Skip to main content

Table 3 Overview of the different control schemes. The set-points (SP), upper and lower boundaries (UB and LB, respectively), the parameters in the proportional (\(Kp\)), integral (\(Ki\)) and differential (\(Kd\)) control algorithm (PID) and the process variable (PV) conditioning according to the set-point are shown for each fermentation

From: Promoting the co-utilisation of glucose and xylose in lignocellulosic ethanol fermentations using a data-driven feed-back controller

Fermentation (controller) Set-point and boundaries (g/L) Regulatory layer
SP UB LB If PV > SP If PV = SP If PV < SP
PV conditioning PID parameters PV conditioning PID parameters PV conditioning PID parameters
\(Kp\) \(Ki\) (h−1) \(Kd\) (h) \(Kp\) \(Ki\) (h−1) \(Kd\) (h) \(Kp\) \(Ki\) (h−1) \(Kd\) (h)
1 (1) 10 12 8 PV = PV + SP 1.00 0.50 0.05 PV = PV 1.00 0.50 0.05 PV = PV − SP 1.00 0.50 0.05
2 (2.1) 4 10 2 PV = PV + SP 0.31 0.56 0 PV = PV 0.31 0.56 0 PV = PV − SP 0.31 0.11 0
2 (2.2) 4 10 2 PV = PV + SP 0.62 0.56 0 PV = PV 0.62 0.56 0 PV = PV − SP 0.62 0.11 0
2 (2.3) 4 10 2 PV = PV + SP 0.62 0.56 0 PV = PV 0.62 0.56 0 PV = PV − SP 0.80 0.11 0
3 (3) 4 10 2 PV = PV + SP 0.72 0.56 0 PV = PV 0.72 0.56 0 PV = PV—SP 0.72 0.11 0
4 (4)a 4b 10 2 PV = PV + SP 0.72 0.56 0 PV = PV 0.72 0.56 0 PV = PV − SP 0.72 0.11 0
  1. aThe regulatory layer in Fermentation 4, included a rate controller to prevent abrupt responses caused by noisy measurements
  2. bThe PLS model predicted the glucose concentration with a bias of 5 g/L, resulting in a real SP of 9 g/L and not 4 g/L