Making Things That Matter

Kim Phelan

Visit Liberty Utilities’ Powersite Dam in Forsythe, Missouri––one of the first original hydroelectric dams in the state––and you’ll come away with two different perceptions depending on where you spend your time: Picnic and play outdoors at the site’s park-like grounds and feel refreshed, or, for insiders only, tour the inner-workings of the 100-plus-year-old dam structure and wonder if you’ve been transported to Alcatraz, minus the cells. 

Either way, Taney County’s dam on the White River, which, upon its completion, created Lake Taneycomo in 1913, is a historic landmark cherished by the community, especially locals like Mike Renfrow who value the preservation of old things that tell stories of the past. Renfrow himself, general manager at AFS Corporate Member Monett Metals, is now a part of the dam’s story, having worked with a team that unlocked an ancient casting conundrum and reverse-engineered a worn cast-iron crosshead ready for replacement. 

In the 1930s, the dam underwent some modernization and the crossheads were replaced during the upgrade; nevertheless, how Renfrow’s metalcasting predecessors produced the complex part remains mysterious––but devising the critical crossheads’ contemporary counterparts needed inside each of the dam’s four generators became a six-month, obstacle-riddled puzzle for both the low-volume job-shop foundry and its customer, Aurora Motor and Machine. As it happens, both companies specialize in very unique types of projects. 

Aurora, also over 100 years old, is one of the country’s rare machine shops that performs most work manually, one piece at a time. Their collection of machinist tools displayed across the walls give Aurora a Smithsonian air, said Renfrow. As for Monett Metals, which employs 75 people in Monett, Missouri, it frequently creates replacement parts for eclectic machines and antique vehicles, including WW2 Cushman scooters. Special among its achievements, the foundry contributed cast metal ribbons (from recovered Twin Towers steel) for a 9/11 First-Responders Memorial in New York, as well as a 20-ft. stainless steel eagle feather for an Osage Native American Veterans Memorial in Oklahoma. In his estimation, Renfrow said, the Taneycomo dam project and the legacy it represents falls into this prestigious category of historically important work. 

“When we were unloading the crossheads at the dam and I was running the hoist, it gave me a lot of pride,” he said. “You know, foundry work is a hard job. But when I have a young employee who I see has potential, I tell him, ‘You can go down the road, and you can work at a call center, or you can go somewhere and put widgets in a box. But here, we make things that matter. When I drive around town, I’m looking for something that we made, and it gives you a sense of pride that you make something that matters.’ That is certainly true about this project.”

Square 1

Joe Johnson, nearly 70, is an Aurora machinist who some would call an artist––he retired from his 35-year career as an electrician at Liberty Utilities, formerly known as Empire, and views his second career for the last nine years as a passion and a hobby. Because of his long connection with the power company, he’s been the go-to source for most part replacement work at the dam, including linkage that opens the gates, which is prone to breaking. Meanwhile, Monett Metals had come to lean on Aurora fairly regularly to flatten out bowed metal pieces, such as end pieces of gear boxes, in the machine shops’s 100-ton hydraulic press. The relationship proved fortuitous just when Johnson was mulling over the newly arrived crosshead job, with nothing more than a “broken and busted old cast iron casting” to go on. 

Having made a rough drawing, Johnson first thought he’d remake the four crossheads as weldments using flat plate and tubular stock. But the execution would be far from simple, and Monett’s frequent presence at Aurora brought the foundry top of mind for a second opinion.

“It would have been so labor intensive to make the weldments,” Johnson said. “We got to thinking about Monett Metals, so we called over there and visited with them; we set out our drawing and they said they would like to do it. Casting was the cheapest and best option.”

A low-volume, high-mix business that does nobake sand molding, Monett also has an investment foundry onsite to do lost wax investment casting. The company’s sweet spot, said Renfrow, is 10–25 pieces in sand; 20–25 as investment casting. 

“But we do one-offs, which gives us a whole niche; we are doing work that no one else wants,” Renfrow said. “For a typical investment foundry, if it’s not 1,000 parts, they don’t want to look at it. So we’ve got a pretty good niche there.”

The 80-year-old crosshead replacements were just the sort of challenge they like to sink their teeth into, he added. 

Engineering Phenomenon

One complexity seemed to lead to another, said Monett Metals Senior Casting Engineer Jesse Friend, and the biggest challenges fundamentally stemmed from the age and missing data of the 1930s part.

“All the original engineers or anyone who knew anything about how the original casting was made for, has been deceased for decades,” he said. “So being asked to accurately measure a part that is this old, broken, and has no drawings whatsoever, was like figuring out how the pyramids were made.”

His first step in reverse engineering the part, in the absence of a scanner, was to painstakingly measure the entire part feature by feature, developing a 2D drawing by hand. From this he constructed a 3-dimensional model of the casting using Solidworks. Once an accurate 3D model was generated, Friend ran an initial simulation of the existing casting geometry to determine the natural progression of solidification. His initial simulation parameters included:

  • Alloy: WCB
  • Casting weight: 410 lbs.
  • Pouring temperature: 2850 F (The thinnest section of the part was about ¾”.)
  • CFS (Critical Fraction Solid) calculations were set to be calculated at 35% solid. This was an overly critical number, he said, chosen because of the critical nature of the component. Normally, for a large sand casting, 50% solid would have been the suggested starting point, which would have displayed less porosity in the simulation results.  
  • Niyama calculations (where temperature gradient and cooling rate are calculated during a simulation) were set to be performed at 40% solid. 
  • Silica sand was simulated as the mold material.

