Robust control system for wastewater treatment plants
Development the robust two cascaded control systems of dissolved oxygen, consists of 12 feed-back valve controllers, 1 feed-back and 1 feed-forward blower controllers and decoupling matrix
Poor control of dissolved oxygen (DO) is caused by the inability to control in real time airflow supplied to individual grids. Elevated DO concentrations cause waste of aeration energy. Improvements of DO control in each grid and ability to optimize DO set points for each grid allow to reduct energy used for aeration. There are two cascaded DO control systems were designed: valve control system and blower control system.
Aeration system model (Plant model) was designed which included (see pic Plant):
-
4 basins - nonlinear models of the aeration system elements (i.e. pipes, diffusers, and valves);
-
2 controlled blowers;
-
12 motorized control valves;
-
12 manually controlled valves;
-
12 flowmeters with measurement noise;
-
pipeline grid;
Each element of the circuit has dynamic and static characteristics and was adjusted individually. Plant model was augmented with the measurement noise.
Highlights of Designed Robust Control System
-
feed-back control of airflow to each grid (12 PI valves controllers) and pressure control of the blower (1 PI blower controller);
-
decoupling matrix for feedback controllers calculated using H-infinity optimization method;
-
feed-forward blower control;
-
most-open-valve control;
-
control of noisy airflow;
Feed-back control of airflow to each grid and pressure control of the blower
The control system consists of decoupling matrix and 13 PI regulators. All coefficients for matrix, 12 PI airflow controllers and 1 blower controller were calculated to satisfy the following requirements:
-
The stability of the process;
-
Fast time response;
-
Minimum error;
-
Minimum overshoot;
-
Minimum oscillation;
The process should be decoupled, i.e. the behavior in one loop should not influence the behavior of other loops, and the variables do not interact with each other. The methods of decoupling are based on the model of the process and optimization algorithms (non-smooth H-infinity) of the controllers. The decoupling of multivariable PID Control is performed by a decoupling matrix, that is used to decouple multi-input – multi-output system. This approach provides the robustness of the control system.
Most-open-valve control
Most opened valve (MOV) logic is a technique developed to eliminate the excess power demand caused by excessive blower discharge pressure.
A goal of an MOV based blower control is to make a blower discharge pressure to be equal to the minimum pressure necessary for delivery of the required air flow to each aeration grid.
The method is implemented by maintaining, at least, one of the aeration flow control valves at maximum opening all the time. This valve is referred to as “the most open valve”. By maximizing a valve opening, the pressure drop across all valves and the resulting main header pressure are minimized.
Feed-Forward control
The performance of the system could be enhanced by adding a feedforward control to the feedback control and the MOV strategy. The feedforward signal can be added to the control algorithm to advance the demand for air from a blower based on the estimated sum of airflow set points for all grids. To implement a feed-forward control, the inverse model of blower must be developed. The inverse model of blower is used in control subsystem and advance the speed of blower.
Stages of project
-
Parameter estimation of the plant model
-
Validating the simulation results by open loop tests in plant model.
Airflow error in the static mode does not exceed 2.6% for a given range of the data.
-
Developing control system;
-
Controllers tuning;
-
Simulation;
-
Validating the simulation results;
Achievements
-
Oscillations of aeration valves and blowers can be reduced by applying decoupling control method to valves control.
-
The most opened valve (MOV) method is a better alternative to blower control then a constant pressure control. Robustness of the MOV method can be enhanced by using decoupling of the blower control loop with valve control loops.
-
Feed-forward control of the blowers reduces the time necessary to reach new set points, i.e. it improves the agility of the control system without jeopardizing accuracy.
Designed methods allow to optimize DO set points for each grid and reduct energy used for aeration.
Flow in Grids #3, #5 Flow in Grids #6, #7 Total Flow in Basin
Flow in Grids #3, #5 Flow in Grids #6, #7 Total Flow in Basin
Blower pressure [psi] Most open valve control (ref=0.7)
Flow in Grids #3, #5 Flow in Grids #6, #7 Total Flow in Basin
