Axial flow turbines

Details Axial flow turbines Axial flow turbines

2000-08-25

2002-07-09

Lopez, F. Daniel (Department: 3745)

Fluid reaction surfaces (i.e., impellers)

Specific blade structure

Irregular, flanged or channel forming blade surface

C415S914000

Reexamination Certificate

active

06416289

ABSTRACT:

FIELD OF THE INVENTION
This invention relates to axial flow turbines, whether driven by gas or steam, of the reaction type.
BACKGROUND OF THE INVENTION
Reaction type turbines have aerofoils with profiles that cause acceleration of the working fluid along at least a leading region of the suction surface of each aerofoil. The flow over that leading region is laminar but, depending upon the Reynolds number of the aerofoil airflow, the boundary layer further downstream may undergo a transition to turbulent flow and/or there may be transition or separation bubbles formed, which can result in large energy losses.
It is known to give aerofoil surfaces in a turbine, in particular in the low pressure output stages of a gas turbine, a transition-promoting configuration to reduce these losses. In U.S. Pat. No. 4,822,249 (Eckardt et al) a continuous spoiler edge is located closely behind the point of maximum surface velocity on the suction surfaces of the blades of a turbine wheel and extending over substantially the entire radial length of the blades. The function of the spoiler is to promote rapid transition from laminar to turbulent boundary layer flow on the suction surfaces without the formation of laminar separation bubbles. In GB 580806 (Griffith) it is proposed to roughen the entire aerofoil suction surface of reaction type blading for compressors and turbines so as to produce a very thin layer of more or less uniformly disturbed flow over that surface without disturbing the main flow beyond the boundary layer.
The topography of a surface with roughness is complex and there is no single definitive measure of roughness. A widely used basic perimeter is used “average roughness” (Ra), defined as the arithmetic average of the absolute values of the measured profile height deviations of the surface from the surface profile centreline within a given sampling length. This definition is also valid for previously used alternative terms “arithmetic average roughness” (AA) and “centreline average roughness” (CLA). Typical values of Ra for turbomachinery components are 125 microinches (3.2×10
−3
mm) for material as cast and 25 microinches (6.3×10
−4
mm) for polished components. Thus, in GB 580806, it has been proposed that the required roughening of the suction surface can be achieved by using sand-cast blades which are not given any polishing or smoothing treatment. In U.S. Pat. No. 5,209,644 (Dorman) which also proposes roughening aerofoils in the output turbine section of a gas turbine that operate in the range of exit Reynolds numbers of 80,000 to 200,000, the surface roughness is in the range 120 to 200 AA microinches (3×10
−3
to 5×10
−3
mm). Here the roughening is applied to both the suction and pressure surfaces over the entire chord and over most of the aerofoil span and is intended to reduce separation of the boundary layer and formation of recirculation zones or bubbles in the boundary layer.
In the case of modern low pressure turbine blading which operates at low Reynolds numbers (e.g. 70,000-250,000) with highly loaded aerofoil sections, the formation of boundary layer separation bubbles towards the rear of the suction surface cannot be avoided. Steady flow design methodology focusses on ensuring that transition occurs within a bubble, causing it to reattach to the surface as a turbulent boundary layer before the aerofoil trailing edge.
FIG. 1
shows the steady flow isentropic velocity distribution in an annular blading row for a conventional aerofoil (the diamonds plot) and for a high-lift aerofoil (the circles plot) having a lift coefficient approximately 20% greater than the conventional aerofoil. The ordinates are normalised velocities, that is to say, the ordinates are given by the ratio of local flow velocity to exit velocity, and the abscissae are chordal distances normalised as a fraction of the aerofoil chord length measured from the leading edge. The upper pair of plots are for the suction surface and the lower pair are for the pressure surface. A feature of the velocity distribution of high-lift aerofoils, such as that shown in
FIG. 1
is the continuous acceleration of flow on the suction surface over the region A from the leading edge to a peak velocity point B typically downstream of the geometric throat in the blade row which will be located at C. The suction side boundary layer is laminar all the way to the peak velocity point B. Deceleration begins after the peak velocity point and in steady viscous flow the boundary layer separates shortly after the start of the deceleration, forming a separation bubble which shows as a plateau up to transition point D. Because transition is reached, the separation bubble reattaches before the trailing edge, resulting in a sharp pressure recovery.
The effectiveness of the high-lift aerofoil design relies on the reattachment of the bubble before the trailing edge because an open separation bubble gives very high losses. Reattachment of the bubble should not occur too early, because that allows unwanted growth of a turbulent boundary layer on the final region of the suction surface which also increases losses.
The preceding discussion, and prior art examples referred to above which seek to avoid the formation of separation bubbles, are all based upon a consideration of aerofoils operating in steady flow. However, in the typical turbine the flow is not steady. There is interaction between succeeding aerofoil rows because the wakes from one row will impinge periodically on the aerofoils of the succeeding row.
A comprehensive review of researches on wake passing effects on separation bubbles is given in “Blade Row Interactions in Low Pressure Turbines”, H P Hodson, von Karman Institute Lecture Series 1998-02, Blade Row Interference Effects in Axial Flow Turbomachinery Stages (1998). The Hodson study considers high-lift, low Reynolds number aerofoils which have been developed for low pressure turbines in order to reduce weight and manufacturing costs. As already mentioned, these aerofoils have regions of significant deceleration on their suction surfaces which can result in the formation of substantial separation bubbles in the absence of wake-passing effects.
The interaction between succeeding aerofoil rows in an axial flow turbine is shown schematically in
FIG. 2
(Binder et al) which illustrates how the wakes are transmitted and distorted through downstream rows. The wakes W leaving a first rotor row 2 are chopped by the following stator row 4 and the chopped segments W′ of the original wakes are further distorted in the flow through the stator row. The dashed lines 6 indicate the stator wakes. Relative to the moving aerofoils of the following rotor row 8, the wake segments are arranged in avenues 10, (indicated by the chains of circles) and if the second rotor row 8 has a different number of blades from the preceding rotor row 2, the respective avenues of wake segments will enter the second rotor row 8 at different phases to the blades of the row.
As is discussed in more detail in the Hodson study, turbulent flow appears in turbomachines typically by bypass transition because of the high levels of turbulence that exist. In this process at points within the boundary layer some distance from the leading edge turbulent spots can form and spread downstream and laterally. Immediately following the rear of a turbulent spot a calmed region is formed having laminar-like characteristics with a very full velocity profile, with a trailing edge travelling at about 30% of the freestream velocity. The unsteady flow of passing wakes can initiate this mechanism to have a beneficial influence on profile losses.
The flow pattern at an optimum wake-passing frequency is illustrated in
FIG. 3
which is a space-time diagram of the flow over an aerofoil in which the distance along the aerofoil chord from leading edge to trailing edge is given along the abscissa axis and time values (t/&tgr;) given along the ordinate axis have been normalised by the period of wake passing (T) over the aerofoil.

Affiliated with

Also associated with

Rating

Axial flow turbines does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Axial flow turbines, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Axial flow turbines will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2855796