Abstract
The efficient and economic production of landfill gases (LFG) by optimally adjusting LFG production settings is of high interest as a promising source of biomass energy. A key obstacle in LFG production optimization is the large-scale and complex system with overwhelming uncertainty and heterogeneity. We propose a simplified ensemble-based optimization (EnOpt) method to solve the LFG production optimization problem when constraints are not a concern, where the gradient information is obtained from an ensemble of realizations of the system. For constrained optimization, a novel parameterless genetic algorithm is proposed and successfully applied to the simulated LFG process. The effectiveness of the proposed (EnOpt) method and the parameterless genetic algorithm is demonstrated with the simulation of a landfill and gas generation and transport therein, using a parallel computation strategy. © 2014 American Institute of Chemical Engineers AIChE J, 60: 2063-2071, 2014 © 2014 American Institute of Chemical Engineers.
Original language | English |
---|---|
Pages (from-to) | 2063-2071 |
Number of pages | 9 |
Journal | AICHE Journal |
Volume | 60 |
Issue number | 6 |
Early online date | 1 Feb 2014 |
DOIs | |
Publication status | Published - Jun 2014 |
Externally published | Yes |
Bibliographical note
The computation work was carried out using the Linux cluster at the High-Performance Computing and Communications (HPCC) of the University of Southern California.Funding
Financial support from Bayeco (Beijing) Environmental Technology Ltd. as a member of the Texas-Wisconsin-California Control Consortium is gratefully acknowledged.
Keywords
- Ensemble-based optimization
- Landfill gas system
- Parallel computation
- Production optimization