Future of Energy Storage

 

Source- MIT Energy Initiative (Future of Energy Storage Study )

A decarbonized electric sector will require replacement of existing fossil fuel resources with intermittent renewable energy resources. Energy storage technologies are a key constrain for this transition. The energy storage technologies and products available now are not in the scale that is needed for the decarbonation. A study estimates than upto 12 TWh of grid storage will be needed by 2050, this is more than 500 times the existing 24 GWh of storage that exists in the US.

Battery Storage Technologies

The characteristics of energy storage can be classified into two types- power (kW) vs. energy (kWh).  Based on the chemistry battery storage technologies used, each of them can be better suited for providing large power or providing large volume of energy storage. For example, of the three types promising energy storage technologies, lithium-ion batteries are more for providing high power, but not suited for storing large volume of energy as the materials is expensive, Redox-Flow batteries are neutral, while Metal-to-Air batteries can provide cheaper volume but not large power. Looking at the grid need, to replace the existing gas-powered plants, we need long duration battery storage that can provide multiday energy storage.

Li-ion is not the sole solution for grid need- Based on the costs, the energy storage technology can only compete with natural gas system if priced at $20/kWh. This is more than 10 times more expensive than the existing lithium-ion battery storage price of $250/kWh. To get to this lower price will require finding cheaper battery chemistry than lithium-ion, which inherently is a more expensive element. Even while the costs of lithium ion have decreased astronomically in the past years, to meet the grid need just with lithium-ion will require more than 20% annual growth for the next 20 years. This growth will be challenging to sustain. While li-ion will continue to play an pivotal use in EV and for high power applications, li-ion is not the solution for long duration battery energy storage applications.

Redox flow batteries are better suited for long duration storage- Redox flow batteries are essentially a electrochemical reactor. The flow batteries where different chemicals that have different electric potentials, and these chemicals are passed through a electrochemical reactor that allows the passage of electric charge creating a energy source. The energy produced is proportional to the size of the tanks of electrolyzes, and the power is depended on the size of the reactor. The challenge with the redox is that they have lower energy density than compared to li-ion. The state of art is the vanadium redox flow battery. There are companies such as Largo Clean Energy and Cell Cube are developing this technology. However, the there is only Vanadium as it is a rare element.  A solution should rely on working on using the elements that are abundant like sulphur and iron, or engineering new organic compounds.

Metal-air batteries hold promise for long duration storage- Metal air batteries – Aluminum Air, Iron Air, Zinc Air- chemistry works where the metal or the metalloid is oxidized (essentially the chemistry works as rusting-discharging and reducing the rust-charging) and can used to produce electricity. Air electrodes are currently inefficient; however the lower cost of electrode metals hold promise that this technology could lead to development of cost-effective large duration battery storage. Form Energy is developing Iron-Air battery storage, Noon Energy is working on Carbon-Air batteries. This technology needs more research and development.

Mechanical (Non-battery related) energy storage technologies

There are several promising technologies that are being developed that do not use battery chemistry for energy storage. These technologies include use of hydrogen storage, thermal storage, compressed air, and gravity storage technologies.  

Pumped hydro- The most mature long duration energy storage is the pumped storage hydro that has more than ten hours of energy storage capabilities. This pumped hydro storage essentially requires a selective geographical terrain where water can pumped up (stored energy) that can be later discharged when needed. In US, the areas geographical areas where the pumped hydro facilities have been build has already been build out.

Compressed Air- Compressed air technology compresses atmospheric air to high pressure, which creates considerable heat and compressed air. The heat is stored at a thermal energy storage. The compressed air is pumped underground caverns for storage. To discharge, the compressed air is pulled from underground and send through a turbine (and uses the previous thermal energy for are expansion) to generate electricity when needed. If the thermal energy is not stored, then later in the discharging process, the system will require external fuel to create heat to decompress the compressed air.

Thermal Energy storage- Thermal energy uses heat as storage and use the heat later for generating electricity. The thermal energy is stored upwards of 1200 degree Celius. There are designs that retrofit  the existing coal or natural gas plants where these steam turbines are run through thermal storage rather than burning coal or natural gas. This technology has a potential for  long duration energy storage.

Hydrogen- The hydrogen storage can use as either electrolyzer or used as combustion. The challenges with the using hydrogen is two fold- ability to cost effectively store hydrogen, and second the challenge of using hydrogen as hydrogen is orderless, colorless, and highly combustible. However, the commercial

North American Generation Plan Model Validation (NERC MOD) Testing

 

Garbage-In, Garbage-Out. This is why it is important to ensure that in any modeling the inputs and functions are modeled correctly, and the model outputs are validated and reassured before being used for practical applications.

