An electrochemical energy storage device is called a battery. The term covers both rechargeable batteries (accumulators) as well as batteries that are not rechargeable. There are a number of chemical variants, such as lithium-ion, lithium-iron phosphate and lead batteries. Impedance spectroscopy can be used to analyse battery characteristics, regardless of the battery’s chemistry.
The (nominal) power capacity of a battery is the amount of electrical charge that can be stored in it. It is given in either ampere hours or watt hours. Both specifications are equal; the specification in watt hours includes the battery’s nominal voltage. How much capacity can actually be drawn depends on the how often and much the battery is used.
State of Health (SoH)
SoH describes the state of health of a battery as the ratio of its remaining capacity to the original capacity of a brand-new battery. The capacity of brand-new batteries will decrease irreversibly over time, with various factors such as production quality, number and type of charging cycles, usage as well as calendar and sudden ageing contributing. Thus, the state of health indicates the percentage of capacity still available (compared to nominal capacity). Even with batteries of a single type, this value can differ significantly between individual batteries.
State of Charge (SoC)
SoC describes the state of charge of a battery. It indicates the percentage of the available charge compared to a fully charged battery (SoC 100% = fully charged). Since chemical processes take place inside the battery even when it is idle, the state of charge depends on usage and is not a clear value.
Impedance spectroscopy (electrochemical impedance spectroscopy, EIS) is used, among other things, in the field of battery storages. A defined current is fed into the battery at a certain frequency range (spectrum) and the voltage response is recorded. The complex impedance (alternating current resistance) is then calculated using this data, enabling statements to be made about the state of charge (SoC), the state of health (SoH) and even the temperatures inside the battery.
The characteristic curve created when the internal resistance of a battery is measured by impedance spectroscopy is called the impedance spectrum. It can be used, among other things, to determine the state of charge (SOC) and the state of health (SOH). At NOVUM, the impedance spectra are evaluated by neural networks and the results can be used to generate customer-specific forecasts of usage and service life.
State of charge reserve
The state of charge reserve (also known as the state of charge buffer) is the excess capacity planned in the manufacture of a battery to guarantee available power over a certain period of time. For this reason, the battery’s actual capacity is often significantly (> 20%) more than its specified nominal capacity because the state of health of a battery decreases over time. However, the more predictable the battery’s behaviour, the lower these reserves can be kept. This reduction enables a battery type to be optimised or a smaller battery model to be designed with the same capacity.
The charging of a battery is not a linear process, but instead divided into two phases. It is charged with constant current (CC) in the first phase. When the end-of-charge voltage is reached, charging continues with constant voltage (CV). The second charging phase takes considerably more time, which can be significantly shortened through battery management.
Self-reinforcing effect of battery wear
As batteries age – each differently – their internal resistance increases and their usable capacity decreases. This increases a battery’s sensitivity to maximum current and voltage values. This raises the temperature, which in turn has a negative effect on the battery’s health – its ageing accelerates and the cycle starts again. In order to prevent sudden loss of capacity, batteries are usually replaced at 80% of their original capacity (SoH). Precise determination of battery status allows various measures to be taken to slow the cycle down, thereby increasing battery service life.
A virtual image of a battery storage device, battery module or battery cell created using real-time data from the original battery is called a digital twin, and the pair age together. The twin accompanies the original throughout its entire lifetime, constantly learning and therefore becoming increasingly similar to the original. The digital twin can be used in a variety of applications, such as predictive maintenance and simulation of different usage scenarios.
Predictive maintenance refers to the forecasting of undesired conditions, such as loss of capacity or failure of modules or battery storages. These forecasts are made using neural networks, based on available data from the batteries (or the digital twin). This significantly reduces the risk of an unplanned failure, lowers maintenance costs and simplifies the planning for battery replacement in a battery storage device. Recommendations for structuring existing battery storages are also possible.
Machine learning refers to the acquisition of new knowledge through algorithms that are trained to be able to recognise relevant patterns in large amount data. Using the same scheme, new data can be assessed after a learning phase. Machine learning is based on neural network technology.
Neural networks can efficiently evaluate large amounts of data and extract relevant information from it. This type of information processing connects individual artificial neurons (elements that deliver a specific response depending on the incoming signal and according to defined rules) to one another as nodes in a network architecture, with the aim of creating structures able to solve complex problems. The networks adapt themselves to the problems to be solved but can as well be supported by a human “trainer”.
An application programming interface is an interface linking a software system to another so that data can be exchanged.