Mutation particle swarm optimization (M-PSO) of a thermoelectric generator in a multi-variable space. (15th November 2020)

• Record Type: Journal Article
• Authors: Wang, Xi; Ting, David S-K; Henshaw, Paul
• Is Part Of: Energy conversion and management. Volume 224(2020)
• Publication Date: 2020-11-15
• Subjects: TEG -- Optimize -- Multi-parameter -- Algorithm; Direct energy conversion -- Periodicals; Energy storage -- Periodicals; Energy transfer -- Periodicals; Énergie -- Conversion directe -- Périodiques; Direct energy conversion; Periodicals; 621.3105
• DOI: 10.1016/j.enconman.2020.113387
• Languages: English
• ISSNs: 0196-8904

Abstract

Highlights: A TEG was optimized using a M-PSO algorithm with constant mutation factor. M-PSO is more effective than the traditional PSO algorithm for a TEG optimization. Increasing the mutation factor or particle population can improve the accuracy. Acceleration constants affect the convergence speed for all population sizes. Abstract: With the recent development of thermoelectric materials, thermoelectric generators (TEGs) have become a technology with a huge potential in the energy recovery field. A TEG module consisted of 199 cascade couples was analyzed by a one-dimensional steady-state thermodynamic model. Meanwhile, using the thermodynamic analysis results, two discrete objective functions for the output power and efficiency of the TEG module were built on a multi-variable searching space where working conditions and geometric structures were varied. Mutation particle swarm optimization (M-PSO) was selected to conduct the optimization of the output power and efficiency of the TEG module because it converges more accurately and quickly than PSO. Under temperature differences below 40 K, the optimal output power and efficiency were 23.6 W and 4.05%, respectively. It is necessary, however, to simultaneously consider the output power and efficiency as important targets in the application of TEG technology. In this way, a weighted approach was applied to this study to establish a multi-objective function for the TEG module. Then, a multi-objective optimization was conducted using the M-PSO algorithm for the combined output power and efficiency in the same searching space. With the optimization, the output power and efficiency reached 6.69 W and 3.99%, respectively, when the weight factor for the efficiency was 0.8.