Blog

Round-Trip Efficiency as a performance guarantee for Battery Energy Storage Systems

Round-Trip Efficiency accounts for the energy losses and therefore energy costs during BESS operation phase. But what does RTE mean and how is it measured? Read this article to have a clearer picture of the energy cost -related OPEX of a BESS investment.

Introduction

Round-Trip Efficiency (RTE) is a widely used performance metric in Battery Energy Storage Systems (BESSs). It is considered important when planning an investment, as it has an intuitive meaning in an investment calculation: it accounts for the energy losses incurred by the BESS, which translate into energy costs as OPEX.

RTE is expressed as a percentage, such as 88%, and indicates how much of the energy charged into the BESS can later be discharged at the same point of measurement. In other words, with the RTE of 88%, 12% of the energy would be lost on the way.

The above is how RTE is intuitively understood. However, understanding the intuitive definition is still far from understanding the operation-point-dependent nature of this metric and how it is measured in real life. The objective of this text is to elaborate on how the different losses of a BESS account to an RTE are defined and measured.

Definition

The author’s personal experience in the BESS market is that not everyone adheres to a clear definition of RTE or a clear measurement method to verify performance. I interpret this as due to a general lack of knowledge about this metric.

There is often an inherent incentive for suppliers to provide a higher RTE value.

Since a higher value is interpreted to mean lower OPEX and theoretically more positive cash flows, the buyers often favor higher RTE promises. This can mean, for example, scoring higher points in a tender or generally favoring the supplier giving the highest RTE percentage.
Due to the above, promising a higher number by using a non-standard calculation method or RTE definition (instead of a standard definition), or a non-standard definition of how the RTE is measured or verified.

The buyers, of course, want to make sure that they get what is promised to them. In delivery contracts, RTE promises are often backed by contractual penalties, such as liquidated damages, that are proportional to the deviation. Also, there might be a contractual minimum for the RTE that gives rise to contract termination or triggers a make-good-type remedial action.

The IEC-62933-2-1 standard provides a clear and detailed definition of the RTE. The first, and maybe most important, lesson arising from this definition – which will be a theme throughout this text – is that RTEs of two different BESSs cannot be directly compared; instead, they are system-design-specific metrics. They could be compared if the building blocks, Point of Connection, and design specifications were exactly the same. The intent of the IEC methodology is not to model typical operation, but to ensure repeatable and comparable worst-case testing. The first relevant point in approaching RTE is to define the Point of Connection (PoC, as used in the above-referred IEC standard), the contractual point between the buyer and the supplier at which the RTE is measured. In the graph below, the main powertrain of a high voltage connected is illustrated.

Typical values for RTE-% are between 84-92% measured on the AC side, depending on the exact location of the PoC, rated power, storage duration and the battery type. Typical values for the RTE-% for only the batteries measured at DC connection are between 90-96%, again depending on the rated power it is measured with.

For the exact same batteries, converters, and transformers, the RTE would differ depending on how the PoC is defined. For example, there would be a different result for the RTE measurement if the PoC is defined to be in High Voltage at an electrical substation, or if the RTE is defined, for example, at the Medium Voltage transformer terminals.

The above considers only the main circuit power train (charging and discharging from battery to grid and back), but the IEC standard clearly states that BESS auxiliary power use must also be taken into account in its definition and measurement. The IEC-standard-based measurement for RTE is performed by first fully discharging the BESS. Then the energy measurement is started at the PoC, and the BESS is charged and discharged fully for at least two complete charge and discharge cycles (potentially more) with the rated charge and discharge power. The discharged and charged energy is measured, respectively, and the RTE is calculated based on their ratio, i.e.

After that, the average result of the RTE of these cycles is calculated to determine the RTE performance value. It is typical that RTE varies a bit between the measurement cycles based on thermal variations, etc.

It is important to note that the test is carried out at the rated power, which means the losses are at their maximum, and so is the auxiliary power use (dominated by the battery cooling system). This provides comparability between different supplier’s approaches to the same specification, but is only applicable when you charge and discharge the BESS at full power all the time.

For example, in frequency regulation applications or multi-market optimization with the BESS, the average BESS charge or discharge power is typically considerably lower than the rated power (where RTE is defined), and thus it yields a too conservative estimate of practical energy losses.

The operation point-dependent nature of RTE

The energy losses of each component vary at different operating points, leading to different charge and discharge power. Of course, for example, the cooling system’s efficiency is also different at different ambient temperatures and starting points (cooling warm batteries that have been in use or cooling batteries that have already been cooled down), but we’ll neglect these details in this analysis.

The Power Conversion System (PCS) losses in % are typically different at each operating point than at their rated power or peak efficiency power, where the losses are typically defined by the manufacturer. And depending on the system design, the RTE is not necessarily measured at the power where the PCS efficiency is indicated in the manufacturer’s datasheet.

