience

This work was part of a big project that studied the human dimensions of energy systems. Some of this work centered around modeling people’s willingness to participate in demand response programs from utilities. My project thrust team’s focus was to understand the impact that humans have on the energy system (e.g. increased energy use during heat waves and winter storms), and the impact of extreme weather and power outages on humans.

To evaluate the impact of weather on the energy, we chose Winter Storm Uri as our case study. This winter storm caused widespread power outages in Texas. The electric grid was extremely overwhelmed by the sudden demand for much more energy, while generation was simultaneously struggling due to frozen wind turbines (early on) and frozen gas infrastructure (later). The high demand caused the network frequency to fluctuate out of range, which caused cascading failures. Eventually, utilities had to resort to load shedding: they disconnected significant numbers of customers from the grid in order to restore network frequency into a stable range.

Seven time steps showing a progression of temperatures and outages over a span of 10 days. The outages are lagging about a full day behind the cold wave. Below the map views, there are timelines of outages by county, total unavailable electricity, and temperatures. — Weather and outage visualizations from different stages of the winter storm, combined into an illustration. The bottom section shows parts of the timeline view (selecting the 10% most affected counties above the red line), the generation outages view, and the median temperature for the entire timeseries, with vertical lines marking seven significant times. The upper section shows the temperatures and outages in each of the time steps.

My Contributions

Co-led a team of 5 researchers for 1.5 years of the project.
Obtained and processed power outage data and weather data (temperature, precipitation) and fused it with data extracted from FERC reports.
Conducted a case study of the 2021 Winter Storm Uri power outages in Texas using weater data, reported outages from EAGLE-I, and documented load shedding and grid failures from FERC.
Mentored a postdoctoral researcher and guided development of interactive visualization tools for the relation between power outages and weather, and their impact on human travel behavior.

An overview of the visualization tool. The top portion of the image consists of two parts: a map of Texas counties (a) and a list of counties with purple bands indicating outages (b). The map is colored in different shades of purple (outages). Some counties are outline in red, and the ERCOT boundaries are represented by a green line which encompasses most counties. Below the band chart, a diagram shows generation shortages and load shedding (c), and at the bottom, a line chart shows temperature over multiple weeks (d). The bottom left shows several sliders and other UI elements. — Sample analysis using the OutVis analysis tool. (a) shows the choropleth map of the outage data for a selected state at a specific time, with an optional temperature data overlay. Users can select counties in view (b) and highlight them with red outlines. The Texas Interconnection boundaries are displayed in green. (b) displays the temporal outage severity for every county of the state. Counties can be sorted by severity at the specified time (black bar). The red bar indicates the threshold of the n % most affected counties, as selected by the user. (c) shows power generation shortages (pink) and load shedding (orange) and (d) displays the median temperature over time. Finally, the tool provides data exploration options (e) to select the time range, the specific time, the aggregation level for severity, or set the threshold to automatically highlight the most severe outages, control of temperature overlay in (a), and state selection.

Hydrogen Energy

Hydrogen is one of the emerging fields in clean energy production. When used as a fuel, it only produces water. It can be produced by renewable energy sources when there is an energy production surplus. This is great, so why isn’t it used everywhere? There are multiple factors at play. The technology is still relatively new and not widely used at-scale yet (i.e. hundreds of Gigawatts)

There are various different fuel cell technologies that can turn hydrogen into energy, each with its own strengths and weaknesses.

Automated Balance of Plant and Cost Estimates for Hydrogen Fuel Cells

In one of my projects, we wanted to find out if the type of hydrogen cells (Proton Exchange Membrane) that is used for semi-trucks could be used to build a cost-effective plant. The idea is to take a lot of these fuel cells and rack them up along with power infrastructure to get to the desired size.

My Contributions

Set up a shared configuration for use by the cost model and footprint analysis.
Developed a methodology to automatically generate geometry for a hydrogen plant footprint from a config file (geopandas, shapely).
Produced two different configurations for power infrastructure: one with a pad of inverters between transformers and fuel cells, and one with pods of fuel cells directly changing an inverter transformer.
Determined physical sizes for buildings, pads, cables, pipes, roads, and support infrastructure.
Integrated footprint generation into cost model.

