Efficient Battery Storage Optimization

• Efficient battery storage optimization is a framework that models subsystem-level power, SoC, and aging to maximize economic returns and extend BESS lifespan.
• The methodology integrates nonlinear, degradation-aware models with scenario-driven multi-objective optimization to address heterogeneous subsystem aging and technical constraints.
• Scenario analysis demonstrates that fully informed, degradation-cost-aware dispatch can boost revenue up to 21% per unit SOH loss, illustrating its significant economic impact.

Efficient battery storage optimization encompasses the rigorous design and operation of Battery Energy Storage Systems (BESS) to maximize economic returns, reliability, and lifespan under physical, market, and aging constraints. Modern frameworks integrate detailed subsystem models, degradation-aware objectives, scenario-based revenue analysis, and computationally tractable solvers. Explicit accounting for heterogeneous subsystem aging and degradation costs is essential for optimizing real-world multi-string BESS dispatch, as ignoring internal variability degrades both feasibility and economic performance (Grane et al., 7 Jul 2025).

1. Subsystem Variables and Aging-Aware Modeling

Efficient BESS operation demands the explicit modeling of power, energy, health, and aging at the subsystem level. Each string $i$ (indexed $i \in \{A,B\}$ for a two-string example) is described by:

• $p_{ch,i,t},\ p_{dis,i,t}$ (kW): string charge/discharge power at time $t$
• $p_{i,t} = p_{dis,i,t} - p_{ch,i,t}$: net DC-side injection at $t$
• $SoC_{i,t} \in [SoC^{min}, SoC^{max}]$: per-unit state-of-charge
• $SOH_{i,t} \in (0, 1]$: remaining capacity fraction at horizon start
• $q_{loss,cal_{i,t}},\ q_{loss,cyc_{i}}$: incremental calendar and cyclic losses

Parameters include string nominal capacity $Q^{nom}_i$ (kWh), inverter efficiencies $(\eta_{ch}^{inv}, \eta_{dis}^{inv})$, power rating $p^{max}_i$, market price $c_t^{id}$ (€/kWh), aging cost $c^{aging}$, and end-of-life SOH threshold $EOL$.

Aging is captured via nonlinear, scenario-dependent models: for cyclic aging,

$$q_{loss,cyc_{i}} = \frac{((a_2\,C_{rate} + b_2)\,(c_2\,(DOC-0.6)^3 + d_2))^2}{2\,(1\!-\!SOH_{i,0}^{cyc})}\,\Delta FEC$$

and calendar aging, using SoC and temperature dependent expressions. This multivariate structure supports both near-term dispatch and long-term health estimation (Grane et al., 7 Jul 2025).

2. Optimization Problem Formulation

Efficient storage management is cast as a scenario-driven multi-objective optimization:

Objective Function

Maximize net economic return minus explicit degradation cost:

$$\max \left( \sum_{t \in T} \sum_{i} p_{i,t}\,c_t^{id}\,\Delta t - \sum_{i} C_{deg,i} \right)$$

where $C_{deg,i}$ encapsulates cyclic and calendar losses scaled by $c^{aging}$ and $Q_i^{nom}$ relative to $1-EOL$.

System Constraints

For all $i, t$, constraints include:

• Power balance: $$p_{i,t} = \eta_{dis}^{inv}\,p_{dis,i,t} - p_{ch,i,t}/\eta_{ch}^{inv}$$
• Rated power: $0 \leq p_{ch,i,t}, p_{dis,i,t} \leq p^{max}_i$
• SoC evolution: $$SoC_{i,t} = SoC_{i,t-1} + (\Delta t / (Q^{nom}_i\,SOH_{i,t-1}))\,[\eta_{ch}^{inv}\,p_{ch,i,t} - p_{dis,i,t}/\eta_{dis}^{inv}]$$
• SoC bounds: $SoC^{min} \leq SoC_{i,t} \leq SoC^{max}$
• SOH update: $$SOH_{i,t} = SOH_{i,t-1} - \sum_{\tau \leq t} q_{loss,cal_{i,\tau}}$$

Degradation-aware dispatch penalizes unnecessary cycling and deters infeasible schedule assignments to aged strings.

3. Scenario Analysis and Performance Metrics

Four optimization scenarios formalize the technical and economic impact of subsystem aging and degradation cost:

Scenario	SOH Model	Degradation Cost	Cycle Limit	Revenue/SOH Loss (€/kWh-loss)
I. Baseline	Homogeneous	None	2 FEC/day	430,203
II. Het.-Aware	Heterogeneous	None	2 FEC/day	469,122 (+9%)
III. Deg.-Cost	Homogeneous	Included	None	492,329 (+14%)
IV. Fully Informed	Heterogeneous	Included	None	523,102 (+21%)

Key metrics:

• Operational revenue $R$
• Power schedule mismatch $S^{mismatch}$
• Missed revenue $R^{missed}$
• SOH loss $\Delta SOH_i$
• Revenue per SOH loss $\eta_{rev/\Delta SOH}$

Scenario IV yields a 21% higher revenue per unit SOH loss, demonstrating that correct string-level SOH modeling and explicit wear-cost inclusion significantly enhance operational and long-term economic efficiency (Grane et al., 7 Jul 2025).

4. Dispatch Feasibility, Aging Trade-Offs, and Adaptive Control

Naive optimizers (Scenario I) over-allocate dispatch to aged strings, causing ~9.7% schedule mismatch and 1.6% missed revenues. Heterogeneity-aware models (Scenario II) nearly eliminate infeasibilities and unlock additional value. Degradation-cost-aware operation (Scenario III, IV) aligns cycling with market opportunities, steering dispatch toward high-profit, low-wear periods.

Fully informed optimization (Scenario IV) dynamically underutilizes aged strings unless arbitrage margins justify wear. The optimizer avoids marginal cycles where expected degradation cost exceeds market profit and ensures all technical limits (SoC, voltage) are respected, preventing early cutoff or asset damage.

5. Practical Guidelines for BESS Operators and EMS Developers

Efficient battery storage optimization for multi-string systems requires:

• Real-time tracking of string-/module-level SOH for all subsystems
• Inclusion of calibrated per-unit wear costs in the economic objective
• Elimination of ad hoc cycle limits in favor of explicit economic trade-offs

Heterogeneity- and degradation-aware dispatch improves both near-term revenue and long-term asset value. Operators should calibrate models to subsystem-specific characteristics, implement rolling-horizon optimizations that re-assess SOH at each cycle, and account for schedule infeasibility risks. The value of subsystem-level precision grows with storage age and asset diversity.

6. Implications and Research Frontiers

Ignoring subsystem heterogeneity is not merely a refinement but a technical necessity in operational BESS serving arbitrage markets over extended lifetime horizons. As installations age and subsystems diverge in health profiles, optimization accuracy directly translates into increased economic returns and asset sustainability.

The methodology is extensible to systems with more than two strings or modules, higher-dimensional state representations, and can accommodate advanced degradation functions and market products. Ongoing research explores coupling with secondary services (grid support, frequency regulation) and stochastic market models to further leverage aging-aware optimization for systemic energy storage integration (Grane et al., 7 Jul 2025).