Some of the features in the internal passageway were inherently difficult to measure, he said. 

“For example, some features appeared to be somewhat spherical, almost egg shaped, and other features I simply couldn’t get a good measurement because I just couldn’t get my hands in there! 

“Adding to the complexity was the fact that the original part was cast iron,” he continued. “Our part would be steel.  When it comes to casting engineering, the geometry that works with cast iron does not work with steel––they are completely different animals and solidify completely different. Right up front, I knew I would have to make adjustments to the design of the internal passageways of the casting.” 

Simulation Saves the Day

His next challenge was determining the geometry that would have flexibility based on part function, and this was hinged on clear communication with Johnson at Aurora Motor and Machine to understand what mattered and what didn’t. 

Deciding how to core the internal passageways was also problematic. In the end, Friend concluded the only way to create them was to create a core within a core, which required modeling the negative space of each core separately, then modeling the corebox around it, making sure everything was drafted correctly, he said, so the sand would pull out of the box during molding. 

“Running a quick draft analysis in Solidworks showed that the main internal passageway would require removable pieces in the corebox that would come out with the sand,” Friend said. “You’re almost doing two forms of engineering at once because you have to add draft in such a way that your molders don’t run into issues, while at the same time ensuring that the draft does not cause solidification issues in key areas. I think it was the most involved I have ever been in the actual design of the pattern itself.”

Friend did eight total simulations that drove decisions on pouring temperature and height, risers, insulated sleeves, gating and sprue, runner and “slagtrap,” stainless steel chills, geometry adjustments for ideal solidification, zircon core, and more. The final four crossheads were 20.5 in. long by 15.5 in. in diameter (with 4-flats), weighing 350 lbs.––with a 736-lb. pour weight (48% yield). 

Small Part, Big Role

The crosshead is critical to the opening and closing of the dam’s 36 gates, which are held closed by the river itself. Johnson estimated each crosshead bears up to 1,500 ft. lbs. of torque.

Here’s how it works, from a mechanical perspective: 

Each of the four crosshead castings has a rack/gear interface on its top and bottom. Like a piston, the casting fits inside a sleeve, which on its top and bottom has worm gear apparatuses. By manually operating a yellow wheel handle, an operator can shift the position of the crosshead inside the sleeve/tube thanks to the racked features on the casting.

The crosshead is attached to two connecting arms protruding out of the green sleeve/tube, and each arm is attached to a downward shaft that goes down a few floors––the arms activate the gates that open and close like a Venetian blind to provide water flow to the generator waterwheel and also regulate the speed of the generator in proportion to the load. Each gate resembles an airplane wing, tapered to a point at one end. 

Over time, the dam’s original equipment has been updated, retrofitted and digitized. Hydraulic rams have been added to one end of each crosshead so they are controlled with hydraulics, but they can still be manually adjusted. Today, a bar attached to the crosshead digitally sends a signal between crossheads and gates. In the dam’s control room, old, original dials remain side by side with a new, state-of-the-art control system that displays real-time data.

Getting Creative With Machining

When the original crossheads were cast over 80 years ago, they were likely machined with dial indicators and hand wheels on the machines; machinists may have used a fly cutter to cut the gear racks, Johnson speculated. Nothing was a standard size, he said––off just enough to “mess things up.” For example, the bore hole through the center and the cross hole were 55 thousandths smaller than the standard size, he said.

“So we bought a 4-in. reamer,” he said, “and we did business with Southwest Grinding in Joplin, Missouri, so they sized that 4-in. reamer to finish the hole at the correct size. I also took one of the old crossheads over there and they ground us some cutters to match the gear rack on the old crosshead.” 

Back at Aurora, cutting each gear rack down was a six- to seven-hour manual machining job, but another problem had to be confronted in order to do it. 

“We did not have a machine in the shop that we could do this with,” Johnson continued. “So we bought a 1950s jig mill to reach the length that we needed to reach, to cut the gear racks and to make the bores. I guess that tells you how unique this project was.”

Found on the internet in Cincinnati, the 24,000-lb. antique mill was a $10,000 investment, and the only button pushing Johnson did for his work was on the one to activate a digital readout to set his spacing. “Everything else was manual,” he said. 

“Once we got that done, everything else was standard tooling,” Johnson added. “It took about two weeks per crosshead to machine it all.”

In May 2021, pattern construction was completed, and the first casting sample was approved in August––the finished, machined part was installed at the dam by Johnson. A month later, the remaining three crossheads were completed then installed at the end of September, and Renfrow was glad to be part of the “backstage tour” and hands-on experience of putting his company’s work into place as part of the dam’s continued legacy.

“Now that we’ve been successful with this and other hydroelectric dam work, we’ve become part of a Rolodex of providers,” Renfrow said. “But this one’s uniquely different. Because we hit a home run and the guys at the dam were excited and thankful, I think it will open doors for us to be pulled in on more of this type of work.”    

Click here to view the article in the January 2022 digital edition.