Publications

Simulating Onset and Evolution of Thermal Runaway in Li-ion Cells using a Coupled Thermal and Venting Model 

Abstract:

Control volume analysis for conservation of species and conservation of energy. A coupled, thermal and gas generation/venting model has been developed for simulating the onset and evolution of thermal runaway in 18650 format Li-ion battery cells. The model simulates heat and gas generation during external heating of an electrically isolated cell that results in thermal runaway. Gas generation within the cell leads to pressure build up until the point at which the vent mechanism opens and relieves the internal pressure.  Compressible flow of gases is modeled through the vent cap as a function of pressure ratio across the vent. The energy balance of the battery cell includes: heat generated from decomposition reactions and electrical short, external heat transfer to the surroundings, heat absorbed with vaporization and melting processes, as well as the energy loss as material is vented from the cell. The model was able to capture features of the temperature evolution of the battery cell well and generate detailed information about the progression of thermal runaway. The model was exercised to simulate time-to-venting and time-to-thermal-runaway for various changes in cell design parameters such as: amount of free liquid electrolyte, external convection coefficient, and electrolyte evaporation rate after vent opening.

Reference:

Ostanek, J., Li, W., Mukherjee, P., Crompton, K.R., Hacker, C. (2020).  "Simulating onset and evolution of thermal runaway in Li-ion cells using a coupled thermal and venting model," Applied Energy.  DOI: 10.1016/j.apenergy.2020.114972

 


Conjugate Heat and Mass Transfer Model for Predicting Thin-Layer Drying Uniformity in a Compact, Crossflow Dehydrator

 

Abstract:

Modeling results showing temperature and relative humidity along the midplane of a compact dehydrator for drying corn and other food products.A conjugate heat and mass transfer model was implemented into a commercial CFD code to analyze the convective drying of corn. The Navier–Stokes equations for drying air flow were coupled to diffusion equations for heat and moisture transport in a corn kernel during drying. Model formulation and implementation in the commercial software is discussed. Validation simulations were conducted to compare numerical results to experimental, thin-layer drying data. The model was then used to analyze drying performance for a compact, crossflow dehydrator. At low inlet air temperatures, the drying rate in the compact dehydrator matched the thin-layer drying rate. At higher temperatures, heat losses through the external walls resulted in temperature and moisture variations across the dehydrator.

 

Reference:

Ostanek, J.K. and Ileleji, K. (2019). “Conjugate Heat and Mass Transfer Model for Predicting Thin-Layer Drying Uniformity in a Compact, Crossflow Dehydrator,” Drying Technology.  DOI: 10.1080/07373937.2019.1590394.

 


Overcharge and Thermal Destructive Testing of Lithium Metal Oxide and Lithium Metal Phosphate Batteries Incorporating Optical Diagnostics

 

Abstract:

Schlieren images of 18650 battery failure showing spray of flammable gases and liquid droplets.Lithium batteries have a tendency to fail violently under adverse conditions leading to the rapid venting of gas. Overcharge, thermal heating, and a combination of the two conditions are applied here to investigate the gas venting process. A test chamber has been constructed with data recordings including chamber pressure and temperature, battery voltage, current, and surface temperature as functions of time throughout the charging and failure processes. High-speed imaging and schlieren flow visualization are used to visualize the gas venting process. A direct comparison between lithium iron phosphate based K2 26650 and lithium nickel manganese cobalt oxide LG 18650 cells is made through a test series of the three failure methods. Failure under thermal, overcharge, and thermal-overcharge conditions are generally similar in terms of the gas venting process, but are observed to have increasingly energetic failures. The thermal-overcharge abuse condition demonstrates an ability to reconnect via internal short circuit even after an initial electrical failure seen as the refusal to accept charge. This reconnection is associated with a secondary, more energetic failure which can produce weak shock pressure waves. ​

 

Reference:

Mier, F., Morales, R., Coultas-McKenney, C., Hargather, M., and Ostanek, J.K., 2017, "Overcharge and thermal destructive testing of lithium metal oxide and lithium metal phosphate batteries incorporating optical diagnostics," J. Energy Storage. vol. 13. pp. 378-386.

