An increase in battery energy storage system (BESS) deployments reveal the importance of successful cooling design. Unique challenges of lithium-ion battery systems require careful design. The low prescribed battery operating temperature (20° to 25°C), requires a refrigeration cooling system rather than direct ambient air cooling. The narrow allowable temperature variation, no more than 5°C between hottest and coldest battery, requires near perfect air distribution. And, the rapid changes in power with time require tight control. Without proper thermal management, overheating cells will degrade, malfunction or even catch fire [2] [3].
Several modeling tools aid in the design process to ensure good designs. These include:
- Overall mass and energy balance equations for selection of cooling equipment sizes;
- Transient FEA for studying the transient behavior and control methods; and,
- CFD software for evaluating the air flow distribution and the resulting temperature variances.
Sizing the cooling system with overall energy and mass balances
Getting the right capacity and turn-down capabilities for BESS thermal control will result in better performance and a longer life for the batteries. Under-sizing the cooling system can lead to battery overheating. Over-sizing the cooling system can lead to short cycling of the cooling system and large air temperature swings when the unit turns on and off.
Heat generated by the batteries
Heat generation from the batteries is the largest load and therefore the most important to predict accurately. Heat results from reaction entropy and activation energy losses, electrical and ionic resistance, and chemical transport. [4] [5] [6] The heat generated is typically flat between 20% and 80% state of charge (SoC). Heat generation increases significantly as a discharge approaches 0 SoC and when the charge approaches 100% SoC, as shown in Figure 1.
Figure 1. Heat generated through a cycle at 1 C-rate showing rise in heat at the end of a discharge and end of a charge.
Li-ion battery heat generation typically follows I2R behavior, which can be rearranged to a dimensionless form:
The parameter α is constant for a battery type and independent of the array size and how the battery strings are arranged.
Because of degradation, the heat generation increases over the life of a project. Increases of 35% to 70% have been reported by battery companies. Heat generation increases at lower temperatures, approximated by the α/αref = √Tref /T within 10° to 50°C based on a reference temperature at 25°C.
Other heat loads
The high-voltage DC bus and cabling system generates resistive heat according to G = I2R, and is typically designed to be about 0.25% of the maximum DC power. Heat generated by lighting, communications equipment, power supplies and controllers remains relatively constant with time at about 500 W to 1 kW of heat.
Heat also comes in from the external environment. Commercial HVAC sizing software accurately calculates the environmental heat load. A general approximation that works for single-story buildings located in temperate climates is 1 ton of cooling per 500 ft2.
Sum of heat loads
Using the calculations described above, the cooling requirements for a lithium-ion BESS system per MWh of batteries at different C-rates are shown in Table 1.
Table 1. Representative heat loads for a 1-MWh enclosure at various C-rates.
HVAC equipment selection
HVAC capacity ratings are based on a set of nominal conditions [7]. The actual capacity of the HVAC is lower than the nominal rating in most energy storage applications due to hotter outdoor temperatures, lower interior set point temperatures and lower humidity. If operating in a hot desert, the actual capacity of an air conditioner may only be 50% of the nominal capacity, or even less as the evaporator and condenser coils get soiled. As a general rule of thumb for energy storage, the HVAC equipment nominal rating should be 150% larger than the sensible cooling load required based on the calculations above.
Heat transfer between the cooling air and battery modules
Today, most stationary BESS systems use air as the medium to cool batteries. In addition to a properly sized cooling system for the enclosure, the modules must also have a properly designed method for locally transferring heat into the cooling air. The steady state heat transfer equation for local cooling of the module can be described by the equation:
Where U’ [W/°C] is the overall heat transfer coefficient that includes conduction of the heat from the cells to the surface, and convection of the heat from the surface to the air. The cooling surface may be internal to the module and/or the external casing of the module.
In order to keep the battery below a maximum temperature, two conditions must be met. First, the overall heat transfer coefficient must be sufficient by meeting the equation:
Second, the air flow volume (passive or active) past the modules must be sufficient to absorb the heat generated:
These two conditions must be checked and if they fail, then no amount of HVAC capacity can keep the battery cool. If the limitation is conduction of the heat through the module to the surface and from the surface to the air, then the module must be redesigned for better heat transfer.
Transient considerations for sizing the cooling system
Figure 2 shows examples of load profiles for different BESS applications. A transient analysis studies how the HVAC system effects the battery temperature as the load changes.
Figure 2. Examples of daily cycles (24-hr) for stationary BESS systems.
