Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
21~'~3~3
CONTINUOUS MONITORING OF REINFORCEMENTS IN STRUCTURES
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention is directed to methods and systems for monitoring the
integrity of structures and, in particular, to a method and system for
monitoring
structural reinforcement or securing members, such as post-tensioning cable,
in
concrete. More particularly still, breakage of such cables is detected,
distinguished from background noise, and located by means of subsequent
analysis or on a real-time basis. The source of the breakage is located by
known
triangulation techniques. In addition, cost saving space multiplexing of
sensors
is used.
2. Belate~LA~
United States Patent 3,949,353, granted April 6, 1976 to Waters et al, and
titled
"Underground Mine Surveillance System", discloses a system for maintaining a
continuous log of activity in and around an underground mine where seismic
energy in the area is continually monitored, processed and classified into
meaningful relative data indications; and which system further includes
selectively deployable seismic energy monitoring equipment providing more
specific data in the event of mine catastrophe. This system utilizes
permanently
disposed seismic energy detectors and/or emergency detectors placed in
2
accordance with the particular exigency, and detected seismic energy return is
continually processed to maintain a data log indicative of general type and
location of mine activity, with particular capability for isolation of unusual
seismic events by comparison with statistical data constraints of
predetermined
character.
United States Patent 4,386,343, granted May 31, 1983 to Shiveley, and titled
"Acoustic Emission Intruder Alarm System", discloses an acoustic emission
burglary detection system for detecting physical attacks made on a protected
structure such as a vault, safe or the like. Sensors (13, 14) mounted on the
protected structure detect acoustic emission stress wave signals produced by
an
attack and provide an event signal of a corresponding frequency and with an
amplitude and duration dependent upon those of the stress wave signals. Event
signals exhibiting a frequency less than 50,000 Hz are much less likely to
have
been originated by a physical attack upon the protected structure and are
filtered
out. The remaining event signals which exceed a predetermined level, are
integrated over a predetermined time period. If the resulting value exceeds a
predetermined level, an alarm is activated. Means are provided for testing the
detection circuit by providing electrical pulses through one of the sensors to
cause it to generate mechanical stress wave signals in the protected structure
which can be detected and processed by the detection circuitry.
United States Patent 4,649,524, granted March 10, 1987 to Vance, and titled
~~7~3'~3
3
"Integrated Acoustic Network" discloses an integrated acoustic network system
to provide warning of impending groundfall in underground mines. The system
includes a plurality of geophones which derive acoustic signals by which the
source of seismic disturbances is located, and an array of high frequency
piezoelectric sensors which pick up signals from small ground disturbances
which
precede groundfall. A warning system is provided both at the scene of mining
operations and at a central location of impending groundfall and of the
location
of its occurrence.
In United States Patent 3,949,353, unusual seismic events are isolated by
comparison with stored statistical data constraints. In United States Patent
4,386,343, event signals exhibiting a frequency above 50,000 Hz are integrated
over a predetermined time period and, if the resulting value exceeds a
predetermined level, an alarm is activated. In United States Patent 4,649,524,
the number of seismic events is counted and the cumulative amount of energy is
estimated for every minute and the ratio of energy/event is calculated.
United States Patent 4,535,629, granted August 20, 1985 to Prine, and titled
"Method and Apparatus for Structural Monitoring with Acoustic Emission and
Using Pattern Recognition", discloses an acoustic emission monitoring system
used for monitoring fatigue crack growth in metal or other materials such as
occur, for example, in highway bridges during normal traffic loading. The
transducers are placed on the plates to be tested to allow detection of
acoustic
21'~83'~~
4
emission from a particular site. By applying specific recognition methods to
the
acoustic emission AE, detection of flaws can be detected from a random noise
background. The pattern recognition technique first subjects the received AE
energy to an energy window test and if the energy is within the window, it is
subjected to a rate test and if the energy exceeds predetermined rates, it is
passed
to a location test so as to locate the position of flaws.