 


Visualization Study of Vapor Formation during Pool Boiling under Modified Confinement Surfaces

 

Abstract:

Vapor formation under different confined surfaces and heat flux levels.Understanding and predicting the eff ects of confinement on pool boiling is important for increasing the power density of electronic systems using two-phase, thermosyphon cooling systems. In comparison with spherical bubble formation in unconfined pool boiling, the presence of a confining surface near the heated surface will result in the formation of deformed/elongated bubbles. The present visualization study investigates the effects of confinement through high resolution, still photography taken from underneath a transparent heater test surface. The heater was comprised of an ITO-coated, glass substrate. Saturated pool boiling experiments using a dielectric refrigerant, RE347mcc (HFE-7000), were performed at atmospheric pressure for a flat, upward-facing surface. Unconfined tests were performed as a baseline, and three confining surfaces were tested: a plain surface, a surface with a rib, and a surface with relief holes drilled in it. The visualization study shows details of vapor structures growing underneath the confinement surfaces and escaping around the edges of the confinement surfaces. As the vapor escapes the confining surface, incoming liquid is drawn in to replenish the heater. The incoming liquid would occasionally displace any pre-existing vapor pockets, resulting in finger-like structures forming within vapor pockets. In some cases, a growing vapor pocket would initiate during the liquid replenishment, adjacent to existing vapor pockets. The newly formed and expanding vapor pocket was found to displace the surrounding liquid and the pre-existing vapor pockets. As heat flux is increased, the size and shape of the large vapor structure changes as well as the wetted area between the structures. For the plain surface, a vapor layer was observed on the test surface at a heat flux of 5 W/cm^2 while the rib and drilled surfaces remained wetted.

 

Reference:

Ostanek, J.K. (2017). “Visualization of Vapor Formation during Pool Boiling under Modified Confinement Surfaces,” J. Flow Visualization and Image Processing.  vol. 23(1-2), pp. 117-136.

 


Measurement Sensitivity Analysis of the Transient Hot Source Technique Applied to Flat and Cylindrical Samples

 

Abstract:

Schematics of various transient hot source sensor configurations.The transient source measurement technique is a nonintrusive, nondestructive method of measuring the thermal properties of a given sample. The transient source technique has been implemented using a wide variety of sensor shapes or configurations. The modern transient plane source (TPS) sensor is a spiral-shaped sensor element which evolved from transient line and transient hot strip (THS) source techniques. Commercially available sensors employ a flat interface that works well when test samples have a smooth, flat surface. The present work provides the basis for a new, cylindrical strip (CS) sensor configuration to be applied to cylindrical surfaces. Specifically, this work uses parameter estimation theory to compare the performance of CS sensor configurations with a variety of existing flat sensor geometries, including TPS and THS. A single-parameter model for identifying thermal conductivity and a two-parameter model for identifying both thermal conductivity as well as volumetric heat capacity are considered. Results indicate that thermal property measurements may be carried out with greater measurement sensitivity using the CS sensor configuration than similar configurations for flat geometries. In addition, this paper shows how the CS sensor may be modified to adjust the characteristic time scale of the experiment, if needed. 

 

Reference:

Ostanek, J.K., Shah, K. and Jain, A., 2017, “Measurement Sensitivity Analysis of the Transient Hot Source Technique applied to Flat and Cylindrical Samples,” J. Thermal Sci. Eng. Appl. vol. 9(1).  pp. 011002 (1-12).

 


Reducing Cell-to-Cell Spacing for Large-Format Lithium Ion Battery Modules with Aluminum or PCM Heat Sinks under Failure Conditions

 

Abstract:

Heat spreading from a failed battery cell to neighboring cells showing different heat sink materials and cell-to-cell spacingsThermal management is critical for large-scale, shipboard energy storage systems utilizing lithium-ion batteries. In recent years, there has been growing research in thermal management of lithium-ion battery modules. However, there is little information available on the minimum cell-to-cell spacing limits for indirect, liquid cooled modules when considering heat release during a single cell failure. For this purpose, a generic four-cell module was modeled using finite element analysis to determine the sensitivity of module temperatures to cell spacing. Additionally, the effects of different heat sink materials and interface qualities were investigated. Two materials were considered, a solid aluminum block and a metal/wax composite block. Simulations were run for three different transient load profiles. The first profile simulates sustained high rate operation where the system begins at rest and generates heat continuously until it reaches steady state. And, two failure mode simulations were conducted to investigate block performance during a slow and a fast exothermic reaction, respectively. Results indicate that composite materials can perform well under normal operation and provide some protection against single cell failure; although, for very compact designs, the amount of wax available to absorb heat is reduced and the effectiveness of the phase change material is diminished. The aluminum block design performed well under all conditions, and showed that heat generated during a failure is quickly dissipated to the coolant, even under the closest cell spacing configuration.

 

Reference:

Coleman, B., Ostanek, J.K., and Heinzel, J., 2016, “Reducing cell-to-cell spacing for large-format lithium ion battery modules with aluminum or PCM heat sinks under failure conditions,” Applied Energy. vol. 180.  pp. 14-27.

