Date of Award

Fall 2013

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Mechanical Engineering

First Advisor

Suresh Garimella

Committee Chair

Suresh Garimella

Committee Member 1

Timothy Fisher

Committee Member 2

Eckhard Groll

Committee Member 3

Rakesh Agrawal


Concentrating solar power (CSP) is a renewable and demonstrated technology for large-scale power generation but requires multiple engineering advancements to achieve grid parity with conventional fossil fuels. Part of this advancement includes novel and inexpensive thermal energy systems to decouple daily power production from intermittent solar collection. Dual-media thermocline tanks, composed of molten salt and solid rock filler, offer low-cost storage capability but the concept has experienced limited deployment in CSP plants due to unresolved concerns about long-term thermal and structural stability. The main objective of the present work is to advance the understanding of thermocline storage design and operation necessary for future commercial implementations. A multiscale numerical approach is conducted to investigate tank behavior at both a device level for comprehensive short-term analysis and at a system-level for reduced-order long-term analysis.

A computational fluid dynamics (CFD) model is first developed to simulate molten-salt thermocline tanks in response to cyclic charge and discharge modes of operation. The model builds upon previous work in the literature with an expanded study of the internal solid filler size as well as added consideration for practical limits on tank height. Reducing the internal filler size improves thermal stratification inside the tank but decreases the bed permeability, resulting in a design tradeoff between storage performance and required pumping power. An effective rock diameter of 1 cm is found to be the most practical selection among the sizes considered. Also of interest is the structural stability of the thermocline tank wall in response to large temperature fluctuations associated with repeated charging and discharging. If sufficient hoop stress is generated from storage cycles, the tank becomes susceptible to failure via thermal ratcheting. The thermocline tank model is therefore extended to predict wall stress associated with operation and determine if ratcheting is expected to occur. Analysis is first performed with a multilayer structure to identify stable tank wall designs. Inclusion of internal thermal insulation between the porous bed and the steel wall is found to best prevent thermal ratcheting by decoupling the thermal response of the wall from the interior salt behavior. The structural modeling approach is then validated with a simulation of the 182 MWht thermocline tank installed at the historic Solar One power tower plant. The hoop stress predictions are found to show reasonable agreement with reported strain gage data along the tank wall and verify that the tank was not susceptible to ratcheting.

The preceding use of commercial CFD software for thermocline tank simulation provides comprehensive solutions but the ease of application of this approach with respect to different operating scenarios is constrained by high computing costs. A new reduced-order model of energy transport inside a thermocline tank is therefore developed to provide thermal solutions at much lower computational cost. The storage model is first validated with past experimental data and then integrated into a system model of a 100 MWe molten-salt power tower plant, such that the thermocline tank is subjected to realistic solar collection and power production processes. Results from the system-level approach verify that a thermocline tank remains an effective and viable energy storage system over long-term operation within a CSP plant. The system-level analysis is then extended with an economic assessment of thermocline storage in a power tower plant. A parametric study of the plant solar multiple and thermocline tank size highlights suitable plant designs to minimize the levelized cost of electricity. Among the cases considered, a minimum levelized cost of 12.2 cent/kWhe is achieved, indicating that cost reductions outside of thermal energy storage remain necessary to obtain grid parity.

As a sensible heat storage method, dual-media thermocline tanks remains subject to low energy densities and require large tank volumes. A possible design modification to reduce tank size is a substitution of the internal rock filler with an encapsulated phase-change material (PCM), which adds a high density latent heat storage mechanism to the tank assembly. The reduced-order thermocline tank model is first updated to include capsules of a hypothetical PCM and then reintegrated into the power tower plant system model. Implementation of a single PCM inside the tank does not yield significant energy storage gains because of an inherent tradeoff between the thermodynamic quality (i.e., melting temperature and heat of fusion) of the added latent heat and its utilization in storage operations. This problem may be circumvented with a cascaded filler structure composed of multiple PCMs with their melting temperatures tuned along the tank height. However, the benefit of a cascade structure is highly sensitive to appropriate selection of the PCM melting points relative to the thermocline tank operating temperatures.