A rectangular plot of land (170x90 meters). Along the South border, there are a cooling pad with cooling units, fuel cell buildings with fuel cells and inverters, and a high-voltage substation. They are connected with hydrogen and cooling pipes, and cables. Along the North border, there a control building, parking, a pond, and security checkpoint. The components are surrounded by access roads. — Example of a generated 100MW plant. All infrastructure and labeling is parametrized to ensure physically possible layout. In addition to the rendered elements, the small pipes and cables connecting each pod of fuel cells are computed.

Selected Publications

Tech Report on using Proton Exchange Membrane Fuel Cells for Stationary Power Plants: coming soon
Code for Automated Balance of Plant and Cost Estimates: coming soon

Technoeconomic Analysis

In another project, we performed a technoeconomic analysis for a utility partner. We were tasked with evaluating the feasibility, requirements, and cost for the conversion of a 340MW gas power plant to a hydrogen plant.

My Contributions

Performed market research on 4 different hydrogen fuel cell technologies and their readiness for deployment in power plants, including contact with Original Equipment Manufacturers (OEMs) and a utility customer.
Familiarized myself with regulations for hydrogen storage safety to assess sizing and footprint requirements.
Performed market research on different kinds of electrolyzers and their readiness for large-scale deployment, which
Conducted a footprint analysis for each of the final selected fuel cell models. The footprint included fuel cells, access pathways, and steam methane reformers with hydrogen storage for fuel cells which operate on pure hydrogen to fill the gap until pure hydrogen becomes available. Power infrastructure was expected to be placed on top of the fuel cell structure.

Two large blue, irregular polygons, separated by a gap of consistent with. The larger, upper polygon contains some text and a gray rectangle with many small teal rectangles, indicating PEM fuel cells. The lower polygon is roughly triangle-shaped and contains a large gray rectangle labeled Steam Methane Reformer, and a smaller gray triangle with teal rectrangles indicating hydrogen storage. — Generated layout for the installation. The gray rectangle in the large polygon is the available space on the project site. The teal markers inside the rectangle are individual fuel cells. The text is automatically filled in based on individual produc sizes. The smaller polygon is land that is presently not in use. This is where hydrogen generation and storage could be placed.

Other Successes

This work resulted in the customers choosing a hydrogen-based solution for their future plant.

Power Systems Modeling and Monitoring

This section includes several smaller projects related to the Advanced Research on Integrated Energy Systems (ARIES) research platform.

Many households are switching to electric vehicles, and there is ongoing research for super fast charging infrastructure. This will be an unprecedented load for the electric grid. Testing these upgrades in a testbed allows researchers and stakeholders to understand and mitigate potential impact. However, even the biggest testbed cannot get close to real cities in scale. Emulated energy use allows researchers to add virtual loads to the network, emulate the interaction of large loads with the grid, and test the integration of distributed energy resources.

My Contributions

Supported the planning and feasibility analysis for the deployment of an Internet of Things (IoT) cluster of 600+ Raspberry Pi devices which will serve as virtual loads for the testbed.
Mentored junior staff on the development of transfer functions representing load profiles of different energy end use applications which will run as virtual loads.
Compiled a commercial buildings dataset for all metropolitan and micropolitan areas (MSAs) in the US by fusing over 10 data sources. This dataset serves as a basis for realistic emulation of distributed energy resources (DER) for whole metropolitan areas as it allows an accurate representation of different commercial building type counts for individual MSAs.
Planned and built a Grafana dashboard for real-time observation of data from the ARIES testbed, including system state, power flow and measurements of energy generation, consumption, and storage for 4 energy sources and 6 research assets.
Ensured relevance and validated correctness of dashboard elements through collaboration with domain experts.

Interoperability of Legacy Devices

As systems around us evolve, there’s an increase in sensors and controllers all around us: in building management systems, traffic systems, and the energy sector. This diverse ecosystem of devices typically has a wide range of features and abilities in their communications, data formats, and controllability, which poses challenges in interoperability across devices and systems.

This project’s goal was to get a better understanding of the specific challenges in different domains, and identify solutions which will work across sectors.

Spoiler alert: Standardization was at the top of everyone’s wish list.

My Contributions

Defined the scope of the workshop, and invited over 20 researchers from buildings, transportation, and the electric grid.
Planned the structure of the workshop, which consisted of 4 plenary kick-off presentations and small-group breakout sessions to discuss each of the topics and a plenary summary session.
Organized a half-day workshop with a selected group of experts from each domain.

Publications

Tech Report on Interoperability of Legacy Devices: coming soon