 


Heat Generation Rate Measurement in a Li-ion Cell at Large C-Rates through Temperature and Heat Flux Measurements

 

Abstract:

Experimental setup and validation for internal heat generation rate measurement.Understanding the rate of heat generation in a Li-ion cell is critical for safety and performance of Li-ion cells and systems. Cell performance, cycle life, and system safety all depend on temperature distribution in the cell, which, in turn, depends on heat generation rate within the cell and on heat removal rate at the cell surface. Despite the existence of a number of theoretical models to predict heat generation rate, there is not much literature on experimental measurement at high C-rates. This paper reports measurement of heat generation rate from a Li-ion cell at high discharge rates, up to 9.6C, using measurements of cell temperature and surface heat flux. As opposed to calorimetry-based approaches, this method can be applied in situ to yield measurements of heat generation rate in laboratory or field use provided that at least one a priori test is performed to measure the temperature gradient within a cell in the same ambient condition. This method is based on simultaneous determination of heat stored and heat lost from the cell through heat flux and temperature measurements. A novel method is established for measurement of the internal temperature of the cell. Heat generation measurements are shown to agree with well-established theoretical models. The effect of actively cooling the cell is briefly discussed.

 

Reference:

Drake, S.J., Martin, M., Wetz, D.A., Ostanek, J.K., Miller, S.P., Heinzel, J.M., Jain, A. (2015). "Heat generation measurement in a Li-ion cell at large C-rates through temperature and heat flux measurement," J. Power Sources. DOI: 10.1016/j.jpowsour.2015.03.008.

 


An Experimentally Validated Transient Thermal Model for Cylindrical Li-ion Cells

 

Abstract:

Experimental validation of analytical thermal model.Measurement and modeling of thermal phenomena in Li-ion cells is a critical research challenge that directly affects both performance and safety. Even though the operation of a Li-ion cell is in most cases a transient phenomenon, most available thermal models for Li-ion cells predict only steady-state temperature fields. This paper presents the derivation, experimental validation and application of an analytical model to predict the transient temperature field in a cylindrical Li-ion cell in response to time-varying heat generation within the cell. The derivation is based on Laplace transformation of governing energy equations, and accounts for anisotropic thermal conduction within the cell. Model predictions are in excellent agreement with experimental measurements on a thermal test cell. The effects of various thermophysical properties and parameters on transient thermal characteristics of the cell are analyzed. The effect of pulse width and cooling time for pulsed operation is quantified. The thermal response to multiple cycles of discharge and charge is computed, and cell-level trade-offs for this process are identified. The results presented in this paper may help understand thermal phenomena in Li-ion cells, and may contribute towards thermal design and optimization tools for energy conversion and storage systems based on Li-ion cells.

Reference:

Shah, K., Drake, S.J., Wetz, D.A., Ostanek, J.K., Miller, S.P., Heinzel, J.M., Jain, A. (2014). "An experimentally validated transient thermal model for cylindrical Li-ion cells," J. Power Sources. DOI: 10.1016/j.jpowsour.2014.07.118.

 


Modeling of Steady-State Convective Cooling of Cylindrical Li-ion Cells

 

Abstract:

Effect of cell aspect ratio on maximum temperature rise within the cellWhile Lithium-ion batteries have the potential to serve as an excellent means of energy storage, they suffer from several operational safety concerns. Temperature excursion beyond a specified limit for a Lithium-ion battery triggers a sequence of decomposition and release, which can preclude thermal runaway events and catastrophic failure. To optimize liquid or air-based convective cooling approaches, it is important to accurately model the thermal response of Lithium-ion cells to convective cooling, particularly in high-rate discharge applications where significant heat generation is expected. This paper presents closed-form analytical solutions for the steady-state temperature profile in a convectively cooled cylindrical Lithium-ion cell. These models account for the strongly anisotropic thermal conductivity of cylindrical Lithium-ion batteries due to the spirally wound electrode assembly. Model results are in excellent agreement with experimentally measured temperature rise in a thermal test cell. Results indicate that improvements in radial thermal conductivity and axial convective heat transfer coefficient may result in significant peak temperature reduction. Battery sizing optimization using the analytical model is discussed, indicating the dependence of thermal performance of the cell on its size and aspect ratio. Results presented in this paper may aid in accurate thermal design and thermal management of Lithium-ion batteries.

Reference:

Shah, K., Drake, S.J., Wetz, D.A., Ostanek, J.K., Miller, S.P., Heinzel, J.M., Jain, A. (2014). "Modeling of steady-state convective cooling of cylindrical Li-ion cells," J. Power Sources.  DOI: 10.1016/j.jpowsour.2014.01.115.

