A Gain-Scheduling PI Control Based on Neural Networks

Table 3

Programmed plant tests for Case .

Schedules	Set-point changes	Disturbance changes	Time
Nominal; (0.000, 0.000)	Δ1y = (0.000, −0.150); Δ2y = (−0.150, 0.000)	Δ1d = (0.700, 0.000); Δ2d = (−0.700, 0.000); Δ3d = (0.000, 0.700); Δ4d = (0.000, −0.700)	@Time = (50, 100); @Time = (150, 200); @Time = (250, 300); @Time = (350, 400)
Maximum gain; (−0.303, 0.242)	Δ1y = (0.000, −0.150); Δ2y = (−0.150, 0.000)	Δ1d = (0.700, 0.000); Δ2d = (−0.700, 0.000); Δ3d = (0.000, 0.700); Δ4d = (0.000, −0.700)	@Time = (50, 100); @Time = (150, 200); @Time = (250, 300); @Time = (350, 400)
Sign change; (0.213, −0.245)	Δ1y = (0.000, −0.100); Δ2y = (−0.100, 0.000)	Δ1d = (0.700, 0.000); Δ2d = (−0.700, 0.000); Δ3d = (0.000, 0.700); Δ4d = (0.000, −0.300)	@Time = (50, 100); @Time = (150, 200); @Time = (250, 300); @Time = (350, 400)