Electric Vehicles

Electric Vehicles

How Effective?

  • Post-impact safety issues were discussed by a ‘group of interested experts’ in 2009, with regards to useful and appropriate amendments to relevant United Nations Regulations.

  • UN Regulation 100 specifies requirements for the safety of electric vehicles ‘in-use’ and the protection of users against electric shock, and is now mandatory for EU type-approval. Amendments to UN Regulations 94 (front impacts) and 95 (side impacts) were completed in 2010 and include requirements for post-impact electrical safety. These have since been adopted by WP.29 (the World Forum for Harmonisation of Vehicle Regulations, a subsidiary of UNECE). Other amendments include UN Regulation 13 and 13-H (emissions), 85 (engine power), and 101 (CO2 emissions).

  • A proposal made at the 155th session of WP.29 (the World Forum for Harmonisation of Vehicle Regulations, a subsidiary of UNECE) is concerned with the development of a UN Global Technical Regulation containing safety provisions for the electrical safety. This includes post-impact issues, such as:

    • Electrical isolation

    • Battery integrity

    • Best practices or guidelines for manufacturers and/or emergency responders

    • Battery discharge procedures.

(Visvikis, 2012)

  • Electric propulsion technology inherent in EVs carries some potential risks which are unique from conventional ICE vehicles, and which are important for post-impact vehicle safety issues. Rechargeable energy storage systems (RESS) are used to provide electrical energy for the electric propulsion of EVs, and may be realised through batteries, capacitors and electromechanical flywheels. Hazards associated with batteries have been given particular attention, which can include:

    • Electrolyte (or other material) spillage due to cell casing damage;

    • Chemical reaction to extreme temperature or fire, and;

    • Electrical risks such as short circuit, over voltage and voltage reversal.

(Visvikis et al., 2010)

  • If an electric vehicle is involved in a collision, the impact could compromise the integrity of the electrical systems and increase the risk of electrical shock. For example, if electrical isolation is damaged such that the positive and negative terminals in the circuit come into contact with the vehicle bodywork, then subsequent human contact with the bodywork will result in electric shock. Protective mechanisms exist to disengage the high-voltage electrical system in the event of a crash thereby reducing the risk of this event. This is also an important mechanism for reducing the risk of fire or explosion following an impact.

(Visvikis, 2012)

  • Additional hazards are presented by EVs powered by lithium ion batteries. Lithium is a highly reactive substance and electrolytes are highly flammable. As a safety measure, manufacturers of lithium ion batteries incorporate internal and external protection mechanisms to reduce the risk, such as fuses, PTCs, built-in weak spots and flame retardants added to the electrolyte.

(Tjus, 2011)

  • The safety of batteries may be assessed via appropriate abuse tests. Issues related to lithium ion batteries have included thermal runaway, electrolyte leaks, smoke, venting, fire and explosion as a result of several abuse tests including overvoltage, overdischarge, heating arcing, crush, nail penetration and internal and external short.

(Pesaran et al., 2009)

  • Lithium ion batteries have also been investigated elsewhere using abuse tests, such as perforation, mechanical crushing, external short circuit, overcharging/overdischarging, overheating, fuel fire immersion and water immersion. The tests are used to characterise the tolerance levels of lithium ion batteries but it has been argued that they do not necessarily offer a measure of battery safety. For example, perforation and overheating abuse tests have been shown to lead to cell venting, but this event does not result in reduced safety of a battery pack which has sufficient protective devices. The authors of this work also argue that few serious incidents relating to lithium ion batteries have been reported despite extensive worldwide use over the last decade. Further advances in battery technology such as iron phosphate positives or lithium titanate negatives have produced greater abuse tolerance thereby increasing safety when used in EVs.

(Kalhammer et al., 2009)

  • However, the risk associated with electric batteries was highlighted in one incident in the USA with a Chevrolet Volt. Three weeks after a pole impact test, the electric caught fire as a result of damage to the battery. Coolant leaked onto electronic components following the impact, and because the battery was not properly discharged the vehicle caught fire whilst it was placed in storage. This illustrates the importance of managing risks associated with battery safety, even after an impact has occurred.

(Smith, 2012)

  • There are a number of properties of hydrogen which are relevant to hydrogen fuel-cell vehicle safety.

    • Hydrogen gas is odourless and colourless meaning it is harder to detect than natural gas and gasoline vapour.

    • Hydrogen burns with a blue flame which is almost undetectable in daylight.

