Thermo-Electro-Hydrodynamic Investigations Of All-Vanadium Redox Flow Battery Using Lumped Model And Numerical Simulation

Abstract

Renewable energy technology is needed urgently due to crisis and challenges against conventional energy's. However, intermittent and random nature of the non-conventional energy sources like solar and wind leads to low quality electricity output and a poor stability in the grid. The electricity energy storage (EES), storing and releasing electricity, is available temporarily to resolve those problems. Among all the EES technologies all-vanadium redox ow battery (VRFB) is a promising option. The VRFBs have many advantages including long cycle life, active thermal management, elimination of electrolyte cross-contamination, high energy e ciency, low cost for large energy storage systems, etc. Most important feature of redox ow batteries is its power exibility. Also, capacity design of a battery is not coupled making more e cient design of the battery con gurations. Hence, the redox ow batteries are preferred for the applications of peak shaving, load leveling, grid integration and frequency regulation. Compared to other ow batteries, VRFB is more preferred since it does not su er from the problem of electrolyte cross-contamination due to the fact that both the half cells of the battery employ di erent species of vanadium in the electrolyte. As a result, electrolyte lifetime is increased signi cantly. It has been observed that the disposal of vanadium ions from VRFB do not create any environmental issues in comparison with conventional lead acid battery. However, a few challenges have to be taken into account to implement this echnology. As of now, there are only a few lumped and comprehensive models developed for VRFBs which explain system performances and characteristics at steady and unsteady states under di erent operating conditions. The present thesis employs lumped dynamic model for the all-vanadium redox ow battery, using the conservation equations of charge and mass combined with the e ects of major resistances, electrochemical reactions and recirculation of the electrolyte through external reservoirs. The same model is extended by adding the e ects of crossover of vanadium ions through membrane and mass transfer. This model is able to predict the cell voltage variation and capacity loss of cell for di erent membrane materials. One more lumped thermal model is used, which is based on conservation of energy to predict the temperature variation of cell and reservoir as a function of time, surrounding air temperature and battery structure. A comprehensive numerical model is used for two-dimensional, three-dimensional steady and unsteady simulations. The model is based on the conservation laws along with the fundamental modes of transport and kinetic model for reactions involving the vanadium species. The processes in the redox ow batteries are described by the equations of electrodynamics, electrochemistry and uid mechanics. The electrolyte ow through porous electrodes and di usion of proton ions through the membrane can be modeled by the laws of conservation of mass and momentum. By solving the Poisson equations, the distributions of electric potential for VRFB can be determined. Using Nernst equation and Butler-Volmer law, the electrochemical interaction of species in the cell can be obtained. Combination of these processes makes the solution of equations more complicated. The same model is also extended by including the concentration over potential and ohmic losses and predicts various e ciencies accurately for di erent parameters. There are many issues which need to be addressed in order to make it commercially viable application. These issues include optimization and scale-up which encompasses ow geometries assemblage and operation conditions, membrane fouling, development of electrode materials resistant to oxidation and improvement in electrolyte stability. Modeling and simulation are cost-e ective methods for solving these problems, which can minimize the time and costs of the laboratory experiments. In addition to this, modeling and simulation can reveal the detailed information about fundamental processes inside the vanadium redox ow battery, which is useful for optimizing the design and operating condition. Moreover, there is a need to develop control oriented models that can capture the performance accurately. Keeping these motivations in mind, this doctoral thesis focuses on development of lumped and comprehensive numerical models. The objectives of this work are two folds. Firstly, it is to develop a lumped model which includes lumped dynamic model incorporating with and without crossover e ects and thermal modeling. The other objective is to develop two and three-dimensional steady and transient numerical model encompassing theoretical analysis which will be helpful for fundamental understanding of the underlying mass and charge transport characteristics in the porous electrode. A primary focus is to gain more detailed coupled characteristics of mass and charge for the operation of vanadium redox ow battery to optimize the performance of the battery. The detailed research objectives are given below. A lumped dynamic model simulation has been carried out. The model is able to predict the e ects of ow rate, applied current density, concentration, porosity, temperature on the performance of an all-vanadium redox ow battery. It is observed that higher electrolyte ow rate gives longer charging time and higher cell performance. The lower applied current density yields better potential di erence than higher applied current densities. The higher vanadium concentration shows increase in cell voltage which leads to high performance. The lower porosity of the electrode gives little less potential di erence than higher porosity electrode during charging, while discharging it gives higher potential di erence, but overall lower porosity electrode gives better performance. The higher operating temperature leads to poor performance of the cell. These parametric studies help to optimize the performance of the all-vanadium redox ow battery. The above model is extended by adding the e ects of mass transfer and crossover of vanadium ions through membrane. The model predicts the capacity loss for di erent membrane materials. The simulation results show that reaction rate constants and di usion coe cients depends on temperature and it a ects cell performance. The model is able to predict the capacity loss of the battery due to the di usion of vanadium ions across the membrane over many cycles at di erent temperatures. The e ect of temperature and porosity on concentration change is studied for three membrane materials such as, Selemion AMV membrane, Selemion CMV membrane and Na on 115 membrane and the temperature ranging from 10 C and 40 C. The e ects of temperature on di usion coe cients and reaction rate constants have also been studied. It is observed that the capacity loss for Selemion AMV membrane and Selemion CMV membrane shows linear variation with number of cycles and it does not stabilize. In the case of Na on 115 membrane, capacity loss is experienced initially up to 77 cycles, then stabilizes with increase in number of cycles. The simulation results show that crossover and mass transfer e ects have signi cant impact on the performance of the cell potential response. A thermal model is developed for the vanadium redox ow battery system based on the conservation of energy for transient conditions. The model has three energy balance equations. One energy balance equation for the battery stack and the other two energy balance equations for electrolyte reservoirs are solved. One stack with 19 cells and two electrolyte reservoirs are considered for simulations. The model formulation considers the pump power loss, chemical power loss and power loss due to cell internal resistance. The dynamic model is able to predict the stack and tank temperatures for various working conditions. The results show that the stack and reservoir temperature are very much sensitive to the e ects of ow rate, surrounding air temperature. Therefore, this model can be more useful to design the battery system for speci c location of the installation...