Checklist for Best PID Performance
There is an incredible offering of PID features and options enabling maximum disturbance rejection and setpoint response but also the coordination of loops and unit operation optimization. While the full aspects of the PID capability are book worthy, this checklist and an excerpt from an ISA Automation Week 2012 paper can get you started on the right path.
If you don’t get the valve action and control action right, nothing else matters. The controller output will ramp off to an output limit. The valve action (inc-open and inc-close) can be set in many different places, such as the PID block, analog output (AO) block, splitter block, signal characterizer block, current to pneumatic (I/P) transducer, or the positioner. Make sure the valve signal is not reversed in more than one location for an inc-close (fail open) valve. Once the valve action is set properly, the control action is set to be the opposite of the process action. The control action is reverse and direct if an increase in the PID output causes the PID process variable (PV) to increase or decrease, respectively. Verify with process engineer the valve action, process action, and resulting control action required.
The following checklist is not intended to cover all the configuration requirements but some of the major application details to be addressed for PID controllers. Most of the power of the PID is untapped. For more information see appendices for InTech 2012 Jan-Feb article "PID tuning rules" and the ISA book Good Tuning: A Pocket Guide 3rd Ed
- Does the measurement scale cover entire operating range including abnormal conditions?
- Is the valve action correct (inc-open for fail close and inc-close for fail open)?
- Is the Control action correct (direct for reverse process and reverse for direct process if the valve action is set)?
- Is the best f"Form" selected (ISA standard)?
- Is the "obey setpoint limits in cascade and remote cascade mode" option selected?
- Are the "back calculate" signals correctly connected between blocks for bumpless transfer?
- Is the "PV for back calculate" selected in secondary loop PID?
- Is the best "Structure" selected (PI action on error, D action on PV for most loops)?
- Is the "setpoint track PV in manual" option selected to provide a faster initial setpoint response unless setpoint must be saved in PID?
- Are setpoint limits set to match process, equipment, and valve constraints?
- Are output limits set to match process, equipment, and valve constraints?
- Are anti-reset windup (ARW) limits set to match output limits?
- Is the module scan rate (PID execution time) less than 10% of minimum reset time?
- Is the signal filter time less than 10% of minimum reset time?
- Is the PID tuned with an auto tuner or adaptive tuner?
- Is rate time less than ½ deadtime (rate typically zero except for temperature loops)
- Is external-reset feedback (dynamic reset limit) enabled for cascade control, analog output (AO) setpoint rate limits, and slow control valves or variable speed drives?
- Are AO setpoint rate limits set for blending, valve position control, and surge valves?
- Is integral deadband > limit cycle PV amplitude from deadband and resolution?
- Can an enhanced PID be used for loops with wireless instruments or analyzers?
The setting of all options and parameters must be verified as applicable. Simulations representative of the dynamic behavior of the process and the field automation system along with the actual configuration to form a virtual plant is advisable for testing and confirmation plus training and opening the door to process control improvement. See ISA Interchange post "What are the types of dynamic simulators and the pro and cons" for details. The following excerpt on PID Structure from the ISA Automation Week 2012Paper "Effective Use of Key PID Features" may help with checklist item 8.
Table 1 Major PID Structure Choices
- PID action on error (ß = 1 γ = 1)
- PI action on error, D action on PV (ß = 1 γ = 0)
- I action on error, PD action on PV (ß = 0 γ = 0)
- PD action on error, no I action (ß = 1 γ = 1)
- P action on error, D action on PV, no I action (ß = 1 γ = 0)
- ID action on error, no P action (γ = 1)
- I action on error, D action on PV, no P action (γ = 0)
- Two degrees of freedom controller (ß γ adjustable 0 to 1)
The structure choices use b and g set point weighting factors. A controller that has both factors adjustable is called a "two degree freedom controller." Other structures have the b and g factors set equal to 0 or 1. The user can also omit a mode entirely to get P-only, I-only, ID, and PD control with various assigned factors. PI control is achieved by simply setting the derivative (rate) time to zero. In general, the user must not set the controller gain equal to zero in an attempt to get I-only or ID control and not set the integral (reset) time to zero in an attempt to get P-only or PD control. Note that the use of P-only or PD control requires additional choices of how to set the bias and its ramp time. Table 1 lists 8 choices offered by a major DCS supplier.
