Enhancing EV Safety & Performance: The Essential Role of Temperature Monitoring

Author : Joel Sylvester, CTO of Dukosi

12 March 2024

The continuous monitoring of voltage, current and temperature are crucial for electric vehicle (EV) battery packs to maintain predictable performance and for the constituent cells to stay within their safe operating area (SOA). However, maintaining safety requires frequent and accurate monitoring. By reassessing how we approach monitoring EV battery temperatures, could we enhance safety and performance?

A well-engineered battery management system (BMS) needs to monitor the internal cells as well as managing charging network connections - all while accounting for the impact of internal cell and external environmental temperature. Therefore, accurate cell monitoring is of utmost importance throughout the entire operation of a vehicle. This becomes especially critical during the charging process, as the influx of sustained high current raises battery pack temperature. However, the cells inside can have manufacturing differences that mean some become hotter than others. Therefore, it is necessary to carefully balance charging speeds to uphold safety.

A typical EV battery pack consists of several modules, with each generally holding 12 to 16 cells. It is common though for such modules to incorporate just a couple of temperature sensors - a limitation that can impact on both performance and safety. Consequently, the inclusion of more temperature sensors offers an opportunity to enhance the monitoring process, thereby improving safety, as well as potentially enhancing both vehicle and charging performance.

The importance of temperature sensors
So why is it that the number of temperature sensors is currently significantly less than the number of cells in each module? Primarily, it is because the incorporation of a network of sensors, wiring and connectors into a pack introduces additional weight, material expenses and conflicts in relation to minimising the likelihood of short circuits. These economic and packaging limitations significantly influence the quantity of temperature sensors that can be accommodated within a battery pack.

As a compromise, designers are usually required to place temperature sensors at strategic points of interest, such as the side of 12-16 cell modules or the end of modern cell-to-pack designs. However, it must be noted that these sensors are often located quite far from the monitoring boards.
Consider a scenario where multiple prismatic cells are arranged linearly inside a single module, with the assumption of 1 temperature sensor for every 8 cells (as is commonplace). If any of these cells sustains damage, its temperature can start to rise, potentially exceeding the manufacturer’s maximum figure. However, in this scenario, the damaged cell lacks a dedicated temperature sensor, and the nearest sensor is located several cells away. There is thus a potential delay in the BMS detecting a thermal limit breach. In order to maximise safety, it is imperative that abnormal behaviour is detected before a thermal runaway event.
Thermal runaway propagation
Some packaging configurations might facilitate more efficient distribution of sensors, however, even under the most favourable circumstances, the vast majority of cells will remain without a sensor - instead relying on transmission of temperature increase from the faulty cell to its neighbours for detection. 

Figure 1: Thermal runaway in a cell can lead to similar occurrences in adjacent cells
Figure 1: Thermal runaway in a cell can lead to similar occurrences in adjacent cells

When cell temperature exceeds a threshold, it can trigger thermal runaway. In this particular scenario, the rising temperature leads to the melting of the separator between the electrodes. Consequently, the cell can release combustible gases and ignite, posing a genuine risk of fire spreading to adjacent cells. A simulated example of this is shown in Figure 2.
During the experiment, cell 1 was exposed to a heat source for a duration of approximately 4mins, resulting in a swift elevation of its temperature to approximately 170°C. Even when the heat source was taken away, thermal runaway had already commenced. This can be seen in the subsequent exponential temperature rise occurring in the following 10mins. An almost instantaneous rise in temperature to nearly 380°C is then shown, with the remaining energy in the cell being released all at once as it ‘burnt out’. Meanwhile, the temperature of cell 2 started to rise as a result of heat conduction from cell 1.

Approximately 10mins following the activation of cell 1, cell 2 also entered a state of thermal runaway, ultimately reaching burnout at around 27mins. By then cells 3 and 4 were thermally engaged too, leading to temperatures nearing 400°C and their eventually burnout.

