|
Issues Impacting Bridge Painting: an Overview
Chapter 7. Task F - Productivity Improvement
Table of Contents
INTRODUCTION
The objective of this task was to apply state-of-the-art
technology with the goal of improving productivity and enhancing
quality control. After discussions with various DOTs, contractors,
consultants, and paint suppliers, it was determined that
the application of sensors to the maintenance, repair, and
replacement of coating systems offered an immediate opportunity
to improve productivity through increased performance of
coating systems. Although estimates vary, it is believed
that over 70 percent of coating-system failures are the result
of improper surface preparation or poor coating application
techniques. It is apparent that improvements in quality control
of the painting application process would reduce coating
failures. In addition, it would also have a major impact
on the reduction of life-cycle costs for the bridge maintenance
system through increased durability of a quality paint application.
This chapter presents brief discussions of several techniques,
including both monochrome and color-visible techniques as
well as near-infrared imaging. The increased availability
of low-cost digital imaging devices coupled with the availability
of rugged low-cost powerful computers make these techniques
practical for their application in the fieldt. The utilization
of sophisticated signal-processing techniques will allow
these imaging devices to provide easily interpreted results
that can be generated by inspection personnel with minimal
technical training.
The techniques described below are global
in nature, allowing inspection of significant areas of
the bridge structure.
They have the potential to provide a quantifiable assessment
of the quality of the paint prior to surface preparation
and, subsequently, the quality of the application of the
paint system. Finally, the examples presented are the results
of application of the techniques in the laboratory on coated
steel panels and work in the field.
DEGREE OF SURFACE RUSTING
ASTM has developed a standard method for the evaluation
of the degree of rusting on painted steel surfaces.(10)
A scale of rust grades ranging from 0 to 10 that correspond
to 100-percent rust for grade 0, and no rust or less the
.01 percent of surface rusted for grade 10. The included
photographic standards -- labeled 4, 6, 8, and 9, which correspond
10-, 1-, 0.1-, and 0.03-percent rust, respectively, were
used in this study. The application of this standard requires
that the inspector judge the rust grade by visual comparison
of the actual bridge with these standards. It is relatively
simple to apply image-processing techniques to a digitally
recorded image of the surface and quantitatively measure
percentage of rusting. This concept was explored using a
monochrome digital camera to record the images of the photographic
standards discussed above. The application of a simple threshold
(assigning 0 or 1 based on gray scale) followed by a count
of picture elements (pixels) could be used to quantitatively
measure percentage of rust. The primary difficulty with this
method is the sensitivity, where a shift in threshold results
from shifts in the gray scale of the obtained image. A better
approach would be to compute the contour of each rusty point
and then measure its size, followed by a total count of the
contoured areas. This approach tends to be more tolerant
to gray scale variations. An example of the results obtained
using the Steel Structures Painting Council (SSPC) photographic
standards as test samples is shown in FIGURES
72, 73, 74,
and 75. A contouring algorithm
was used to compute the percentage of rust in an area and
the corresponding rust grade is automatically printed in
the upper left-hand of the recorded image. This approach
can be made to easily sort the photographic standards. After
calibration to the ASTM D610-85 photographs, verification
of the technique was performed on actual steel panels used
as an equivalent National Association of Corrosion Engineers
(NACE) standard. How well it will function on actual bridge
areas where the contrast and the recorded gray scales may
not be as good as these photographic standards needs to be
determined. This technique needs to be applied to actual
bridge surfaces or pieces of bridges where comparison to
manually measured rusted areas can be performed. Tests of
this type would allow evaluation of performance under conditions
that include aged or faded paint, dark colors, and surface
contaminants such as grease or dirt stains. Tests also need
to be performed under typical field conditions to determine
minimum lighting requirements.
EVALUATION OF BLAST-CLEANED
SURFACE
The degree of cleanliness of a steel surface prior to paint
application is critical to the durability of the paint. This
quality in the maintenance painting progress is frequently
not maintained. A means of objectively evaluating the blasted
surface prior to painting would be a valuable aid in improving
the quality control of the process. Current practice makes
use of visual degree of cleanliness standard. (11)
Ultimate responsibility lies with the inspectors and is critically
dependent on their skills and dedication. If a paint failure
occurs, there is no traceable data other than what has been
documented by the inspector. A stored digital image should
be capable of providing a quantitative real-time and permanent
record of the surface condition. The simple monochrome techniques
described previously may have differentiation problems associated
with the subtleties of the cleaned surface. A technique based
on color measurement would be more applicable. An example
of this approach can be seen in FIGURES
76, 77, and 78.
These histograms show the distribution of pixel values for
the red, blue, and green elements of images recorded from
the A SP-10 and A SP-5 standard photographs of blast-cleaned
surfaces. The vertical axis in the figures represents the
number of pixels that have a specific value (intensity).
The x-axis represents intensity with 0 being black and 300
saturation. The histograms of the A SP-10 and A SP-5 standards
show that the centers of the distributions are distinctly
shifted for each color component. Furthermore, the green
component, FIGURE 78, shows
a clear difference between the peaks of the distributions
for the two test samples. These results show that color measurements
have promise for detecting subtle differences in cleaned
surfaces. The approach should also allow rust to be measurable
despite the presence of paints having similar coloring.
QUANTITATIVE MEASUREMENT
OF DAMAGED AREA
A visual evaluation procedure has been developed
(based on the red/green/blue techniques described above)
and tested
that can determine the percentage of rusted and/or damaged
paint. In this procedure, the image is acquired by the of
a color CCD camera (FIGURE 79).
