Are We Misleading Our Operators?

The current industry practice for graphic displays may be deceiving our operators, leading to wrong decisions. Our obsession with numbers has resulted in digital displays of process variables that are typically primary process variables (PV) on piping and equipment graphics as the principal method of conveying process understanding. The overriding problem is fundamental and extensive.

The digital display coupled with human impatience and inability to take into account deadtime leads to premature and incorrect actions particularly for slow loops. The most important slow loop is often temperature since temperature is used for the control of composition and quality besides the energy balance. I have seen the problem where operators get obsessed with the digits and have no sense of direction of where the PV has been or where it is going. The problem goes beyond decimal places. A faceplate or dial with no fractional display still has a problem. Consider the example How-Digital-Displays-Mislead.pdf. The temperature PV is 48 and the SP is 50. Should the steam or chilled water valve be open? Looking at the faceplate or digital value or even a dashboard dial on their graphic display operators and engineers say the steam valve should be open. Interesting enough, my experience is technicians may not be fooled as easily. If you look at a trend chart with the proper time frame, you realize the chilled water valve should be open to prevent overshoot due to the acceleration of temperature trajectory. Today the typical trend chart available from the faceplate is next to useless because the PV span and time frame are not intelligent. A short time frame would not have shown the acceleration. The cited example was actually reported to me from 2 control rooms as something wrong with PID algorithm. I got the report only because I knew the people at the plant very well. I would venture the same problem is experienced to some degree in almost every control room. It leads to manual tuning of too much reset action and not enough gain action because the integral mode has no sense of direction and would work to open the steam valve even if overshoot was eminent, whereas the proportional mode would work towards getting the chilled water valve open as the temperature increases. The lack of understanding of trajectories explains why most manually tuned temperature loops have too much reset action (too fast of a reset time) and not enough gain or rate action.

I think the operator would benefit from a trend whose Y axis span (PV, SP, OUT span) and X axis span (time span) can be adjusted and optimized that shows trajectory over the span and time frame of interest. The identification of the process deadtime and maximum rate of change of the process variable used in the short cut tuning method as discussed in Deminar #6 PID Tuning for Near-Integrating Processes could be used for initialization of the trend’s span and time frame.

I also think XY plots with time trajectories (e.g. batch worm plots and compressor surge map plots) would be helpful. The user should be able to zoom in on these trends. For further discussion on this wish of mine see my entry XY Plots for Operations and Process Analysis

Technical and Maintenance support staff need to see digital values but not beyond the resolution of the device. I think PV, SP, and Out values displayed should not be more precise than the automation system. Since the measurement values displayed are in engineering units the number of decimal places if any depends upon the span besides the type of measurement. Measurement resolution ranges from 0.01% for Elite Coriolis and Turbine flow meters, premium resistance temperature detectors (RTD) with integral mounted transmitters, and pH electrodes with small fixed liquid junction potentials to 1% for other technologies. Similarly, the PID output to a secondary loop depends on the secondary measurement resolution. The PID output to the field depends on the resolution of the final control element. For control valves, the resolution ranges from 0.1% to 5% of stroke. Some special digital actuated valves rarely seen in the process industry a have resolution of 0.02%. Variable Frequency Drive speed resolution ranges from 0.05% to 0.4%. Heaters “duty cycle” resolution range from 0.1% (phase fired SCR) to 3% (electromechanical contactors). I think you could default to a digital display resolution of 0.1% for support staff provided it is adjustable. Noise suppression is essential for digital values displayed for operators.

Noise is a problem. I think technical support staff should see the noise so they can work on reducing but operators should not so the noise doesn’t confuse them. However there should be a noise band parameter that would prevent updates of the last digit from noise for both the operator and the PID to reduce the dither from noise that wears out packing and inflicts disturbances on this and other loops.

A far out idea would be a rate of change PV created by passing the PV through a deadtime block to create an old PV one process deadtime in the past to provide a better understanding of where a measurement is going and how fast. The deadtime bock solves some of the noise and resolution problems but there would still be a decision on what resolution is useful and what filtering and noise band parameter may be needed. The idea is the result of the idea for universal tuning and rapid modeling where the process deadtime and maximum rate of change of the PV is identified. The absolute key to this identification is the deadtime block to create an old PV one deadtime in the past.

Finally, the lack of trajectories of the second variable (SC), third variable (TV), and fourth variable (FV) available in HART signals for the operator means that the very first question when there is a problem “What maintenance was done or needs to be done” wisely noted in the Control Talk Column Drowning in Data, Starving for Information, 4 is relegated to Automated Management Systems.

These ideas could be prototyped and tested in addition to the many possibilities discussed in the Control Talk column Operators Unleashed.