Battery Augmentation Planning: Keeping Your BESS SOH Above 70% for 25 Years

A 100 MW / 400 MWh LFP BESS with an 8,000-cycle warranty needs its first augmentation in year 9 at 0.5C with 1 cycle/day cycling — but delay it to year 12 with a higher DoD and the NPV impact is -$4.2 million due to lost energy throughput. That $4.2 million gap between the optimal and a one-cycle-delayed schedule is the kind of financial exposure that goes unnoticed when augmentation planning relies on rules of thumb instead of data-driven modeling.

This post is a deeper follow-up to our Battery Augmentation Planning guide, which covered the fundamentals of SOH triggers, bulk vs phased replacement, and financial modeling. Here, we go beneath those fundamentals to examine the optimization problem itself — how degradation tables translate into augmentation schedules, how different chemistries shift the optimal timing, and how pre-commissioning SOH loss reshapes the trajectory from day one. All figures below draw from Energy Optima's platform, which tracks degradation across 16,068 data points and supports configurable SOH thresholds for augmentation triggers.

SOH Trajectory Fundamentals

Every augmentation schedule begins with a State of Health (SOH) trajectory — the curve that maps time (or cycles) to remaining capacity. The trajectory is the output of a degradation model, typically a semi-empirical equation fitted to manufacturer cycle-life test data and augmented with calendar aging terms.

A standard SOH trajectory for an LFP cell follows a two-phase shape:

• Phase 1 (years 1-5): Rapid initial degradation driven by solid-electrolyte interphase (SEI) formation and anode structural reorganization. SOH drops from 100% to roughly 94-96% in this period, depending on C-rate and temperature. This is not linear — roughly 40% of the cycle-life SOH loss occurs in the first 20% of cycle life.
• Phase 2 (years 6-20+): A shallower, near-linear degradation slope driven by lithium inventory depletion and cathode cracking. The slope in this phase determines whether the system needs one, two, or three augmentations over the project life.

Degradation tables — whether from manufacturer datasheets, independent lab tests, or in-situ monitoring — provide the raw data that defines this trajectory. The critical step is translating those tables into a year-by-year SOH projection that accounts for the specific cycling profile of the project. For example, a 0.33C, 1.5 cycle/day profile degrades an LFP cell roughly 1.7x faster (in calendar time) than a 0.5C, 1 cycle/day profile, even though the latter has higher instantaneous C-rate, because the former accumulates 50% more cycles per year.

Energy Optima tracks 16,068 degradation data points spanning manufacturer test data, independent laboratory aging studies, and field measurements from operating assets. These data points feed a semi-empirical model that projects SOH trajectories for any combination of C-rate, DoD, temperature, and cycle frequency. The model outputs a year-by-year SOH table that directly feeds the augmentation scheduler.

Trigger Thresholds: Warranty vs Throughput Covenants

Augmentation is triggered by reaching a threshold. But not all thresholds are created equal — some are hard contractual cliffs, others are soft economic signals. Understanding which applies to your project is the first step in setting the schedule.

Warranty Milestone Triggers

Manufacturer warranties define the boundaries of what the vendor covers. The most common warranty triggers for augmentation are:

• SOH = 80% (end-of-warranty): The standard contractual end-of-life threshold. Most LFP warranty papers guarantee at least 80% SOH after a specified number of cycles or years. If the system drops below 80% before the warranty period expires, the manufacturer is obligated to replace cells at their cost — but this typically requires the project to operate within defined parameters (temperature, C-rate, DoD).
• SOH = 70% (functional end-of-life): Some warranties include a secondary threshold at 70% SOH, below which the battery is considered functionally end-of-life regardless of cycle count. This is the lower bound for the vast majority of BESS projects — operating below 70% SOH generally means the battery cannot meet its original power and energy ratings simultaneously.
• Throughput covenants: Some advanced warranty structures include energy throughput guarantees — e.g., "the battery will deliver at least X MWh of cumulative throughput before reaching 80% SOH." This throughput covenant can be more restrictive than the cycle-count warranty because it accounts for both cycle depth and RTE degradation.

