So Smooth

Is smooth good? We like smooth trend charts but is that what is really going on in the process? Do we want a smooth talking measurement or the straight story?

My first clue dates back to a startup of a world class intermediate plant when smooth temperature recordings were traced back to sand in thermowells from when the pipelines were sandblasted during construction. Then, in a downstream plant a report came in that temperature sensors on extruders were to be now installed in large blocks of metal rather than the melt because the trends were smoother. Many years later in a lab, a biochemist proudly showed how he had smoothed out his temperature recording on the bioreactor by partially retracting the sensor in its thermowell. Finally, I heard horror stories about thermocouples installed in glass lined thermowells on exothermic reactors.

The concern is not restricted to temperature. Rugged (thick glass) and most high temperature electrodes are extremely slow. pH electrodes installed in overflow lines and behind baffles in a vessel have an environment so still that process buildup makes the electrode smooth out changes even when the flow restarts or the agitator speed is increased. Just a 10 millimeter film on a pH electrode can increase its response time from 10 seconds to 100 seconds. Coated electrodes are slow electrodes. Multiple electrodes should expose the foul up but then again the smoothest response I have seen was for 3 electrodes all installed with their protective caps still on.

The easiest way to slow down a measurement is to increase the filter time constant in the DCS. Here the sky is the limit particularly for pressure systems that blow their rupture disks. For some fast gas pressure systems, putting in a faster transmitter will make the trend recordings look worse even though the pressure loop is doing a better job because it is seeing the disturbances better.

For pressure, flow, and inline composition and temperature control, the measurement time constant is probably already the largest time constant in the loop. An increase in this time constant due to coatings or filter times not only makes the trend chart smoother but allows the user to increase the controller gain which furthers the deception. You and the controller are seeing an attenuated version of the real world.

Other time proven ways to make trend charts look smoother to impress friends and relatives is to increase the process variable scale range, decrease the time scale range, and increase the compression, update time, and exception trigger for data reporting.

A smooth loop could be good news or bad news, which leads me to my Top Ten List.

Top Ten Good News Bad News

(10) The good news is that smart instrumentation has been approved. The bad news is it is a dumb installation.

(9) The good news is that the control valves are not oscillating. The bad news is the loops are all in manual.

(8) The good news is the new project manager is a process control engineer. The bad news is you are the project manager.

(7) The good news is that all the process variables are drawing a straight line. The bad news is they are off scale.

(6) The good news is that digital positioners have been added to all of the control valves. The bad news is the position measurement is a “smooth talker”.

(5) The good news is that the loops are no longer oscillating in automatic. The bad news is the plant is shutdown due to the loops being in automatic.

(4) The good news is that your group’s name has “advanced” in it. The bad news is the name is “Advanced Aged Engineers”.

(3) The good news is that you have reached the level to work through “others”. The bad news is there are no “others”.

(2) The good news is that you will have a creative new office. The bad news is it has virtual walls.

(1) The good news is that you have been offered a retirement package. The bad news it is a gift certificate.