United States Patent 4,565,964, granted January 21, 1986 to Matthews, et al.,
and titled "Cable Integrity by Acoustic Emission", discloses a system for
monitoring the integrity of a cable, for example a cable following a variable
depth sonar body through the ocean. Such a cable comprises a core of
electrical
wires surrounded by load bearing wires which are secured to the sonar body
through a terminator. For various reasons cracks can appear in the load
bearing
wires and in extreme cases one or more of the wires may break. The monitoring
system includes a transducer located near the terminator where the wires are
most
likely to crack or break. Acoustic emissions caused by the incidence of cracks
or breaks are picked up by the transducer. The resulting electrical signals
are
amplified and passed up the electrical core of the cable to the towing vessel
where they are processed. Novel aspects of the system are the water coupling
between the location of the cracks or breaks and the transducer and special
processing circuitry which enables breaks, cracks and electrical noise to be
distinguished from each other.
CA 02178373 2006-08-22
United States Patent 4,738,137, granted April 19, 1988 to Sugg, et al. , and
titled
"Acoustic Emission Frequency Discrimination", discloses an acoustic emission
signal processor that selectively sorts acoustic signals on the basis of
frequency
content, rather than just the frequency. The processor allows rejection of
some
5 signals having a particular frequency content, or can provide for separate
counting or other processing of these signals.
SUMMARY OF THE INVENTION
The present invention endeavours to solve the problem of distinguishing and
locating a single non-recurring event, namely, the breakage of reinforcing
element embedded in concrete. For this reason, neither integrative, cumulative
nor statistical techniques of the prior art would work.
Strengthening of concrete structures, such as bridges or concrete floors of
modern buildings, is often accomplished by means of highly tensioned cables
which are extended through conduits embedded in the concrete. Post-tensioning
cables sometimes corrode and break, thereby impairing the integrity of the
stnzcture. Often these broken cables remain undetected. The monitoring of
these
inaccessible structural reinforcements to measure their integrity has long
been a
problem. Conventionally, the cables are visually inspected involving drilling
a
217373
6
view port into the concrete at each cable location. Visual or electrical
inspection
of the cable is then performed to determine if the cable is still bearing
load.
A method and system for monitoring reinforcing elements of a structure non-
destructively is provided. Monitoring is performed on a continuous basis by
means of acoustic or seismic indicators. When a tensioned reinforcement within
a structure breaks, energy, specifically in the form of acoustic and seismic
energy, is released into the surroundings. Appropriate detectors responsive to
these forms of energy are positioned on or near the structure to detect the
energy
emissions.
According to a broad aspect of the present invention, a method is provided for
monitoring the failure of tensioned reinforcements in a structure comprising:
positioning a plurality of acoustic/seismic detectors about the structure in a
known arrangement and in close proximity to the structure; processing signals
from the detectors; and identifying signals due to the failure of tensioned
reinforcements in the structure.
According to a further broad aspect of the present invention there is provided
an
apparatus for monitoring the failure of tensioned reinforcements in a
structure
comprising: a plurality of acoustic and/or seismic detectors positioned about
the
structure in a known arrangement and in close proximity to the structure,
means
for processing signals from the detectors; and means for identifying signals
due
CA 02178373 2006-08-22
7
to failure of tensioned reinforcements in the structure.
In another method aspect of the present invention, the steps of processing and
identifying signals from the detectors further comprise: analyzing signals
from
the detectors for their frequency content; and identifying the signals as
being due
S to a failure event only at times (i.e. time slices or snapshots) when the
frequency
content exhibits frequencies above a predetermined frequency-threshold and
voltages above a predetermined voltage-threshold.
Accordingly, the method of analysis uses these quantities; frequency,
power or energy, and time.
In a further method aspect, the step of processing further comprises
converting
the signals from the detector to their Fourier transforms.
In a yet further aspect of the present invention, it has been found that
rupture of
an unbonded post-tensioning strand is identifiable upon detecting a plurality
of
nearly simultaneous acoustic emissions originating from sources located
substantially on a line along the course of the ruptured strand or cable.
This is possible because the speed of sound in the steel cable is faster than
the
speed of sound in concrete. The rupture of the cable or a wire in the cable
produces a sound wave which travels along the length of the cable, and allows
~17~~'~
g
several places on the cable to act as a source of energy propagating into the
concrete slab.
According to another aspect of the invention, a method is provided for
distinguishing between signals arising from events such as the breakage of a
tensioning cable and signals arising from other events, such as an object
falling on
a concrete surface.