The transient thermal model presented has been used to evaluate many design decisions, such as:
- Determine the amount of cooling required in a particular usage case based on the application’s load profile;
- Determine if cooling staging is required and, if so, what levels of staging;
- Determine the effects of different temperature control methods;
- Evaluate the effectiveness of adjusting thermal masses in both the module and the enclosure;
- Evaluate the effectiveness of adjusting air flows in the enclosure and in the modules.
As an example, ISO New England publishes a transient load profile for simulating frequency regulation dispatch for energy storage systems [8]. A transient model assumed a 2-MW/1-MWh BESS component with forced-air-cooled lithium-ion batteries. The plots in Figure 3 show a 24-hour portion with the resulting HVAC cycling and air temperatures in the enclosure bases on a specific enclosure and HVAC design. The model shows that although 60 kW of heat may be generated by the batteries for brief periods of time, not more than 21 kW of cooling is ever required to maintain the air set-point temperature. Based on the transient analysis, the HVAC size could be reduced to one-third of the maximum instantaneous heat load.
Figure 3. The resulting heat flows and temperature profile for a 2-MW/1-MWh BESS system with multiple stages of 3-RT on air conditioning for cooling over a period of 1 day in a heavy use case with New England ISO data.
Achieving balanced air distribution with CFD
Another key aspect of the cooling system is how the air is ducted through the enclosure to keep the batteries within an allowable temperature variance. CFD analysis excels at calculating spatial values for temperature, static pressure, air velocity and air flow direction.
CFD analysis helped make critical design decisions, such as:
- verify sufficient cross-section area for air flow throughout the enclosure,
- select locations for air barriers and deflectors,
- determine optimum size, shape and register locations for supply ducts,
- provide direction for optimizing the vane positions in supply registers,
- determine significance and location for air barriers between or around equipment, and
- determine the effectiveness of additional fans/blowers in the enclosure to augment the HVAC.
A CFD analysis was used to analyze air flow and the resulting temperatures for an enclosure containing batteries with wall-mount HVAC units on both ends. The analysis used battery data and HVAC data from the respective manufacturers. Figure 4 displays graphical results of the temperature profiles of three planes in the x, y and z axis. These profiles detail the air flow through the enclosure and the resulting temperatures. It is easy to see how the cold air (in blue) is introduced and how it distributes. This information was used to help improve the ducting design.
Figure 4. CFD results showing planes with temperatures and air flow vectors.
Conclusion
Batteries generate heat like other electrical equipment, however, manufacturer performance warranties require a low temperature and a very narrow window in which the batteries can operate. Although designing the thermal management system for a battery energy storage enclosure presents these unique challenges, the tools presented in this paper are being used with success.
Nomenclature
References
[1] P. P. X. Z. G. C. J. S. C. C. Qingsong Wang, “Thermal runaway caused fire and explosion of lithium ion battery,” Journal of Power Sources, vol. 208, pp. 210-224, 2012. [2] S. G. T. F. F. Todd M. Bandhauer, “A Critical Review of Thermal Issues in Lithium-Ion Batteries,” J. Electrochemical Society, vol. 158, no. 3, pp. R1-R25, 2011. [3] M. O. L. L. J. L. X. H. G. Liu, “Analysis of the heat generation of lithium-ion battery during charging and discharging considering different influencing factors,” J Therm Anal Calorim, vol. 116, pp. 1001-1010, 2014. [4] S. G. T. F. F. Todd M. Bandhauer, “Temperature-dependent electrochemical heat generation in a commercial lithium-ion battery,” Journal of Power Sources, vol. 247, pp. 618-628, 2014. [5] S. F. Ashkan Nazari, “Heat generation in lithium-ion batteries with different nominal capacities and chemistries,” Applied Thermal Engineering, vol. 125, pp. 1501-1517, 2017. [6] “Battery Heat Generation Data,” Confidential Battery Manufacturer, 2016. [7] ASHREA STD 16, Method of Testing for Rated Room Air-Conditioner and Packaged Terminal Air-Conditioner Heating Capacity, 2016. [8] ISO New England, “Simulated Automatic Generator Control (AGC) Setpoint Data,” [Online]. Available: https://www.iso-ne.com/isoexpress/web/reports/grid/-/tree/simulated-agc. [Accessed 2016]. [9] D. F.-F. M. G. Brett Simon, “U.S. Energy Storage Monitor,” GTM Research/ESA, 2018.
Source: solarpowerworldonline
