"From Gas Turbines to Tornadoes," ME Online Web Exclusive, April 3, 2003

for 4/03/03

From Gas Turbines to Tornadoes

by Joseph T. Hamrick

Results of gas turbine research may be applied to the understanding of how tornadoes are formed. Despite many years of research by meteorologists in the National Oceanic and Atmospheric Administration, the method by which tornadoes are spawned remains a controversial subject. Efforts in the 1980s by the National Oceanic and Atmospheric Administration to understand tornadoes met with little success.

Research at the National Advisory Committee for Aeronautics Lewis (now NASA Lewis) Flight Propulsion Laboratory on air flow behavior in gas turbines coupled with cyclone flow research performed at the Iowa State University provides insight in flow behavior that may lead to a better understanding of how tornadoes are formed. It is the purpose here to discuss the contribution that gas turbine research can offer as to the understanding of the generation of tornadoes and a possible solution to reducing their intensity.

In 1984, the National Oceanic and Atmospheric Administration made an effort at understanding the birth of a tornado by use of P-3 airplanes and radar to probe live storms. The prime goal of the project, which was led by Peter Ray, was "to relate large scale patterns of air circulation in a storm system to the formation of the intense vortex of a tornado." In a telephone conversation with Walter Sullivan of the New York Times, Ray stated, "In spite of many years of research, this relationship is still unknown."

A publication entitled "Tornado" provided by the U.S. Department of Commerce and the National Oceanic and Atmospheric Administration attributed tornado formation to "forces set up by the imbalance created when cool air overrides warm air. The imbalance is compensated by upward movement convection from the lower layers of warm air, which becomes a rotary flow and forms the tornado vortex."

The vagueness of the explanation of tornado formation prompted many aero-dynamicists to step forward and promote research into the aerodynamic aspects of formation. Aerodynamic research into the mechanism of tornado funnel formation was subsequently funded by the National Science Foundation and the State University Foundation for research at Iowa State University. A report on the research was published in "The Physics of Fluids" in 1976. A major finding of the experimentation involved the ground effects on a vortex formation. The report states "It has been shown by velocity measurements that a highly concentrated vortex is formed due to the interaction of an atmospheric vortex with a ground plane. It is surprisingly interesting to observe that this vortex cannot be formed with the removal of the ground plane."

Some years earlier in the 1950s, gas turbine research in part at the NASA Lewis Propulsion Laboratory in Cleveland was devoted to determining the effects of secondary flow in centrifugal impeller passages, turning elbows, and cascades of axial flow compressor passages. A review of the research results led to the conclusion that aerodynamicists engaged in gas turbine research could make a contribution to the explanation of tornado formation.

The research at Lewis Laboratory was concerned with effects of friction that occurs between the air and the plane walls for flow being forced to turn in an elbow or cascade of blades. At the wall, the velocity of a particle touching the wall is effectively zero. The frictional effect is such that there is a zone in which the velocity of the air increases from a low rate near the wall to the velocity of the free stream. The air in that zone is referred to as the boundary layer. The static pressure across the entire boundary layer is the same as that in the free stream, a situation that gives rise to cross flows in the boundary layer near the plane walls of an elbow or cascade of blades. The direction of these flows in a cascade of blades is illustrated in Figure 1, which was taken from a NACA report by Hansen, Herzig and Costello.

Figure 1 - Secondary-flow components in annular nozzle cascade.

The explanation for formation of these boundary layer cross flows is that in the turning flow the free stream velocity is that in the turning flow, the free stream velocity is lower at the concave face than at the opposing concave face to the convex face. As a consequence, there is a gradient of decreasing pressure from the concave face to the convex face. In the free stream, centrifugal force offsets the pressure differential between the two faces and there is no cross flow. However, at the plane walls where the velocity in the boundary layer is low, there is insufficient centrifugal force to offset the pressure differential, and some cross flow occurs as illustrated by the arrows at the walls in Figure 1. These flows are commonly referred to as secondary flows. Further, the variation in through flow velocity in the boundary layer results in proportionate variations in cross flow velocities. These variations give rise to formation of vortices in the cross flows. The rapidity with which these cross flows move from the concave face over to the convex face is shown by the smoke traces in Figure 2 (from report by Hansen, et al), which shows four vortical streams of smoke moving from the concave face to the convex face in the boundary layer.

Figure 2 - Flow deflection in wall boundary layer of one passage.

In experiments with an elbow that were conducted under the direction of J.D. Stanitz at the Lewis Laboratory concurrently with the cascade experiments by Hansen, et al, a similar result was achieved. Smoke traces resulted from injections into the boundary layer near the plane wall of the elbow. To evaluate effects of boundary layer thickness on vortex formations in the elbow, spoiler screens were placed on the walls ahead of the plane surfaces of the elbow. Spoilers of varying height were used to determine the effect of a thickening boundary layer.

Flow measurements showed that thick boundary layers produced less distinct vortex formations. Stanitz concluded, "As the inlet boundary layer on the wall increases, a rather sudden difference occurs in the secondary flow pattern, perhaps associated with the reduced importance of viscous effects because of the smaller velocity gradients in thick boundary layers."

For a mesocyclone in contact with the Earth, the boundary layer at the Earth's surface produces continuous formations of vortices that move to the central core of the whirling air mass in the same manner as shown for the cascade and elbow. Upon reaching the center of the mesocyclone, the vortices merge, precessing upward and transferring their vorticity to the core vortex. It is deduced that the constant feeding of these boundary layer spawned vortices into the central core produces a tornado. Thus, the strength of a tornado is logically dependent upon the strength of the mesocyclone from which it is spawned and the configuration of the terrain over which it has traveled.

The dependency upon the shape of the terrain is well demonstrated by occurrences in the Appalachian Mountains. While all of the conditions, i.e., penetration of warm air by a cold air mass and forming of mesocyclones, are in place for formation of devastating tornadoes, they seldom occur. Instead the thick boundary layers formed at the mountainous interface between the mesocyclone and Earth so emasculate the secondary flow pattern that a strong central core does not emerge. While strong cyclonic wind patterns develop in the mountainous areas and often uproot large trees and damage buildings, the phenomenon is seldom credited to cyclonic action because of the absence of an identifiable central core or evidence of a twisting pattern in the vegetation.

If the foregoing mechanism of tornado formation is correct, the question arises as to what, if anything can be done to reduce their intensity. At this point, it appears that there is the possibility of altering mesocyclonic boundary layer formations at the Earth's surface in such a way as to reduce the intensity of tornadoes. It is shown clearly in the results obtained by Stanitz that thickening the boundary layer causes as significant change in the vortices generated in the boundary layer of a turning air stream. Further, the absence of strong tornadic formations in mountainous terrain suggests the possibility that altering the boundary layer in level areas can reduce tornado intensity. Planting of randomly located forests of tall-growing, deep-rooted trees or creating earth barriers at strategic locations should achieve some reduction.

Joseph T. Hamrick is an ASME Life Member. He is president of Aerospace Research Corp. in Roanake, Va.