Modeling and Control

Around 1984, there was a breakthrough in use of simulation for checkout and training. Software packages, such as MIMIC and its predecessor SIMVOX, automatically generated tieback simulations from the configuration and the input and output (I/O) cards and emulated the serial communication between the simulation and the DCS. These packages enabled the simulation to read all of the DCS outputs and send back all of the corresponding DCS inputs. Besides inherently providing a test of the I/O channel assignments, the simulation was separated from the DCS and expanded to cover the entire plant. The tieback simulation sent back the proper motor run contacts for the valve limit switch positions for discrete I/O that was particularly critical for batch operations. For control loops, the process variables was the PID controller output multiplied by a process gain and possibly delayed and filtered to simulate process dynamics. For indicators, fixed values were entered. A method was developed to switch these fixed values and to zero out loop process variables based on whether a flow path was established. A 1 or 0 status of each pump and valve in the piping path were multiplied together to determined the status. Ramps triggered by path status were added to simulate batch and startup responses. Batch operations could be run 100 times faster than real time, and be reset. Failures could be introduced. In Monsanto, these customized tieback simulations were credited with reducing the time to checkout and startup a DCS by 60% or more. By 1986 all Monsanto projects used the software package and its associated methodology and by 1994 nearly all of Monsanto’s 100 operating units were controlled by a DCS. The rapid deployment of the DCS had immediate benefits in terms of safer and more efficient operation plus provided a basis for a program of process control improvement over the next 6 years that lead to 4% further reduction in the cost of goods.

The tieback simulations with pathway logic and custom ramps achieved rapid education of the operators on how to effectively use displays and configuration. To develop better process understanding, the tieback simulations were in some cases enhanced by first principle process models. While the lack of a standard methodology resulted in custom process models of limited scope that were difficult to keep updated, the concept of a process model being connected to an actual DCS forever changed the landscape of process simulation. Up until this time process simulations for operator training used very expensive emulations of the control system at a cost of 200 thousand to 2 million dollars. Most nuclear power plants and some chemical plants and refineries went this route. However, it was not practical to include the detailed features of the control loops (e.g. structure, form, modes, and feedforward), sequences, batch executives, and the operator interface (e.g. displays and historian). Attempts to match and maintain were costly and prone to over simplification. The use of the actual DCS allowed the dynamic simulation to focus on the modeling of the process. The development of packages such as DeltaV Simulate Pro provided the ability to download the actual configuration and displays to a personal computer creating a virtual plant eliminating the need for the DCS console and controllers without any emulation or translation of the control system for training.

Plants are losing experienced operators and engineers so there is an even great potential benefit from periodic operator training. How can we provide training systems that wow decision makers when there may be no one left in the plant to support or even appreciate process simulations for operator training? I don’t have all the answers but here are some key aspects based on my experience:

(1) Live demos of virtual plants for key processes
(2) Online process metrics
(3) Expansion of audience beyond operations to process, control, and maintenance
(4) Modular and generic framework
(5) Ability to run slow processes much faster than real time
(6) Focus on process dynamics and interactions
(7) Readily increasing levels of fidelity for flexible cost and performance

For a perspective of the importance of the operator and some possibilities of online process metrics, check out the Dec 28 entry in the “Tuning and Control System Performance” Category. Next week we will look at some approaches to make the first principle process model more flexible in terms of cost and performance.

The virtual plant offers a break through in training and knowledge discovery but its potential depends upon the ability to develop dynamic simulations that capture the process relationships and response important for process understanding and control.

The best practice of practical real time simulation could easily fill a book but I need to wind this up and move on to other opportunities so here are a few ideas on how to make a process model more flexible in terms of cost and performance and maintainability. It is important to realize the art of simulation is simplification to what is essential.

A significant portion of the time is spent trying to decipher the intricacies of a plant’s DCS configuration and displays. If there is an accurate P&D with the relative location of every pump, fan, valve, and measurement noted along with the complete DCS tag name, and there is browser access to each tag name to assign DCS outputs as process inputs and process outputs as DCS inputs in the model, the need to dig into the configuration is vastly reduced. Note that special DCS I/O such as pulse counts must still be identified and separately addressed.

