Open-Loop Testing of Process Response

This chapter describes how to characterize the response of a process variable with a controller in manual output mode, that is, with no feedback from changes in the process variable. Then tuning rules based on mathematical models of the response are used to estimate tuning constants for the controller. 

---

Ziegler and Nichols presented an open-loop test method in their classic paper on tuning control loops. The method consists of starting with the control loop at steady state, putting the controller in manual output so that there is no feedback response from any change in the process variable, and introducing a step change in the controller output (valve position). Then the results are used to characterize the process response by an apparent... 

To make use of empirical tuning relations, one approach is to obtain the so-called process reaction curve. We disable the controller and introduce a step change to the actuator. We then measure the open-loop step response. This practice can simply be called an open-loop step test. Although we disconnect the controller in the schematic diagram (Fig. 6.1), we usually only need to turn the controller to the manual mode in reality. As shown in the block diagram, what we measure is a lumped response, representing the dynamics of the blocks Ga,... 

One reason why this approximation works is that process unit operations are generally open-loop stable, and many are multi-capacity in nature. Reminder Underdamped response of the system is due to the controller, which is taken out in the open-loop step test. 

For slow-response loops (e.g., certain temperatnre and composition control loops), field tuning can be a time-consuming procednre that leads to less than satisfactory resnlts. Step test results can be used to generate FOPDT models, and tnning parameters can be calcnlated from a variety of techniques. This approach suffers from the fact that it takes approximately the open-loop response time of the process to implement a step test, and during that time, measured and unmeasured disturbances can affect the process, thus corrupting the results from the step test. In addition, it is unlikely that the selected turfing approach will result in the proper balance between reliability and... 

Understand some procedures for testing the open-loop response of processes for characterization and mathematical modeling... 

An integrating process response, for example, the level in a reflux drum or distillation column base, will eventually overflow or run empty in an open-loop step test without further manual intervention. One example of an open-loop step test... 

An open-loop step test on a level controller would start with the liquid level and controller output at steady-state values with a constant flow rate of liquid into the vessel. Then, increase the controller output by a step change, and observe the rate of change in the liquid level. For example, with the initial liquid level at 50% and the controller output at 70%, put the controller in manual output mode and increase the controller output to 80%. Observe the rate of change of the process variable, say, a decline of 1.0%/min. This can be used as the maximum process response rate, MPRR, for calculating a controller gain, K,., by the Z-N equation in Table 10.2. =... 

The responses for yi, y2, and so on are then observed. All loops are kept open during this test that is, no feedback controllers are operational. Then step changes can be made in the other inputs, one at a time, and open-loop response data can be obtained for all the controlled variables. The steady-state gain depends only on the final value of each y, from which the change in y, Ay, can be calculated. Thus, the individual process gains are given by the formula (see Chapter 7) ... 

To test the hypothesis that dead time has no effect on the open-loop process attenuation for a capacity-dominated process, again perform a frequency response test using the sine wave input. Run the 40 min (frequency 0.025 min ), 25°C ainilitude disturbance through the process of Figure W3.5 with the tank level set to 50 per cent. Has the attenuation changed from the value you calculated earlier ... 

---