Bus Fire Safety: Safer battery systems in electric buses

Li-ion batteries are widely used in various consumer products and are beginning to be utilised in various types of electrified vehicles (xEV). xEVs have the potential to be safer than conventional combustion engine vehicles, simply because they have less or no flammable gasoline/diesel onboard^1,2. Additionally, xEVs can potentially increase safety aspects due to, for example, the increased freedom of design enabled by the small size of the electric motors and the various ways that the batteries can be positioned in the vehicle. However, new technologies may also introduce hitherto unknown risks. It is important to study these risks in order to properly redress them for vehicle design.

Electrified buses include hybrid electric buses, plug-in hybrid electric buses and fully electric buses. Fully electric buses have an electric motor and a large battery pack for propulsion – no additional combustion engine for propulsion is used. In the case of hybrids and plug-ins, both an electric motor and a combustion engine is used in tandem where the battery pack has less energy capacity compared to that of fully electric buses. Fully electrical propulsion of buses provides several benefits: zero tail-pipe emissions; more silent driving and higher efficiency. Different bus applications, such as school buses, city buses and long-haul buses, have different ‘electrification needs’. For example, a long-haul fully electric bus would probably require a battery that is too heavy to be able to carry sufficient load in the present transport systems and would therefore need continuous charging, i.e. electric roads, or other forms of energy storage (e.g. hydrogen and fuel cell) or a conventional combustion engine driveline. City buses, on the other hand, require moderate sized batteries because they typically have a lower average speed and travel a shorter distance per route, with the capability of charging the battery at each end of the route as well as at bus stops.

SP Technical Research Institute of Sweden (SP) is involved in the project ‘Safer battery systems in electrified vehicles – development of knowledge, design and requirements to secure a broad introduction of electrified vehicles’, together with Atlas Copco, Chalmers University of Technology and Elforsk and with financial support from the Swedish Energy Agency. The project includes various abuse (destructive) tests on commercial Li-ion battery cells to study the cell response in terms of variables such as temperature, gas emissions, fire and explosion.

Li-ion is the family name of many types of different electrode materials (anode and cathode) all utilising lithium ions travelling between them. The characteristics of different types of Li-ion cells can vary quite significantly; regarding energy and power densities, life time and safety, for instance. Besides different electrode materials, there are other components that vary, for example, electrolyte composition and separators. Li-ion cells for the automotive industry are typically produced with higher quality techniques and materials (more pure raw materials) for improved performance, life time, and safety, etc. However, there are unfortunately no intrinsically safe Li-ion cells with sufficient usable properties (life time, energy/power densities) available today. All Li-ion cells have a safe window in which they operate. If they are outside this window, they can self-heat and eventually go into what is called a ‘thermal runaway’ and potentially cause fire. In order to protect the Li-ion cells from various abusive conditions, such as those shown in Figure 1 on page 00, several safety techniques are employed in a battery pack, some of which are schematically shown in Figure 2 on page 00. A battery management system (BMS) monitors and controls the voltage of each cell, current, temperatures and electrical isolation protection, etc. Mechanical crash structures are used to protect the battery pack from being deformed. Disconnectors (contactors) are used to shut-down a battery pack in regular use (i.e. when a bus is parked and turned-off) and in case of a crash it can also disconnect the battery pack. A fuse is present to protect against short circuiting the battery pack. However, some situations – for example, internal cell short circuiting (on micrometre scale) – are very difficult to completely protect against, despite significant efforts being made in manufacturing. The experience of the consumer market shows that there is a small probability (ppm-level or less) of internal short circuiting in Li-ion cells, potentially resulting in a thermal runaway and a battery fire. With a few-cell-battery (commonly used in consumer products) and with the low probability, the risks associated with internal cell short circuits are general relatively low. However, in a large battery pack, with many cells, the probability of a single cell thermal runaway will of course increase due to the large number of cells, and the potential consequences with such a large battery pack will also increase. This leads to an increased risk of a cell safety incident occurring, making it important to minimise its impact on the rest of the battery pack as well as the electrified bus. In case of a battery cell failure it is essential to have early detection and quick notification to the driver and passengers, since it usually takes approximately two minutes to evacuate all passengers from a fully occupied bus – and possibly longer for passengers with disabilities.

It is important to hinder or delay propagation of a thermal runaway from one battery cell to adjacent cells, or from one battery module to adjacent battery modules. The cell-to-cell propagation of a thermal runaway in a single cell to adjacent battery cells can be significantly affected by battery pack design, however it may add cost, weight and volume. In a battery system ‘fire walls’, for example, between battery modules can be used to delay/stop propagation. The integration and placement of the battery in the bus can also affect propagation.

The numbers of fires in buses with conventional fuels are a concern in several countries today. Common fire sources are excessive heat igniting fuel or oil leakage, wheel fire, and electrical short circuits in the 24 V system. Several of these potential fire sources will still be present in electrified buses. However, for fully electric buses some significant heat ignition sources will be removed because the hot combustion engine and some of its hot subsystems are not needed. Today there are relatively few electrified buses in operation, although the numbers are increasing. There have been a few fire incidents with electric buses but still the statistics are too limited in order to perform an adequate statistical investigation of the probability of a fire in an electrified bus.

The electrical drive line in electric buses uses relatively high electrical voltage systems, both DC (direct current) and AC (alternating current). The battery system voltage is typically about 600 VDC (voltage direct current). The electrical hazards are well-known and are minimised by the vehicle manufacturers by proper design. The use of an electrical two pole system – the so-called ‘floating ground’¹ – significantly increases the electrical safety in an electric bus. In practical use with correct handling by rescue and service personnel, for example, the electrical hazards for the voltage systems are considered to be low.