 


Measurement of Anisotropic Thermophysical Properties of Cylindrical Li-ion Cells

 

Abstract:

Experimental setup to measure thermal conductivity and heat capacity of Li-ion cellsCylindrical Li-ion cells have demonstrated among the highest power density of all Li-ion cell types and typically employ a spiral electrode assembly. This spiral assembly is expected to cause large anisotropy in thermal conductance between the radial and axial directions due to the large number of interfaces
between electrode and electrolyte layers in the radial conduction path, which are absent in the axial direction. This paper describes a novel experimental technique to measure the anisotropic thermal conductivity and heat capacity of Li-ion cells using adiabatic unsteady heating. Analytical modeling of the method is presented and is shown to agree well with finite-element simulation models. Experimental measurements indicate that radial thermal conductivity is two orders of magnitude lower than axial thermal conductivity for cylindrical 26650 and 18650 LiFePO4 cells. Due to the strong influence of temperature on cell performance and behavior, accounting for this strong anisotropy is critical when modeling battery behavior and designing battery cooling systems. This work improves the understanding of thermal transport in Li-ion cells, and presents a simple method for measuring anisotropic thermal transport properties in cylindrical cells.

Reference:

Drake, S.J., Wetz, D.A., Ostanek, J.K., Miller, S.P., Heinzel, J.M., Jain, A. (2014). "Measurement of anisotropic thermophysical properties of cylindrical Li-ion cells," J. Power Sources. DOI: 10.1016/j.jpowsour.2013.11.107.

 


Comparison of Pin Surface Heat Transfer in Arrays of Oblong and Cylindrical Pin Fins

Abstract:

Comparison of pin-average heat transfer in third row of the array.Pin fin arrays are most commonly used to promote convective cooling within the internal passages of gas turbine airfoils. Contributing to the heat transfer are the surfaces of the channel walls as well as the pin itself. Generally the pin fin cross section is circular; however, certain applications benefit from using other shapes such as oblong pin fins. The current study focuses on characterizing the heat transfer distribution on the surface of oblong pin fins with a particular focus on pin spacing effects. Comparisons were made with circular cylindrical pin fins, where both oblong and circular cylindrical pins had a height-to-diameter ratio of unity, with both streamwise and spanwise spacing varying between two and three diameters. To determine the effect of relative pin placement, measurements were taken in the first of a single row and in the third row of a multirow array. Results showed that area-averaged heat transfer on the pin surface was between 30 and 35% lower for oblong pins in comparison to cylindrical. While heat transfer on the circular cylindrical pin experienced one minimum prior to boundary layer separation, heat transfer on the oblong pin fins experienced two minimums, where one is located before the boundary layer transitions to a turbulent boundary layer and the other prior to separation at the trailing edge.

 

Reference:

Kirsch, K.L., Ostanek, J.K., Thole, K.A. (2014). "Comparison of Pin Surface Heat Transfer in Arrays of Oblong and Cylindrical Pin Fins," J. Turbomach. vol. 136, pp. 041015 (1-10).

 


Improving Pin-Fin Heat Transfer Predictions using Artificial Neural Networks

 

Abstract:

Schematics of pin-fins in a gas turbine blade and ANN model.  Comparison of semi-empirical model performance to the ANN model. In much of the public literature on pin-fin heat transfer, the Nusselt number is presented as a function of Reynolds number using a power-law correlation. Power-law correlations typically have an accuracy of 20% while the experimental uncertainty of such measurements is typically between 5% and 10%. Additionally, the use of power-law correlations may require many sets of empirical constants to fully characterize heat transfer for different geometrical arrangements. In the present work, artificial neural networks were used to predict heat transfer as a function of streamwise spacing, spanwise spacing, pin-fin height, Reynolds number, and row position. When predicting experimental heat transfer data, the neural network was able to predict 73% of array-averaged heat transfer data to within 10% accuracy while published power-law correlations predicted 48% of the data to within 10% accuracy. Similarly, the neural network predicted 81% of row-averaged data to within 10% accuracy while 52% of the data was predicted to within 10% accuracy using power-law correlations. The present work shows that first-order heat transfer predictions may be simplified by using a single neural network model rather than combining or interpolating between power-law correlations. Furthermore, the neural network may be expanded to include additional pin-fin features of interest such as fillets, duct rotation, pin shape, pin inclination angle, and more making neural networks expandable and adaptable models for predicting pin-fin heat transfer.

 

Reference:

Ostanek, J.K., 2014, “Improving Pin-Fin Heat Transfer Predictions Using Artificial Neural Networks,” J. Turbomach. vol. 136, pp. 051010 (1-9).

