Emerson Exchange 365

by Dan Hebert

Poorly behaving process control loops often indicate the need for a tune-up of the controller’s proportional, integral and derivative (PID) algorithm terms. But what do you do when no amount of tuning will bring the intended process value consistently back in line with its desired setpoint?

On Thursday at Emerson Exchange, Mark Coughran, senior process control consultant with Emerson Process Management, illustrated how to look beyond the PID parameters in order to fix a variety of troublesome loops. Although many loops can be made to perform satisfactorily by correct tuning and setting of PID parameters, some cannot, Coughran explained. Common problems with these untunable loops include incorrect process variable readings, improper interaction between the output and the controlling device, and communication issues.

“PID loop troubleshooting starts with adjustment of PID parameters to achieve the desired response,” said Coughran. But when tuning efforts aren’t sufficient, the tuning exercise itself can reveal information about the loop’s behavior and assist in locating the root cause of the problem.

One of the examples that Coughran presented involved a process plant that had a problem tuning a pressure control loop in a NOX unit. The loop used the pressure in a heater as the process value, and then sent an output to adjust a damper valve based on the difference between the setpoint and the actual pressure.

The main problem with the loop was that the process value could not reach the setpoint for any value of the output. “Our consultants worked with the plant operator to put the controller in manual mode and make steps on the output, but still saw no recognizable pattern of the process value responding to the output,” explained Coughran. 

After examining the components outside the controller, the problem was found to be improper operation of the damper valve. The fix was to repair the valve so it responded to the positioner pressure, which restored the loop to its correct working order. The control system was also reconfigured to read the HART information from the damper including actual position of the valve, which in the future will immediately indicate positioner or valve problems.

In another instance, the cascade control loop in an acid gas removal had a master pressure control loop and a slave flow control loop. “The cascade loop wasn’t responding to our PID tuning efforts. We checked out the flow control valve and it was operating correctly, as was the entire slave control loop,”  Coughran related.  “The master pressure control loop tuning was correct, but upon further examination we found the high limit setting for the master pressure loop to be incorrect.”

In yet another case, a fuel pipeline was experiencing a nonlinear response of a flow control valve. This was causing dangerous oscillations in pressure and prevented the pipeline company from using its new delivery station. “After several unsuccessful attempts to tune the loop, the difficulty was found to be a malfunctioning actuator coupling,” explained Coughran.

“As with many troubleshooting issues, these fixes were simple once the exact nature of the problem was indentified. But finding the faulty component outside the loop requires careful procedures to be followed, lest one continue to adjust PID parameters for an inordinate length of time when the problem in fact resides outside the loop,” he concluded.