In the table below, the nature of the components causing losses and how they change based on the operation point are explained.

Component in the BESS power train	Power losses’ dependence on the operation point
Transformers	Transformers’ losses are formed by no-load losses and load losses, which are typically defined in the transformer datasheet. No-Load losses are magnetizing currents, etc., that will be produced regardless of the operating point and can be considered fixed. Load Losses are typically defined at the transformer’s rated power; if the actual loading is lower or higher, the losses are scaled accordingly. Transformers’ load losses can primarily be considered ohmic in nature, and therefore they scale in a quadratic relationship to the load power, i.e. Because the entire charge/discharge power flows through the system’s transformers, they can significantly influence the overall RTE. EcoDesign transformers are preferred for minimizing losses.
Cables	Unless there are really long cables in the system (hundreds of meters), the cables’ contribution to overall RTE is negligibly small. Typically, BESSs are designed so that the low-voltage cabling (which would produce most losses due to higher current) length is minimized, i.e., batteries, PCS, and transformers are all close to each other. Longer cabling is typically used for medium-voltage applications. Cable losses can be easily calculated from the cable datasheet values (resistance per km of cabling). The losses are ohmic in nature, and therefore, the above equation would be applicable.
Power Conversion System	Power Conversion System losses are caused mainly by the semiconductors’ switching (and ohmic) losses. These losses can vary somewhat between manufacturers and different converter topologies. Typically, manufacturers state only one value (peak efficiency), but depending on the topology, efficiency can vary a lot. Therefore, if, for example, a 2 MW PCS converter is used at 1 MW power, it is necessary to verify what the efficiency is at that operation point. The losses change with the operating point and can be roughly modeled as linear. Typically, there is a minor difference in PCS efficiency in charge and discharge direction.
Batteries	The batteries’ internal power losses can be roughly modeled to be mainly ohmic in nature, which means that the above equation for losses scaling quadratically is applicable. It is also important to note that there is a difference in battery efficiency between charge and discharge, which causes asymmetric losses; the majority of the battery losses during a full round-trip are accumulated in the discharge direction. There are other minor loss mechanisms, such as entropic heat losses, etc., but modeling the battery as an ohmic load provides a reasonable approximation.
Auxiliary power	Instantaneous auxiliary power losses are very difficult to estimate, as the auxiliary power use typically has a time delay to the changes in the charge/discharge power, unless the power has been constant for a long time. For example, if the batteries are charged with the rated power of the system, but prior to charging the batteries have been in rest, due to the large mass and thermal constant of the batteries, it will take a while before the battery cooling system (which would produce most of the losses in proportion) kicks in and starts drawing full power. Therefore, auxiliary power losses should be averaged or, better yet, considered purely energetically when evaluating RTE. Some auxiliary power use (such as control systems, etc.) is a fixed load and therefore can be considered fixed irrespective of the load. However, the main auxiliary load – the cooling system for PCS and batteries – follows the same logic as the heat losses for the aforementioned components, but considers HVAC/chiller efficiency.

Note that the above table only considers power losses (MW), but Round-Trip Efficiency is about energy. When instantaneous power losses are considered over the charging and discharging times of the BESS, we can calculate the energy losses. In this text and the provided RTE calculation sheet, the power (and the losses) is assumed constant during charge and discharge processes. This is an approximation and is not exactly true due to changing State of Charge (SoC) of the battery and related change in battery terminal voltage, but when the power is averaged over the whole charge or discharge process, the approximation is sufficiently precise.

The losses’ scaling becomes different when energy is considered, because the charging and discharging times change according to the operation point of the BESS. This, in turn, changes how long power losses occur.

For example, in the above table, the batteries’ losses were indicated to be ohmic in nature. Consider that a battery of 20 MWh is charged full or discharged empty with its rated power, which we simplify to be 10 MW. Let’s assume that the efficiency for discharge direction is, for example, 95 %, meaning that in discharging, the power losses are 500 kW, and the discharge would take 2 hours. This means that the energy losses would be 1000 kWh.

Now, let’s halve the discharge power to 5 MW. With the assumption of the ohmic nature of the losses, they would now be

This looks considerably smaller, but consider that the discharge time is now four times longer (our hours). Therefore, the energy losses during the discharge process would be 500 kWh. This is a simplified calculation, but it shows that even if the batteries’ instantaneous power losses change quadratically when the operation point changes, energy losses scale approximately linearly when the operation point (power) changes. Thus, although instantaneous losses decrease quadratically with power, total round-trip energy losses decrease approximately linearly as operating power is reduced.

Understanding the operation-point dependence is important in system design and RTE calculations, because when the system is assembled, it is rare that all components operate exactly at their datasheet ratings (rated power or peak efficiency), as they are discrete components. For example, if a BESS with system ratings of 10 MW / 20 MWh were built from 3 pieces of 4 MVA transformers, 3 pieces of 4 MW converters, and 4 pieces of 6 MWh batteries, none of these individual components would work at their nominal datasheet ratings.