    • The high buoyancy of hydrogen means that it will disperse in the air around 4 times and 10 times faster than natural gas and gasoline vapour, respectively. When mixed with air there is a risk of fire.

    • The small molecular structure of hydrogen gas means it can more easily leak through porous structures.

    • Hydrogen is flammable between 4% and 75% by volume with air. In comparison, gasoline is flammable between 1% and 8% by volume with air.

    • ‘Hydrogen embrittlement’ can occur following extended exposure of some materials, such as steel, to hydrogen. This can cause leakage or catastrophic failure of that material.

(Visvikis et al., 2010)

  • As a result of the chemical composition of hydrogen, hydrogen fuel-cell vehicles carry additional risks. These may be classified as:

    • Physiological (e.g. asphyxiation, thermal burns, frostbite, hypothermia and overpress injury)

    • Physical (e.g. component failures due to low temperature deterioration, thermal contraction and hydrogen embrittlement)

    • Chemical (e.g. burning or explosion)

  • The odourless, colourless and tasteless nature of hydrogen gas means that leaks cannot be detected by humans without the use of hydrogen sensors. Unlike natural gas which has added odorants to enable easy detection, current odorants in existence have been shown to contaminate fuel cells, and so are not appropriate for use in vehicles.

(Rigas and Amyotte, 2013)

  • Post-crash limits for liquid fuels exist and have been used to benchmark the leakage limits for hydrogen in industry standards, relevant to the use of hydrogen fuel-cell vehicles. However, due to differences in the chemical properties of hydrogen compared with conventional liquid fuels, this benchmark may not be appropriate.

  • In one study investigating the properties of hydrogen, the authors concluded that the volume of hydrogen leaked is more important than the leak rate when considering safety implications. Specifically the study recommended:

    • Accumulation of hydrogen in the passenger compartments should be avoided.

    • Multiple sensors may be required to alert passengers of a hydrogen leak.

    • Post-accident devices which vent the passenger compartments are effective.

  • The study provides some information on the the likely impact of hydrogen leakage and accumulation in a vehicle, however, as the authors state, “the study is not indicative of how a hydrogen fuel system would perform in a crash” since the tests involved a simulation using conventional vehicles and ‘vehicle compartment simulators’.

(Hennessey et al.)

  • An analysis of fire hazards for the existing vehicle fleet and the Emerging Fuel Vehicle (EFV) fleet found that, compared with a rate of 350 fire-related deaths per year as a result of traditional fuel vehicles (ICE), there were 420, 910 and 1300 predicted fire-related deaths per year for petrol-electric HEVs, compressed gas vehicles and hydrogen fuel-cell vehicles, respectively. This analysis was performed via predictive methods of risk assessment and was not based on empirical data, however the model suggests an increased risk of fire-related deaths with EVs over conventional ICE vehicles.

(Levy, 2008)

  • One study from the USA focuses on some of the safety issues which the fire service may encounter. Electric propulsion systems are associated with new and unanticipated hazards compared to conventional ICE vehicles:

    • One of the biggest challenges for emergency responders is identification of EVs when attending an incident. The exterior body of some EVs and HEVs is not easily distinguishable from conventional ICE vehicles, therefore an accurate risk assessment of the associated hazards may be difficult.

    • The addition of a high-voltage electrical system in EVs is a crucial difference from ICE vehicles which is important for vehicle extrication procedures because of the added risk of electrocution. Cabling for the high-voltage electrical power systems in EVs and HEVs are often coded via bright colours to facilitate clear identification.

    • After isolation, it takes a fixed-period of time for a high-voltage system to fully dissipate. It is therefore recommended that at least 10 minutes is allowed before the system can be deemed safe.

    • Due to their ability to function silently, it is crucial for emergency responders to be certain that they have fully disabled an EV. It is recommended that emergency responders always assume an EV is powered-up, even in the absence of vehicle noise.

    • Extinguishing a high-voltage battery fire normally requires either a large volume of water or a ‘defensive approach’ whereby the battery is allowed to burn and consume itself (as long as there a no further exposures to heat). Battery fires can lead to electrolyte spillage so adequate personal protective equipment (PPE) is imperative.

(Grant, 2010)

Risks associated with low vehicle noise emission

  • Some authors have argued that due to the predicted increase of electric and other low-noise vehicles in urban areas, engine noises in these locations are likely to be significantly reduced. It is argued that there may be an increasing number of road traffic accidents between these vehicles and vulnerable road users (with particular emphasis on the elderly, children and those with visual impairments who rely on both visual and auditory signals for perceiving the environment around them), although no evidence of this increase is cited. To reduce the risk of traffic accidents the authors suggest it may be necessary to develop acoustic warning signals, to replace missing engine sound, especially in urban areas.

