If the purpose of regulatory control is to maintain the mass and heat balances of a process, then the purpose of advanced process control (APC) is to manipulate the mass and heat balances to achieve optimum performance. As far back as 1970, APC experts and analysts have repeatedly claimed that the benefits of implementing APC were a 2% to 6% increase in production. A recent APC provider company’s press release claimed that the use of APC could enable capacity-restrained plants to increase production by 2% to 3%.
In a typical hydrocarbon processing plant, such an improvement would increase bottom line hundreds of thousands of dollars per year or more. The folks at AspenTech say, depending on the processing unit, the benefit numbers might be much greater. (See Table 1)
TABLE 1: Estimated Benefits using Dynamic Matrix Control in Hydrocarbon Processing
|Estimated Refinery Unit Benefits||Benefit in $s Per Year|
|Crude Unit 150,000 BPSD||$2,250,000 to $4,500,00|
|FCCU 70,000 BPSD||$4,200,000 to $10,500,000|
|Cat Reformer 50,000 BPSD||$1,500,000 to $4,500,000|
With so much upside potential for improving a business’ bottom line, one would think the phones of APC provider companies would be ringing off the hook, but they aren’t—although a recently released market study from Frost & Sullivan predicts that by 2012, APC sales will be nearly double 2005 sales of $600 million. While that prediction is encouraging, when adjusted for inflation, the forecasted growth hardly reflects the much-touted APC upside potential. Something seems to be missing. What is it about APC technology that is taking businesses so long to embrace it?
A Bit about APC
The need for APC likely began shortly after John Ziegler and Nathaniel Nichols published their breakthrough paper on loop tuning in 1942. Once the PID (proportional–integral–derivative) control loop could predictably be brought under control or tuned, process control technicians and engineers rapidly accelerated the use of closed-loop control. However, what immediately became apparent to early adopters was that single-variable, closed-loop PID wasn’t always capable of getting the control job done. Users found that minimizing the interactions of other controlled and uncontrolled variables required the addition of controller modifiers (e.g., lead, lag, bias, gain, feed-forward, cascade, etc.).
By the 1950s, the development and availability of control modifiers had advanced so that entire unit control strategies for boilers, ammonia processes and distillation columns were being successfully developed and implemented.
While working for Shell, Charlie Cutler and his colleagues developed the forerunner of today’s model-predictive control (MPC) techniques. In 1979, he presented his landmark paper, “Dynamic Matrix Control—a Computer Control Algorithm,” which remains one of the most important advances in modern APC.
Cutler’s success kicked off a flurry of additional APC initiatives, including Edgar Bristol’s work on multivariable adaptive feed-forward and Bob Otto’s and Kelvin Erickson’s work on multivariate control.
People tend to think of APC as a complex control algorithm, when sometimes it’s simply soving old problems by looking at them through new glasses. In a 1986 article, “Analysis of Valve-Position Control for Dual-Input Processes,” authors Chen Ching Yu and William Luyben introduced a completely new approach to tuning the integral-only function of a controller in order to achieve a valve response fast enough to reject large process upsets.
Terry Blevins of Emerson Process Management says, “People tend to confuse valve position control with valve positioners or digital valve controllers. The tuning of the integral-only control function for valve position control is much more critical than most people realize. The technique described by Yu and Luyben helps eliminate the limit cycle at the split range point caused by the increase in nonlinearities and the decrease in resolution imposed by backlash, backfilled pipes and dip tubes, rangeability limits and friction, particularly associated with beginning a flow from zero. This technique also eliminates the conceptual and tuning problems with valve position control.
The cost-effectiveness of computing power and memory kicked off another round of R&D spending by major control companies intent on developing commercially feasible APC applications, including controller-based fuzzy logic and neural networks, self-tuning, and multivariate adaptive feed-forward.
By the late 1990s, users had a wide variety of sophisticated APC applications available from which to choose. However, despite an increasing number of conference presentations and seminars heralding all sorts of benefits to be derived from using APC, the market remained relatively stagnant.
All About the Benjamins
Many APC implementations have been very successful, at least in the eyes of the engineers that installed them and the operations folk who work with them.
One anonymous poster to the APC-Network.com web site says a multivariable controller has been operating with a 98% uptime since Dec. 1999, had a seven-month ROI, and produced $730,000 of annual benefit to the company—$5.1 million of savings in the past seven years.