The recorded images are enhanced by state-of-the-art techniques
pioneered by NASA and the military. These techniques have
been integrated into a software package designed to assess
the percentage of damaged area in real time.
All image pixels that have values that are
lower than a selected threshold are considered representative
of damaged
area and are assigned a value of 1. The remaining image
pixels that have values that are greater than or equal to
the threshold
value are assigned the value of 0. The resulting binary
image shows the damaged area of the original image. The percentage
of damaged area is calculated by first counting the number
of pixels that have a value of 1 and then dividing that
sum
by the total number of pixels. A portion of a girder (FIGURE
80) from the IDOT storage yard was used to evaluate
the technique and the results are shown in FIGURE
81.
INFRARED THERMOGRAPHY
Introduction
As demonstrated above, visual techniques can
quantify damaged areas within the visible spectrum. However,
the ability to
detect delaminations, voids, and other non-visual phenomena
requires a technique to monitor an expanded electromagnetic
spectrum. One such technique is thermography. Infrared (thermal)
imaging is often referred to as thermography, an optical
technique for remote detection of a scene's thermal radiation.
Physically, it is based on thermal radiation laws and technically,
it is similar to infrared thermometry. A completely non-invasive
technique, it does not requires any contact with an object
and can be used to measure the temperature distribution of
the remote or moving targets. Various ranges of spectral
response and optical configuration gives this technique a
considerable level of flexibility in adapting it to a wide
spectrum of applications. It is a unique tool for measurement,
visualization, and analysis of various steady-state and transient
heat-transfer phenomena. Analysis of the emissive and heat-transfer
properties often provides a unique "signature" of various
physical structures, processes, or objects. These capabilities
have made infrared thermography an indispensable diagnostic
tool for variety of industrial, military, and scientific
applications, including materials non-destructive evaluation;
aerial and vehicle-based testing of structures and buildings;
online inspection and non-destructive testing of electrical
installations and nuclear, chemical, and petrochemical industrial
complexes. In many applications, infrared imaging allows
early detection and quantitative diagnostic analysis of faults
and failures of structures and materials, thus making feasible
establishment of preventive maintenance scheduling.
In the last few years, applications of infrared
thermography have been even further broadened due to immense
developments
and enhancements in the field of infrared sensing, image
acquisition, and computer-processing hardware.
DISBONDMENT OF EXISTING
PAINT SYSTEMS
The evaluation of the strength of the bond
of paint and the identification of delaminations and voids
present a complex
problem. Spot measurements may be made by using magnetic
dry film thickness gauges to assess the thickness of the
original paint system. Similarly, the determination of its
bond strength may be made by using one of several standard
tests. The basic problem is that these tests give a reading
only at the point of measurement with no given assurances
that they are an accurate gauge of any zone or of the entire
structure. The need exists for a global method that will
detect the differences in the bond condition of the overall
surface and direct the inspector to suspicious areas where
additional point measurements can be made of the coating
bond strength. In this case, thermography makes use of the
fact that the bond strength affects the heat-transfer properties
of the insulating paint bonded to a conductive substrate.
This technique involves applying heat or cooling to the surface,
either in the form of a continuous source or as a transient
pulse, and observing differences in temperature from one
location to another with an infrared sensor as the surface
either heats up or cools down. The choice of heat source
depends on a combination of conditions, including ambient
temperature and the heat transfer properties of the coating
and the substrate. On a bridge, the heat source might be
a heat lamp, a torch, or flash lamp. One interesting possibility
would be to use the thermal gradients resulting from solar
heating. In this case, observations would be made either
during sunrise or sunset on the areas of the bridge that
are of interest. An imaging infrared sensor would allow temperature
differences to be mapped. Hot areas, where the cooling rates
were slower, would indicate areas of poor bond strength.
These areas could be marked for later reference and review.
An important factor to keep in mind with this technique is
that absolute measurements are not required and only temperature
gradients are required to pinpoint suspicious areas.
Thermal-wave imaging was evaluated to determine
its ability to detect debonded paint areas. The thermographic
system
is schematically represented in FIGURE
82 and is employed not only as an infrared sensor,
but also as a high-resolution color camera. The camera
is used
to identify an area and act as a reference for the thermographic
image. The panels used in this evaluation were obtained
from a bridge girder supplied by IDOT. The surfaces were
in a
bad state of degradation. FIGURE
83 shows a typical sample area. As can be seen, there
are areas of bare metal, rust, millscale and blister, and
intact original paint. After the panel was heated, the
image was obtained and enhanced by computer software. The
areas
of damaged and delaminated paint are easily seen (white
and light shaded areas (FIGURE 84).
In the preceding example, a continuous heat source was
employed. However, if it had been replaced by a pulsed
source and the
gradient of heat flux monitored, it would be possible to
identify and differentiate between voids and delaminations
and also to determine relative bond strengths.
CONCLUSION AND RECOMMENDATIONS
Preliminary evaluation has demonstrated the applicability
of digital image analysis for paint inspection, including
distinguishing and measurement of damaged paint areas. Image
texture and color information were successfully used for
evaluation of the damaged area, and the results have shown
the sensitivities of color image features to certain paint
characteristics. The image analysis technique based on accurate
color image acquisition can be a useful and effective tool
for laboratory as well as field paint assessment and can
provide a permanent record of results. However, it is essential
to apply the knowledge of paint experts to the analysis of
color images, and additional work with a greater number of
paint samples is needed in order to establish consistent
correlations between color image features and paint characteristics. |
