Optimum Behavior of a Third-Order Process under Feedback Control

Abstract

sorption, this VCM level during drying may become much lower as the degree of steam superheat during regeneration increases. ConclusionsBased on this work, the following processing scheme is recommended. The carbon should have some residual

Optimum controller settings for processes with dead time: effects of type and location of disturbance

“…For equal time constants and a small delay (L/Tx = 0.1), the value of TRopt for step load changes is 1.7P" or 2.2PU, depending on the location of the disturbance, compared to 0.8PU when there is only one time constant. These high values of TR/PU are similar to those recommended for processes with two large and one small time constant (Hazebrock and van der Waerden, 1950, Harriott, 1964, Weber and Bhalodia, 1979), since the frequency response near the critical frequency is quite similar when the delay or the third time constant is much smaller than the other lags. Using half the maximum gain for these systems does not give much phase Table I.…”

Controller tuning of a third-order process under proportional-integral control

“…Furthermore, in the format of Figure 4, one is not constrained to the standard CCM parameters. As pointed out in the previous paper of Weber and Bhalodia (1979), ß values between 0.1 and 0.6 are often better than the value of 1.2. Having chosen a value for ß, one can then use Figure 4 to determine the stable region boundary and, hence, an appropriate value of cx through the use of a gain margin such as 1.7.…”

“…Feasible Regions and Unstable Regions of the -Plane In the earlier paper of Weber and Bhalodia (1979), a third-order process was characterized in terms of parameters that are related to the CCM tuning method and to a Bode diagram. The maximum gain Kmax is the reciprocal of the normalized amplitude ratio of a Bode diagram at the frequency where the phase lag is 180°.…”

“…The earlier paper, Weber and Bhalodia (1979), was confined to overdampled cases, whereas the present paper includes the situation in which two of the elements combine to exhibit underdamped behavior.…”

Feedback Control of a Two-Input, Two-Output Interacting Process

“…The integral square of the error (ISE) was chosen as the performance index because it is easily calculated. Furthermore, this function has been used in optimization studies of single loops by Jackson (1958) and Weber and Bhalodia (1979). Therefore, the behavior of the two-loop model here can be related to comparable single loops.…”