Richard Heinrichs holds BS degrees in electrical engineering and physics and an MS in electrical engineering from M.I.T. (Cambridge, MA). His master's thesis research involved stimulated Raman scattering in hydrogen. He also has a PhD in physics from the University of Massachusetts, where he studied superfluid He films. He was a postdoctoral researcher in the physics department at U.C. Santa Barbara, studying nonlinear fluid dynamics, until the fall of 1986, when he joined the technical staff of M.I.T. Lincoln Laboratory. As part of the Laser and Sensor Applications group he has worked on a variety of projects, including nonlinear optics for atmospheric turbulence compensation, optical pumping of mesospheric sodium, optical detection of wake vortices, and angle-angle-range imaging lidars. He was interviewed by Frederick Su.
Why do we need this technology?
The primary reason for developing this technology and doing these studies has to do with increasing airport capacity during instrument landing conditions--that's inclement weather with poor visibility. Every plane, when it flies, will generate a pair of counter-rotating horizontal tornadoes called wake vortices. When planes fly close behind one another, wake vortices can be a hazard. If a small plane flies into the wake of a much larger plane, the wake can actually flip the small plane over. In fact, there have been several fatal crashes in the last few years due to small aircraft encountering the wakes of larger leading aircraft.
Right now, during instrument landing conditions, the FAA mandates separations for aircraft based on their size according to three categories: heavy aircraft, such as the DC-10s and the 747s; large, which run the gamut anywhere from 757s down to Lear jets; and small, which are the Cessnas. The mandated separation is six nautical miles for a small plane behind a heavy. When planes are landing, that's about three minutes. If you have a system to monitor wake vortices and that enables you to predict short-term behavior, say half an hour, then by also using some predictive capability of the meteorology, you can possibly reduce these spacings without compromising safety.
I should qualify this a bit. In the United States, under visual flight rules--or good visibility conditions--there are no real wake vortex spacings. It's simply dependent on the pilot of the following aircraft to maintain a safe spacing or, especially for small aircraft, not fly below the glide path of the leading aircraft. But, under instrument landing conditions it's really the controller's obligation to maintain these spacings. In that case, all the aircraft fly one behind the other. That's really where the increase in capacity could be achieved.
The Lincoln Laboratory effort in this program has been, so far, a measurements program in support of the NASA-Langley Terminal Area Productivity Program. Our goal is the development of a system that automatically spaces planes based on wake vortex hazard and predicted meteorology. As part of this measurements program, we have constructed and fielded a wake vortex lidar as well as a whole suite of meteorological instrumentation.
How does this technology work? What kind of laser? What are you bouncing the beam off?
This is a coherent laser radar. It's a cw CO2 laser at 10.6 *m with 20 watts of power. At that wavelength, we basically scatter off aerosols--dust particles, small water droplets, whatever happens to be out there. In an airport environment, you have a large number of particles in the atmosphere because of the aircraft engine exhaust.
We send out the beam and we collect the backscatter with the same 13-inch telescope. Then we mix the return with a nearby second CO2 laser that is frequency locked to the first and offset in frequency by 10 MHz.
What we really measure then is the beat frequency between the backscattered return from these dust particles and our local oscillator laser. That beat frequency varies as the Doppler shift of the particles and, so effectively, by measuring that frequency we measure the velocities of these particles. And since they're entrained in the air, we effectively measure the velocity of the air at that location. Even though this is a cw beam, if we focus the beam, it turns out that our backscatter comes primarily from the focal region. A typical width of this sensitive region around the focus might be about 6 m. at a range of 100 m. It's related to the depth of focus of the system. The width of the sensitive region is proportional to the square of the range. For our particular system, we have effective ranges out to 300 m. We could probably measure vortices farther than that. But 300 m. is a practical limit for this device, which is what works best for small phenomena such as wake vortices.
Does your system work well in inclement weather?
We can operate in the rain if we can keep our scanning mirror dry; we have a special cover designed for this. In fog that is not too thick our signal will increase because the greater backscatter will be stronger than the attenuation of the laser beam at short ranges.
So you're measuring two signals coming back, one from the region of interest and one from a local laser?
Yes, that's right. We're measuring the frequency difference between the backscattered signal and a local reference laser. People have built similar kinds of laser radars in the past; the first one built and used to measure wake vortices was probably over 20 years ago. One thing that makes ours different from the originals is that they used the same laser for their local oscillator. They had basically split off a little of the initial laser beam to mix with the return to get a beat signal between the two. The problem was that you couldn't tell the difference between velocities going away from you and velocities coming toward you. There's a phase shift between the two, but you don't have anything to reference that to. So, effectively, all you see is just the magnitude of the velocity. You can't see the direction. With our system, because of the local oscillator, Doppler shifts from particles that come toward us are upshifted in frequency. Our beat frequency then would be greater than 10 MHz. Particles going away from us would have their scattering downshifted in frequency, and so would be shifted in frequency below 10 MHz. At 10 µm, a 1-MHz shift corresponds to a 5-m/s velocity, either positive or negative.
How far away is the setup from the runway?
We measure wake vortices in three different regimes. The first one is the out-of-ground regime. In that case, we may be 3 to 4 km from the end of the runway, underneath the approach path, and looking straight up as the planes fly overhead. Under those circumstances, the typical approach path is about 20 to 1, meaning that if you're out a kilometer, then the altitude of planes at that level is about 1/20th of a kilometer, or 50 m. At 4 kilometers, the planes are at about a 200-m altitude. The out- of-ground here refers to the fact that the vortices are high enough in the air that they don't feel the presence of the ground.
