Modeling and simulation of two-phase flow evaporators for parabolic-trough solar thermal power plants.

Abstract

The main goal of this work was to contribute in modeling horizontal twophase

flow boiling channels (evaporators) for parabolic-trough solar thermal

power plants. The evaporator model is essential for the design of control

schemes. Computational fluid dynamics were applied in order to model and

discretize the evaporator.

Finite volume and moving boundary models were explored in this work.

However, some issues in the dynamic simulation of finite volume models, i.e.

chattering, did not allow for taking full advantage of dynamic simulations.

Furthermore, none of the current moving boundary models considered dynamic

switching between all possible flow configurations in evaporators for

parabolic-trough solar thermal power plants.

The developed evaporator models were obtained from physical principles.

The object-oriented equation-based modeling paradigm, which was considered

for the design of the dynamic evaporator models, contributed to model

maintenance, reusability and decoupling. The equation-based modeling

paradigm increased reusability further, by not fixing the causality and thus

making the models suitable for a wider range of experiments and simulations.

In order to validate the evaporator models, experimental data from a

direct steam generation parabolic-trough solar thermal power plant was used

– the DIrect Solar Steam (DISS) facility owned by Centro de Investigaciones

Energéticas MedioAmbientales y Tecnológicas (CIEMAT)-Plataforma Solar

de Almería (PSA), a Spanish government research and testing center.

This work makes contributions to the field of modeling and simulation of

dynamical systems. A chattering study of finite volume homogeneous twophase

flow dynamic models is presented. The general solutions employed

to solve the chattering problem were analyzed, and particular approaches

were implemented: Mean Densities and the Heuristic approaches, enabling

discretized models to be simulated effectively. New mathematical moving

boundary models were developed in order to support dynamic switching

between all possible flow configurations for evaporators and condensers, and

an equation-based object-oriented library was also implemented to test the

integrity and stability of the moving boundary models.

Approaches to solve chattering in finite volume models and new moving

boundary models were validated against experimental data taken from the

DISS facility. For this purpose, different DISS models were implemented, by

considering finite volume and moving boundary models. Some unknown parameters

were calibrated using a multi-objective genetic algorithm. Finally,

these DISS models were compared against experimental data from the DISS

facility in terms of accuracy and performance. In order to facilitate these

tasks, simulation and calibration frameworks were defined. All the developed DISS models, composed of finite volume and moving

boundary evaporator models, were exposed to a wide range of operating

conditions and disturbances, proving that they can take full advantage of

dynamic simulations and can be used for the design, testing and validation

of advanced control systems.