Throughput Covenant Triggers

Throughput-based covenants are more common in tolling agreements and energy storage service contracts. Instead of specifying an SOH floor, the contract specifies a minimum cumulative energy throughput (MWh delivered) that the system must achieve. This type of covenant is sensitive to both degradation and round-trip efficiency loss — as the battery ages, both capacity and RTE decline, reducing the throughput per cycle.

In practice, throughput covenants create an earlier augmentation trigger than pure SOH-based warranties. A system might meet 80% SOH at year 12 but have already failed its throughput covenant at year 10 because cumulative throughput fell below the contractual trajectory. The augmentation optimizer in Energy Optima handles both trigger types simultaneously — the schedule is set by whichever threshold is hit first.

PPA and RA Triggers

Revenue contracts add a third layer of triggers. A Resource Adequacy (RA) agreement might require the system to demonstrate at least 90% of nameplate capacity during summer peak hours. A PPA might require 85% minimum deliverable capacity in any settlement interval. These contractual triggers often force augmentation earlier than the asset-level economic optimum, which is why modeling the interaction between all three trigger layers (warranty, throughput, revenue) is essential.

Replacement Strategies Compared

Once a trigger threshold is reached, the project owner must decide how to add capacity. There are three primary strategies:

Bulk Replacement

At a predetermined year, the project replaces all battery containers that have fallen below the SOH trigger in a single campaign. This is operationally the simplest approach — a single mobilization, one commissioning cycle, one period of lost revenue. But it leaves value on the table by replacing containers that still have 3-5 years of useful life at 75-80% SOH.

Best for: Small projects (<50 MWh), remote sites where mobilization cost is very high, projects with homogeneous container aging, and situations where warranty terms require uniform fleet composition.

Phased (Rolling) Replacement

Containers are replaced individually or in small groups as each reaches its SOH trigger. Over a 25-year project, a phased strategy might execute 6-8 small augmentation campaigns rather than 2-3 large ones. This approach minimizes upfront capital by replacing only what is needed exactly when it is needed, and it takes maximum advantage of declining battery prices (later replacements are cheaper).

Best for: Large projects (>200 MWh), fleets with heterogeneous container aging, merchant projects where revenue loss during outages is material, and any project where battery prices are expected to decline rapidly.

Hybrid Strategy

The project schedules one or two large bulk replacements at critical SOH thresholds (e.g., first augmentation when fleet-average SOH hits 82%, second when it hits 75%) and uses annual mini-augmentations between them to top off individual underperforming containers. This balances the mobilization efficiency of bulk replacement with the capital efficiency of phased replacement.

Best for: Most large-scale projects. The hybrid strategy typically yields a total NPV cost that is 4-8% lower than pure bulk and 10-15% lower than pure phased when accounting for both capital cost and revenue loss during outages.

The Augmentation Optimization Problem

At its core, augmentation optimization answers one question: At what year does the cost of adding capacity equal the revenue lost from degradation?

Formally, the optimization minimizes the net present value of total project cost (augmentation expenditure + revenue loss from degradation) over the project life:

Minimize: NPV = Σ[Caug,t × (1+r)-t + Rloss,t × (1+r)-t]

Where:

• Caug,t = cost of augmentation in year t (battery cells + logistics + labor + commissioning + revenue loss during outage)
• Rloss,t = revenue lost in year t due to degradation (energy throughput shortfall vs nameplate capacity)
• r = discount rate (typically 6-10% real for project finance)
• t = year of project life

The constraints on this minimization are:

• SOH must never fall below the hard warranty threshold (typically 80% or 70% depending on the warranty contract)
• Cumulative throughput must meet any throughput covenant minima
• PPA/RA capacity commitments must be satisfied in every commitment period
• The augmentation schedule must respect site operational constraints (seasonal outages, permitting timelines, equipment availability)

The optimization is non-trivial because the cost of augmentation is itself a function of time — battery cell prices are projected to decline 6-12% per year in real terms through 2035. This declining-cost curve creates a powerful incentive to delay augmentation, even as revenue losses mount. The optimal schedule is the year (or years) at which the marginal benefit of waiting one more year (lower battery cost) equals the marginal cost of waiting (lost revenue).