(Brand et al., 2012)

  • An experiment undertaken by the Dresden University of Technology in Germany investigated how visually impaired, blind and sighted participants responded to the sound produced by internal combustion engine (ICE) vehicles, electric vehicles (EV) and hybrid-electric vehicles (HEV). Three groups of pedestrians (sighted, blind and visually impaired) detected the sound of EVs significantly later than ICE vehicles, with the ICE being detected from a distance of approximately 36m and EVs at approximately 14m. The study also looked at manufactured sounds on EVs and found that synthesised noise, despite being 7dB quieter, produced similar reactions to recorded vehicle sound. The study provides useful data on reaction times to different noises, albeit using a fairly small sample (27 sighted subjects and 10 visually-impaired or blind subjects). However, the study was performed in a laboratory and subjects were asked to imagine they were standing on a pedestrian footway waiting to cross the road, so did not accurately replicate conditions found in the real-world. In addition, no information was provided on the speed of the moving vehicles from which sound was recorded, so application of the results to real-world scenarios is not straight-forward.

(Altinsoy, 2013)

  • There is an important trade-off between optimal noise levels for health and social well-being and optimal noise for the safety of pedestrians, cyclists and other road users who rely on vehicle noise as a warning sound. From a safety perspective, EVs may sometimes be considered too quiet. Any synthetic sound which is applied to EVs must be able to convey vehicle speed, acceleration/deceleration, and driver intentions. Current research is investigating the best sounds to achieve this.

(Chamard et al., 2012)

  • The National Highway Traffic Safety Administration (NHTSA) investigated the rate of crashes between pedestrians and HEVs and ICE vehicles. Accidents involving pedestrians and cyclists involving HEVs and ICE vehicles tended to occur on roadways, in low speed limit areas, during daytime and in clear weather. The study found a higher crash rate with HEVs, with these vehicles being twice as likely to be involved in a slow moving or manoeuvring crash with pedestrians as equivalent ICE vehicles. HEVs were also involved in more slow speed crashes with cyclists than ICE vehicles. At low speeds the noise differences emitted by the ICE and HEV are most different (low tyre or wind noises mean the engine noise – or lack of it – is dominant). Whilst it is possible that noise contributed, the cause of the accidents was not investigated in this study so it cannot be concluded that the higher crash rate attributed to HEVs was due to the lack of engine noise. Furthermore, other than controlling for speed limit, there is no discussion of the location of the collisions investigated so it is possible that the likelihood of encountering pedestrians and cyclists was greater in some scenarios.

(NHTSA, 2009)

  • An in-depth review of research literature, standards and legislation concluded that whilst the detectability of electric vehicles moving at low-speed may be lower than for conventional ICE vehicles, there is currently little strong evidence to suggest an increased risk to vulnerable road users (such as cyclists, pedestrians and the visually-impaired). However, it was argued that the number of electric vehicles on the road could rise in the future which may increase the risk to vulnerable road users. Synthetic noise has been developed to provide a warning to road users in the absence of engine noise; artificial engine sounds appear to be favoured over other types of noise.

(Visvikis et al., 2010)

  • A research project conducted by the Transport Research Laboratory, commissioned by the Department for Transport, reviewed vehicle accident statistics to determine whether the accident risk posed by EVs differed from that posed by traditional ICE vehicles. The key findings were:

    • Considering passenger cars and car-derived vans, the likelihood of being involved in a collision with a pedestrian was equal for EVs/HEVs and traditional ICE vehicles.

    • It appears that whilst the relative number of EVs/HEVs involved in any accident is smaller than ICE vehicles, proportionately more of the EV/HEV accidents involve pedestrians. This may be due to the usage patterns of EV/HEV which gravitate towards urban areas.

    • Only two EV/HEV accidents involving a pedestrian with a disability were identified, so conclusions about risk for visually-impaired road users are not possible.

    • It was not possible to include information on vehicle speed, location or manoeuvre nor was it possible to differentiate between EVs and HEVs. Hence, whether or not HEVs involved in accidents with pedestrians were running on electric power or the conventional ICE is unknown, therefore it cannot be concluded that low noise emission was a contributory factor to the accidents.

