|
Wear in Lean-Phase
Pneumatic Conveying Pipelines
Guest article by Richard Farnish, Consulting Engineer
Pneumatic conveying pipelines are used
throughout industry as an efficient means of transporting materials
between storage points and processes.
Lean phase exists where a
pipeline is being operated with a high air flow rate which induces
sufficient turbulence that particles become entrained in the air, and are
conveyed along the pipeline at a relatively high velocity (typically 15 to
18 m/s is sufficient as a minimum conveying velocity). These systems
operate at a relatively low pressure, and have a low capital cost
associated with their installation. However, running costs tend to be
relatively high as a result of the level of power consumption and
maintenance required. Wear of pipelines is often the biggest headache with
lean phase systems, which is caused by the high impact velocities of
particles against the pipe walls (also creating breakage of the particles
if they happen to be friable).
The Wolfson Centre for Bulk
Solids Handling Technology, has recently undertaken a programme of
research to develop a more effective means of predicting bend life than
has until now been available to engineers. In order to undertake an
analytical study upon which to develop improved numerical techniques for
predicting bend life, it was necessary to construct a rotating disc
erosion tester in which samples of wall material could be subjected to
controlled levels of erosion. The apparatus which was used at the
laboratory was constructed to hold ten samples of wall material, which
could be mounted such that the angle of particle impact could be fixed to
any desired angle (see Fig.1). Similarly, variables such as particle
velocity and intensity of particle impacts per unit area could also be
finely controlled.
Figure 1: Schematic
view of the rotating disc accelerator erosion tester.
Tests were initially
carried out using a structural mild steel for the targets and olivine sand
as the erosive medium. The results obtained from the initial trials
clearly illustrated the importance of the particle velocity vector
(direction, i.e. angle of impingement and magnitude), and particle
concentration on the amount of erosion damage caused (see Fig. 2)
Figure 2: Results
of erosion tests on mild steel using olivine sand in the rotating disc
accelerator.
A simple power law type
model was developed to predict the erosion damage that occurred for these
given conditions,
E = a Vn
where V is the particle
velocity magnitude in m/s and both a and n can be shown to vary in a
predictable manner with angle of impingement. (The full discussion of this
model can found in “An investigation of the low velocity / low
concentration solid particle erosion of structural mild steel using a
centrifugal erosion tester”, 5th International Conference on Bulk
Materials Storage Handling and Transportation, The University of
Newcastle, Australia, 1995. ISBN 0-85825-627-4 pg 191-198.)
In order to obtain
supplementary data for the numerical model a pneumatic conveying test loop
was constructed. This pipeline was constructed from 50mm bore mild steel
pipe and ran for a length of approximately 45m. A complex abrasive feed
mechanism consisting of a pressure vessel, screw feeder and eductor were
used to accurately meter the flow of material into the test loop. As in
the case of the erosion tester, olivine sand was used and replaced after
each test run in order to eliminate the effects of particle degradation
which would inevitably occur during one pass through the test loop.
The key element of the test
loop was a 90° bend (having a r/d ratio of 14) which was situated midway
between two 10m lengths of straight pipe (the equal lengths of pipe
serving to ensure that the stream of entrained material striking the bend
would not be influenced by any turbulence set up by the preceding bend and
the section after the bend to minimize the influence of any back pressure
after the bend). The test bend was machined the enable the placement of an
ultrasonic thickness transducers at 45 points, which would enable data
indicating the localized loss of wall material to be obtained. (see Fig.
3).
Figure 3: Photograph
showing the location points for the ultra sonic transducers
The processed data obtained
from these measurements appears in Fig 4 below. It is important to note
that although the chart does not clearly indicate the point of puncture,
this can be attributed to the fact that puncture occurred between two
transducer mounting points and therefore was not detected.
Figure 4: Bend
penetration results at puncture area.
The data was accumulated
over 22 separate test runs during which time 5.5 tonnes of olivine sand
was conveyed through the system. Observations made as a result of this
test programme indicated that puncture occurred in the region where the
particles struck the pipe wall for a second time and not at the point of
primary impact. A large amount of material wastage was found to occur in
the primary impact area, but because of the large area over which the
erosion occurred, puncture did not occur.
