Study On Thermal Parameters Measurement For Pouch Cells

Lithium-ion batteries(LIBs)are strongly thermo-electrochemical coupled systems.Their thermal parameters including specific heat capacity and thermal conductivities are of importance to characterize the thermal property.However,the structure with multiple layers in wet environment of LIBs poses challenges to measurement of thermal parameters,calling for a both fast and precise approach to meet increased demand from academia and industry.Focusing on thermal parameters measurement of laminated pouch cells,nondimensionalization and sensitivity analysis are selected as methodology.This thesis conducts systematic research on analysis of model resolution,optimization of measurement precision and surrogate model solution,which provides theoretical basis and approaches to improve thermal parameters measurement.Based on the fact that all the thermal parameters measurement is correlated with a heat conduction model,the modified transient plane source(MTPS)method is chose as an example to conduct research.Considering the universality of obtained results,the mathematic model in MTPS method is first nondimensionalized.The corresponding dimensionless variables naturally become the inlet to optimize this method.Afterwards,dimensionless sensitivity is utilized to quantitatively define the model resolution.From the point of view that the model resolution should completely tolerate the experiment deviation,the correlation between model resolution and the desired precision of a method is obtained.By this correlation,the threshold for model resolution is given as an instruction for optimizations.According to the correlation between model resolution and the desired precision of a method,it should be noted that improving precision of a method equals to finding some factors that can enhance the dimensionless sensitivity of the model.How both experimental and geometric factors affect the dimensionless sensitivity of the MTPS method are investigated.The result shows that the higher heat flux in safe ranges,the closer location to the center of cell surfaces,the longer heating duration in the conditions of meeting the dimension reduction assumption and the lower thickness-width ratio can help to improve the dimensionless sensitivity of the MTPS method.An example is given to demonstrate the utilization of these findings above.To remove the barriers in computing efficiency and economic cost when a thermal parameter measurement method applying finite element model(FEM),a dimensionless surrogate model solution is proposed.First,following the instruction of precision optimization,the high quality database for better model resolution is established.Then the diagram of the neural network is determined after check in nodes numbers.The testing result suggests the surrogate model can reproduce the output from FEM,reflecting remarkable generalization of it.Finally,the parameter recognition results given respectively by MTPS method using FEM and surrogate model are compared,showing a deviation within the range of temperature measurement but 17 times efficiency improvement when using surrogate model.