Fiber Optic Sensing Technologies for Battery Management Systems and Energy Storage Applications

This review addresses how fiber optic (FO) sensing can enhance battery management systems (BMS) by enabling accurate internal measurements of key state variables (temperature, strain, pressure, refractive index, and gas evolution). The context is the rapid growth of Li-ion batteries in EVs, heavy-duty trucks, and grid-scale storage, where current BMS rely largely on external electrical sensors (voltage, current, thermistors/thermocouples) that are susceptible to EMI and cannot probe internal states. The purpose is to survey FO sensor technologies and recent demonstrations for battery monitoring, assess their advantages and challenges, and map sensing modalities to application scales. The importance lies in improving SOC/SOH/SOL estimation accuracy, enabling early detection of thermal runaway, and potentially reducing pack oversizing and cost while improving safety.

The paper reviews FO sensing modalities and recent advances relevant to batteries: (1) Single-point sensors: Fiber Bragg Gratings (FBGs) including uniform, chirped, and phase-shifted variants; evanescent wave (FOEW) sensors with U-shaped/tapered/D-shaped geometries and functional coatings (polymers, plasmonic nanocomposites, MOFs); fluorescence-based sensors (lifetime, quenching, anisotropy) for temperature and gas sensing; Fabry–Perot (FP) interferometers (intrinsic/extrinsic). (2) Quasi-distributed sensors: multiplexing schemes (WDM, TDM, FDM, CDM) for arrays of FBGs and interferometers; integration strategies (FBG-FP hybrids) and mapping of scheme to application scale. (3) Fully distributed sensors (DFOS): Rayleigh, Raman, and Brillouin scattering with OTDR/OFDR interrogation, emphasizing Rayleigh-OFDR and Raman-OTDR for temperature/strain mapping. The review also surveys recent battery-focused studies over the past five years on internal/external temperature, strain, pressure, refractive index, acoustic signals, distributed temperature mapping, and vent gas monitoring (especially CO2) for early failure detection.

This is a narrative review synthesizing recent research on FO sensors for battery monitoring. The authors: (1) Identify three application scales (passenger EVs, heavy-duty trucks, utility-scale storage) and typical cell formats (cylindrical, prismatic, pouch), quantifying cell counts per system scale. (2) Classify FO sensors by spatial topology (single-point, quasi-distributed, fully distributed) and sensing mechanisms (wavelength, phase, intensity, polarization changes), and summarize interrogation approaches (WDM/TDM/FDM/CDM; OTDR/OFDR). (3) Compare FO vs conventional electrical sensors for temperature measurement, including sensitivity, accuracy, robustness, and estimated costs. A cost model estimates sensing point costs factoring interrogator fixed cost, number of FBGs, and fiber length, with example total system costs: approximately $10,725 (EV), $15,500 (electric truck), and $1,100,000 (grid-scale), highlighting interrogator dominance. (4) Review and analyze battery-focused demonstrations: internal vs external temperature sensing (FBGs, FP), strain-temperature discrimination approaches (reference FBG, hybrid FBG-FP), strain/SOC and strain/SOH correlations (internal/external), pressure and refractive index sensing (FBG variants, custom waveguides), FOEW intensity-based sensing embedded in electrodes, distributed temperature mapping (Rayleigh-OFDR, Raman-OTDR, FBG arrays), and vent gas sensing (CO2, CH4, CO, HF) using fiber-based methods (TDLAS, colorimetric). (5) Propose mapping of multiplexing schemes to application scales (WDM for EV modules; WDM/FDM for truck packs; TDM for large-scale sites) and discuss future cost-reduction pathways (photonic ICs, LED sources, FBGs in MMF).