The conclusions drawn from
this phase of the research were:
Rebound of the particles
from the point at which they first strike the bend wall tended to induce
the particles to move towards the axis of the pipe bore. This behaviour
causes an increase in the intensity of particle impacts in the region of
secondary impact. Therefore accelerating the rate of puncture in this
region.
Inter-particulate collision effects in
the region of primary impacts tend to reduce the amount of erosion
damage as the distance around the bend is increased.
Owing to observations
that the majority of particles are in the lower half of the pipe bore,
wear of the primary impact area tends to occur in the lower section of
the pipe.
Because the majority of
the mass of particles is in the lower half of the pipe bore the momentum
of these particles tends to be greater than those in the upper half of
the pipe bore. This may have the effect of causing the puncture point in
the region of the secondary impact location to occur in the upper half
of the pipe bore.
It was subsequently
observed that the application of the model shown in equation 1 did not
result in an accurate prediction for the degree of bend wall penetration
that occurred. Primarily this was attributable to the model not taking
into account the particle dynamics in the conveyor bend, by not accounting
for the build up of particles at the pipe bore centre line in the region
of the secondary impact position. In order to refine the model it was
necessary to determine a factor to emulate the dynamic effects of the
particles, this undertaking formed another aspect of the investigation.
When the wear pattern
within the failed pipe bend was viewed it could be seen that the ripples
formed by the impacting material in the bore a resembled the curve of
aberration for a concave cylindrical mirror (see Figs. 5 & 6).
Figure 5: Photograph of the worn out conveyor bend
Figure 6: Schematic diagram illustrating the key features of the curve of aberration
Use of a multiplication factor based upon
the area beneath the curve of aberration as a correction factor to account
for the increase in the intensity of particle impacts in the area
surrounding the puncture point. The form of the equation used to predict
the wear life of the pneumatic conveyor bend under consideration was
refined into the following form:
Mass conveyed = W A
k E cos((p/2)-f)
where E (mm3/kg) is the
erosion rate predicted using the power law erosion model, f is the angle
of impingement, k is the multiplication factor derived from the geometry
of the curve of aberration, W is the pipe bend wall thickness (m) and A is
an arbitrary area of the pipe bend wall that is under consideration (m2).
The use of this model
indicated that puncture of the bend would occur after 6.08 tonnes of
material had been conveyed - which was an 11% overestimate. It is
suggested that any inaccuracies in this prediction can be reduced by
accounting for other effects including inter-particulate collisions.
However, even with the existing degree of error the model offers a
substantial improvement on previous predictive methods.
A comparative study using
five different materials was undertaken to obtain comparative data on
wear. The materials used for this stage of the research were: mild steel,
cast basalt, alumina ceramic, nitrile rubber and ultra high molecular
weight polyethylene. The erosion of all of these materials is dependant
largely upon the velocity vector, angle and magnitude of particle
collisions. In general terms ceramic materials tend to exhibit larger
amounts of wear at angles of impingement approaching 90°, while metals
and thermoplastic polymers exhibit most wear at oblique angles of impact
(i.e. 10° - 40°). See Figs 7 & 8.
Figure 7: Results from
the erosion at 30° impingement angle
Figure. 8: Results for
erosion at 90° impingement angle
The assessment technique
that has been covered by this article offers a promising way of quickly
assessing the wear life of a radiused pneumatic conveying bend. It should
be noted that the following provisions need to fulfilled if the technique
is to applied effectively:
The wear behaviour of the
material combination (abrasive and wall material) must be known or
measured,
The particle size is such
that the individual particles are not dramatically affected by the gas
flow (i.e. not extremely fine).
The Wolfson Centre for Bulk
Solids Handling Technology
at the University of Greenwich, London, UK
Telephone: +44 (0)20 8331 8646,
Fax: +44 (0)20 8331 8647,
Email: R.J.Farnish@gre.ac.uk,
Website: http://www.gre.ac.uk/directory/wolfson
Help others by posting your comments, suggestions and
experiences with bulk solids feeding or any other materials handling concerns you may have
on our On-Line Help Forum. For
past Ask Joe ! Articles, visit the Ask Joe! Archived Articles.
Guest articles for the Ask Joe! Column are always welcome,
for more information please contact Joe Marinelli directly at his email address:
joe@solidshandlingtech.com.
|