 


Effect of Streamwise Spacing on Periodic and Random Unsteadiness in a Bundle of Short Cylinders Confined in a Channel

 

Abstract:

Images of flowfield and decomposed flowfield showing contribution of periodic motions and random motions.While flow across long tube bundles is considered classical data, pin-fin arrays made up of short tubes have become a growing topic of interest for use in cooling gas turbine airfoils. Data from the literature indicate that decreasing streamwise spacing increases heat transfer in pin-fin arrays; however, the specific mechanism that causes increased heat transfer coefficients remains unknown. The present work makes use of time-resolved PIV to quantify the effects of streamwise spacing on the turbulent near wake throughout various pin-fin array spacings. Specifically, proper orthogonal decomposition was used to separate the (quasi-) periodic motion from vortex shedding and the random motion from turbulent eddies. Reynolds number flow conditions of 3.0e3 and 2.0e4, based on pin-fin diameter and velocity at the minimum flow area, were considered. Streamwise spacing was varied from 3.46 pin diameters to 1.73 pin diameters while the pin-fin height-to-diameter ratio was unity and the spanwise spacing was held constant at two diameters. Results indicated that (quasi-) periodic motions were attenuated at closer streamwise spacings while the level of random motions was not strongly dependent on pin-fin spacing. This trend was observed at both Reynolds number conditions considered. Because closer spacings exhibit higher heat transfer levels, the present results imply that periodic motions may not contribute to heat transfer, although further experimentation is required.

 

Reference:

Ostanek, J.K. and Thole, K.A., 2012, “Effect of Streamwise Spacing on Periodic and Random Unsteadiness in a Bundle of Short Cylinders Confined in a Channel,” Exp. Fluids. vol. 53(6). pp. 1779-1796.

 


Wake Development in Staggered Short Cylinder Arrays within a Channel

 

Abstract:

Instantaneous flowfield for pin-fin wakes at different Reynolds numbers. Staggered arrays of short cylinders, known as pin–fins, are commonly used as a heat exchange method in many applications such as cooling electronic equipment and cooling the trailing edge of gas turbine airfoils. This study investigates the near wake flow as it develops through arrays of staggered pin fins. The height-to-diameter ratio was unity while the transverse spacing was kept constant at two cylinder diameters. The streamwise spacing was varied between 3.46 and 1.73 cylinder diameters. For each geometric arrangement, experiments were conducted at Reynolds numbers of 3.0e3 and 2.0e4 based on cylinder diameter and velocity through the minimum flow area of the array. Time-resolved flowfield measurements provided insight into the dependence of row position, Reynolds number, and streamwise spacing. Decreasing streamwise spacing resulted in increased Strouhal number as the near wake length scales were confined. In the first row of the bundle, low Reynolds number flows were mainly shearlayer-driven while high Reynolds number flows were dominated by periodic vortex shedding. The level of velocity fluctuations increased for cases having stronger vortex shedding. The effect of streamwise spacing was most apparent in the reduction of velocity fluctuations in the wake when the spacing between rows was reduced from 2.60 diameters to 2.16 diameters.

 

Reference:

Ostanek, J.K. and Thole, K.A., 2012, “Wake Development in Staggered Short Cylinder Arrays within a Channel”, Exp. in Fluids. vol. 53(3).  pp. 673-697.

 


Flowfield Measurements in a Single Row of Low Aspect Ratio Pin-Fins

 

Abstract:

Instantaneous flowfield showing horseshoe vortex system in pin-fin arrays.Pin-fin arrays are commonly used as compact heat exchangers for cooling the trailing edge of gas turbine airfoils. While much research has been devoted to the heat transfer characteristics of various pin-fin configurations, little work has been done to investigate the flowfield in pin-fin arrays. Such information may allow for further optimization of pin-fin configurations. A new pin-fin facility at large scale has been constructed to allow optical access for the use of nonintrusive measurement techniques such as laser Doppler velocimetry and time-resolved, digital particle image velocimetry. Using these techniques, the flow through a single row of pin fins having a height-to-diameter ratio of 2 and span-to-diameter ratio of 2.5 was investigated. Results showed that the length of the wake region decreased with increasing Reynolds number. At higher Reynolds numbers, Karman vortices developed closer to the pin fins than for single, infinitely long cylinders. Transverse fluctuations correlated well with endwall heat transfer indicating that the Karman vortices play a key role in energy transport. 

 

Reference:

Ostanek, J.K. and Thole, K.A. 2012,“Flowfield Measurements in a Single Row of Low Aspect Ratio Pin-Fins,” J. Turbomach., vol. 134. pp. 051034 (1-10).