4

As described in Table 4-6 on the previous pages, the PID block reads 13 user parameters and uses the rest of the 40 word RefArray for internal PID storage. Normally you would not need to change any of these values. If you are calling the PID block in Auto mode after a long delay, you may want to use SVC_REQ #16 to load the current PLC elapsed time clock into %Ref+23 to update the last PID solution time to avoid a step change on the integrator. If you have set the Override low bit of the Control Word (%Ref+14) to 1, the next four bits of the Control Word must be set to control the PID block input contacts (as described in Table 4-5 on the previous pages), and the Internal SP and PV must be set as you have taken control of the PID block away from the ladder logic.

PID Algorithm Selection (PIDISA or PIDIND) and Gains

The PID block can be programmed selecting either the Independent (PID_IND) term or standard ISA (PID_ISA) versions of the PID algorithm. The only difference in the algorithms is how the Integral and Derivative gains are defined. To understand the difference, you need to understand the following:

Both PID types calculate the Error term as SP – PV, which can be changed to Reverse Acting mode PV – SP if the Error Term (low bit 0 in the Config Word %Ref+12) is set to 1. Reverse Acting mode may be used if you want the CV output to move in the opposite direction from PV input changes (CV down for PV up) rather than the normal CV up for PV up.

Error = (SP – PV)

or (PV – SP) if low bit of Config Word set to 1

The Derivative is normally based on the change of the Error term since the last PID solution, which may cause a large change in the output if the SP value is changed. If this is not desired, the third bit of the Config Word can be set to 1 to calculate the Derivative based on the change of the PV. The dt (or Delta Time) is determined by subtracting the last PID solution clock time for this block from the current PLC elapsed time clock.

dt = Current PLC Elapsed Time clock – PLC Elapsed Time Clock at Last PID solution

Derivative = (Error – previous Error)/dt

or (PV – previous PV)/dt if 3rd bit of Config

 

Word set to 1

The Independent term PID (PID_IND) algorithm calculates the output as:

PID Output = Kp * Error + Ki * Error * dt + Kd * Derivative + CV Bias

The standard ISA (PID_ISA) algorithm has a different form:

PID Output = Kc * (Error + Error * dt/Ti + Td * Derivative) + CV Bias

where Kc is the controller gain, and Ti is the Integral time and Td is the Derivative time. The advantage of ISA is that adjusting the Kc changes the contribution for the integral and derivative terms as well as the proportional one, which may make loop tuning easier. If you have PID gains in terms or Ti and Td, use

Kp = Kc

Ki = Kc/Ti

and

Kd = Kc/Td

to convert them to use as PID User Parameter inputs.

The CV Bias term above is an additive term separate from the PID components. It may be required if you are using only Proportional Kp gain and you want the CV to be a non-zero value when the PV equals the SP and the Error is 0. In this case, set the CV Bias to the desired CV when the PV is at the SP. CV Bias can also be used for feed forward control where another PID loop or control algorithm is used to adjust the CV output of this PID loop.

GFK-0467K

Chapter 4 Series 90-30/20/Micro Instructions Set

4-179

Page 260
Image 260
GE 90-30/20/Micro manual PID Algorithm Selection Pidisa or Pidind and Gains