Signal Selectors

Instead of the error (PV- SP) being calculated in the controller it is calculated in each of die two bias algorithms. The two errors are compared in the signal selector and the selected 

However, we also have to consider the tuning of the PID controller. If the dynamics of each constraint are different then the controller will require different tuning depending on which constraint is selected. A better approach is to move the controller gain from the PID to the biases. Process dynamics are obtained as usual by steptesting the feed flow SP. The dynamics of the fuel valve position are used to develop a set of tuning constants and the 

In this configuration each constraint has its own PID controller — a pressure controller on burner inlet and a valve position controller on the fuel valve. In this example both constraint controllers are configured as reverse-acting. The DCS must support certain features. The signal selector must include anti-reset windup to prevent the output of the unselected controller from saturating. Secondly the PID controllers must be of the incremental type, otherwise bumpless initialisation is difficult to achieve. These features are present in most DCS, but should be checked. 

Measurement noise can create a problem with this configuration. With the full position form of the PID controller any noise superimposed on the output will have little impact on signal selection. But with the incremental form signal selection is based on the much smaller change in MV. The change in controller output caused by noise can readily exceed the change made to correct a deviation from SP. 

---

The idea for safe control is to have the air lead the fuel on increases in demand and fuel lead the air on decreases in demand. On load increases, the air is increased ahead of the fuel. On load decreases, the fuel is decreased ahead of the air. This is accomplished with high- and low-signal selectors. 

The reader can easily determine how a low-signal selector works for the fuel flow controller. It would compare the signals from the steam pressure and the air flow. A flue gas oxygen analyzer should be installed to continuously monitor or even trim the air flow. 

The functions h), i) and j) are used when a two-shaft drive system needs to be simulated. When applied they usually require a special signal selection block to be incorporated just before the fuel valve or governor. The purpose of this signal selector is to automatically choose the lowest value or its two input signals, so that the least fuel is passed to the combustion system. This contributes to... 

The control signals from the two pressure controllers go to a high-signal selector switch. When the pressure set by the wide-range system reaches a certain point, then, the narrow-range system takes over control. Control response charts are in Fig. 11.38. The situation is similar to that described for the chlorine system and shown in Fig. 11.20. 

Signal selectors are used in the most basic form of multivariable control, i.e. multi-input single-output (MISO) applications. The fired heater described also has a maximum limit on burner pressure. This is also approached by increasing feed rate. In order to operate at maximum feed rate the controller must be able to continuously identify which is the more limiting constraint. Figure 8.3 illustrates one possible configuration. 

This has both a flow controller and a pressure controller. The flow controller permits the addition of the compensation and feedforward schemes. The pressure controller provides burner protection. In this case their outputs are compared in a low signal selector. This provides protection against too high a pressure. If the pressure exceeds the maximum, as entered as SP in the pressure controller, the controller output will reduce to close the valve... 

Conversely, on decreasing demand, the controller output is passed by the low signal selector to the SP of fuel flow controller. The fuel flow measurement provides an input to the high signal selector. If the fuel flow controller does not respond then this will override the output sent to the air flow controller. Thus air is not permitted to reduce until the fuel flow reduces. 

We use the term override for the most part to refer to the use of two or more controllers connected to a control valve through high or low signal selectors. Logic is built in to enable one controller to override (i.e., take over fi om) the other controller or controllers ... 

These devices select either the higher or the lower of two input signals. As many input signals as desired may be accommodated by arranging selectors in series. Multiple-input low-signal selectors are now available from vendors. In practice one or more selectors are inserted between the output of a normal controller and its final control element, usually a valve. The outputs of the override controllers are also connected to these selectors. As process constraints are approached, one of the override controllers will take over or override the normal controller and drive the final control element in the proper direction— either to force the process away from a constraint, or to hold it a safe distance from a constraint. 

Since the gain 4 rday is to put out 15 psig at the 25 percent level, and its output goes to a low signal selector, it clearly will exercise no control action at a level of 25 percent. On the other hand, if level is dropping, the output of the gain 4 relay will drop below that of other controllers at some value of level above zero. We are never sure, therefore, at what point the low-level override will take over, but we know positively that it will be between the zero and 25 percent levels. 

---