Aerospace Engineer Proves Staggered Engine Start Theory
Those watching a Delta IV Heavy vehicle launch from Vandenberg Air Force Base in January 2011 noticed something was not quite right when a fireball engulfed the rocket during liftoff, setting insulation on fire as the launch vehicle cleared the tower.
This was the first Delta IV Heavy launch out of Vandenberg, and the fireball took everyone by surprise.
“When we launch these rockets, failure is very costly,” said Dr. Ejike Ndefo, director, Fluid Mechanics Department. “The burning insulation stayed on fire for a very long time. That was the first time we had seen this behavior.”
Although the launch went off without any further complications, it was a problem that needed to be solved to prevent a repeat during the next Vandenberg Delta IV Heavy launch.
“The concern was for the thermal issues,” said Dr. Jin Wook Lee, engineering specialist, Aerothermal Analysis, Fluid Mechanics Department. “The avionics, structure, and mechanical components on a rocket are vulnerable to heat. Insulation material, once burned, releases particulates that could contaminate these multibillion-dollar satellites we are launching.”
Lee requested and was awarded an Independent Research and Development project to develop a capability that can investigate a very large-scale combustion problem. The ultimate goal of the project was to verify a previously suggested theory – a staggered engine start as an ignition plume mitigation option.
He realized the fireball reached the heat and height it did, not because of a flaw in the rocket, but as a result of the design of the Vandenberg launch pad, Space Launch Complex-6.
The SLC-6 was originally built to launch the Titan III, and later modified to launch Air Force space shuttle missions and the Titan IV. This is why the design of the pad is different from its counterpart pad at Cape Canaveral Air Force Base, SLC-37B. These two launch pads are designed with trenches that collect hydrogen expelled during the preparation for engine start. This gaseous hydrogen ignites in the trenches and is vented out through duct openings.
Lee said there could be several reasons the plumes from the first Delta IV Heavy launch out of SLC-6 were more intense than the ones that occurred out of SLC-37B. The most likely is that the larger pad window and deeper ducts of SLC-6 can vent out a larger volume of ignition plume, which then spread to encompass the vehicle surface. Deeper ducts allow more hydrogen gas to collect under the launch table, enabling more combusted gas to rise, creating the fireball effect.
Some plumes were expected and planned for. This plume, however, exceeded the expected heating levels when the flaming hydrogen was forced up through the pad window instead of out through the vents, pushing the flames up far higher.
To stagger an engine start, one engine fires alone, then the other two fire moments later simultaneously. The suction created by the first engine start would hopefully pull the hydrogen from the next two engines down and out instead of up.
“Changing the engine start sequence is a big deal,” Lee said. “It affects a lot of people – propulsion, acoustic environment, structure, and even guidance. It’s a big decision to make. My job was to ensure this was the right decision.
He suggested using computational fluid dynamics software to make accurate calculations and simulation of a staggered engine start launch.
Lee had one problem though. There was no way for him to accurately simulate such large-scale combustion with the computational tools at hand. He reached out for help from Professor Ed Luke, Mississippi State University, and Dr. Rex Chamberlain, Tetra Research Corporation.
“One role of the FFRDC is to leverage the expertise of academia and national laboratories,” Lee said. “We developed the capability based on their tools.”
The team set themselves a definite two-year deadline in which to perform a successful CFD simulation, before an August 2013 national security launch.
“They had to model everything; the vehicle, the pad, how hydrogen would be affected, et cetera,” Ndefo said. “That process takes a large computer a lot of time.”
The team requested and was granted high priority use of the Department of Defense high performance computing resources. They ran three separate simulations, two using the DoD computing center and one using Aerospace technical computing, to be sure of the eventual real-life results.
It took 512 central processing units four months to complete each simulation.
“With one computer, it would have taken decades to do these calculations,” Ndefo said.
Lee was able to see his work in action when a staggered engine start was used for a 2013 national security launch, with complete success.
“That flight proved how accurate the work was,” Lee said. “The simulations predicted very accurate heating levels, and the numbers matched with the flight data. All the hard work paid off.”
“The customer really pressed us, asking ‘Is it going to be OK?,’” Ndefo added. “He (Lee) worked very hard and got us comfortable for the launch.”
While the project is complete for the West Coast, Lee’s work is not finished yet. He is currently running similar calculations for staggered engine starts on SLC-37B launch pad. He said it is widely beneficial to have a consistent launch procedure across the board. He is currently on schedule to complete his simulation prior to the first East Coast staggered engine launch in January 2015.