The graph below shows how the RTE-% changes for an example 2-hour BESS (e.g., the10 MW/20 MWh mentioned above) depending on the point at which it is measured.

The graph shows that the RTE improves slightly when the power is lower than the rated power, as the instantaneous losses in various components are scaled in a quadratic fashion. Charging and discharging time increases, but asymmetrically, and therefore efficiency increases slightly at lower power. Below ~50% of the rated power of the BESS, the fixed loads (no-load losses, auxiliary consumption, etc.) start to play a larger proportion of the total losses, which causes the RTE to drop. Important takeaways are:

The peak RTE is not necessarily at the peak (rated) power of the system
Fixed losses dominate at low load
Quadratic losses dominate at high load

Proportion of different components contributing to BESS losses

The BESS losses – regardless of the exact system ratings – are majorly dominated by the internal heat losses of the batteries and the auxiliary power used for battery cooling. As shown above, the battery heat losses change heavily based on the point of operation, whereas RTE is indicated only at the rated power of the BESS – therefore, assuming charging and discharging with rated power at the PoC, all the time.

Below is an example distribution of energy losses over a full charge-discharge round-trip for a 2-hour BESS system.

The distribution would differ between 1-hour and 4-hour systems due to the different emphasis on battery power losses within the system, but the battery losses still play a major role in RTE. Consequently, the RTE is mainly determined by the battery selected for the BESS and the efficiency of its cooling system. Considering that cooling plays such a major role in losses, it is worthwhile to note that, in practice, this is also related to ambient conditions, which could influence the RTE measurement result. This is the IEC standard mentioned above also considers standardized ambient conditions.

Financial and contractual considerations

Since we have established that RTE only describes BESS losses at one operating point – and, by definition, at the rated power where losses are at maximum – it does not accurately model energy costs for BESS use.

It is not a typical use scenario for a BESS in commercial operation to simply charge and discharge at the rated power; therefore, RTE does not necessarily provide a realistic, albeit possibly very conservative, evaluation of energy losses and the consequent energy costs. In the author’s personal experience, the actual losses in the commercial operation of a BESS are about 30-40% of the energy losses forecasted by RTE. This is simply because the average power the BESS is used at, due to multi-market optimization, is almost always well below its rated power on average.

The above analysis should also illustrate that the RTE is not primarily a performance metric for a system integrator, but rather a consequence of a given system design. RTE is primarily a function of selected battery chemistry, system topology, cooling architecture, and measurement boundary – rather than an adjustable performance lever of the system integrator. This is why the owners should be interested in how the RTE is defined by the BESS system integrator and what it is based on, rather than discussing only one number that is not comparable across different system designs.

A given system integrator has limited influence on the RTE, as it is primarily defined by the selected battery and its RTE. The cables and transformers on the market differ little, but PCS efficiencies vary between manufacturers.

To accurately model and be sure of energy losses and energy costs during the lifetime of the project, the exact RTE value and the contractual penalties (LDs) for not meeting the promised performance value might not be as important as the fact that the owner and the supplier understand, define, and measure the RTE the same way.

How to model energy costs more accurately

Instead of using RTE-% as the metric to model energy losses, modeling the actual operational profile of a BESS (based on the forecasted optimization strategy, or something close to it) and its charge/discharge power over a representative time period would provide a more accurate basis for analysis.

The above simulation, with the knowledge presented in this article, shows how losses change with operating point, indicating that losses are accurately analyzed based on real-life operating profiles rather than maximum losses at maximum power. This would give the right idea of the metric’s meaningfulness.

Conclusion

This article has thoroughly reviewed the nature of RTE from various perspectives and, hopefully, illuminated what lies behind a single number. To conclude, Round-Trip Efficiency as a performance metric is:

Technically meaningful
Financially relevant
Standardized for comparability (if measured according to the IEC standard)

But there are caveats:

It is operation-point dependent
It is not comparable between systems built in a different way
It represents a rated-power worst-case scenario
It is highly dominated by battery physics and thermal behavior, and therefore, less of a performance metric
It is boundary-dependent

The most important factor is not the absolute RTE percentage alone, but understanding how it was defined, measured, and modeled. Acquiring this understanding would mean more refined loss modeling and, therefore, refined financial modeling.Understanding how different suppliers evaluate their RTE and what is it based on also makes for more savvy purchase decisions.

RTE calculation template

For the above purpose, we are offering a RTE calculation template where the calculation is based on real component-level input data and the RTE is calculated based on the IEC-62933-2-1 standard’s philosophy, taking the operation point of each part of the system into account. We hope that it helps buyers, consultants and engineers understand and evaluate RTE more accurately.