Thermal runaway is normally defined as the increase in charge or float current that occurs as a result of the increase in cell temperature from the initial applied constant potential. If left unchecked, the currents can reach high values and, ultimately, lead to the destruction of the cell. This definition does not explain why all cells floated at constant potential do not suffer from thermal runaway. The aim of this paper was to investigate and explain the cause of this. Thermal runaway is normally defined as the increase in charge or float current that occurs as a result of the increase in cell temperature from the initial applied constant potential. If left unchecked, the currents can reach high values and, ultimately, lead to the destruction of the cell. This definition does not explain why all cells floated at constant potential do not suffer from thermal runaway. The aim of this paper was to investigate and explain the cause of this transition from normal stable behaviour to unstable thermal runaway.A series of 6 V, 100 A h, valve-regulated lead-acid (VRLA) batteries were overcharged at potentials of up to 2.65 V per cell and the currents, temperatures and gas-evolution rates measured during thermal runaway. From these results, it was concluded that separator dry-out was the critical parameter that controls thermal runaway behaviour. This conclusion was reinforced by other data for the effect of saturation on the resistance, the normal float behaviour and the gas transport in VRLA separators.A model of the structure of partially saturated separators was developed to explain the observed behaviour, and was used to predict possible improvements in separator structure to increase resistance to runaway.••BatteryLead-acidSaturationSeparator dry-outThermal runawayValve-regulatedThe problem of thermal runaway in sealed nickel–cadmium cells is well known and chargers either use constant current or some form of modified constant potential to avoid the problem. There is no such widespread problem in valve-regulated lead-acid (VRLA) cells. Nevertheless, thermal runaway has been recognised as a possible failure mode in VRLA cells although its incidence is small,,. There is little experimental evidence in the open literature to assist the understanding of the important parameters that result in runaway. It is generally believed, however, that float potential, separator dry-out, temperature and insulation are of importance,,.Thermal runaway is usually considered to be the result of positive feedback of current and temperature when a cell is placed on float charge at constant potential. The initial float current flowing through the cell causes an increase in cell temperature, this causes an increase in current that further increases the temperature until both current and temperature reach high values. Berndt suggests that the phenomenon arises because heat generation has an exponential relationship with temperature but heat dissipation has a linear relationship. This explanation does not indicate why cells under normal float conditions do not experience runaway. There must be some parameter, or set of parameters, that changes normal well-behaved float behaviour into runawa. All experiments were carried out with single VRLA batteries (6 V, 100 A h) that were designed for standby applications. Most trials were conducted on new batteries but, for comparison, some old batteries that had seen 10 years service and gave 80% rated capacity (i.e., at end-of-life) were also used. For each test, the battery was placed on float at 2.28 V per cell (Vpc) and at the required test temperature until the float current was stable. This was normally 24 h. The float potential was then raised to the test potential and the current allowed to rise without any limit. During the experiment, the current was monitored at regular intervals. In some tests, the case temperature was monitored with a thermocouple and the evolved gas measured by collection over water. The composition of the evolved gas was determined by gas chromatography. A full list of the tests is given in Table 1.Table 1. Experimental conditions and measurements taken3.1. Current profilesThe current profiles for new batteries during the first 100 days of the trial are given in Fig. 1. The data reveal several important characteristics. The currents generally show a slow increase for a period, followed by a rapid rise to thermal runaway. Peak currents are of the order of 50 A. This is followed by an even more rapid decrease to virtually zero current. Applied potential was the main driver for thermal runaway; all batteries at 2.62–2.65 Vpc exhibited thermal runaway, whereas those at 2.40 Vpc showed a much lower and more uncertain trend to runaway. In fact, at this lower voltage, it was taking in excess of 50 days for the current to reach high levels and it is debatable if this can be termed thermal runaway. It is interesting that temperature is shown here not to be a cause of runaway. It does, however, act as an accelerating factor, i.e., all the batteries at 2.65 Vpc suffered runaway, but those at the higher ambient temperature were the first to display the effect. The final point of interest worth noting is the length of time it took to achieve runaway. Even under the most severe conditions of 2.65 Vpc and 60 °C, almost a day elapsed before runaway occurred. At a potential of 2.40 Vpc, the time involved becomes very long indeed. Data for the full test period of 340 days are presented in Fig. 2. Even at this stage, the 2.4 Vpc/40 °C battery displayed no sign of runaway. It is theref.
