Online battery remaining-useful-life (RUL) predictions can be optimistic when cycles from the same cell cross data splits, predictors use future observations, or right-censored records are labeled as failures. We present a leakage-resistant, censor-aware evaluation using a 209-cell pool from NASA, CALCE, HUST, and MATR. Of these, 86 cells formed the internal pool; 21 observed-endpoint MATR development cells froze the external model; 21 official MATR primary-test cells with a strict-80% state-of-health (SOH) endpoint provided the primary point evaluation; and 82 official MATR test cells (21 events and 61 censored) supported a survival sensitivity. Predictions are causal, cells are independent units, and censored cells receive no pseudo-endpoint. A benchmark-informed heuristic routes among ridge, LSBoost, and compact XGBoost using outer-training descriptors; it is a hypothesis, not a validated selector. Internal fold-macro RMSEs for NASA, CALCE, HUST repeated grouped, and HUST protocol-family holdout were 12.15, 38.91, 137.16, and 126.39 cycles. On the 21 primary-test cells, protocol-aware XGBoost achieved 26.98-cycle battery-macro RMSE and 92.92% coverage for nominal 90% intervals. Protocol-aware features reduced battery- macro RMSE by 13.36%; this improvement remains suggestive (p = 0.0868) rather than statistically conclusive. A matched audit showed XGBoost significantly outperformed ridge but not LSBoost or random forest. A HUST learning curve improved from 20 training cells to the full set, so the routing cutoff is not a sample-sufficiency threshold. At the strict-80% SOH endpoint, nonlinear XGBoost-AFT did not improve significantly over penalized linear AFT. This work introduces a leakage-resistant, censor-aware evaluation framework for capacity-only online lithium-ion battery remaining-useful-life (RUL) prediction, rather than proposing a novel standalone prediction model. Within this framework, existing base predictors including ridge regression, LSBoost, compact XGBoost, TCN, and GRU are reused as model backends. Protocol-aware metadata features serve as an optional add-on feature branch for ablation comparison and are not mandatory inputs for the core capacity-only framework. This study performs algorithm-oriented offline evaluation using ground-truth capacity/SOH readings from public cycling datasets, rather than testing on embedded BMS hardware. The prediction performance will deteriorate when fed with noisy SOH/capacity estimates output from real battery-management systems. The benchmark-informed routing scheme serves as a heuristic hypothesis, not a universally validated or preregistered model-selection criterion. The evidence supports rigorous cell-level evaluation and cautious, representation-aware model choice, while strict-endpoint sample size and early-life uncertainty remain unresolved. This study conducts algorithm-oriented offline evaluation using ground-truth capacity and state-of-health (SOH) readings from public battery-cycling datasets, rather than performing tests on embedded BMS hardware. Prediction performance will degrade when fed with noisy SOH/capacity estimates output by real-world battery-management systems.
The effect of safety valve orientation on the thermal runaway behavior of prismatic lithium-ion batteries may vary with state of charge (SOC), yet this combined influence remains insufficiently understood. In this study, 15 Ah prismatic ternary lithium-ion batteries at 0% SOC, 50% SOC, and 100% SOC were externally heated with the safety valve oriented upward or downward. External venting phenomena, surface temperature, temperature rise rate, voltage response, gas concentrations, mass loss, and venting temperature were systematically characterized. The results showed that the overall thermal response weakened with decreasing SOC, while the influence of safety valve orientation varied across SOC levels. At 100% SOC, the orientation difference was mainly reflected in rapid heating intensity, whereas at 0% SOC, sustained rapid heating and thermal runaway occurred only in the downward orientation. Two-way ANOVA showed that SOC and safety valve orientation significantly affected both maximum temperature and maximum temperature rise rate. Their interaction was significant for the maximum temperature rise rate but not for the maximum temperature, indicating a stronger combined influence on the dynamic heating process. These findings provide experimental evidence for evaluating the influence of cell orientation on thermal runaway risk and offer practical guidance for battery installation, safety valve arrangement, and vent path design in battery systems.
