Advanced regulatory versus model-predictive control

Some currently practicing control engineers may be regarding the features and capabilities of advanced regulatory control (ARC) as inferior to model-predictive control (MPC)—but are they really?

03/03/2015


A contributor to a recent LinkedIn APC blog made the following statement about the difference between advanced regulatory control (ARC) and model-predictive control (MPC):

“In a general and very simple way, an advanced regulatory controller acts based on an error, while a predictive controller uses a model in order to predict what is going to happen and acts in consequence to avoid it.”

As a long-time proponent of ARC for the solution of many industrial control problems, I would like to correct a common misconception held by probably 99% of current practicing control engineers regarding the features and capabilities of ARC.

During the 1970s, and with the advent of digital control systems and process control computers, those of us involved in the evolution of ARC developed two very powerful ARC techniques, referred to as feedforward control and decoupling control. The reason for developing these techniques was quite simple: feedback control, regardless of how sophisticated, was insufficient by itself for keeping important control variables close to setpoint when disturbances occurred.

Feedforward control was developed initially as the name suggests—to make adjustments to independent variables (MVs) to keep important dependent variables (CVs) close to setpoint when the feed rate to a unit operation changed. How was this done? With models, of course! Open loop response data was first gathered to observe the response of the CVs to changes in the feed rate. This response was typically fit with a dead time and lag model. The next step was to “invert” this model so as to develop the feedforward pattern needed to adjust the MV used to control the CV so as to keep the CV close to setpoint during the feed rate change.

So, for example, on a distillation column where the overhead temperature is being controlled with reflux flow, the challenge is to apply dead time and lead/lag compensation to the feed rate, passing the incremental trajectory changes (with proper gain) to the reflux flow, so that the reflux is adjusted in coordinated fashion, such that the disturbance is cancelled out. If the feedforward compensation is perfect, no feedback correction is required at all!

MPC provides more or less identical “model-predictive” control action for feedforward variables (FFVs). ARC feedforward control action can actually be made superior to MPC feedforward control action. The “proper gain”—that is, the amount the MV must be adjusted to keep the CV on setpoint—is not a constant ratio. For example, the reflux/feed ratio required to keep the top temperature on setpoint may be different today from what it was yesterday, due to feed composition changes, etc. ARC can incorporate dynamic, adaptive feedforward gain compensation to account for process changes. MPC cannot.

The other model-based control feature of ARC, decoupling control, was developed for similar reasons, and initially for distillation columns. On many fractionators where high purity products are being produced, the heat and material balances are highly coupled, such that moves made on one end of the column also affect the other end. We recognized this phenomenon early on and realized that simultaneous decoupling moves could be made on both ends of the column so as to minimize heat and material balance upsets. For example, if the reflux needs to be increased to increase the purity of the overhead product, then a similar decoupling move needs to be made to the reboiler heat so as not to disturb the bottom product purity. The decoupling moves are determined from a model developed in a manner similar to the one described earlier.

Sounds to me like these ARC techniques developed over 40 years ago:

“Use a model in order to predict what is going to happen and act in consequence to avoid it.”

Which is to refute the misconception that ARC is not model-predictive. It certainly is as implemented today, and has been for the last 40 years. In addition, it can be made multi-variable; that is, feedforward and decoupling control action can be implemented simultaneously for several FFVs and DCVs acting on several MVs to control several CVs. And, finally, in general, ARC is cheaper to implement, is more durable, is more easily understood by operators, and requires less maintenance, when compared to MPC. It should always be considered first as the go-to technology for solution of most common industrial control problems.

This post was written by Jim Ford. Jim is a senior consultant at MAVERICK Technologies, a leading automation solutions provider offering industrial automation, strategic manufacturing, and enterprise integration services for the process industries. MAVERICK delivers expertise and consulting in a wide variety of areas including industrial automation controls, distributed control systems, manufacturing execution systems, operational strategy, business process optimization and more.



No comments
The Engineers' Choice Awards highlight some of the best new control, instrumentation and automation products as chosen by...
The System Integrator Giants program lists the top 100 system integrators among companies listed in CFE Media's Global System Integrator Database.
The Engineering Leaders Under 40 program identifies and gives recognition to young engineers who...
This eGuide illustrates solutions, applications and benefits of machine vision systems.
Learn how to increase device reliability in harsh environments and decrease unplanned system downtime.
This eGuide contains a series of articles and videos that considers theoretical and practical; immediate needs and a look into the future.
Sensor-to-cloud interoperability; PID and digital control efficiency; Alarm management system design; Automotive industry advances
Make Big Data and Industrial Internet of Things work for you, 2017 Engineers' Choice Finalists, Avoid control design pitfalls, Managing IIoT processes
Engineering Leaders Under 40; System integration improving packaging operation; Process sensing; PID velocity; Cybersecurity and functional safety
This article collection contains several articles on the Industrial Internet of Things (IIoT) and how it is transforming manufacturing.

Find and connect with the most suitable service provider for your unique application. Start searching the Global System Integrator Database Now!

SCADA at the junction, Managing risk through maintenance, Moving at the speed of data
Flexible offshore fire protection; Big Data's impact on operations; Bridging the skills gap; Identifying security risks
The digital oilfield: Utilizing Big Data can yield big savings; Virtualization a real solution; Tracking SIS performance
click me