Structure 1 (PID action on error) provides the fastest approach to a new setpoint by virtue of a spike in the controller output from the setpoint change. For small setpoint changes and low controller gains, the spike can help get through significant valve backlash and stick-slip to get the valve moving. However, this spike is often seen as disruptive by operators when they make setpoint changes. If the burst of flow through the control valve does not affect other users of the process or utility fluid, the spike is more of a psychological than a process concern. The spike can be made to be more of a bump by decreasing the gamma factor (γ). At any rate, the reduction in rise time (time to reach setpoint whether PV is increasing or decreasing) is marginal for good control valves or higher controller gains and larger setpoint changes.
The PV overshoot of SP that results from load disturbance tuning can largely be eliminated by adding a setpoint filter equal to the reset time as discussed in the Setpoint Filter section with a marginal increase in rise time. The use of setpoint feedforward can slightly reduce the overshoot and rise time. The advantage of setpoint feedforward is more significant when the controller gain is relatively low. The biggest improvement is offered by a relatively simple Bang-Bang logic for smart sequenced positioning of the final control element. The output is first positioned at the appropriate output limit and held until the rate of change of the PV multiplied by the deadtime is near the setpoint. At this point the output is positioned at a final resting value and held for a loop deadtime. The final resting value is captured from previous batches or startups of continuous processes. For more info see the Control May 2006 article "Full Throttle Batch and Startup Response".
Structure 2 (PI action on error, D action on PV) eliminates the spike from derivative action for a setpoint change by setting gamma to zero (γ=0). The increase in rise time going from structure 1 to 2 is negligible for the more important loops, such as column and vessel temperature where derivative action is used, because the step in the output from the proportional mode on a setpoint change is large because of the high controller gain.
Structure 2 is the structure most often used. Increases in process gain or deadtime will increase the overshoot unless the controller gain is decreased accordingly. If the elimination of setpoint overshoot is much more important than rise time, then structure 3 may be best.
Structure 3 (I action on error, PD action on PV) eliminates overshoot but with quite a sacrifice in speed of approach to the setpoint. The rise time may be 5 times larger than the rise time for structure 1 for the same tuning settings and process dynamics.
Structure 4 (PD action on error, no I action) is used on processes adversely affected by integral action. The temperature control of severely exothermic polymerization reactors use structure 4 because integrating action in the controller increases the risk of a runaway. To be safe, the reset time should be increased by a factor of 10 for these positive feedback processes, but users may not be aware of this requirement leading to overshoot that can trigger a runaway. The bias here is set equal to the normal PD controller output when the PV is at setpoint.
Structure 4 is used for total dissolved solids (TDS) control of boiler drums and vessel level control to eliminate the slow reset cycles from too small of a reset time and to prevent slow rolling oscillations. For drums, the boiler blow down may be discontinuous making control of the TDS integrating response more difficult with integral action. For reactors, the increase in level from proportional-only control for a decrease in reactant feed flow provides a more constant residence time. However, the setting of the bias in these applications is confusing to the user.
Structure 4 is used on batch processes that respond in only one direction. For example, in bringing a batch pH up to a setpoint by the addition of a base where the base is not consumed in a reaction, the batch will only respond in the direction of increasing pH. The pH will overshoot setpoint if integral action is used. If split ranging is added with an acid reagent, there will be some wasted reagent due to cross neutralization of reagents and limit cycling across the split range point from stiction that is greatest near the closed position. For structure 4 and a single reagent, the bias is set for zero reagent addition when the PV is at setpoint.
Structure 5 (P action on error, D action on PV, no I action) is used for the same reasons as structure 4. As with structure 2, the spike from rate action for setpoint changes is eliminated.
Structure 6 (ID action on error, no P action) is used for valve position controllers (VPC) to eliminate the interaction with process controller whose valve position is being optimized. The VPC could be optimizing the coarse adjustment from a large control valve in parallel with a small control valve manipulated as fine adjustment by the process controller. The VPC could be optimizing utility supply pressure or temperature to minimize energy use or optimizing feed rate to maximize production rate. Normally, the rate time is zero. The tuning of this integral-only controller is problematic. Tuning rules often cited are the integral time should be larger than 10 times the product of the gain and reset time of the process controller and 10 times the residence time of the process to eliminate interaction. However, the VPC response for this tuning is too slow to prevent the process controller from getting into trouble for large disturbances. Feedforward action can be added to help, but a more flexible and easier to tune solution is to use external-reset feedback, AO block directional rate limits, and a PID developed for wireless. This solution is described in the Setpoint Rate Limits section.
Structure 7 (I action on error, D action on PV, no P action) is used for the same reasons as structure 6. As with structure 2, the spike from rate action for setpoint changes is eliminated.
Structure 8 (two degrees of freedom) is used to provide a balance between a fast rise time and minimal overshoot. A setpoint filter equal to the reset time is a simpler solution.