For generator modeling, NERC has set specific standards and specifications to ensure that the generator models are accurate.

There are different types of timescale at which the machine response incurs- from the fast electromagnetic, to electric mechanical, and to steady state. For each of these different analysis, people use different types of analysis.

For steady state analysis, we do power flow analysis, essentially solving an ohms’s law. For dynamic behavior analysis, we look into time domain and review transient stability analysis. NERC and FERC have provided standards for steady state analysis. These standards are important to understand, protect, validate controls and models for transmission system. For example, when there was blackout in WECC in the early 1990, the power modeling actually showed that the grid should be functioning fine. This showed the need to ensure that the power models should be validated to ensure accuracy.

The question is how accurate is accurate enough? The test vs. simulation generally should be accompanied with  an error threshold.  NERC Reliability Standards in the US, however does not provide performance standards, but states the accuracy target is effectively achieved by checks/balances afforded by all parties having to reach agreement on the acceptability of associated models. The US systems provide this leeway because systems are complicated with multiple variables, they call can be modified to get an result.

NERC includes models for positive sequence models, the new standards and guidelines are moving towards validation of electro-magnetic transient (EMT) point on wave models. There are several kinds of model validation for generation services. These can include – i) Open Circuit Generator Saturation Curve Measurement, ii) Open Circuit Automatic Voltage Regulator Voltage Reference Step Test, iii) Normal Synchronization, iv) V-Curves Measures, v) Reactive Capability tests, and others.

These standards help to ensure that the models behavior the way the actual physical assets behave. Quality-In, Quality-Out

EV Charing and Power Distribution Infrastructure Implications

 

In China 50% of the new vehicles sold are either full EV or Plug-in hybrids. In US, the growth is slower around 15% of new vehicles sold. The growth continues to increase as with more advances in EV technology and cost reductions through economics of scale in battery technology. With the technological breakthrough of battery technology, the range of the EV cars will likely continue to increase, which will lead to increase in adoption of EVs. All the EVs means it will require the electric grid to be able to handle the charging.

Battle of the Charging Connectors- The most common one charging type is the CCS1 type charging port which includes both low voltage AC level 1), high voltage AC (level 2), and DC charging. In US, we use the model SAE J1772 charging. Tesla has its own charging standard called the NACS (North America Charging Standard). There was a competition between the standardization of the chargers between CCS1 and NACS. The Tesla’s NACS eventually won out, with all the EVs starting to use NACS charging port regardless of which company makes the EV. The NACS standard will change the J172 standard to SAE standard of J3400. For older models, there are adaptors that can be use vice versa. With the transition, all the charging stations in the US will eventually will have the NACS.

Levels of charging- There are three types of charging levels- Level 1 that is single phase AC either 120 V or 240 V. The utility source is connected to the EVSE. These EVSE does not include any power electronics, its only job is interface, communication, and protection. The EV itself includes a AC to DC inverter that the EV uses to charge the battery. Level 3 or the DC Fast Charging puts out the DC directly into the EV, and skips the AC to DC conversion.  DC fast chargers can charge at 400V to 800 V. The DCFC usually are connected to 3 Phase AC utility source. The Tesla super chargers have liquid colling, sensors, and connectors.

Charging limits- The latest version of the Tesla DC fast charges can charge at 800V+  upto 350kW maximum charging. This charging capacity is anticipated to increase to make the charging faster. In real charging scenarios, the EVs cannot be charged at the high power 250KW level the whole time. The fast charging works best for depleted batteries. The new ChargePoint chargers can charge upto 500KW. Tesla envisions that their charging go upto 1 MW. Tesla is using these 1MW charger (Mega charger) to charge their tesla semi trucks. There are other companies that are building DC lower chargers at 20-50kW chargers. These could be used at businesses, hotels, car dealerships instead of AC charging for fasting charging to electric cars.

Installing chargers- The NEC limits that the maximum output can be only upto 80% of the circuit breaker rating, i.e. if you have a 100 amp circuit breaker, then at 80% limit, you can only power at 19.2kW at 240Volts. For level 1, and cars can charge at 1.5kW at 120Vs. Unlike DC charging curves, the level 2 charging stays flat for most of the time when it is charging. Because if it is a continuous flat load, the distribution planners have to think about how to meet these types of load. For homes that don’t have a 200Amp panel, the chargers can be connected directly to the electric service utility meter. This is upstream from the service panel, and reduces the need to change the electric upgrades to install level 2 system.