    • Measurements of vehicle noise were also conducted with EVs, HEVs and conventional ICE vehicles at a range of speeds in the study. It was concluded that “peaks in the pass-by noise spectra related to exhaust noise” were the only differences in sound profile identified between the various types of vehicle.

(Morgan et al., 2011)

  • Two field studies in Germany examined the perceptions of low noise emission EVs of 70 participant drivers. Interviews and questionnaires were used to gauge driver opinions before driving an EV for the first time and after three and six months of driving an EV. Incidents related to low noise emission were rare, and typically occurred at low speeds. Responses also revealed driver perception of risk due to low noise reduced with increasing driving experience, whilst the perception of driver comfort due to low noise increased over the 6-month driving period. The study provides valuable information on driver perceptions of EV noise emissions through investigation of real-world EV driving experience.

(Cocron and Krems, 2013)

  • One study investigated EV test drivers and their experiences relating to the lack of noise emission. A naturalistic driving study followed 40 participants driving an EV over a 6-month period and found that:

    • Drivers were aware of the low noise emission and aware of the associated potential dangers.

    • Drivers adapted their driving as a result of the low noise emission.

    • The few incidents which were reported typically involved pedestrians and cyclists and occurred at very low speed with no injuries sustained. At higher speeds there are other sound cues from the vehicle (e.g. tyre noise, wind etc.), so low noise emission is largely an urban issue in environments with low speed limits.

  • Whilst a naturalistic driving study is advantageous in that it allows investigation of driver behaviours and experiences in a real-world setting, it is limited in that there is no control of confounding factors. For example, it is not possible to conclude that low noise emission was a contributory factor to the reported incidents between EVs and pedestrians and cyclists. It is also possible that involvement in the research raised the awareness of safe driving practices in drivers, leading them to adopt more cautious behaviours whilst participating in the study.

(Cocron et al., 2010)

  • Three studies were designed by Wogalter et al. (2013) to investigate concerns related to low vehicle noise emission.

    • In the first study, 378 participant questionnaires indicated “substantial positive interest” in EVs, but approximately half of respondents indicated that a silent vehicle would bother them as a pedestrian. The most commonly recommended type of sounds to add to silent vehicles were ‘engine’ and ‘hum’ sounds, but there was also a high degree of acceptance for no sound to be added.

    • The second study investigated preferences for types of sounds in greater detail. 316 participants rated 14 sounds in terms of their acceptability for adding to low noise emission vehicles. As with study 1, the highest rated sounds were ‘engine’ and ‘hum’. ‘No added sound’ was rated as significantly less preferable than engine and hum sounds.Studies 1 and 2 examined EV and sound preferences using only questionnaires and descriptive terms; no actual sounds were heard by participants.

    • The third study offered greater ecological validity by showing 24 participants a video of a moving 2007 Toyota Prius coupled with one of six categories of sound: engine, horn, hum, siren, whistle and white noise (it should be noted that these conditions differ from real-world pedestrian experiences of EV noise). The preferences for each synthetic sound addition were again investigated, and findings revealed that, like the earlier studies, engine noise was most preferable as an auditory cue, followed by hum sounds and white noise.

(Wogalter et al., 2013)

  • A study was performed to investigate the responses of blind participants to a passing electric vehicle with a) no noise added; b) idling engine noise added, and; c) a repeating bell and engine noise added. Compared with a traditional ICE vehicle, the loss of engine noise emitted from an electric vehicle significantly reduced the ease with which blind participants were able to identify approaching vehicles. However, the addition of synthetic engine noise was found to be effective at aiding vehicle identification. All conditions tested in the study involved the vehicle travelling at 15 mph and passing participants at a distance of 15 feet. The performance of blind participants measured here cannot be easily extrapolated to other distances or vehicle speeds: it is possible that the effect of vehicle noise would be even more important at higher speeds but this was not investigated in this study.

(Goodes et al., 2008)

  • Fourteen adults with visual impairments were asked to respond to three different passing vehicles in a randomised trial. The three vehicles tested were: a standard HEV, an HEV with added Vehicle Sound for Pedestrians (VSP), and an ICE vehicle. In both a forward-moving and backward-moving vehicle detection test, the detection distances for the HEV with added VSP and for the ICE vehicle were significantly longer than for the standard HEV with no additional noise. These findings suggest that participants were not as effective at detecting the HEV as a standard ICE vehicle, but were aided by the addition of synthetic vehicle noise (i.e. VSP). Interestingly, there were no differences in detection distances between two different test sites with low and high levels of background noise.

