As I said in my first column in December 2012, proportional/integral/derivative (PID) solutions seem to require an expertise that is in the hands of a few. But here I want to tell you about a specific kind of PID solution I alluded to in that first column: external-reset feedback, also known as the “dynamic reset limit.” Here’s some history of how it was developed, how it was lost for decades, and why it’s useful now.
>> Full Power of PID: Greg McMillan started his discussion on PID in the December issue. Visit http://bit.ly/awslant57
The technology that enabled this solution was an implementation of integral action that uses a filter in a positive feedback path. The filter input is either the controller output or the external-reset feedback signal. This technology was developed in 1925, but the use of an integrator (as depicted in presentations, publications and standards) for integral action in the PID was not possible in the days of pneumatic controllers. The advantage of positive feedback technology, therefore, was not recognized then. When electronic controllers were developed, integral action in many systems employed an integrator, available as a standard circuit component, to implement the integral mode depicted in the literature.
While the basic implementation looks simple for a series PID, the positive feedback path can have you going in circles. Also, I have not seen any block diagram that shows how feedforward signals and different PID forms and structures are incorporated. So, I developed a block diagram showing how all the major PID options could be implemented (see illustration).
I can’t confirm the pictorial validity of this diagram because there is nothing to compare it to. The block diagram was not easy to conceptualize and required a breakthrough in thinking enabled by the review of my protégé Hector Torres from the ISA Mentor program.
External-reset feedback enables the use of setpoint filters to coordinate the timing of flows, for blending or reaction, without the need for retuning the secondary flow loops. The primary PID honors the setpoint filter time setting to ensure there is no transient composition or stoichiometric unbalance for a change in production rate.
A new opportunity for external-reset feedback involves directional move suppression. Analog output blocks have separate setpoint rate limits for the up and down direction. This feature can suppress interactions; protect the equipment, environment and process from disturbances during optimization; and stop split range oscillations.
To prevent unnecessary crossings of the split range point, the rate limit can be made slower in the direction toward the split range point. This capability can increase bioreactor yield, for example, by stopping the pH PID from making unnecessary additions of sodium bicarbonate that cause an increase in cell osmotic pressure and consequential premature cell death from membrane rupture.
The technique in general helps reduce oscillations that often exist at the split range point due to the discontinuities of making a transition from one control valve to another, and from the increased friction from sealing and seating surfaces as one valve closes and another opens. The option also stops the limit cycles from the ensuing pronounced stick-slip near the closed position.
Directional move suppression provides a slow approach to a normal or optimum operating condition and fast getaway from an abnormal condition. For surge control, the analog output rate limit provides fast opening and slow closing of surge control valves. This is extremely useful for recovery from surge and for preventing re-entry into surge.
This has been done crudely for decades by quick exhaust valves on the actuators, which typically led to overreaction and oscillations. For RCRA 2 and 12 pH limits in waste treatment systems, for example, the analog output rate limit provides a slow approach to the limit (to conserve reagent use) and a fast recovery from a disturbance in the direction of the limit (to prevent a recordable RCRA environmental pH violation).
External-reset feedback also enables a much-easier-to-tune and responsive valve position control (VPC). Traditionally, the VPC has used slow, integral-only control to reduce interactions. Astute users have employed customized gap action control to deal with disturbances. The directional move suppression and the elimination of limit cycles inherently reduce the adverse effects of interaction, backlash and stiction, and make VPC tuning easier.
This brings us to the final advantage of directional move suppression for now (many more advantages will evolve). By controlling the transfer of variability from the controlled variable to the manipulated variable, we can suppress interaction. The least important and most disruptive PID would have analog output setpoint rate limits introduced to avoid upsetting the other loops.
>> Greg McMillan, [email protected], is a retired Senior Fellow from Monsanto-Solutia and an ISA Fellow. Greg received the ISA “Kermit Fischer Environmental” Award for pH control in 1991, and received the ISA Life Achievement Award in 2010. He is also the author of numerous books on process control. His most recent book 101 Tips for a Successful Automation Career was inspired by the ISA mentor program he started at ISA Automation Week 2011.