Energy Optima's augmentation optimizer solves this problem using a dynamic programming approach. It evaluates augmentation in every possible year (or every 6-month interval) and computes the full NPV of each candidate schedule, then selects the schedule(s) with the lowest total cost. The optimizer can also produce a "cost curve" showing how total NPV changes as the first augmentation year is shifted earlier or later — making visible the $4.2 million gap cited in the opening example.

Worked Example: 200 MWh Across 3 Chemistries

To illustrate how different battery chemistries change the optimal augmentation schedule, consider a 100 MW / 200 MWh (2-hour) system operating at 0.33C with 1.5 cycles per day. The system is located in a temperate climate (avg 22°C), operating at 90% DoD, and has a 25-year project life with an 8% real discount rate. Battery pricing assumes a starting cost of $120/kWh at COD, declining 8%/year real through 2035, then 3%/year thereafter.

Energy Optima's optimizer computed the optimal augmentation schedule for three chemistries using the platform's degradation model (16,068 data points across LFP, NMC-811, and sodium-ion cells):

Metric	LFP (8,000-cycle)	NMC-811 (4,000-cycle)	Sodium-Ion (6,000-cycle)
Year of 1st augmentation	11	7	10
Capacity added (1st)	28 MWh (14%)	38 MWh (19%)	32 MWh (16%)
Year of 2nd augmentation	19	13	18
Capacity added (2nd)	22 MWh (11%)	32 MWh (16%)	28 MWh (14%)
Year of 3rd augmentation	—	19	—
Capacity added (3rd)	—	26 MWh (13%)	—
Total capex (NPV)	$3.1M	$5.8M	$3.6M
Revenue loss from deg. (NPV)	$1.4M	$2.1M	$1.6M
Total cost (NPV)	$4.5M	$7.9M	$5.2M
Augmentation events	2	3	2

The data reveals several important patterns:

LFP dominates on augmentation economics. At $4.5M total NPV cost (capex + revenue loss), LFP is 43% cheaper than NMC-811 over the project life. This is driven by LFP's flatter degradation curve after year 5, which postpones both augmentation events and reduces the capacity that needs to be replaced.

NMC-811 requires three events. The steeper degradation slope forces a first augmentation at year 7 rather than year 11, and the battery has degraded enough by year 13 to need a second round. A third event at year 19 is marginal — the model found it was slightly cheaper to add capacity than to accept the revenue loss from operating at 68-70% SOH for the final 6 years.

Sodium-ion is competitive but not yet dominant. With 6,000-cycle rated life and slightly higher cell cost ($95/kWh projected vs $75/kWh for LFP at first augmentation year), sodium-ion sits between LFP and NMC. The total NPV of $5.2M is 15% higher than LFP but 34% lower than NMC-811. As sodium-ion cycle life improves (8,000-cycle variants are in development), it could match LFP's augmentation economics.

The first augmentation year is the single largest lever. Shifting the first LFP augmentation from year 11 to year 9 (more aggressive) increases total NPV by $1.8M because the added capacity costs more than the revenue saved. Shifting it to year 13 (more passive) increases total NPV by $2.4M because revenue loss exceeds the savings from delaying. The optimum is a well-defined convex minimum, not a flat region.

The FAT-to-COD Effect

One of the most commonly overlooked factors in augmentation planning is the SOH loss that occurs between Factory Acceptance Testing (FAT) and Commercial Operation Date (COD). This "FAT-to-COD" gap typically accounts for 0.3-0.7% SOH loss, depending on the time elapsed, storage conditions (temperature, SOC), and the number of cycles performed during commissioning.