The other regime is what we call near-ground effect. In that case, we might be 1 km or a little bit less, say, 700 m, from the end of the runway or the touchdown site. Instead of being right underneath, we'd be off to the side, perhaps 50 m, looking at the vortices.
The third location would be at the touchdown site or maybe just 50 m back. We'd be just off to the side looking directly at the vortices. In that case, what we're primarily interested in is the interaction of vortices with the ground, including the generation of secondary vorticity that may result in a subsequent rise of the vortices after they're generated.
Let me take one step backwards. As a plane flies through and generates this pair of counter-rotating vortices, velocity fields are produced that extend beyond the distance between the vortices. The rotating flow from one vortex pushes its neighbor down. And the rotation from the other vortex, in turn, pushes its neighbor down. Consequently, the vortices descend toward the ground. At the ground, you have a boundary condition that says the perpendicular component of the velocity field has to go to zero at the ground plane. You can satisfy this boundary condition by imagining a pair of virtual vortices under the ground with equal but opposite circulation to the real vortices. At ground level, the field from these virtual vortices will affect the real vortices and push them outward. After the vortices split apart at the ground and move outward, secondary vortices can be generated that can cause the original vortices to rise. This was observed by the Germans about five years ago at Frankfurt Airport where they have a concern with parallel runways. We've since observed the same effect.
What's the timeframe for all this happening?
Say the vortex is generated at 200 m above the ground. Descent rates might be 1 to maybe 2 m/sec. So we're talking about 60 to 90 seconds.
Then they rise back up?
In cases where the crosswinds are low, they split apart and can rise back up. This subsequent rise of the vortices does not always happen. Since the generation of secondary vorticity is a nonlinear effect, the actual process may be quite sensitive to details of the local meteorology and even the conditions of the local terrain.
What's the rise rate? About the same?
No. It's far less and they typically just rise back up just a little bit.
If the planes are 200 m high, it's not a real concern. But if the planes are much closer to the ground--50 m, 20 m, and you have parallel runways, then these vortices can come down, go to the side, and rise back up again. There are situations where they can actually go right into the glide path of a neighboring runway. That's a real concern. We've actually measured vortex lifetimes in excess of 2-1/2 minutes, close to 3 minutes. That's long compared to typical time separations for many landing aircraft.
What should be the minimal separation, then, for parallel runways?
At Frankfurt airport in Germany parallel runways are separated by about 500 m (1700 feet). In the U.S. the criteria for independent runway operation is that runways must be at least 3400 feet apart. In cases where the runways are closer together, the arrivals and departures must be synchronized due to, among other considerations, wake vortices.
Tell me about your system.
Another thing that distinguishes our system from earlier wake vortex lidars that have been created is that as we scan and collect the data, a computer processes the data, analyzes it, and recognizes a vortex signal. It will actively track the vortices and maintain the scanning of the laser beam, both in angle and in focus range, to continuously follow the vortex as it moves. In tracking the vortex, the computer has a model of how the vortex should move and can use measurements that the lidar itself has made of the wind to help predict where the vortex is moving. Our system can also measure the vortex strength.
Our lidar measurements of wake vortices are made in conjunction with simultaneous measurements of the atmospheric state. These are collected with a variety of sensors we have set up, including a 150-ft instrumented tower, acoustic profilers, and balloon launches. The goal here is to understand how wake vortex behavior is affected by the local meteorology.
For example, one simple effect of meteorology is if you have crosswinds, winds perpendicular to the runway, they will have a tendency to blow the vortices out of the glide path where they're much less of a hazard.
A few years ago, everybody was talking about wind shear. Is wind shear a natural phenomenon as opposed to wake vortices? Can you tell them apart?
Wind shear refers to a whole class of natural phenomena, the most prominent ones being microbursts and gust fronts. Microbursts are downdrafts of cool air from thunderstorms that spread as they hit the ground and can be fatal to aircraft operating near the ground. The downdraft is caused by the combination of falling wet air and negative buoyancy caused by evaporative cooling of the falling air mass. Gust fronts are the shallow masses of cool air pushed out in front of storms, usually representing regions of convergent surface air and abrupt wind shifts.
Will they create vortices also? And can you distinguish between these and the manmade ones?
Both microbursts and gust fronts can have vorticity associated with them near the ground, but their danger to aircraft lies not in the vorticity that may be present, which is weaker than wake vortices, but in the effects that the rapidly changing winds associated with these phenomena create on aircraft lift. They can easily be distinguished from wake vortices by their length scales. Wake vortices occur over tens of meters, while microbursts and gust fronts occur over hundreds of meters to a few kilometers.
Do you have anything else to add?
We have, so far, fielded our system twice at Memphis International Airport. Memphis was chosen mainly because Lincoln Laboratory already has an integrated terminal weather system site there and because there is a good variety of weather conditions. So we were already collecting a lot of meteorological information at that location. What really makes this measurement program different from what many people have done before is that this has been set up to do behavior measurements of wake vortices as well as the associated meteorology over a variety of weather conditions. People, in the last several years, have come to appreciate that meteorology is very important if you're going to understand how these things transport and decay. We have to understand the effect of meteorology on the behavior of the wake vortices.