- FO sensor advantages: immunity to EMI/RFI, electrical insulation, lightweight/flexible, embeddable at cell level, high sensitivity, multiplexable, capable of multi-parameter sensing (temperature, strain, pressure, refractive index, acoustic, gases). - Cost and scalability: Interrogator cost dominates FO systems. Estimated example total costs: ~$10.7k (EV), ~$15.5k (electric truck), ~$1.1M (100 MWh grid-scale). Cost per sensing point can drop with scale (e.g., illustrative $165/pt EV; $31/pt heavy-duty truck; $11/pt grid-scale) using multiplexed FBGs. Low-cost point sensors (FOEW, FP) and low-cost interrogation (photonic ICs, LEDs) are critical. - Sensor performance (examples): FBG temperature sensitivity ~10–13 pm/°C; reported accuracy as good as ±0.05 °C with suitable interrogators. Other FO point sensors near ~10 pm/°C sensitivity with ±0.2 °C accuracy. - Internal vs external temperature: Internal cell temperature can exceed external by 10–20 °C under load; specific demonstrations show ~2 °C gradients internally not captured by surface sensors and up to 10 °C internal vs ambient differences (coin cells at C/20). Accurate internal sensing improves model-based SOC/SOH estimation. - Strain–temperature discrimination: Cross-sensitivity in FBGs addressed via (a) reference strain-isolated FBGs, (b) dual-sensor approaches (one strain-sensitive, one strain-isolated), and (c) hybrid FBG–FP single-fiber sensors solving for both strain and temperature. - Strain correlations: Strain correlates strongly with SOC and with SOH (capacity fade). Internal FBGs embedded in anodes show higher strain amplitudes than centrally located references; machine-learning/time-series methods (e.g., dynamic time warping, EKF) achieve >99% SOC estimation in static tests using strain after temperature compensation. Strain signatures reveal electrode intercalation stage transitions and residual strain at high SOCs (80–100%) linked to degradation. - Pressure and refractive index: Operando pressure sensing with FBGs identifies SEI formation during first charge; refractive-index sensing via custom cladding waveguide gratings correlates with electrolyte composition, offering a path to infer internal resistance (a key SOH component). - FOEW sensing: Embedded FOEW probes in electrodes track optical transmittance changes correlated with SOC and capacity; transmittance slope features align with anode phase transitions and capacity fade indicators. - Distributed temperature sensing: Rayleigh-OFDR maps internal temperature/strain gradients across cells; Raman-OTDR and quasi-distributed FBG arrays detect hotspots and abnormal thermal events. FBG networks demonstrated hotspot detection in small packs and recorded thermal runaway peaks up to ~477 °C during abuse tests. - Vent gas sensing and early failure: CO2 is a dominant vent gas (60–75% by volume in abuse), often generated early (~70–90 °C) during SEI decomposition, making it a strong early-warning indicator. Fiber-based approaches (TDLAS, colorimetric) show feasibility for real-time CO2 and other gas detection (CH4, CO, HF) with ppm-level accuracy. - Application mapping: WDM for tens of points (EV modules), WDM/FDM for tens–hundreds (heavy-duty trucks), and TDM for hundreds–thousands (utility-scale) to balance multiplexing capacity and complexity.

The review demonstrates that FO sensors can directly access internal battery states that are weakly observable via external electrical measurements, thereby improving SOC/SOH/SOL estimates and enabling earlier detection of thermal anomalies. Internal temperature sensing corrects biases in OCV-based estimators arising from thermal gradients, while strain sensing provides a physically rich proxy for SOC and mechanical degradation processes linked to SOH. Multi-parameter discrimination (strain vs temperature) unlocks robust use of FBGs in cells. Distributed and quasi-distributed FO sensing is effective for thermal hotspot localization and early warning of abnormal events at module/pack level, which is crucial for preventing thermal runaway propagation. Gas sensing, particularly for CO2, offers a complementary early indicator of failure, preceding thermal runaway onset. However, practical deployment must align sensor/interrogation architectures with application scale to manage costs. The significance to the field is the clear path to integrate FO sensing in high-value applications (e.g., grid-scale storage, electrified aircraft) where improved reliability and safety justify the investment, and where fewer, larger cells favor internal instrumentation. Continued advances in low-cost interrogation and photonic integration are poised to broaden adoption.

This review consolidates recent advances in fiber optic sensing for battery monitoring, mapping sensing modalities to application scales and demonstrating how FO sensors can measure internal temperature, strain, pressure, refractive index, and vent gases to enhance BMS capabilities. It highlights effective strain–temperature discrimination, strong correlations between strain and SOC/SOH, and the utility of distributed sensing for hotspot detection and early failure warning. The primary barrier is interrogator cost, especially for cell-level monitoring in EVs. Promising cost-reduction strategies include photonic integrated circuit-based photodetectors, LED light sources, and inscribing FBGs in multimode fibers to relax alignment and source requirements; increased use of lower-cost FO point sensors (FOEW, FP) is also encouraged. Future work should prioritize: long-duration cycling and abuse testing of fiber-instrumented cells to validate durability and electrochemical stability; development of robust, low-cost interrogation platforms; integration of gas sensing (e.g., CO2) with thermal/strain sensing for early fault detection; and application to next-generation chemistries (e.g., solid-state Li-metal) to monitor O2 evolution, interfacial stability, and dendrite-related strain/pressure signatures.

- High interrogator/demodulation cost limits practicality, particularly for cell-level EV monitoring; fully distributed systems add complexity and expense. - Cross-sensitivity in FO sensors (e.g., FBG) necessitates discrimination methods, increasing system complexity and potential calibration burden. - Electrochemical stability and mechanical durability of embedded fibers require further validation over long cycling and abuse conditions; potential intrusiveness and sealing challenges remain. - Ultra-weak backscatter signals in DFOS and spatial resolution–sensitivity trade-offs (OTDR) constrain some distributed approaches. - Real-time in situ gas sensing demonstrations in battery packs are limited; integration into BMS for actionable early-warning thresholds needs further development.