What You Need to Know About Roll-Cooling Design

Roll cooling of sheet, film, coatings and more is an inherently unbalanced process.  I have more 25 years experience designing and manufacturing cooling rolls but did not realize until I retired, became a consultant and got involved in purchasing rolls for my clients that there was often no heat transfer or mechanical bending analysis performed by the manufacturer. I found that many were simply fabricated to meet the outer dimensions and not analyzed for cooling or mechanical performance. 

I realized this when it dawned on me that my clients rarely questioned me about matters such as heat transfer requirements, nip loading, sheet thickness, pump capacity and more when requesting quotes for new rolls. The only thing that seemed to matter to them was the envelope dimensions. But when I had been involved in the design and manufacturing of cooling rolls, we used all that information in the actual design.

Many of our rolls went into large, high-performance extrusion lines and had to handle the thermal and mechanical requirements that went with such large equipment. It would have been a costly error to supply cooling rolls up to 12 feet face length and 60 inches in diameter that could not handle the extrusion capacity.  

Cooling rolls might not look like high-tech machines when you compare them to electronics, but the fact is there is a surprising amount of technology involved in their design, manufacture and use. They have two basic functions: They calendar or size the products that pass through them, and then remove heat from the plastic so the thickness and uniformity is maintained after the material exits the rolls. That involves knowing the forces required to flatten various types and thicknesses of plastics as well as the mechanical bending design of the roll under load. 

Obviously if the rolls bend, it would be impossible to form a perfectly uniform profile. The nip loading of a roll is typically carried by both the inner and outer shells, with the spiral wraps transferring the load between them. The spiral wraps must be rigid enough and spaced properly to transfer the load from the outer shell to the inner shell without deflection. That’s done to keep the outer shell as thin as possible in order to maximize heat transfer into the coolant. 

Understanding Roll-Bending Calculations

Although roll-bending calculations are common in mechanical engineering, they must be understood and properly applied to rolls, as they can carry very large loads without bending. The nip load can be several hundred pounds per inch of sheet width. Using the hydraulic cylinder capacity for the roll-closing force is the best way to determine that, as operators tend to raise the pressure to achieve flatness, which actually makes the bending worse.  Also, many sheet stacks use lever arms to multiply the closing force, so it’s important to determine if they are being used.

Due to the many variations in extrusion processing, there is no “universal” mechanical design. Unfortunately, my more recent experience has indicated that rolls are often ordered without the specific process load requirements. This makes it impossible to design the exact construction that leads to a general “fits all” design. The mechanical loading calculations are well known and conform to composite beam-loading calculations.  The cooling design is generally less understood and seldom used in the roll designs, although it’s the entire purpose for the rolls’ ability to match the extrusion capacity.  

The three-roll stack is the most common sheet cooling/forming arrangement, although more rolls can be added if necessary to handle the cooling load. In a “down” stack (as shown in the accompanying illustration), the top and middle roll form the “nip” and do most of the calendering and resultant mechanical loading. 

An “up” stack can also be used as it positions the nip point closer to the floor for utility of workspace and height of the extrusion die. The contact area at the nip point is very narrow between two round rolls and requires very little heat transfer of the polymer heat to the top roll. Its primary cooling requirement is to prevent the plastic from sticking to it as well as set the calendar gap. The middle roll absorbs the largest portion of the overall heat load because it has essentially 180 degrees of contact and has the most heat in the plastic.

Due to the many variations in extrusion processing there is no “universal” mechanical design of roll stacks.

Cooling requires the transfer of heat through the polymer sheet itself, then through the steel outer shell and finally into the coolant. As a result, the outer shell thickness is a important variable — the thicker the shell the poorer the heat transfer. But often more important is the liquid cooling heat absorption, which is highly dependent on the coolant being both inadequate in volumetric flow as well as being in turbulent flow. 

