The complete guide to BESS optimisation

The complete guide to BESS optimisation

Learn how battery energy storage optimisation works, from battery science, state of charge and state of health to rack imbalance and marginal cost trading.

Battery energy storage systems (BESS) have become one of the fastest-growing assets in the global energy transition. As renewable generation increases and electricity networks become more dynamic, batteries are able to absorb electricity when it is abundant and release it when it is needed most.

But owning a battery is not the same as optimising one. A BESS asset sitting idle, or dispatched on simple rules, captures only a fraction of its potential value and can degrade faster than necessary. Getting the most out of a battery requires understanding what is actually happening inside it, second-by-second and cycle-by-cycle, and using that understanding to make commercially-informed decisions.

This guide explains how BESS optimisation works in practice, from the underlying battery science, through state of charge, state of health and rack imbalance, to the economic principle that ultimately governs every dispatch decision: marginal cost trading.

Why battery storage matters

Electricity systems must keep supply and demand in balance every second of every day. As wind and solar generation grow, that balance becomes harder to hold, because renewable output is variable and does not always align with when electricity is needed.

Batteries solve a problem that no other technology can solve as effectively: they can store electricity when it is plentiful and cheap, and release it when it is scarce and expensive, all within milliseconds if required. This makes BESS assets valuable across almost every layer of the electricity system, from second-by-second frequency control to seasonal capacity adequacy.

For asset owners, developers and investors, this creates an opportunity, but it also creates a challenge: a battery’s value depends on how the asset is operated over its operating life.

What is BESS optimisation?

BESS optimisation is the process of deciding when a battery should charge, discharge or stand idle, to maximise the commercial and operational value it delivers, while protecting its long-term health.

It sits on top of the physical battery and its Battery Management System (BMS). Where the BMS keeps the asset safe, optimisation software decides what the asset should be doing at any given moment across every available electricity market, and how hard it should be pushed to get there.

Optimisation typically involves:

• forecasting electricity prices, grid conditions and system needs
• tracking the real-time technical state of the battery, including state of charge, state of health and any internal imbalance
• deciding which market, or combination of markets, offers the greatest value at any given moment
• dispatching charge and discharge instructions automatically, within the battery’s technical and contractual limits
• verifying performance and settling revenue

BESS optimisation turns a battery from a static piece of infrastructure into a continuously-managed, revenue-generating asset.

Battery science: how a BESS actually works

To optimise a battery well, it helps to understand what is physically happening inside it.

System

A utility-scale BESS is built from a large number of individual lithium-ion cells, arranged in a hierarchy:

• cells: the smallest electrochemical unit, where energy is actually stored and released through the movement of lithium ions between a positive and a negative electrode
• modules: groups of cells connected together, usually in series, to increase voltage
• racks (or strings): groups of modules connected in series, forming the basic building block of the system
• containers: groups of racks connected to a shared Power Conversion System (PCS), which converts DC battery power to AC grid power and back

Because cells within a rack are connected in series, they behave like a chain: the weakest cell in the chain sets the limit for the whole rack. This single fact underpins most of the technical challenges in BESS optimisation, including rack imbalance, which is covered later in this guide.

Chemistry

Most utility-scale batteries use one of two lithium-ion chemistries:

• Lithium Iron Phosphate (LFP): increasingly the dominant chemistry for stationary storage, valued for its thermal stability, longer cycle life and lower cost, though it produces a very flat voltage curve across much of its operating range
• Nickel Manganese Cobalt (NMC): offers higher energy density, making it attractive where space is constrained, with a voltage curve that changes more clearly as the cell charges and discharges

The choice of chemistry has optimisation consequences. LFP’s flat voltage curve makes it harder to estimate state of charge from voltage alone, which is one of the reasons advanced monitoring and optimisation platforms have become essential rather than optional.

Metrics

• capacity (MWh): how much electricity the system can store
• power rating (MW): how quickly the system can charge or discharge
• C-rate: the relationship between power and energy capacity; a higher C-rate means faster charge and discharge relative to the battery’s size, generally at the cost of faster degradation
• round-trip efficiency: the proportion of electricity recovered on discharge compared with what was used to charge the battery, after accounting for conversion and thermal losses
• depth of discharge (DoD): how much of the battery’s capacity is used in a given cycle

Every optimisation decision is made within the boundaries these metrics define. A platform that ignores them, and simply chases the highest price in the market, will win short-term revenue at the cost of long-term asset health.

State of Charge (SoC)

State of Charge describes how much usable energy is currently stored in a battery, expressed as a percentage of its capacity. It is the single most important real-time variable in BESS optimisation, because it defines what the battery can do right now: how much it can still charge, and how much it has left to discharge.

