Controllers need tuning, i.e. the setting of the controller parameters to sensible values that makes the controller response effective. Take the example of finding the optimum values of PID controller, i.e. K, Ti and Td. Good control performance can be achieved with a proper choice of the controller parameters, but poor performance and even instability can result from a poor choice of values. Controller tuning methods based on the dynamic performance have been used for many decades.
Generally speaking, we have the following qualitative considerations to make concerning the settings of the controller parameters:
6
6 If the gain K is too small, then the control action will become too slow and the
controller is not acting as quickly as we wish;
6
6 If K is too large, then the controller is too sensitive to disturbances and may overreact.
In some cases the controller will cause the system to be unstable, that is, the process variable will oscillate with increasing amplitudes. This is illustrated in Figure 5.13 (left);
6
6 If the integral time Ti is made short, then the regulator may overreact. It is too “ambitious”
to make the error equal to zero, and the control system may be unstable. On the other hand, if Ti is too large, then the error may approach zero too slowly. See Figure 5.13 (right);
6
6 Ti should be longer than the typical time constant of the individual process unit, so that the error e gradually approaches zero.
Figure 5.13 Left: a P-controller with various K values. Large K values lead to instability (oscillations).
0 2 4 6 8 10 12 14 16 18 20
Time
1.5
1.0
0.5
0
Sensor signal
1.4 1.2 1.0 0.8 0.6 0.4 0.2
00 2 4 6 8 10 12 14 16 18 20
Time
Manual tuning of PID control is surprisingly common. The tuning procedure relies on the specific properties of the three terms described above. We restrict our discussion to P and PI control because D control is a little specialised.
The systematic tuning goes as follows:
1. Determine whether the priority for the closed-loop system is reference tracking (i.e. a setpoint that is changing) or load disturbance (when the setpoint is constant);
2. Determine whether steady-state accuracy is essential to the control system performance;
3. P control tuning: introduce proportional action by increasing the value of the proportional gain K until the speed of the response is acceptable. See Figure 5.13;
4. I control tuning: If steady-state accuracy is considered important, then introduce integral action into the controller by slowly decreasing the integral time from Ti = ∞ (when no integral control takes place) to smaller values. The Ti should be decreased so that an acceptable settle time is achieved;
5. Balancing the controller terms: increasing K may increase the overshoot. To compensate for this, K has to be decreased. A little fine tuning between K and Ti will be necessary to achieve acceptable time responses.
The manual procedure is a trial and error process. It is hoped that steps (4) and (5) will eventually converge to an acceptable solution. Basically, while K changes the speed of response, changing Ti alters the settling time, with a tendency to introduce overshoot.
Excessive overshoot usually needs to be avoided, but the settle time must be reasonably short so that the desired output level is reached.
Naturally, the manual tuning method is somewhat laborious, time-consuming and an ineffective use of resources. We can do better by using some system knowledge.
There are smarter methods for tuning controllers, most notably the two Ziegler–
Nichols open and closed loop methods.
Additionally, there are a number of autotuning methods available that will tune the controller based on analysis of the reaction response curve automatically and while the control loop is in
190 Smart Water UtilitieS: Complexity made simple operation. Most commercial control systems today are supplied with autotuning features.
auToTuning
Today most commercial computer control systems are delivered with PID controllers including automatic tuning, or autotuning. The identification of the process model and the tuning of the controller are made automatically.
An identification experiment is automatically performed after a specific request by the operator and the values of the PID parameters are updated at the end of it. For this reason the overall procedure is also called one-shot automatic tuning or tuning-on-demand.
The design of an automatic tuning procedure involves many critical issues. The choice of the identification procedure is usually based on an open-loop step response or on a relay feedback experiment. In the latter case the controller is simply (automatically) replaced by an on-off controller that will cause the process output to oscillate slightly. The process model parameters are found from this identification experiment.
Finally a tuning rule is executed. At the end the tuned PID controller will automatically take over the control.
The autotuning methods have been developed over the last two decades and show very good results in the process industry. 6
There are literally hundreds of textbooks on control, so we select a couple of them from this large collection. Automatic control is sometimes called the hidden technology. It appears everywhere around us and we do not even think about it. It is used in the cruise controller in the car, in the temperature control of a room, as well as in the hundreds of thousands of controllers in water and wastewater operations that make sure that flow rates, pressures, levels and temperatures are right. Our own body contains a multitude of control systems, for example, to keep our body temperature at 37°C, despite variations in the surrounding environment. Even if control is applied in completely different areas, there is a common theory that is independent of the applications.
A superior textbook has been published recently:
Åström, K. J. and Murray, R. M. (2014) Feedback Systems: An Introduction for Scientists and Engineers.
The authors are world leaders in the development of control theory and engineering. The complete book is available for free on the web, called FBSwiki (http://www.cds.caltech.edu/~murray/amwiki/
index.php/Main_Page). On this site you will find the complete text of the book as well as additional examples, exercises, and frequently asked questions.
Copyright in this book is held by Princeton University Press, who has kindly agreed to keep the book available on the web. Having studied this excellent book you are a long way to understanding control and its applications.
There is a vast literature on control of water and wastewater treatment systems. Olsson, G.
(2012) ICA and me – a subjective review. Water Research, 46 (6), 1585–1624, available online at doi:10.1016/j.watres.2011.12.054. is a subjective review of the development of ICA in wastewater treatment during four decades.
The state-of-the-art-report, Olsson, G., Nielsen, M., Yuan, Z., Lynggaard-Jensen, A. and Steyer, J.P.
(2005) Instrumentation, Control and Automation in Wastewater Systems. IWA Publishing, London.
Presents quite a comprehensive view of ICA in wastewater treatment systems, leading up to 2005, and can still be said to contain all essential information for smart water systems.
The textbook Olsson, G. and Newell, B. (1999) Wastewater Treatment Systems. Modelling, Diagnosis and Control. IWA Publishing, London, contains a more comprehensive account of model building, measurements, monitoring, detection, diagnosis and control of wastewater treatment systems.
SCADA systems are available from several vendors and the best way to get updated information about them is to consult various web pages, for example ABB, Siemens, Schneider Electric, Rockwell Automation.
There are also interesting lectures on SCADA systems and water and wastewater automation on Youtube. Search for “SCADA” or “water automation”
on youtube.com
To learn more about strategic thinking, consult articles in Harvard Business Review or some of the leading strategic thinkers, i.e. Henry Mintzberg, Michael Porter, Robert Kaplan or Bruce Henderson.
See for example Strategy Safari: The Complete Guide through the Wilds of Strategic Management by Mintzberg, H., Ahlstrand, B. and Lampel, J., Free Press, 2006, Competitive Strategy by Porter, M., Free Press, 1998 and Balanced Scorecard:
Translating Strategy into Action by Kaplan, R. and Norton, D. P.Harvard Business Review Press, 1996 Simon Sinek has written a book on the big why called Start with Why: How great leaders inspire everyone to take action, Penguin, 2011.