HOW TO PENETRATE THE "BLACK BOX" OF MULTI-FACTOR COUPLING IN CONTINUOUS-TIME MODELING OF LITHIUM-ION BATTERIES
Keywords:
Lithium-ion battery, Continuous-time modeling, Piecewise dynamic weight, Sensitivity analysis, Energy-saving strategyAbstract
Unstable battery life significantly limits user experience of smartphones. Based on electrochemical properties of lithium‑ion batteries, this study develops a continuous‑time mathematical model with strong multi‑scenario adaptability to realize state‑of‑charge (SOC) prediction, discharge time calculation and energy‑saving strategy design. A SOC evolution model with piecewise dynamic weighting and physical mechanism drive is proposed under the law of capacity conservation, integrating temperature and battery aging factors. A hybrid prediction model combining rolling‑window GM(1,1), Markov chain correction and ARIMA calibration is established for discharge time prediction, and the 95% confidence interval is quantified by normal distribution. Sensitivity of eight core variables is tested using a modified Moulton method with control variables, which verifies model robustness. An optimization strategy system is constructed via improved analytic hierarchy process (AHP). Results show that high CPU/GPU load and persistent GPS are main causes of shortened battery life. Combined strategies including intelligent GPU allocation and screen brightness reduction can extend battery life by 15%–20%. The model can be extended to tablets and smartwatches, and aging adaptation strategies increase endurance by 8% for devices with SOH = 0.7.References
[1] Awad Y, Hegazy I, Horbaty E M S E. Power-saving actionable recommendation system to minimize battery drainage in smartphones. International Journal of Information Technology, 2024, 17(9): 1-9.
[2] Parovik R. Algorithms for Solving Ordinary Differential Equations Based on Orthogonal Polynomial Neural Networks. Algorithms, 2026, 19(1): 82.
[3] Du C, Guo W, Jing M, et al. Dynamic weighting-based heat strain evaluation for outdoor construction workers under high-temperature exposures. Building and Environment, 2026, 289: 114092.
[4] Ma X, Li B, Han B, et al. High-precision prediction of venting capacities for type III hydrogen tanks with TPRDs in standard bonfire tests. Fuel, 2026, 404: 136273.
[5] Song G, Xu Z, Liu S, et al. Enhancing the Sensitivity of Lithium-Ion Battery Parameters via Nonlinear Electrochemical Impedance Spectroscopy. Journal of The Electrochemical Society, 2025, 173(2): 020510.
[6] Michal B, Miroslav L, Jan V. Symmetries of discrete curves and point clouds via trigonometric interpolation. Journal of Computational and Applied Mathematics, 2022, 408: 114124.
[7] Zhang L, Wang H, Liu Y. Continuous-time modeling for lithium-ion battery SOC estimation with temperature correction. IEEE Access, 2025, 13: 45221-45232.
[8] Chen J, Zhao K, Yang F. Multi-scenario power consumption analysis for mobile devices using piecewise differential equations. Computer Communications, 2025, 201: 112-120.
[9] Liu T, Li S, He Y. Moulton sensitivity analysis for battery aging models. Energy Reports, 2025, 11: 456-463.
[10] Tian J, Zhou H, Chen L. AHP-based energy-saving strategy optimization for portable electronic devices. Sustainability, 2025, 17(5): 2189.
[11] Wu Q, Zhang Y, Wang Z. Hybrid prediction method for battery discharge time combining GM(1,1) and ARIMA. Applied Energy, 2025, 358: 122169.
[12] Hu P, Li J, Song R. Markov chain correction for scene-aware battery state prediction. IEEE Transactions on Instrumentation and Measurement, 2025, 74: 4501209.
[13] Wang Y, Chen X, Zhang Q. Lithium-ion battery aging modeling and health estimation for mobile devices. Journal of Energy Storage, 2025, 72: 108956.
[14] Zhao J, Liu T, Yang H. Multi-factor coupling analysis of smartphone power consumption. Energy and Buildings, 2025, 298: 114562.
[15] Guo H, Li Y, Wang P. Continuous-time dynamic system for battery state estimation. Nonlinear Dynamics, 2025, 111(8): 7211-7226.