0

Full Content is available to subscribers

Subscribe/Learn More  >

Study on Nonlinear Identification SOFC Temperature Model Based on Particle Swarm Optimization-Least Squares Support Vector Regression

[+] Author Affiliations
Jinwei Chen, Huisheng Zhang, Shilie Weng

Shanghai Jiao Tong University, Shanghai, China

Paper No. GT2016-56236, pp. V003T06A005; 12 pages
doi:10.1115/GT2016-56236
From:
  • ASME Turbo Expo 2016: Turbomachinery Technical Conference and Exposition
  • Volume 3: Coal, Biomass and Alternative Fuels; Cycle Innovations; Electric Power; Industrial and Cogeneration; Organic Rankine Cycle Power Systems
  • Seoul, South Korea, June 13–17, 2016
  • Conference Sponsors: International Gas Turbine Institute
  • ISBN: 978-0-7918-4974-3
  • Copyright © 2016 by ASME

abstract

A nonlinear autoregressive network with exogenous inputs (NARX) identification model is employed for predicting the Solid oxide fuel cell (SOFC) operating temperature dynamics fast and accurately in a Solid oxide fuel cell–gas turbine (SOFC-GT) hybrid system. At the same time, the least squares support vector regression (LSSVR) method with radial basis kernel function (RBF) which uses particle swarm optimization (PSO) to optimize the LSSVR’s parameters is applied to establish the NARX model. The major factors which affect the cathode and anode outlet temperature of the SOFC-GT hybrid system are the inlet flow rate of cathode and anode. Therefore, the inlet flow rates of cathode and anode are taken as inputs of the NARX model, cathode and anode outlet temperature as outputs. With the training data sampled from the mechanism model which is derived from conservation laws, a SOFC temperature the NARX model based on the LSSVR is established. Investigations are conducted to analyze the effects of training data size and fitness function of PSO on the accuracy of the NARX model. And by comparing the temperature behaviors with the results collected form the mechanism model, the accuracy of the NARX model based on the LSSVR is verified with enough accuracy in predicting the dynamic performance of the SOFC temperature. Furthermore, in the aspect of simulation speed, the NARX model is much faster than the mechanism model because the NARX model avoids the internal complex computation process. For large size training data, the training time of the NARX model is only about 1.2s. For running all 20,000s of simulation, the predicting time of the NARX model is only about 0.2s, while the mechanism model is about 36s. In consideration of the high speed and accuracy of the NARX model, it can be applied to design valid multivariable model predictive control (MPC) schemes with high reputation.

Copyright © 2016 by ASME

Figures

Tables

Interactive Graphics

Video

Country-Specific Mortality and Growth Failure in Infancy and Yound Children and Association With Material Stature

Use interactive graphics and maps to view and sort country-specific infant and early dhildhood mortality and growth failure data and their association with maternal

NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In