A unified approach to PID Control has been found that enables a common and simplified method for setting PID tuning parameters. Key features can be used to eliminate the need for retuning to deal with different dynamics and objectives. Here are steps 1-5 in a methodology that integrates a unified tuning approach and key features to minimize implementation and maintenance efforts.
- Set the output limits to keep the manipulated setpoints in the desired operating range. For variable speed drives set the process PID low output limit so the speed cannot cause the discharge head to approach the static head in order to prevent excessive sensitivity to pressure and to prevent reverse flow. In general, set the anti-reset windup limit to match the output limit. If the output scale is engineering units, the output limits and anti-reset windup must be based on the output scale range and units.
- Choose the best structure for your application. Generally the best choice is a structure with PI on error and D on PV although a structure of PID on error may be useful for small setpoint changes to help get through valve deadband. For a single direction response (e.g. batch heating or neutralization), use a structure such as P on error and D on PV or PD on error so that there is no integral action. For a highly exothermic reaction, you might want this structure to help prevent a runaway from integral action.
- Set the signal filter noise just large enough to keep the controller output fluctuations from exceeding the resolution limit or deadband of the final control element so that the valve or variable speed drive does not respond to noise.
- For near-integrating, true integrating, and runway processes use the lambda integrating process tuning rules. To maximize the transfer of variability from the process variable to the manipulated variable, set the lambda (arrest time) equal to the maximum dead time and use the largest integrating process gain for all possible operating conditions in the tuning. Because of unknowns a more practical lambda is twice the maximum dead time. The rate time is set equal to secondary time constant or 1/2 the total loop dead time, whichever is largest. To maximize the absorption of variability (e.g. surge tank level) use the minimum required arrest time computed for all possible operating conditions. If you decrease the PID gain, proportionally increase the PID reset time to prevent slow rolling oscillations. For runaway processes, the PID gain must not be less than the inverse of the open loop gain. When changing from the Series Form to the ISA Standard Form in newer DCS, convert the tuning settings based on the differences in Form. In the ISA Standard Form the rate time should not be greater than ¼ the reset time. Be sure to realize and convert any difference in Form and tuning setting units between different PID vintages and suppliers. You must take into account the units and inverse relationship between gain and proportional band and reset time in seconds and minutes per repeat. Not addressing these differences can result in settings off by orders of magnitude.
- For self-regulating processes with the open loop time constant less than 4 times the dead time, use the lambda self-regulating tuning rules. To maximize the transfer of variability from the process variable to the manipulated variable set the lambda (closed loop time constant) equal to the maximum dead time and use the largest process gain and smallest time constant for all possible operating conditions in the tuning. Because of unknowns, a more practical lambda is twice the maximum dead time. Schedule tuning settings preferably with an adaptive tuner with the specified lambda to deal with changes in the dynamics.