The computational requirements, numerical hazards, and data requirements on the piping system and fluid flow of a pressure-flow solver are considerable. If there are flow loops for every throttle valve, then the complexity and cost of a pressure-flow solver may be avoided. Of course, this simplification will not identify improperly sized pumps, valves, and pipes. I propose it would be better to add imbedded flow loops in the process simulation rather then venturing into a pressure-flow solver. This simplified approach uses a combination of flow loops and a pathway methodology where the 1 or 0 status of on-off valves and pumps determine an open piping path. The total flow coming out of a piping tee can be written back as the flow going into the tee. The use of flow loops reduces but does not eliminate the need to simulate valve backlash and stick-slip. If a pressure-flow solver is deemed valuable, than I suggest a sequential modular method to avoid ill conditioned matrices and numerical problems during batch operations and the startup and shutdown of equipment.

If the model starts out with initialized but settable molecular weights, densities, and heat capacities, then levels, temperatures, blending, and temperature can be simulated. If the dissociation constants for bases and pH are added, then pH can be added. For the modeling of vaporizers and evaporators, it may be sufficient to add vapor pressures and boiling points of selected components as a function of composition. For reactors, the standard form of Arrhenius and Michaelis-Menten kinetics may be sufficient. Neural networks may be able identify kinetic rates to provide a simpler and higher fidelity hybrid model. The complexity of a full blown physical property package could be reserved for more complex vapor equilibrium problems such as distillation.

Finally, it is most important to get the dynamics right. The process models from on-demand and on-line tuning packages such as DeltaV Insight and model predictive controllers such as DeltaV Predict can be used to supplement or replace first principle models for specific parts of the process.

For my virtual plant experience and top ten list check out
http://www.controlglobal.com/articles/2007/385.html
http://www.controlglobal.com/articles/2007/359.html
and the “Education” and “Process Simulation” categories on this website.

About every year since 1985, I wrote a book whether I needed to or not. Some of the books were written to entertain my self and hopefully others by adding a humorous view point to what is normally a very a serious profession. In 2006 my contribution to this unique endeavor was The Life and Times of an Automation Professional – An Illustrated Guide. While I wrote most of the material and generated most of the cartoon concepts, I choose to list the authors in reverse alphabetical order to be different and give the cartoonist Ted Williams some long overdue recognition. Right after this book was published I expanded my horizons to an even more fertile ground for humor – retirement. Since all of my friends were partly or fully retired, I had the experience base, but how could I make it different?

Trillions of atoms in my brain went amok to break free of the 4D dilemma or at least make reading about engineers and retirement less boring. The result was a break in the space and time continuum and a book on retirement like no other. For conservation of spin, I am sure there is a book somewhere in the far reaches of the universe with humor moving in the opposite direction. There is also the matter of possibly anti-matter in a parallel universe or Anti Funnier Side of Retirement book. However, these books are light years from being available although radio waves indicate a Starbucks on alpha centurion is offering extra hot lattes.

So if you are ready to laugh about the impending or depending retirement of people you know and love including yourself, check out The Funnier Side of Retirement for Engineers and Other People of the Technical Persuasion published in 2007 by ISA. What will I do for 2008?

Front and back covers:
Funnier Side of Retirement

Link to publisher:
http://www.isa.org/retirement

I tend to be descriptive instead of proscriptive in my discussion of process control and improvement. This may be due to my interest in the scientific method and concepts rather than details that may only be good for specific sets of conditions. However, there are certain rules of thumb based on solid principles that serve as guidance for good practices. It is interesting how the factor 1/5 turns up over and over again. So here is a list where “fifths” rule for best performance. For my drinking buddies, this is not about best bottles.

(1) loop dead time < 1/5 x (process time constant + load time constant)*
(2) measurement filter time < 1/5 x (process time constant + load time constant)*
(3) module execution time < 1/5 x (loop dead time)
(4) module execution time < 1/5 x (process time constant)
(5) measurement scan time < 1/5 x (module execution time)
(6) standard deviation < 1/5 x (allowable control error)
(7) noise in controller output < 1/5 x (required change in controller output)
(8) stick-slip of valve < 1/5 x (required changed in controller output)
(9) half dead band of valve < 1/5 x (required changed in controller output)
(10) secondary loop time constant < 1/5 x (primary loop time constant)
(11) secondary loop dead time < 1/5 x (primary loop dead time)
(12) controller gain < 1/5 x (ultimate gain)
(13) controller rate setting < 1/5 x (process time constant)
(14) historian update time < 1/5 x (oscillation period)

* - your main hope for dead time processes is to slow down the load disturbance.

My final recommendation is the 3-D U2 concert film. I have been in front row seats for BB King, BTO, and Joan Jett but this was even better for seeing and feeling how musicians get into their music. Bono and Edge were 7 feet tall playing right above you. The positive energy from the lyrics and instrumentation was incredible.