Buses are generally heavily used and therefore the battery in a fully electric plug-in bus needs to be recharged frequently by pantograph solutions – either from above or by conductive or inductive plate chargers from beneath – and high charge power is needed in order to shorten the charge time. With higher charge currents the risks are increased. This is, however, well-known and adequate technical design solutions are used to handle the high charge currents.

One of the tasks for a BMS is to avoid overcharging cells. However, it can still happen, for example if BMS fails. Figure 3 on page 00 shows a Li-ion cell before and after an overcharge abuse test, where the cell is charged beyond its limits. The overcharge leads, in this case, to a cell venting and releasing gases. Since the cell was in free air it allowed the cell to swell up significantly. Figure 4 on page 00 shows an example of fire due to overcharge.

It is, however, quite seldom that a fire starts and the best way to measure the heat released by cells in a fire is to expose the cells to a fire source, as seen in Figure 5 on page 00 where battery cells are exposed to a propane flame. The tests showed that higher battery electrical charge levels, state of charge (SOC), gives a more rapid heat release rate (HRR) while the total heat release (THR) is roughly the same for all charge levels. Gas emissions were also measured in these tests. The Li-ion cell contains fluorine that can form toxic compounds such as hydrogen fluoride (HF) that can be released^3,4. The gas emissions are not yet well studied and may pose a risk. However, so far no accidents concerning gas emissions have occurred, potentially showing the risk to be small. Still, it is a safety strategy for the cells to release gases if the pressure increases in the cell in order to avoid possible explosion of the cell and a predesigned strategy of how to handle any vented gases is prudent in order to prevent emissions being released into the passenger compartment. In some situations the release of toxic gas emissions might still pose a risk; for example when rapid gas is released from several cells at an indoor bus stop, in a tunnel, parking garage or a multi-storey car park. These situations also pose a larger risk if the gases are ignited. However, more research on gas emissions is needed.

Performing experimental abuse tests is expensive and therefore thermal simulation is a useful tool. To this end, a simulation model with predictive capabilities is under development. The simulation tool is validated against experiments where the heat transfer in a pack of five cells exposed to external heating was measured. The simulations are performed in the versatile Finite-Element software Comsol Multiphysics and are now extended to accommodate modelling fire propagation in large modules consisting of a large number of cells, as is common for xEVs. The model is intended as a screening tool to quickly evaluate different propagation delay solutions and, therefore, the internal structure of each cell is not modelled. The layouts of commonly used cells indicate, however, that the thermal properties are highly anisotropic and therefore anisotropic simulations are employed. Obtaining adequate data of the battery cells is crucial for the computational model, but usually not so easy to access. Data from the literature is used in combination with sensitivity studies of the thermal parameters in order to overcome this obstacle. The results from the test and the simulations agree relatively well until the adjacent cell begins to react and, thus, this rather simplified method can be used to predict the propagation of a thermal runaway event with accurate material data input.

References

1. Larsson, P. Andersson and B.-E. Mellander ‘Are electric vehicles safer than combustion engine vehicles?’, Chapter 4 in Systems perspectives on Electromobility, edited by B. Sandén and P. Wallgren, Chalmers University of Technology, Goteborg, Sweden, ISBN 978-91-980973-9-9, p.33 (2014), https://www.chalmers.se/en/areas-of-advance/energy/publications-media/systems-perspectives/Pages/Systems-Perspectives-on-Electromobility.aspx
2. Larsson, P. Andersson and B.-E. Mellander, ”Lithium-Ion Battery Aspects on Fires in Electrified Vehicles on the Basis of Experimental Abuse Tests”, Batteries, 2, 9 (2016), http://www.mdpi.com/2313-0105/2/2/9/pdf
3. Larsson, P. Andersson, P. Blomqvist, A. Lorén and B.-E. Mellander, ‘Characteristics of lithium-ion batteries during fire tests’, Journal of Power Sources, 271, 414 (2014).
4. Andersson, P. Blomqvist, A. Lorén and F. Larsson, Fire and Materials, DOI: 10.1002/fam.2359, in press.

Biography

Fredrik Larsson has more than 10 years of experience with Li-ion batteries for the automotive industry, in addition to involvement with several HEV and EV projects. Fredrik was previously a technical specialist at Effpower developing Li-ion battery systems. Fredrik joined SP Technical Research Institute of Sweden in 2012 and works full-time conducting research on safety for Li-ion batteries in electrified vehicles and is the Project Leader of ‘Safer Battery Systems in Electrified Vehicles’. Fredrik received a master’s in Engineering Physics (2006), a degree in Licentiate of Engineering (2014) and a PhD degree is expected in 2017 from the Department of Physics at the Chalmers University of Technology.

Johan Anderson has worked in the Fire Research Department of SP Technical Research Institute of Sweden for more than four years and his research is mostly focused on simulations and computational work of fire dynamics and structures exposed to fire. During recent years Johan has participated in several projects regarding safety of electric and hybrid vehicles. He has a master’s degree in physics from the University of Gothenburg and a PhD from the Chalmers University of Technology.

Petra Andersson has worked with fire research for more than 20 years. She obtained her PhD in Fire Safety Engineering at Lund University in 1997 and has since worked in the Fire Research Department of SP Technical Research Institute of Sweden on various research topics such as fire detection, functional performance during fires, extinguishment and environmental effects. Her research includes both simulations and experiments with most recent research focussing on Electric and Hybrid vehicles. Petra will be Chairing the Scientific Committee of the FIVE 2016 Fires in Vehicles conference in October 2016.

Bengt-Erik Mellander is Professor at the Department of Physics at Chalmers University of Technology. He has wide experience of research on energy related applications, especially related to battery safety, photoelectrochemical solar cells and solid oxide fuel cells.