Oxygen control in a wastewater treatment plant using adaptive predictive controllers

1
Oxygen control in a wastewater treatment plant
using adaptive predictive controllers

• Gregor Kandare, Jozef Stefan Institute, Slovenia

2
Adaptive predictive control
3
Process description

• 6 pools with butterfly valves and 2 dissolved
  oxygen sensors each.
• 4 blowers with diffuser for pressure control.

4
Process description
5
Process description
6
Control issues

• Biological dynamics of the process
• Aleatory operation context
• Lack of process information
• Interactive nature of the process

7
Control objectives

• Maintain the dissolved oxygen signal at its
  setpoint by manipulationg the aeration with
  butterfly valves.
• Maintain the air pressure in the main air conduct
  at a setpoint that minimises energy consumption
  and assures good oxygen control.

8
Control strategy

• 6 controllers one for each pool

9
PID control

• Oxygen and valve opening

10
PID control

• Air pressure and airflow

11
Adaptive predictive control

• Oxygen and valve opening

12
Adaptive predictive control

• Air pressure and airflow

13
Oxygen control evaluation
Reactor PID AP Factor
1 0.3974 0.1863 2.13
2 0.6632 0.2528 2.62
5 0.5221 0.1432 3.65
6 0.9138 0.151 6.05
14
Energy optimisation

• Objectives
• Maintain air pressure at a minimal level which
  still permits satisfactory oxygen control.
• Maintain dissolved oxygen setpoints at a minimal
  level that ensures required effluent water
  quality.

15
Pressure optimisation
16
Pressure optimisation
17
Pressure optimisation
18
Pressure optimisation
19
Change PID adaptive predictive
20
Energy consumption estimation
21
Consumption reduction
Pool Average power consumed with PID control Average power consumed with AP control with pressure optimisation Average energy savings
Pool 1 51.36 42.25 17.74
Pool 2 61.55 34.97 43.18
Pool 5 67.63 52.28 22.70
Pool 6 54.47 41.98 22.93
Total 235.01 171.489 27.03
22
Conclusions

• The adaptive predictive controllers stabilise the
  process and maintain oxygens at therir setpoints
• More stable oxygen control and pressure setpoint
  optimisation decrease energy consumption by 15-23