Bringing the cupola into the computer age.

Cupola modeling software introduces science into an art and dispels some of the mystery surrounding the time-tested iron melter.

"If the cupola is important," said Norman Lillybeck, Deere & Co., Moline, Illinois, "then it has to be computerized."

The cupola is important. The more than 200 cupola melters listening to Lillybeck at the AFS International Cupola Conference in March proved as much. According to another speaker at the conference--Francis Kaiser, Modern Equipment Co., Port Washington, Wisconsin--cupola-melted iron will account for about 64% of all the castings poured in the U.S. this year.

New technology and further understanding of the processes, chemistry and environmental implications of the cupola have more than validated those who stayed with the furnace when other foundries converted to electric melting.

That trend toward electric melting was spurred by the notion that electric worked cleaner and at a lower cost than the cupola. As it turns out, induction and electric arc furnaces do not burn as cleanly as thought. More stringent environmental legislation is beginning to mean that expensive new equipment run on other fuel sources must be added to the furnaces, thus erasing the perceived benefits of electric.

However, thanks to new technology, the cupola not only reuses much of its gas emissions in combustion, but it can burn baghouse dust and save foundries hazardous waste disposal costs. Though both are made from or with coal, coke actually exhibits higher energy efficiency than electricity. The cupola's tolerance for scrap that is less than clean makes charging it cheaper and better for a foundry's internal environment.

Table 1. Cost of Cupola Vs. Electric Furnace Given Basic Current Prices for
Each Energy Source

                              Cupola                       Electric Furnace

Melting                       $16                          $27.50
                              200# Coke
                              550 KWH
Carburizing                   Included                     $4.50
                              1-1/2% x $300
Afterburner                   $0.15                        $2.37
                              (pilot only)                 0.91 mil, BTU
Alloy Use                     $2.25                        --
                              Silicon Differential
Misc. Electric                $0.25                        --
                              Blast Blower
Total Energy cost/ton         $18.65                       $34.73

Yet, for all the associated benefits of cupola use, relatively little is understood about its complex processes by the people who use it. Conference speaker Seymour Katz, General Motors R&D Center, Warren, Michigan, says of the problem: "Until now the cupola, at best, was understood qualitatively. We know, for example, that increases in the oxygen content of the blast will increase the iron temperature, melt rate and alloy content of the iron. But we are uncertain of the exact amount."

There are other problems. Though cupola melting is the lowest cost method of producing iron, cupola energy efficiency is about 35% and alloy losses can range from 30-40%. By the very nature of the process, it is difficult to take a scientific approach to manipulating all the variables.

The AFS Cupola Model, then, can be seen as the introduction of a greater degree of empirical science into a process that has largely been an art. Quantitative analysis of the input and operational variables replaces a lot of guesswork, while opening the door to future improvement.

New Capabilities

The finished product is a software program that can be run on a 386 or 486 personal computer with Windows version 3.1 or later. It captures and illuminates the most common, practical processes and reactions inside a cupola, while attempting to account for all the factors that make exact repeatability and output prediction so difficult in cupola melting.

To the foundryman, the model means that there is now a technological window to the interior of the furnace. The geometric layout of the different components becomes clear, as do the reactions that help or hinder the melt.

The immediate implications are a greater ability to predict things like iron temperature and composition or heat and alloy loss. Capable of simulating all practical situations and operations encountered in cupola use, the model can help the user develop true least-cost charging or compute the cost differences between a refractory-lined and unlined cupola. The software provides an efficient way to experiment without the expensive trial and error of actual test melts and a tool for quick diagnosis and adjustment of problems.

For the future, the model undoubtedly will aid in the development of new technology for the cupola. As the model is subsequently improved it may reveal potentially beneficial changes in the cupola itself, or automate the process to the degree that a desired iron chemistry can be maintained at all times.

At the conference, Lillybeck and Katz led off the presentation of four papers detailing the purpose, development, functions and results of the model.

Research and Development

The product of nearly four years of work and a $1.8 million investment by the U.S. government and a consortium of industrial members, the project was sponsored by the AFS Cupola Committee and directed by AFS staff. The initial chemical research was done by the University of Michigan, while thermal and mechanical research was undertaken by General Motors Research and the Massachusetts Institute of Technology. Model refinement was contracted to software specialists to produce the final operational product.

Katz said the project was undertaken with several well-defined objectives. Upon completion, the model was expected to:

* accurately predict cupola performance from given operating conditions;

* provide customized control charts for practical operations;

* enable the purchase of scrap with the most advantageous cost-to-benefits ratio;

* provide accurate estimates of benefits to modify the cupola structure;

* be used as a diagnostic tool for troubleshooting;

* serve as a control center for an automated cupola operation.

With the current version of the software, the first five of those goals have been fully realized, thanks to the mathematical model of the physical interior of the cupola.

Mathematical Structure

To make it as simple as possible, the model was designed to be one dimensional, according to Vladimir Stanek, Academy of Sciences of the Czech Republic, Prague, Czech Republic. Individual quantities, such as temperatures, concentrations and velocities, are regarded as functions of the vertical distance from the charge door. The cupola is divided into threes parts: the shaft, the coke bed in the well and the metal/slag pool in the well.

All the separate operations and reactions that take place in the melt process are computed using 20 differential and algebraic equations. In each of the three physical parts of the furnace, distinct sets of equations solve for the reactions that originate there. Those quantities then figure into the larger sum to provide the total model.