Utility upgrades- As a power utility, all these EV charging will impact the distribution grid. The primary feeders for the distribution utilities have limits to how much they can power.

Feeders- These feeders have thermal limit on how much power than transmit. Considering the thermal limits- if we put 10 fast chargers charging at 175 kW we get 1750 kW that already limits to the how much the power line can provide. Also in mind that these EV are new loads, the feeders are designed for the prior load. If we had Level 2 load charging at 10kW, we can get up to 1000kW with 100 level 2 chargers on the feeders.  

Distribution Transformers- In upstream, these distribution transformers will be overloaded as well.  The common transfomers are rated at 25 to 50 kVA. These transformer limit will quickly overloaded.

What’s next- As EVs are essentially a huge power pack, the EV can provide Vehicle to Home Power that allows the home to be isolated as a microgrid and power the home during an outage. In the future, the EV could even be bi-directional Vehicle to Grid and sell power back to grid.

Load Forecasting- Foundational Step for Enabling Energy Transition.

 

Source- ESIG- 2023 Long Term Forecasting Workshop

The era of stagnant load growth is over. With the upcoming electrification of heat and transport driven by the need for reducing carbon emissions, our electricity consumption is forecasted to double. This forecasted electricity growth transforms the existing planning process- the electric system will change from summer peaking to winter peaking, it changes periods of resource adequacy, planning on transmission and distribution system, availability of generation resources, and everything in between. The task of identifying that the electric load will be is the task of Long-Term Forecasting process.

This Long-Term Load Forecasting is the foundational process for a utility’s operations. These long-term forecasting often look ahead 10 years in ahead in the future and anticipate what the load will be in the future. Based on this plan, the electric utility then plans for development of its infrastructure to be able to meet the electric demand. In an energy market, the load term forecasting provides the market signal the price for Forward Capacity Market, development of Transmission Planning, process for Generation Interconnection, and other processes.

Long term forecasting consists of econometric models based on historical load growth and weather, which is then adjusted based on exogenous forecasts of various factors including energy efficiency, state policy, building electrification, transportation electrification, and other trends. Proper load forecasting needs a strong understanding of the emerging trends, and how these impact the electric load in the future.

Load forecast provides future energy load, but also future demand. The demand modeling provides what the future electric peak load will look like. Using historical weather history (30+ years), a weekly weather distribution is developed that provides what is the total variance of the weather that will be in each week. For each week a Weighted Temperature Humidity Index (WTHI) is used to forecast the demand for each week.

The forecasts provide 50/50 and 90/10 load. The 50/50 load forecast is the load that has a 50% probability of the forecast will exceed. 90/10 load forecast is that there is a 10 percent probability that the forecast will exceed. In other words, the 50/50 load level is a peak demand that is expected every two years, 90/10 is the extreme weather level that is expected once every ten years. The Transmission Planners use the 90/10 peak demand data to plan the load to plan for meeting system reliability to meet all the expected peak. For other resource adequacy measure planners can use the 50/50 peak.

With the energy transition, there is a wide variability in the future load in terms of forecasted solar PV installs, energy storage system, installed wind energy, electric vehicles, and electrification of buildings. These provide a challenge in terms of being able to forecast exactly what the future load will look like. With electrification of the building, the system transitions into a winter peaking system from summer peaking system. The high adoption of solar PV transitions the daily load curve into a duck curve with a sharp peak during in the evenings. Electric Vehicles double the use of electric use.

Our society's successful energy transition relies on accurate load forecasting, providing critical insights into the future grid. This enables the integration of clean energy, resource balancing, adequate energy reserves, and robust transmission infrastructure. As the energy landscape evolves, embracing advanced technologies, distribution-level forecasting, enhanced weather forecasting, and scenario planning will be essential for accurate and reliable load forecasting.

Energy Storage Design

We owe it to electric power system for enabling our modern society. The  electricity is divisible, transportable, and can be converted to other energy forms which all makes it the most convenient form of energy. The challenge with electric power system has been that the lack of storage ability. This means that for most part the electricity that we use in our everyday lives is generated instantly as we consume it. Large part of the electrical engineering, design, and effort goes into maintain this intricate balance between the supply and demand.  Energy storage technologies are exciting part of the power engineering as it changes the old status quo. However, the energy storage technologies are cost prohibitively expensive for everyday practical use and limited cases.