Turbulent flow can be as important as the coolant  temperature and overall flow rate. Turbulence is generated by the velocity of the coolant, so the geometry of the coolant channels must match the available coolant flow rate to reach sufficient coolant turbulence. Turbulent flow greatly increases convective heat transfer and mass-transfer mixing, and reduces static layers on the walls, all of which are all favorable to increased cooling of the plastic.

Turbulence is determined by calculating the Reynolds numbers. Reynolds numbers (Re) below 1,000 are considered indicative of laminar flow, resulting in poor heat transfer. Reynolds Numbers from 1,000 to 10,000 are considered a transition zone, and above that is turbulent flow. The Reynold’s number is used to determine a Nusselt number.

Comparing the Nusselt number at Re = 1,000 to Re = 3,000 shows an increase of more than two times the convective heat transfer. That will often make or break the cooling capacity.  It would require an enormous change in water temperature to match the effect of turbulence. This calculation is one of the most important in the cooling capacity of the rolls. But it absolutely requires information on the pump capacity for each roll, which is generally not even know by many processors. 

Because the coolant channels are from one side of the roll to the other, the coolant necessarily gets hotter as it traverses the roll, meaning more heat is extracted from the inlet water side than from the discharge side. This must be considered in the overall flow rate through the roll to provide a reasonable balance from side to side. Too much change in coolant temperature across the roll can result in variations in the final sheet properties, such as crystallinity, warpage and poor final product roll configuration. 

Unfortunately, for convenience and cost they are often piped to have the coolant flow to all three rolls from the same side of the roll stack. That is generally not a good idea from a cooling point of view but is done for manufacturing convenience to separate the roll drives to one side of the stack with the cooling entrance being on the opposite side. Of course, the greater flow rates through the rolls the less the effect of the temperature increase is, but there are there are several mechanical factors in the roll design that limit the size of the flow passages.

Turbulent flow greatly increases convective heat transfer and mass-transfer mixing, and reduces static layers on the walls.

But it’s been my experience that some roll fabricators do not take into account the multiple heat-transfer calculations necessary to determine the ideal coolant flow rates for a specific polymer application. Consequently, the exiting sheet temperature will not be uniform across its width. This becomes a greater and greater issue as the thickness increases. Increasing temperature from the coolant inlet to discharge side can cause problems with the final sheet product. For example, this can cause the sheet to billow, stretch and wrinkle after leaving the sheet stack and it could not be evenly wound into uniform rolls or even stacks. This is more troublesome with highly crystalline polymers like HDPE and PP.  

Not a Giant Heat Sink

Chill rolls may look like giant heat sinks that don’t require complex calculations — until you understand the many factors influencing their performance. The cooling function on a chill roll involves heat transfer from the polymer layer into a steel outer shell that is typically hard surfaced with a stainless welded overlay. The steel composite shell then has a specific heat- transfer rate that is a combination of the steel and the overlay plus a layer of chrome plating. The water coolant inside the outer shell has a transfer rate that varies with its temperature and turbulence. There is only minor heat transfer to the inner shell once the system reaches a stable condition and it is generally ignored as it tends to hold close to the average coolant temperature. So, the actual heat transfer requires three separate sets of calculations-polymer/outer shell/coolant. 

With thick sheet, the polymer transfer becomes even more complicated as the polymer heat may not actually have adequate time to transfer from the noncontact side through to the roll shell to the outer shell, requiring a larger roll diameter or slower line speed. An effort is made to model this transfer rate as it immediately affects the heat movement as the plastic transfers from roll to roll.

There is quite a series of variables to consider in the complete design of a chill rolls and the “so-called” general design is not a substitute for a completely designed system based on the required performance. Failure to take the entire performance into account can lead to poor performance of the entire extrusion line.

ABOUT THE AUTHOR: Jim Frankland is a mechanical engineer who has been involved in all types of extrusion processing for more than 45 years. He is now president of Frankland Plastics Consulting LLC. Contact jim.frankland@comcast.net or 724-651-9196.