Internal Model Control (IMC) - Neural Network (NN) Gain Scheduling Untuk Pengendalian Kolom Distilasi

Totok R. Biyanto
Journal article Jurnal Teknik Elektro Universitas Kristen Petra • September 2005 Indonesia

Abstrak

This research is develop the alternative control algorithm using Internal Model Control - Neural Network Gain Scheduling (IMC-NNGS) to control mole fraction of methanol-water distillation column. Distillation column with L-V control strategy has pairing Xd-L and Xb-Qr. IMC performances depend on only ? tuning value or filter time constant. With ? tuning value manipulating IMC could be nonlinear control, where ? tuning value is outputs of NN that had been trained by using error variable, process variable, manipulated variable, and set point variable from plant. Gain scheduling using NN could be increase control system performance and product quality. The best IAE changing value shown at mole fraction feed increase. There are IAE equal with 0,234799 for IMC and IAE equal with 0, 00042 for IMC-NNGS. In other word IMCGS has IAE 559 times better than IMC. Beside that IMC-NNGS has faster response, offset free and robust to overcome set-point and disturbance changes.

Metrics

  • 208 kali dilihat
  • 122 kali diunduh

Jurnal

Jurnal Teknik Elektro Universitas Kristen Petra

Jurnal Teknik Elektro Universitas Kristen Petra is published biannually, in May and September, by... tampilkan semua