Self-tuning control

The development vehicle used to create and test the rule base must be as flexible as possible, allowing easy alterations and expansion of the rule base with whatever displays can convey the most information. The delivery vehicle, however, should be virtually transparent to the user, conveying only as much information as needed to solve the problem at hand. Self-tuning controllers can perform their task without explicitlv informing users, but their output and status is available on demanci, and their operation may be easily limited or interrupted. 

Roliani, S. and Bourne, J., 1990b. Self-tuning control of crystal size distribution in a batch cooling crystallizer. Chemical Engineering Science, 45, 3457-3466. 

Krstic, M., A. Knipadanam, and C. Jacobson. 1999. Self-tuning control of a nonlinear model of combustion instabilities. IEEE Plans, on Control Systems Technology 7 424-36. 

Other self-tuning controllers have been designed which overcome these difficulties, e.g. the Generalised Minimum Variance self-tuning controller (GMV)m and the Generalised Predictive Controller (GPC)m. 

Figure 23. Block diagram of self-tuning control loop for downstream reactors...

F. Cameron and D.E. Seborg. A self-tuning controller with a PID structure. International Journal of Control, 38 401, 1983. 

When the process changes cannot be measured or predicted, the adaptive control strategy must be implemented in a feedback manner. Many such controllers are referred to as self-tuning controllers, or self-adaptive controllers, and a typical block diagram is shown in Figure 13. 

Self-Tuning Control of a Fixed Bed Reactor, by F. Buchholt, K. Clement, and S. Bar Jorgensen, Preprints of 5th IFAC Symposium on Identification and Parameter Estimation, Darmstadt (1979), p. 1213. 

To properly evaluate the goodness of the self-tuning controller, one would have to know the details of the internal controller software. However, suppliers of self-tuning controllers are not likely to give out such information because the software is the main factor that sets one controller apart from another. Thus, the details of the software will likely be treated as proprietary information. Therefore, the best method to evaluate a self-tuning controller is to install the controller on an extrusion line and closely monitor the actual performance. 

Soderstrom, T., Stoica, P., 1989, System Identification. Prentice Hall, Cambridge. Westerlund, T., Toivonen, H., Nyman, K.-E., 1980, Stochastic Modelling and Self-Tuning Control of Continuous Cement Raw Material Mixing System. Modelling, Identification and Control 1, 17-37. 

At this writing, programmed adaptive systems have been used for certain critical applications such as temperature control in once-through boilers and heat exchangers. But there apparently is no published report on a self-tuned controller operating successfully on a critical loop in a process plant. (They have been used in aircraft controls.) From the foregoing discussion it should be evident that, no matter how skillfully mechanized, a self-tuning controller is by no means a panacea. 

Like the self-tuning controller, the self-optimizing controller requires no prior knowledge of plant conditions, but instead, conducts its own search. Its goal is to keep the manipulated variable at the point where process steady-state gain dcfdni satisfies the specification. But before this can be done, the controller must first test the process for its gain at each point in the search. The test may be conducted continuously or intermittently. 

Roifel, B., Vermeer, P.J. and Chin, P.A. (1989) Simulation and bnplementatian of Self-tuning Controllers, Prentice Flail. 

In self-tuning control, the parameters in the process model are updated as new data are acquired (using on-line estimation methods), and the control calculations are based on the updated model. For example, the controller settings could be expressed as a function of the model parameters and the estimates of these parameters updated on-line as process input/output data are received. Self-tuning controllers generally are implemented as shown in Fig. 16.26 (Astrom and Wittenmark, 1995). 

Figure 16.26 A block diagram for self-tuning control.

Two advantages of the self-tuning control approach is that the model in Fig. 16.26 is not restricted to low-... 

Vol. 246 RW. Otter, Dynamic Feature Space Modelling, Filtering and Self-Tuning Control of Stochastic Systems. XIV, 177 pages. 1985. 

---