Overcoming existing economic/technical limitations
Aiming to help engineers prevent such situations, but without the need for running countless temperature sensors across battery modules, Dukosi proposes a new approach to safety monitoring with its DK8102 cell monitor chip. Ideally placed on every cell, these devices not only check cell voltage but also each include a built-in temperature sensor. This enables thermal monitoring of individual cells without the additional cost and design complexity of external thermocouples. Furthermore, DK8102 devices support up to 2 optional external thermistors, making them suitable for larger cells or other critical areas. So, instead of a single temperature measurement covering multiple cells, this chip-on-cell technology has the capability of capturing up to 3 real-time temperature measurements per cell.

Every DK8102 features dedicated on-board processing resources and integrated storage for lifelong data recording. In EVs, the chip-on-cell architecture leverages a single bus antenna and contactless, near field connectivity, along with Dukosi’s proprietary C-SynQ communication - altogether simplifying battery design and enhancing long-term pack reliability.

Figure 2: A graph demonstrating propagation of heat and thermal runaway events occurring between 4 adjacent cells
Figure 2: A graph demonstrating propagation of heat and thermal runaway events occurring between 4 adjacent cells

Thermal runaway scenario
For the example previously covered, if a temperature sensor had been present on cell 1, the vehicle’s control systems could have been alerted to the problem when it reached a 80°C threshold value (which occurred after about 1min). If the neighbouring cell had a temperature sensor, it would not detect a similar failure until 7.5mins later. If the temperature sensor was 2 cells away, the delay would be over 15mins, and 25mins for 3 cells away. 

The example depicts a simplified version of a standard battery pack, whereas a properly designed arrangement would incorporate gas and pressure sensors to detect such a catastrophic failure when the initial cell begins venting gases. However, even this method is significantly slower, taking approximately 10mins more than using a temperature sensor on each cell. In EVs, this problem is well understood and addressed by using specialised materials to contain heat and prevent spread. Nevertheless, it remains true that there could be a delay of several mins before the issue is detected. It is not difficult to see how such delays present a critical safety problem that could be solved by placing temperature sensors on every cell.

Enabling faster charging and increasing available capacity 
The rate at which cells can be charged is often restricted by thermal factors. As current levels rise, cells’ temperatures increase, necessitating dissipation of heat into their surroundings. If the temperature of each cell is unknown, fast charging controllers must be cautious, because cell temperatures can only be estimated for the areas lacking direct sensor data.

Similarly, cold temperatures also pose a significant risk during rapid charging due to increased chances of lithium plating and dendrite growth. This leads to unwanted loss of active lithium and weakens cells upon returning to higher temperatures. To ensure optimal performance, keeping the temperature of a typical EV powerpack within the 15oC to 35oC range is recommended. A study conducted by the US Office of Energy Efficiency & Renewable Energy revealed that EV range can be reduced by as much as 39% in freezing temperatures. Achieving optimal efficiency, reliability and safety is dependent on staying within the prescribed temperature range, as any significant deviations may cause notable performance declines and cell degradation.
As a battery pack charges, the cells generate heat, but manufacturing discrepancies result in temperature disparities. This can lead to some cells becoming much hotter than others (see Figure 3). If a cell does not have a temperature sensor, then abnormal temperature rises might go unnoticed. With a sensor on every cell, errant behaviour can be identified immediately.

Figure 3: An illustration showing differences in temperatures across an EV battery pack
Figure 3: An illustration showing differences in temperatures across an EV battery pack

Keeping battery cells within the recommended operating window is challenging for vehicle manufacturers. But by accurately monitoring the temperature of each individual cell, it is possible to ensure that they all operate within a safe operating range, reducing the risk of overheating and potential damage. As well as enhancing safety, this approach also maximises battery pack lifespan and performance. Furthermore, the ability to record temperature measurements on every cell provides valuable data for diagnostics and preventive maintenance, allowing for early detection of any abnormalities or potential issues. Proactive monitoring can help battery designers mitigate overheating risks, contributing to more reliable and robust energy storage systems.

Contact Details and Archive...

Print this page | E-mail this page