Research On Energy Management Strategy Of Lithium Battery For Power-assisted Bicycle With Neural Network

New energy vehicles powered by lithium batteries,with the advantages of being green,environmentally friendly,convenient,and low-cost,have received increasing attention at home and abroad.Lithium battery energy estimation and energy distribution is one of the hot issues in current research on new energy vehicles.This article mainly studies the energy management strategy of lithium battery for power-assisted bicycle.The main work content is as follows:(1)On the basis of learning SOC estimation method,Quantum-behaved Particle Swarm Algorithm(QPSO)and wavelet neural network,an improved QPSO training wavelet neural network is proposed.In order to overcome the disadvantages of QPSO algorithm,which tends to fall into local optimum,an improved QPSO based on variation idea is proposed.The global optimal position is subjected to a certain degree of mutation operation according to its dispersion degree and multiple iterations are performed.To avoid the algorithm into a local optimum.Construct Wavelet Neural Network based on Improved QPSO to estimation the SOC of lithium battery.The wavelet basis function is used as the transfer function of the neural network and the improved QPSO replaces the traditional gradient descent algorithm.Lithium battery is discharged by the cycle step discharge current,and its discharge data is used as a training sample for the neural network.The BP neural network and wavelet neural network were constructed and comparison with the neural network of this paper through experimental analysis,the results show that the proposed algorithm is more suitable for online,real-time estimation of lithium battery for the power-assisted bicycles.(2)Establish a dynamic model of the power-assisted bicycle,and define the specific riding conditions encountered in riding and provide solutions based on riding experience.According to the established dynamic model and the solution of special riding state,the riding control algorithm of the bicycle and the lithium battery energy distribution strategy are given.(3)Built a smart phone-based energy management test platform for the power-assisted bicycle,and complete the SOC estimation and energy management tests for lithium battery of the power-assisted bicycle.The experimental results show that in actual riding,the SOC estimation error of lithium battery based on improved QPSO wavelet neural network is lower,and the lithium battery energy allocation strategy for bicycle-ride riding control has lower allocation error and gives the rider a better riding experience.