Session: 08-01 Thermal Energy Conversion Techniques

Paper Number: 107306

107306 - Prediction of Thermionic Energy Conversion Performance and Parametric Effects Using Genetic Algorithms to Fit Physics-Inspired Model Equations to Prototype Test Data 

Several recent studies have generated experimental performance data for narrow-gap thermionic energy conversion devices.  This investigation explores the use of genetic algorithm methods to fit existing data from the literature with physics-inspired model equations.  The resulting model equations can be used for performance prediction for system design optimization, or to explore parametric effects on performance.  The model equations incorporate Richardson’s law for current density together with appropriate relations for power delivered to the external load.  The trained model enables performance prediction of a small-gap thermionic energy conversion device with inputs of only emitter temperature and load resistance.

The prototype data considered here tested thermionic energy conversion devices to determine the output current and output power as a function of diode voltage and the emitter temperature. Each of the prototype devices used tungsten emitter and collector materials but differed in the specialized coatings used on each device’s electrodes. In this study, the prototype test data is used with a postulated, temperature-dependent, work function for the emitter and the collector materials. As a result, these postulated functions are substituted into the physics-inspired model, yielding a performance model with three adjustable constants. Optimized values of these constants are determined using a genetic algorithm to best fit the experimentally determined performance data for prototype thermionic conversion devices tested in earlier studies. 

This process yields two insightful results. First, determining the best-fit constants indicates the temperature variation of the work function for the emitter, and a value for the collector material. The magnitude of the work functions, and its temperature dependence, are important to performance modeling and understanding how much performance varies with operating conditions. This data-fitting model development also demonstrates an approach that yields a model that can be a better predictor of performance for varying operating conditions. Parametric effects on thermionic device conversion efficiency are explored using the optimized-fit model to determine the sensitivity of performance to operating temperatures and work function values. 

The thermionic performance data is also used to train an Artificial Neural Network (ANN) model in which the parametric effects of emitter and collector temperature on performance predicted by the ANN model is compared to the corresponding effects predicted by the genetic algorithm model.  These two models indicate similar overall effects of emitter temperature variation on performance. However, the use of the genetic algorithm model makes it possible to distinguish the effects of varying emitter temperature on emitter work function.

Presenting Author: Elizabeth D. Juette University of California at Berkeley

Presenting Author Biography: Elizabeth Juette is a Graduate Student Researcher in the Energy and Multiphase Transport Lab in the
Mechanical Engineering Department at UC Berkeley.