Energy storage can be categorized into two types- inverter based resources and generator based resources. Generator based resources have been the traditional energy storage system such as pumped hydro but they have limited availability. The inverter based resources are the battery storage systems, which are more popular systems.

When considering energy storage technologies for grid use, we need to think about four main parameters – i) voltage, ii) real power, iii) reactive power, and iv) energy capacity. The fourth parameter is unique to energy storage as energy storage system only have limited availability and grid planner needs to consider.

For implementing a energy storage, we could conduct a time series loading analysis. With Energy storage, we need consider both charging and discharge situation as well.

Implementing an energy storage project involves steps below:

1. Identifying Need: This involves planning, time series analysis, and power flow modeling to determine where energy storage can be most beneficial.
2. Conceptual Design: Developing an operational and use case procedure, conducting transient and steady-state analyses to design a system that meets the identified needs.
3.  Initial Engineering Design: Includes land assessment, interconnection study, and protection studies to lay the groundwork for the storage system's integration into the grid.
4. Detailed Engineering Design: Focuses on grounding study, equipment study, and harmonic study to ensure the system's technical and safety requirements are met.
5. Community Outreach: Engaging with the community, obtaining necessary permits, and ensuring stakeholder support are crucial for successful project implementation.

Dynamic Line Ratings

Building new transmission lines can be prohibitively expensive. It's not simply a matter of erecting poles and wires; obtaining essential environmental and land use permits from communities and localities can be especially challenging. Transmission lines dictate how much bulk power flows through the system. In situations where the transmission system is constrained and building new lines is not an option, how can we maximize the use of existing transmission lines? This is where Dynamic Line Ratings (DLR) come into play. As discussed below, DLR represents a new paradigm for assessing transmission line capacity and utilizing them to their maximum potential while staying within safety limits.

Physics of Dynamic Line Ratings
The amount of electricity that transmission lines can conduct is limited by several factors, including conductor type, size, voltage level, distance, and environmental conditions. For existing transmission lines, size, type, distance, and voltage are fixed parameters. One major environmental factor affecting transmission lines is temperature. Each transmission line has a maximum temperature limit; when it heats up, it can sag, reaching a maximum sag limit as well. Environmental temperature fluctuates with time and season. DLR adjusts the capacity of transmission lines based on real-time ambient conditions and actual field measurements. This approach contrasts with previous static line ratings, which were fixed and only varied seasonally without real-time measurements.

While the concept seems straightforward, DLR can be complex because actual conditions depend on multiple factors, including wind, ambient temperature, solar irradiation, and others, all of which influence the line rating calculation. Among these factors, wind has the greatest cooling effect on the lines.

Source- ISGAN- https://www.youtube.com/watch?v=xzWoQkVVhFc

As shown in the image above, the transmission capacity doubles from 1000 A to 2000 A as wind speed increases from 0.5 m/s to 5 m/s. The actual calculation of DLR can be intricate. IEEE 738 provides a thermal model for the application and calculation of real-time monitoring and modeling of transmission lines, used in practical applications. The model considers factors such as clearance, temperature, current, and weather parameters to calculate the gains and limits, providing the dynamic ampacity of the transmission line.

These parameters can be based on field measurements or model-based simulations; however, field-based monitoring generally offers higher accuracy.

Sensor installation technologies can be classified into four types, ranked chronologically:

1. Line-mounted sag sensor: Measures sag, wind, and temperature to calculate conductor temperature and clearances. This method provides the highest accuracy for DLR.
2. Line-mounted clearance sensor: Measures clearances, inclination, and vibration to infer sag and conductor temperature.
3. Line-mounted conductor temperature sensor: Measures conductor temperature and inclination to infer sag.
4. Indirect sensors: Do not measure parameters directly on the transmission lines but use information from weather stations and modeling. This is the least accurate method.

Uses

1. Underground: While DLR is primarily used for overhead lines, there are techniques for applying it to underground lines as well; however, the benefits are greater for overhead lines.
2. AC vs DC: Currently, DC circuits cannot use DLR sensors because these sensors are powered by induction from AC lines. In static DC lines, these sensors cannot be powered with existing methods.
3. Voltage applications: DLR can be applied from 20 kV to 800 kV.

