Dick Keyt wouldn’t look out of place on a horse next to Clint Eastwood in a Western. But planes, not horses, are Keyt’s obsession — one in particular.
Rangy and grey-haired, the softly-spoken Texan is the owner — or, as he puts it, custodian — of the Polen Special. Handbuilt in the late 1960s and early 1970s, the plane is an airshow legend.
It has been the fastest four-cylinder aircraft (in fact, the fastest four-cylinder engine) in the world for over 30 years. Keyt is determined to keep it that way, and he has decided to use the most up-to-date technologies to achieve this.
US Air Force veteran Keyt — a former commercial pilot and aeronautical engineer — bought the Polen from its builder, Dennis Polen, in 1998. ‘Dennis started building in 1968 with two friends — the idea was that they would build a plane each,’ he explained. ‘But the other two dropped out, leaving him to complete the project.’
The plane is based on an early 1950s version of the classic US World War II P51 Mustang fighter. ‘There was a very tiny plane called the Midget Mustang; make it a bit larger and you’ve got a Polen Special.’ When Polen completed the plane in 1973 he was ahead of his time. ‘Then, most handbuilt planes were steel tube and fabric, and here was this all-metal craft.’ Initially, Polen was so keen to show off his creation, he took it to a major airshow and penned an article for the
Part of the legend was due to the striking looks of the plane — streamlined, all metal and painted screaming red and silver. But the rest was the performance. Powered by a 200hp four-cylinder aero engine, the plane is capable of 325mph, and is claimed to outstrip every other aircraft in its class. Or at least, that’s what Keyt thought.
Polen suffered a stroke in the mid-1990s, leaving him unable to fly his plane. Rather than letting it fall into disrepair, he decided to sell it to someone who would be able to maintain it and improve its performance. And that’s where Keyt and his wife Debbie, also a qualified pilot, became involved. ‘When I got the plane, my total knowledge of it were the magazine articles,’ he said.
‘But what is interesting is that the performance figures [in Polen’s original article] are all bogus.’ Virtually every figure, from speed, right down to fuel tank size, didn’t match the article, said Keyt — and Polen has never told him why.
Having discovered this, Keyt decided to devote himself to restoring the plane to its fabled performance.
‘My initial goal was to show it off, get Dennis some of the recognition he’d never really had, and get the performance back to where it should have been. But I had to go through a real education.’
The plane had originally been fitted with a turbocharger, but Polen had removed this because, he said, it made the plane less fun to fly. Keyt decided to restore it, but on removing the cowling, he realised it would be no easy task. ‘The technology was obviously 1960s — the cooling drag, particularly, was very bad, and I could see that this was where most of the improvements could be made.’
Keyt contacted some friends at the
To obtain the best performance from the engine, Keyt wants to reduce cooling drag as much as possible and ensure that the turbocharger is working at peak efficiency.
Cooling drag is caused by the movement of the air drawn into the engine over its cooling fins. Efficiency is related to the speed of the air drawn into the turbocharger. In both cases, the design of the cowling, how the air moves over it, and the placement and shape of the air intakes, is crucial. This, said Keyt, has dictated how his project proceeds.
Before Dennis could start work on the CFD plots, he needed a digitised image of the cowling. For this, Keyt turned to Leica Geosystems, which had asked whether it could demonstrate its T-Scan laser tracker on the plane as a demonstration for Lockheed Martin.
‘Lockheed wanted to find out about Leica’s scanning technology,’ Keyt explained, ‘but couldn’t do the demonstration at Lockheed because so much there is classified. So we took our cowling there for the demonstration.’
The scanning data provided Dennis with an interesting challenge. ‘I’ve done all kinds of flow simulation over aircraft, but not using actual scanned data from something that’s been manufactured,’ he said. Not surprisingly, there were some difficulties.
Dennis’s first task was to turn the ‘point cloud’ data from Leica into a high-quality representation of the surface of the cowling. This, he said, proved to be an ideal task for a group of undergraduates. ‘I gave it to them as a project for honours credit.
‘What we found was that when we got the data points from Leica, there was no data connection information between the points. It was just a file full of XYZ co-ordinates. What we had to do was reconstruct a surface that would fit through all the data points.’
Fortunately, the surface the co-ordinates described was so smooth that the topology wasn’t a problem — but the amount of data was. ‘We had to get rid of every other point,’ said Dennis, ‘then we triangulated what was left and that gave us the surface.’ The data still contained many points which were obviously errors, so the team imported the surface into a CAD package. ‘The students got rid of the triangles that represented the errors — literally smoothing it out by hand.’
The simulated surface was still too low-quality to be used for a CFD plot, however. ‘We had to go through with some of our own tools to remesh the surface into what we call a high-quality surface triangulation,’ he said. ‘From there we developed a volume grid, set up the boundary conditions and ran it through our CFD code to get the flow over this cowling.’
The results are already proving useful to Keyt. One of his tasks now is to design a plenum chamber which will sit on top of the engine which will seal all the air coming in from the front inlets, which the CFD plot has shown are actually too large. ‘The new chamber will be completely embedded inside the cowling. This will optimise the airflow management in the engine, letting us squeeze more power out of it. The more power I have at my disposal, the harder I can push the plane. This is what reverse engineering is all about.’
Keyt and Dennis are also studying the CFD plot to see where to position two new intakes that Keyt wants to introduce to serve the engine intercooler and the turbocharger. The latter suffered badly from the previous configuration of the engine compartment, he explained.
‘The air came in from the front intake and was blocked off by a baffle which fed the air into the turbocharger. But the right rear baffle was above the turbocharger, so it was fed down through an S-shaped duct.’
Effectively, this meant that the turbocharger had to suck air in, which is very inefficient. ‘What I want to do is replace that with an inlet on the right side of the cowling that will let air directly into the turbo. That will give us a ram effect to increase the air pressure and the efficiency.’
Ramming the air in will reduce the discharge temperature of the turbocharger, which will also give Keyt more power. ‘Right now, the intercooler cools the air coming out of the charger from 150oC to 93oC,’ said Keyt. ‘This is very thin air, with comparably little oxygen to burn. The cooler we get the air, the denser it gets, increasing the oxygen content significantly.
‘And the more oxygen there is to burn, the more fuel you can pump into the combustion chamber, which in turn increases the horsepower output. We hope to be able to drop the air temperature further to about 65oC.’
Keyt is currently rebuilding the Polen’s undercarriage after a landing mishap, but once this is complete, he hopes to start work on a new cowling. This will be a composite structure, rather than the riveted metal sheets of the original; currently, the position of air intakes is limited by the position of the edges of the sheets.
‘Polen used some composites on the control surfaces when he worked on the plane, so it isn’t betraying the spirit of the thing,’ said Keyt.
Leica has now completed a digitisation of the entire craft, which Keyt hopes will allow him to make further improvements. Dennis is looking forward to starting work on this data, but the sheer number of points is proving to be difficult. ‘The increase in the amount of data is so large that even opening it up to visualise the points is beyond the capacity of our workstations,’ he said, ‘so we’re having to develop programs to go into that raw data for a reduction so there’s something we can at least open, and then reconstruct the surface.’