基于电热耦合的智能手机电池建模与续航预测

Smartphone Battery Modeling and Endurance Prediction Based on Electro-Thermal Coupling

摘要: 智能手机电池的续航预测涉及负载动态、电池状态与环境热效应之间的耦合，现有研究多将三者分开处理，导致复杂使用场景下的预测精度不足。本文建立了负载–电化学–热耦合模型，从线性放电描述出发，逐步纳入硬件功耗分解、电压反馈、温度反馈、电池参数随工况变化及热耦合效应，构成连续微分代数方程组。在待机、轻度使用、视频播放、3D游戏和导航五种典型场景下对电池续航进行仿真，并对极端低温、负载频繁波动和参数不确定性三类非典型工况进行了蒙特卡洛测试。结果表明，处理器负载和屏幕亮度是影响续航的主要因素，在各自设定变化范围内造成的续航差异分别约为86%和48%。低电量阶段由于电压反馈效应，实际放电速度显著快于线性估算。三类非典型工况下预测偏差均处于可接受范围，未出现精度明显下降。该模型框架适用于其他便携设备的续航预测，可为智能手机功耗管理提供参考。

Abstract: Endurance prediction for smartphone batteries involves the coupling among load dynamics, battery state, and ambient thermal effects. In existing studies, these three factors are mostly treated separately, resulting in insufficient prediction accuracy under complex usage scenarios. In this paper, a load-electrochemical-thermal coupling model is established. Starting from a linear discharge description, the model progressively incorporates hardware power consumption decomposition, voltage feedback, temperature feedback, variation of battery parameters with operating conditions, and thermal coupling effects, forming a system of continuous differential-algebraic equations. Simulations of battery endurance are conducted under five typical scenarios, namely standby, light use, video playback, 3D gaming, and navigation. Monte Carlo tests are further performed under three atypical operating conditions, i.e., extreme low temperature, frequent load fluctuations, and parameter uncertainty. The results show that processor load and screen brightness are the dominant factors affecting endurance, causing endurance differences of approximately 86% and 48%, respectively, within their respective variation ranges. In the low state of charge region, the actual discharge rate is significantly faster than the linear estimation due to the voltage feedback effect. Under the three atypical conditions, the prediction deviations remain within acceptable ranges, with no significant degradation in accuracy. The proposed modeling framework is applicable to endurance prediction for other portable devices and can serve as a reference for power management of smartphones.