Questions and Challenges

Although implementing DLR technologies seems straightforward, there are market and technical challenges to widespread adoption:

• Rethinking market incentives: In market-based systems, a key question is whether the incentives are aligned for transmission system owners to implement DLR. Revenues for transmission system owners vary by region, market, or product. In some cases, higher transmission constraints lead to higher transmission capacity prices. Can this be balanced through increased electricity throughput? Are there other direct incentives for transmission owners to adopt DLR technologies?
• Rethinking grid operations: Traditional grid operation methods use static capacities for transmission lines. Transitioning to a dynamic method with field sensors complicates system operations. Do system operators have the capacity to monitor, control, and settle markets at this level?
• Forecasting methodology: Transmission operations depend on load forecasting, capacity, and generation to meet demand. How does this process evolve with DLR implementation?

DLR is a proven technology already used in places like Germany and Belgium. In the US, FERC is actively looking to implement DLR in the wholesale market. While in 2021, the FERC order 881 required the transmission companies to require dynamic line ratings that accurately reflect the system, that is just the first step. As mentioned above, implementation of DLR has several challenges. FERC is now looking to reform the market rules such that DLR can be effectively implemented in the whole system. With the intent FERC, it is just a matter of time when DLR becomes a norm in the US market, which is a logical step forward for the energy market. 

Grid Forming Inverters Part 2

Grid Forming Inverters (GFIs) are a new generation of inverters designed to address the challenges faced by Inverter-Based Resources (IBRs) in an inverter-dominated field. The traditional goal of existing inverters is to maximize current output to the grid, with minimal additional capabilities. However, as IBRs become more popular, there is a need for additional functions from these inverters, such as automatic voltage control, capability to provide frequency response, fast frequency response, and system stability maintenance. These features are typically provided by traditional synchronous generators.

In a system dominated by synchronous generators, when an IBR generator provides input into the system, the IBRs follow the grid frequency, making them Grid Following Inverters (GFL). However, in a future scenario with fewer synchronous generators and less inertia in the system, the additional GFL IBRs can cause the electric system to become unstable. This instability arises because all the GFL IBRs interact with each other, and since each is designed to follow the other, there is nothing to stabilize the system.

GFMs are voltage source inverters that provide fast controlled current injected into the system to balance the system, rather than operating at maximum power point tracking (MPPT) like GFL inverters. They are designed to mitigate the issues faced in an inverter-dominated field.

 A weak grid refers to a situation where the distribution or transmission system has a low short circuit ratio. In such scenarios, when the impedance value fluctuates, the sensitivity of the voltage fluctuates as well. In weak grid scenarios, GFM IBRs can help stabilize the system.

Types of GFM Control Methods: Several types of GFM control methods are being developed that can be largely classified into phaser domain or time domain approaches. These methods can be Virtual Sync machine, Matching control, Dropped Based control, and Virtual oscillator control.

However from the grid control level should be agnostic of the type of technology used for GFM inverters. Fundamentally, the grid control is concerned about two outputs: Real Power (P) and Reactive Power (Q). In traditional GFL inverters, both P and Q are constant. In basic GFM inverter, P can change based on the situation while keeping Q constant. An advanced GFM inverter can change both P and Q based on system conditions.

Real World Applications: In the near term, GFM inverters are being used in microgrid designs and transmission systems with low fault current and low inertia. In the future, it is anticipated that GFMs will be utilized in the distribution grid, necessitating stable and reliable coordination between these inverters. Examples of current GFM utilization include:

a. Microgrids:

·    Micanopy microgrid in FL, which has 8.5 MW of Battery Energy Storage System (BESS) to support the town of Micanopy and nearby neighbors during grid outages.

·    National Grid NY, with a 20 MW, 40 MWh BESS and a 75 MVA circuit. The system includes 5 substations, a 46 KV sub-transmission line, and 10 feeders that can separate to form an island.

·    Watertown, Canada, where a section of the medium-voltage (MV) feeder operates as a microgrid with 1.6 MW and 5.2 MWh BESS.

b. Grid Islands:

·    Dersalloch Wind Farm in Scotland, which is exploring black start capabilities with wind farms.

While GFM IBRs can help strengthen weak grids, they are not a universal solution for every situation. Other solutions to strengthen weak grids include strengthening the transmission system to increase short circuit strength, re-tuning fast control loops to recognize low short circuit conditions, re-imagining IBR controls to introduce additional flexibility in operation, and the addition of synchronous condensers.