State Of Energy Prediction Of Lithium-ion Batteries In Wireless Sensor Network Nodes

Self-powered wireless sensor network(WSN)can achieve long-term sustainable operation of the entire network by harvesting energy from the environment,which provides a feasible solution to the problem of inadequate energy for wireless sensors.However,the unpredictability of the energy that can be collected from the environment poses challenges to the balance between collection,storage,and consumption units.To maintain consistency in data collection and transmission and maximize the node’s operating time,an efficient and robust energy management unit is needed to regulate the system.Lithium-ion batteries are one of the best choices as energy storage devices for selfpowered nodes in WSNs due to their advantages of no memory effect,high energy density,long cycle life,and being pollution-free after being discarded,ensuring that the sensor nodes maintain high power operation for a long time.The key to an energy management strategy is to monitor the remaining maximum available energy of the internal energy storage device.Accurate estimation of the state of energy(SOE)under the operating status can significantly improve the efficiency of node operation.By estimating the remaining available energy of a single node,a reference basis can be provided for task scheduling of the entire sensor node network,thereby extending the overall working life of the network.The main research contents of this paper are as follows:(1)To address the nonlinear characteristics of lithium-ion batteries in WSN nodes and the key factors that affect them,such as environmental temperature,charge-discharge current rate,and cyclic aging,experiments on open circuit voltage(OCV)at different temperatures,hybrid pulse power characterization(HPPC)working condition at different charge-discharge rates and cyclic aging are conducted to obtain the basic experimental data for the key influencing factors.The experimental data show that temperature,charge-discharge current rate,and cyclic aging have a significant impact on the maximum available energy of lithiumion batteries,providing theoretical support for constructing a temperature-compensated fractional-order equivalent circuit model(FO-ECM)and improving the SOE estimation algorithm.(2)To address the problem of large residual energy estimation errors caused by the nonlinear discharge characteristics of lithium-ion batteries,a WSN power management system model is established based on the topology of the energy collection module,the operating mode of the energy storage unit,and the energy consumption model of the photovoltaic energy collection system.Lithium-ion battery tests are conducted at different stages and workloads under the dynamic stress test(DST)working condition to analyze and evaluate the energy status of the storage unit supply module.Furthermore,verification working condition experiments are conducted at different temperatures,cyclic aging,and charge-discharge current rates to provide experimental data support for verifying the robustness and reliability of the proposed algorithm.(3)Considering the fractional-order characteristics of the WSN systems and the main electrochemical characteristics of lithium-ion batteries during charge transfer,a temperaturecompensated bi-polarization FO-ECM is established,considering the effects of various environmental temperatures and cyclic aging conditions.To address the problem of online identification of fractional-order sequence,a variable bi-order fractional-order forgetting factor least squares method is proposed to achieve full-parameter measurements of the lithiumion battery online,adjusting the fractional-order sequence to fit the impedance spectrum and capacitance degree,improving the accuracy of model parameter identification.The results show that the temperature-dependent bi-polarization FO-ECM can better characterize the dynamic characteristics of lithium-ion batteries,and the proposed online identification algorithm has better estimation accuracy under dynamic working conditions.(4)An improved SOE and maximum available energy co-estimation framework is established for traditional WSN node lithium-ion battery SOE estimation,considering the problem of maximum available energy value decay caused by environmental temperature,battery charge-discharge rate,and aging phenomena.The maximum available energy value is updated in real-time to reduce the SOE error caused by fixed energy values and improve the accuracy of SOE estimation throughout the entire lifecycle.A multi-time scale SOE and maximum available energy co-estimation framework is proposed to address the asynchronous and coupled characteristics of maximum available energy and SOE estimation.The microtime scale parameters are updated at each time step.Meanwhile,the macro-time scale parameters are only updated when the constraint conditions are satisfied,effectively reducing the algorithm’s computational complexity.(5)The experimental validation of the maximum available energy and SOE estimation co-estimation framework under multi-time scale is conducted according to the designed dynamic stress test conditions of lithium-ion batteries in WSN nodes.The experimental results show that the SOE algorithm with maximum available energy correction can effectively correct the divergence problem caused by fixed maximum available energy values and significantly improve the accuracy of residual energy estimation under various working conditions.The multi-time scale co-estimation framework shows that smaller time scales can provide more accurate and reliable maximum available energy correction and higher SOE estimation accuracy,but the computational time cost is higher than that of larger time scales.To balance SOE estimation accuracy and algorithm computational complexity,the appropriate time scale should be selected based on the SOE estimation accuracy and time cost in practical battery management system working conditions.