Comparison of energy, exergy and entropy generation-based criteria for evaluating stratified thermal store performances

•The effectiveness of three criteria for assessing the performance of SSTES are compared.•Energy-based θ¯, θ* and ηe, exergy-based ηx¯, and entropy generation-based Ns, are considered.•θ¯, θ*, and ηe give distinct SSTES evaluations during just a part of the charging duration.•ηx¯ offers rational evaluation for charging periods prior to one tank volume change.•SSTES performance at any time during charging may be distinguished with Ns.

It is well established that stratification improves the performance of stratified sensible thermal energy stores (SSTES). A detailed review is presented of studies in which the contribution of stratification to the performance SSTES is assessed on the basis of energy and exergy analyses, and lately, entropy generation analysis. To obtain clear and useful distinctions between stratified and non-stratified stores, and for stores at different states of stratification, analyses that incorporate second law considerations (i.e. exergy/entropy generation) are required. Also, based on outcomes of computational fluid dynamics (CFD) simulations of SSTES, we present results of energy efficiency, ηe, normalized exergy efficiency, η¯x, and entropy generation number, Ns, assessments of SSTES, corresponding to three possible approaches. Energy efficiencies, are shown to be limited in their ability to quantifying these improvements, while η¯x is introduced as an improved exergy-based performance measure. Comparing η¯x with Ns it is found that η¯x is effective for assessing store performances prior to the exit of the thermocline from the store, while Ns offers useful assessments for the full duration of store operation. The parametric dependence of η¯x and Ns on some significant dimensionless variables (Peclet number, PeD, Richardson number, Ri, and aspect ratio, AR) is such that as PeD, Ri   and AR increase η¯x increases. Ns decreases with increasing PeD and AR but the effect of Ri on it depends on the value of Ri.