SoC is estimated, typically from a combination of voltage, current and temperature readings, interpreted through a battery model. This estimation is more difficult than it sounds, for two reasons:

• chemistry matters: LFP cells in particular exhibit a very flat voltage curve across much of their charge range, which means voltage alone is a poor indicator of SoC; two racks can show almost identical voltage while holding meaningfully different amounts of stored energy
• errors compound: small inaccuracies in SoC estimation, left uncorrected, accumulate over time and can lead an operator to believe a battery holds more, or less, usable energy than it actually does. In some documented cases, SoC estimation errors on individual racks have reached as much as 50%, with the gap between reported and actual capacity translating directly into balancing costs when the shortfall has already been sold into a market

Accurate SoC estimation, using physics-based battery models rather than voltage readings alone, is foundational to reliable optimisation. Without it, a dispatch platform is making commercial decisions on a false picture of what the asset can deliver.

State of Health (SoH)

If SoC answers “what can this battery do right now?”, State of Health answers: “how much of the battery’s original capability is left?”

SoH is typically expressed as a percentage of a battery’s original rated capacity, or as an increase in internal resistance relative to when the asset was new. A battery at 100% SoH performs as designed; a battery at 80% SoH has lost a fifth of its usable capacity, and its power capability, efficiency and safety margins have all shifted as a result.

Batteries degrade through two distinct mechanisms, and optimisation strategy needs to account for both:

• Calendar ageing: degradation that occurs simply through the passage of time, driven mainly by temperature and the average SoC at which the battery is held, independent of how much it is cycled
• Cycle ageing: degradation caused by the act of charging and discharging itself, driven by depth of discharge, C-rate, and the number of cycles completed

A battery held at a high SoC in hot conditions can lose meaningful capacity even while sitting idle, purely through calendar ageing while a battery cycled on shallow, well-managed cycles may degrade more slowly than one cycled less often but more aggressively. Good optimisation does not simply try to cycle a battery as much as possible; it tries to extract the most value per unit of degradation incurred, which is precisely the logic behind marginal cost trading, covered later in this guide.

Tracking SoH trends over time also has a direct operational use: forecasting when a battery, or an individual rack within it, is likely to breach a critical capacity threshold, so that maintenance or replacement can be planned proactively rather than triggered by an unexpected failure.

Rack imbalance

Manufacturing tolerances, uneven temperatures across a container, and minor differences in electrical connections can cause small variations in capacity, internal resistance and self-discharge rate. On their own, these differences are insignificant. Left unmanaged, their cumulative effect is not.

This is known as rack imbalance, and it takes two forms:

• intra-rack imbalance: variation between individual cells within the same rack
• inter-rack imbalance: variation between different racks connected to the same Power Conversion System

Because cells and racks are connected in series, the weakest link sets the limit for the whole string. When discharging, the cell or rack with the lowest SoC reaches its limit first, forcing the BMS to halt the discharge while other cells still hold usable energy. When charging, the cell or rack with the highest SoC hits its ceiling first, stopping the charge while others remain under-filled. This means that capacity that technically exists in the system becomes unusable.

The consequences compound over time. Some documented BESS assets have lost more than 10% of usable capacity to imbalance-related effects. Uneven usage also accelerates uneven ageing. The cells or racks doing disproportionate work degrade faster, widening the imbalance further and creating a self-reinforcing cycle. In some cases, protection systems will disconnect an entire rack to prevent damage, removing that capacity from service.

Because most battery management systems are designed to prevent immediate safety events, such as overvoltage or overheating, rack imbalance is difficult to catch with a standard BMS alone. Effective BESS optimisation depends on detecting imbalance early, using physics-based models that combine BMS data, temperature and behavioural history to estimate the true state of each rack, rather than relying on voltage alone. Catching imbalance before it reaches an alarm threshold protects usable capacity, slows uneven degradation, and prevents a battery from being traded on a level of available energy it can no longer actually deliver.

How BESS optimisation works

Although every deployment differs, effective BESS optimisation generally follows five continuous stages.

• continuous monitoring: an intelligent energy platform connects to the battery’s BMS, PCS and site metering, continuously tracking SoC, SoH, cell and rack-level voltage and temperature, imbalance indicators, and grid or market conditions
• forecasting: the platform forecasts electricity prices, grid stress events, renewable generation, and how the battery’s own technical state, including SoH and cycle history, is likely to evolve
• optimisation: this is where the commercial decision is made. Which market, or combination of markets, offers the greatest value at this moment, how hard the asset should be pushed to capture it, and whether that value exceeds the true cost of dispatch, including degradation
• automated dispatch: once a decision is made, charge and discharge instructions are sent automatically, within the technical limits set by SoC, SoH and any active imbalance constraints, and within any contractual obligations the asset holds
• verification and settlement: delivered performance is verified against what was instructed and what markets require, and revenue is calculated and settled accordingly

Markets where BESS optimisation operates

A battery’s flexibility is valuable to almost every layer of the electricity system, and the highest-performing assets participate across several of these simultaneously:

• frequency response: extremely fast charge or discharge to help hold system frequency within safe limits, often one of the highest-value services available to batteries because of their sub-second response capability
• wholesale energy arbitrage: charging when prices are low and discharging when prices are high, capturing the spread between them
• capacity markets: being paid for guaranteed availability during periods of system stress, rather than for energy delivered
• balancing mechanisms and real-time markets: responding to last-minute imbalances between forecast and actual system conditions
• ancillary services: a broader set of services supporting voltage control, reserve and system restoration

Batteries are particularly well suited to this multi-market approach because they can shift between roles far faster than most other assets. The question optimisation software must continually answer is how to allocate a limited, degrading resource across all of them for the greatest total return.