In addition, the model is comprised of individual submodels for other reactions, heat transfer coefficients, specific surfaces, heat capacities, sulfur transfer holdup of liquid products and heat losses. All component equations are solved to build the overall structure of the model and the complete picture of the working cupola.

Input

Given the extremely complex nature of the calculations the program runs, it is important that using it be as simple as possible for the average foundryman. Comments from foundries that tested it pointed out the need to make the model workable to those most likely to use it.

To assure wide accessibility, the user interface has been set up so that only basic computer skills are necessary. Lawrence Smiley, Finite Solutions, Inc., Cincinnati, Ohio, demonstrated its practical simplicity to the conference, entering data and showing how the Windows format is operated with the use of a "mouse" and standard "pull down" menus.

Input data has been organized and sequenced in logical groupings according to the way a foundryman might think of such information, and in units and measurements commonly used in the foundry.

The screen titled "Model Parameters" lists information fundamental to the model and which should rarely, if ever, be changed. Access to these menus is only granted after the model warns the user that any changes will alter the running of the model. More common data that requires frequent adjustment is offered in a separate "quick menu", so that it can be changed easily from one mn to the next. Basic categories of "quick" input data are:

* Charge materials data--number of metallic constituents and their melt temperatures, size of particles, coke composition, etc.;

* Blast data--rate, temperature, humidity, oxygen enrichment;

* Cupola data--geometry, operation, heat loss;

* Physical, chemistry and auxiliary data--heats of reactions, slag density, metal droplet diameter;

* Numerical solution data--total nodal points, coke and oxygen balances, computer operational specifics, minimum coke particle diameter.

To facilitate easy input and minimize the amount of calculation the user must make, the model is equipped with several comprehensive data bases, including common charge materials and their related properties, alloys and their properties, and cupola geometries.

When entering this input, the chemical and physical attributes of the charge materials can be edited on-screen by the user, depending on the exact specifications of the actual samples. When any alloying elements are entered, the program automatically distributes and applies their effects to the properties of the various charges.

The data base of geometries can similarly be displayed and edited, allowing the foundryman to run a simultaneous model of more than one cupola. Geometric parameters include:

* inner diameter of the cupola;

* lining thickness;

* height from taphole to top of charge;

* height from taphole to center of tuyeres;

* iron dam height;

* and height from center of tuyeres.

Running the Model

Once preliminary entries have been completed, running the model is a simple matter. After selecting the "run model" option from the main menu, the user chooses the analysis to be run from a list of stored input data. If more than one analysis is chosen, the multiple model runs will be performed one after the other. Typically, it takes only a few minutes to run one model.

Feedback

When the model has finished its run, all results of primary practical interest can be accessed under the "Model Results" screen. The user needs only to choose the desired output, select this screen, and relevant quantities are given for all important characteristics.

The model also gives profiles of different output parameters, which can be plotted from within the user interface as functions of height along the cupola. Profiles include temperatures for solids, liquids and gases, heat loss flux, a variety of mole fraction quantities, carbon concentration, sulfur concentrations and Fe mass velocity.

Several profiles can be plotted on the same chart, as can profiles from separate model runs for easy comparison. Also, the charts can be plotted to display the effects of certain operating variables, allowing for quick assessments of the implications of combining certain quantities.

Further Improvements

Though the software will be available to foundries later this year, upgrades and improvements already are being researched. Perhaps due to the resurgence of the cupola or the success of the model thus far, new government funding and the interest of government laboratories and universities have advanced work in transforming the model into an artificial intelligence program that would result in full automated cupola control.

Other new developments that will significantly enhance the model include changing the mathematical operations to provide more speed and accuracy, and the possibility of inverse processing. Whereas now the user plugs in inputs to find out results, the desired results would be entered for the model to predict the operating conditions required to achieve them.

If the model represents a quantum leap in cupola technology, the chance for advancement is just as great. "With this revolutionary power in hand," Smiley said, "there is the potential to enter a whole new era of improved control and reduced cost in cupola operation."

Extended Heats: Promoting Savings and Safety

Considering the last AFS Cupola Conference was convened in 1980, it wasn't surprising that cupola melters had several new developments to discuss when they gathered again this past March.

One of the most important topics was the implementation of extended heats as standard practice in many cupola melting foundries. In his presentation on the practice, Patrick W. Frederick, Auburn Foundry, Auburn, Indiana, detailed the benefits and important requirements involved in running extended cupola campaigns.

"To increase efficiency," Frederick said, "roughly a third of foundries that melt with cupolas utilize extended heats of a week or longer."

The practice entails running campaigns anywhere from a few days to a year without daily cooldown and dropping to repair the inside. Advances in cupola design, such as the use of the water-cooled shell, have made this possible.

Foundries see benefits from extended heats in reduced labor costs associated with frequent tapout and burn downtime, as well as enhanced worker safety with the reduction of hazardous drops. Continuously tapping the cupola for a week or longer eliminates the need for multiple furnaces and simplifies charging and pollution control.

To gain these advantages, Frederick said it is critical that foundries maintain certain optimal conditions. Operators need to be experienced enough to make adjustments to compensate for the different operating parameters caused by slag erosion. The temperature of the bottom and the shell must be carefully monitored to identify areas that might experience iron runout. Most importantly, proper installation and use of the correct refractory lining are needed.

Other considerations, such as having a large enough holding furnace and support equipment, go into a successful long campaign practice.