This study evaluated whether artificial ageing at the complete-pack level can reproduce selected electrical and mechanical characteristics of pouch cells aged in vehicle operation. A two-phase procedure combined Worldwide Harmonized Light Vehicles Test Cycle (WLTC)-derived loads with commanded 2C recharging under active BMS current limitation while retaining the original battery management and liquid-cooling systems. Across 300 charging events, 1860 WLTC-based profiles were completed, corresponding to an estimated 114 energy-based full equivalent cycles and 43,300 km of WLTC-equivalent driving. Extracted cells were compared with fresh and real aged reference cells from 160,000 km of vehicle operation using electrical characterization and quasi-static indentation at 100% state of charge. Five health indicators comprised charge- and discharge-based state of health, coulombic efficiency, and two capacity-to-voltage slopes over the 45–32 Ah and 32–19 Ah intervals. Agreement between the electrical indicators was metric-dependent, with close agreement in charge-based state of health and the slope between 45 and 32 Ah. Artificially aged and real aged cells showed comparable mean maximum indentation forces (43.5 versus 42.5 kN) and ISC-onset displacements (2.100 versus 1.991 mm) under the investigated test conditions. Selected characteristics were reproduced after 27% of the reference mileage, corresponding to a distance-based acceleration factor of approximately 3.7.
Accurate prediction of the remaining useful life (RUL) of lithium-ion batteries is a prerequisite for reliable battery management, predictive maintenance and safe second-life utilization. Existing approaches fall into two families. Physics-based models are interpretable but depend on precisely identified electrochemical parameters and are computationally expensive, whereas data-driven models such as the Transformer capture long-range temporal dependencies yet remain physically unconstrained, offer limited interpretability, and generalize poorly when labeled degradation data are scarce. This paper proposes a physics-inspired Transformer, termed PI-Transformer, in which a shared sequence encoder serves two heads: a primary head that regresses the RUL and an auxiliary head that emits a per-cycle degradation curve on which a first-order decay law is imposed as a soft constraint in the log domain. The constraint therefore regularizes the shared representation through an unsupervised auxiliary task rather than the regression target itself, and it is discarded after training, so that the deployed model carries no additional inference-time cost. Within the BatteryML framework, four variants are compared on the MATR-1 fast-charging dataset using QdLinear features: a Transformer decoder baseline, a Transformer encoder, a causally masked decoder and the proposed physics-inspired decoder. The proposed variant reduces the test RMSE by 20.9% with respect to the unconstrained baseline, from 84.01 to 66.46 cycles; over 10 random seeds it attains 68 ± 8 cycles against 84 ± 9 for the baseline, and the ranking of the four variants is unchanged.
Accurate battery-capacity predictions do not necessarily support reliable reserve-sufficiency decisions. This study jointly evaluates short-window capacity estimation, empirical uncertainty, remaining backup time, and operational decisions under domain shift. Eleven lithium-ion cells from the NASA PCoE Battery Data Set were evaluated using cell-held-out validation across domains differing in battery series, discharge current, and temperature. Available discharge capacity to 2.7 V was estimated from initial discharge windows of 1–15 min, with and without single-reference cell-specific correction; remaining backup time was evaluated under constant-current discharge. In the four-cell primary domain at T=10 min, representative zero-shot Random Forest and one-shot Ridge configurations achieved capacity MAEs of 0.0456 and 0.0335 Ah and backup-time MAEs of 1.472 and 1.135 min, respectively. Across the predefined model grid, lower prediction error did not consistently correspond to better empirical coverage or fewer unsafe permissions. Interval coverage was sensitive to calibration-cell composition and deteriorated under transfer for the representative configurations. At a 30 min reserve horizon, nominal-95% lower-bound decisions reduced observed unsafe permissions from 26 to 0 for Random Forest and from 5 to 0 for Ridge, while increasing false refusals from 28 to 145 and from 4 to 109, respectively, across 632 evaluated episodes. In a separate five-cell Orbital capacity-estimation experiment using models trained on voltage/current-only features, single-reference correction increased capacity MAE for both models on every cell. These findings support separate validation of accuracy, empirical coverage, and reserve-sufficiency decisions, with conclusions limited to the investigated cells and operating conditions.