(Kim et al., 2012)

  • A laboratory experiment was set-up to investigate the effect of adding different types of warning sounds to the recorded noise of a passing EV on detectability compared with a standard diesel ICE vehicle. Sound stimuli were designed to investigate three parameters of timbre (tonal content, frequency detuning and amplitude modulation). Listeners were required to signal when they detected the arrival of each car by pressing a button on a computer. Responses times were highly variable across the group of participants, but the study showed that the addition of some types of sound offered improved detectability of the EV, such that the EV was equally detectable as the conventional ICE. However, other types of sound failed to improve the detectability of the EV, despite offering similar sound pressure levels. This research indicates that the composition of the warning sound, and not the volume of the sound, is the key factor for improving pedestrian safety, although, like other laboratory studies in this area the study lacks ecological validity.

(Parizet et al., 2013)

  • To address concerns of low noise emission increasing risk for vulnerable road users, especially visually-impaired pedestrians, the approaching Vehicle Sound for Pedestrians (VSP) system has been developed by Nissan. The purpose of this system is to produce a sound which improves detectability of the vehicle for pedestrians, whilst also satisfying driver requirements and maintaining a quiet environment for neighbourhoods. Designed for implementation in Nissan’s first mass produced Electric Vehicle, the VSP emits one type of sound during forward motion, another type during ‘take-off’, when the vehicle first begins to move and a third type of sound for reverse motion. The characteristics of the sound profile utilise knowledge of human ear sensitivity, age-related hearing loss, typical background noise conditions and visually-impaired pedestrian feedback and preferences gathered by surveys in Japan and USA.

(Tabata et al.)

  • The addition of warning sounds to low noise emission vehicles may only be necessary during a ‘transition’ phase until the majority of vehicles on the road are classified as ‘low-noise’, since after this point the level of public awareness of low-noise vehicles will alter. A warning sound system designed to replace the absence of an ICE should be able to:

    • Draw the attention of vulnerable road users.

    • Be distinguishable from other audible signals in the environment, i.e. it should differ in sound pressure level and in sound profile from a conventional car horn.

    • Contain information on the vehicle speed, including whether it is accelerating or decelerating.

(Voigt et al.)

  • The impact of adding synthetic sound to low noise emission vehicles on driver stress was investigated by the University of Idaho. Through physiological measures (galvanic skin response and heart rate variance) and self-report questionnaires (Short Stress State Questionnaire – SSSQ) researchers examined driver stress whilst operating a Nissan HEV with one of three sound profiles (no added warning sound, manual addition of warning sound via vehicle horn, and automatic addition of warning sound via Nissan VSP system). The study found that the automated VSP system was preferred by drivers, causing less stress than when reliant on a manual warning for pedestrians via the vehicle horn.

(Cottrell and Barton, 2012)

  • In a study with 15 visually-impaired participants, the addition of a VSP alert sound to a HEV resulted in significantly faster and more reliable detection than an identical vehicle with no VSP system and a conventional ICE vehicle. However, no differences were observed in the ability of participants to discriminate between the intended pathway of the vehicles (i.e. whether they turned right at traffic lights or went straight ahead). The study suggests that whilst the addition of a VSP system may enable visually-impaired pedestrians to quickly detect a vehicle, it may not necessarily reduce the risk of collision, for example where vehicles make a right-turn into a side road with pedestrians present.

(Kim et al., 2012)

  • In a summary of literature relating to low noise emission in EVs, it is noted that the World Forum for Harmonisation of Vehicle Regulations has “determined that vehicles present a danger to pedestrians”, prompting the Working Party on Noise to develop actions necessary to mitigate those risks.

(Visvikis, 2012)

Gaps in the evidence

  • Electric vehicle technology is still relatively new to the market. As a result there are a number of gaps in the literature surrounding the safety of EVs. This includes:

    • A lack of research comparing the risk of occupant injury in electric vehicles on the road with conventional ICE vehicles.

    • There is not enough evidence to confirm that the rate of accidents involving vulnerable road users differs for electric vehicles and conventional ICE vehicles.

    • Current investigation into accident rates cannot confirm that the lack of noise emission is a causal factor in the accidents between vulnerable road users and electric vehicles.

    • Risks associated with high-voltage electric batteries have been demonstrated in controlled impact tests, but there is little evidence of the frequency of this risk in real-world road traffic accidents.

  • Date Added: 03 Apr 2012, 08:22 AM
  • Last Update: 05 Feb 2014, 05:16 PM