Marginal cost trading

Every time a battery charges or discharges, it incurs a cost. Some of that cost is obvious, such as the electricity used to charge the battery in the first place. But every cycle consumes a small amount of the battery’s total lifetime capacity, and that consumption has a real economic value that can impact the lifetime return on investment for the asset.

Marginal cost trading is the principle of incorporating both of these costs into every dispatch decision, rather than trading purely on the electricity price signal alone. The marginal cost of operating a battery at a given moment is made up of two components:

• energy losses: the conversion and thermal losses incurred every time electricity flows into or out of the battery, captured by round-trip efficiency
• degradation cost: the value of the capacity consumed by that specific cycle, driven by depth of discharge, C-rate and the battery’s current state of health

This second component is what separates sophisticated BESS optimisation from simple rules-based dispatch. Academic research into optimal battery dispatch has shown that the marginal cost of degradation is not a fixed number: it is time-variant, changing with the battery’s state of health and with the value of the opportunities available elsewhere in the system.

A battery early in its life, in good health, with limited market opportunity ahead, can reasonably absorb a different level of cycling cost than the same battery late in its life when every remaining cycle is less replaceable. The practical implication is that a battery should only charge or discharge when the value on offer exceeds the true marginal cost of doing so. Chasing every available price spread, without reference to degradation cost, can look attractive in the short term while quietly eroding the asset’s long-term value faster than the revenue it generates justifies.

This is also why SoH and marginal cost trading are inseparable. A degradation cost calculated against a battery’s original nameplate capacity, rather than its real, current SoH, will consistently misprice dispatch decisions, usually by encouraging more aggressive cycling than the asset can profitably sustain. Feeding accurate, real-time SoH data into the dispatch logic is what allows marginal cost trading to reflect the battery as it actually is.

Revenue stacking

Because batteries can serve multiple roles, the greatest value rarely comes from a single market. Revenue stacking is the practice of combining several revenue streams from the same asset. But these opportunities frequently compete with one another and every commitment carries its own marginal cost implications for SoH. Effective optimisation continuously evaluates:

• which market is most valuable at a given moment
• what future opportunities may emerge, based on forecasts
• the technical constraints imposed by current SoC, SoH and any rack imbalance
• the marginal cost, in degradation terms, of committing capacity to one market over another
• contractual obligations already in place across markets

The objective is not to maximise activity, but to maximise total value captured over the asset’s lifetime, allocating a finite, degrading resource to wherever it is worth the most at any given moment.

Why GridBeyond’s approach to BESS optimisation is different

Battery energy storage has moved from a relatively simple asset, dispatched on fixed rules or single-market strategies, into a far more complex optimisation problem, one that spans multiple electricity markets, real-time technical constraints, and the long-term economics of degradation.

GridBeyond’s solutions sit across the full technical and commercial picture of a battery asset, continuously combining real-time state of charge and state of health data, cell and rack-level imbalance detection, and market forecasting into a single, live view of what the asset can safely and profitably do at any given moment.

At the centre of this is GridBeyond’s AI-driven forecasting and trading engine. Rather than dispatching against a single price signal, the platform continuously evaluates every available market opportunity against the true marginal cost of dispatch, including degradation, so that a battery is only cycled when the value on offer genuinely justifies the capacity consumed. Digital twin modelling, built into this same forecasting layer, allows asset performance and degradation to be simulated and validated against real operating data, rather than assumed from nameplate specifications.

Because imbalance and SoH drift gradually and are easy to miss with standard BMS monitoring alone, GridBeyond’s platform is designed to surface these issues early, protecting usable capacity and preventing assets from being offered into markets on a level of availability they can no longer deliver.

For organisations considering new battery investment, GridBeyond’s Designer Studies provide pre-investment feasibility analysis, modelling likely revenue, degradation and market participation scenarios before capital is committed, so that optimisation strategy is considered from the earliest stage of a project rather than retrofitted once an asset is already operating.

The result is a closed-loop approach, in which monitoring, forecasting, optimisation and dispatch continuously inform one another. Each cycle a battery completes generates data that improves the accuracy of the next decision, allowing both forecasting and dispatch to become more precise over time, and allowing asset owners to extract maximum value from their battery investment across its full operating life.