The typical timeline works like this:

• Cells are manufactured and shipped (3-6 months in transit, often at elevated temperature in shipping containers)
• FAT is performed at the integrator's facility (10-20 cycles for capacity confirmation, voltage matching, and BMS calibration)
• Modules are shipped to site (1-3 months)
• Site installation and commissioning (2-6 months, involving additional cycling for string balancing and acceptance testing)
• COD — the system begins commercial operation

By the time the system reaches COD, it has already experienced 0.3-0.7% SOH loss from calendar aging during transit and storage, plus cycle degradation from FAT and commissioning cycles. For a 400 MWh system, that 0.5% SOH loss represents roughly 2 MWh of nameplate capacity that is permanently gone before commercial operation even begins.

How does this shift the augmentation curve? A 0.5% pre-commissioning SOH loss effectively shifts the entire SOH trajectory down by 0.5 percentage points. This means:

• A system that was projected to hit 80% SOH in year 14 now hits it in year 13.4 (roughly 7 months earlier)
• The first augmentation event moves forward by 6-12 months, depending on the slope of the degradation curve
• Total augmentation NPV increases by $0.6-1.2M for a 200 MWh system, depending on chemistry and cycling profile

The FAT-to-COD effect is largest for NMC systems (because NMC has faster calendar aging at elevated temperatures) and smallest for LFP (because LFP calendar aging is largely temperature-independent at moderate SOC). In our worked example above, including the FAT-to-COD effect shifts the NMC first augmentation from year 7 to year 6.5 and increases total NPV by $0.8M. For LFP, the shift is from year 11 to year 10.6 with a $0.4M NPV increase.

Energy Optima's degradation model includes a configurable "pre-commissioning SOH loss" parameter with a default of 0.5% and the option to input project-specific estimates based on the actual FAT-to-COD timeline. This ensures the augmentation schedule starts from the true COD SOH, not the theoretical 100%.

Practical Workflow for Augmentation Planning

Based on the analysis above, here is the practical workflow that Energy Optima users follow to develop and validate an augmentation plan:

1. Define the degradation inputs. Configure the battery chemistry (LFP/NMC/sodium-ion), manufacturer warranty cycle count, C-rate, DoD, average temperature, and cycle frequency. Import (or select from the database) the manufacturer's cycle-life test data. Energy Optima's database of 16,068 degradation data points provides defaults for the most common cell types.
2. Set the trigger thresholds. Configure the SOH hard trigger (e.g., 80% for warranty compliance), the soft economic trigger (e.g., 85%), and any throughput covenants from PPAs or RA agreements. These can be set at the system level or per-container for more granular optimization.
3. Input battery price projections. Use the default price curves (8%/year decline through 2035, 3%/year thereafter) or input project-specific projections. Run a sensitivity analysis at +2% and -2% around the base case to understand the range of optimal schedules.
4. Select replacement strategy. Choose bulk, phased, or hybrid. For projects >100 MWh, the hybrid strategy is the recommended starting point. The platform will generate the container-level replacement schedule for each strategy.
5. Run the optimizer. The dynamic programming optimizer evaluates all candidate augmentation years and selects the schedule(s) with the lowest total NPV. Review the cost curve to understand the penalty for deviating from the optimum.
6. Review output. The optimizer produces a year-by-year projection of SOH, capacity, revenue, augmentation capex, and cumulative throughput. Validate that all constraints (warranty, throughput covenant, PPA/RA) are satisfied in every year.
7. Run Monte Carlo. This step is critical for project finance. The optimizer runs 1,000+ scenarios with varying degradation rates, battery prices, and revenue assumptions. The output is a probability distribution of total augmentation cost, which feeds directly into the project financial model's contingency calculation.
8. Generate the schedule. The final deliverable is a year-by-year augmentation schedule with specific years for each augmentation event, the capacity to be added, and the estimated cost (in nominal and real terms). This schedule becomes part of the project's financing documentation and operations plan.

For a detailed treatment of how degradation is modeled from first principles — including the semi-empirical equations and how they are fitted to manufacturer test data — see our BESS Degradation Modeling Guide. For the system-sizing implications — how augmentation changes the initial capacity decision — see Degradation-Aware BESS Sizing.

And for the fundamentals of augmentation triggers, PPA capacity commitments, and the bulk-vs-phased decision, the original Battery Augmentation Planning guide remains the best starting point.