According to another apparatus aspect of the present invention, there is
provided
an apparatus for detecting the location of failure of a tensioned
reinforcement in
a structure comprising: a plurality of detectors positioned about the
structure in
a known arrangement and in close proximity to the structure, the detectors
being
responsive to acoustic energy or seismic energy or a combination thereof to
produce a signal; central monitoring means coupled to each of the detectors,
comprising means for conveying signals from each of the detectors, means for
identifying a signal due to failure of the tensioned reinforcement in the
structure;
timing means for determining the time of arrival of the energy at each of the
detectors; and, reporting means to indicate the failure of the tensioned
reinforcement.
BRIEF DESCRIPTION OF THE DRAWINGS
The preferred embodiment of the present invention will now be described in
~17~3'~3
9
detail in conjunction with the annexed drawings, in which:
Figure 1 is an illustrative and block schematic diagram of the system
according
to the present invention;
Figure 2 illustrates an embodiment of the present invention for multi-floored
structure;
Figure 3 shows an example Fourier transform showing the frequency spectrum
from one of the detectors shown in Figures 1 and 2;
Figures 4A and 4B are illustrations useful in explaining operation of the
present
invention; and
Figure 5 is a flow chart useful in explaining operation of the system of
Figures
1 and 2.
Figure 6 is a graph of distance from the event against energy for an event
caused by
an object hitting the surface of a concrete slab.
Figure 7 is a graph of distance from the event against energy for an event
caused by
a tensioning cable breaking within a concrete slab.
CA 02178373 2006-08-22
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
Referring to Figure 1 of the drawings, the system for monitoring the
structural
integrity of a concrete floor 10 having therein embedded post-tensioning
cables
11 and 12, comprises central processor 14, an event record monitor 15, as well
5 as acoustic, seismic or acoustic\seismic sensors or detectors DI, DZ and D3,
placed at arbitrary locations close to or contiguous the surface of the floor
10.
In the drawing, there is a breakage in cable 12 at 13 which causes signals to
be
received at detectors Dl, DZ and D3. The central processor 14 comprises a
buffer
and/or interface 16 to which the detectors D,, D2 and D3 are coupled, and
which
10 associates signals detected at each of the detectors D1, D 2 and D 3 with
the
respective real-time of detection by means of clock 17. The interface 16
multiplexes the signals from the three detectors and applies them to high-pass
filter 18 (which may be a filtering routine of the central processor 14)
having an
approximate cut-off frequency of a few kilohertz, preferably in the present
preferred embodiment 6 kHz. The output of the high-pass filter 18 is applied
to
threshold detector and event identifier 19 which decides, on the basis of the
spectral voltage density of the filtered signal, to recognize and identify a
breakage event. The processor 14 preferably has capability for keeping a
statistical record 20, which would be used to aid in predicting future
failures.
A background noise monitor 21 preferably continuously monitors the
background noise detected by the detectors Dl, D2 and D3 and sounds or
displays
an alarm on the monitor 15 should the background noise level fall below a
CA 02178373 2006-08-22
I1
preselected value. Thus, the noise monitor 21 receives its input signal prior
to
filtering and processing.
Referring now to Figure 2, this shows an embodiment suitable for a multi-floor
structure, comprising a plurality of floors, of which three are shown 10A, lOB
and lOC. Each floor has a number of sensors\detectors S1 to S8 and a floor
detector LA, LB, LC, etc. In order to reduce cost and wiring complexity, and
given the very small likelihood of simultaneous cable breakage, the detectors
are
spatially multiplexed. As shown in Figure 2, all detectors S 1 (one on each
floor)
may be at the same respective position for ease of installation and analysis
(but
they need not be as long as their positions are known), All detectors S I are
connected in parallel to a bus 31, all detectors S2 are connected to bus 22,
(not
shown), all detectors S3 are connected to a bus 23, and so forth. Preferably,
the
detectors have high impendence outputs, in order not to represent a
significant
load to each other. Each floor is identifiable by a detector LA, LB, LC etc.
The
outputs of the sensors S1 to S8 are applied via the respective buses 31, 22,
23
etc. to a detector A/D (analog-to-digital) converter 24, while the outputs of
the
floor detectors, LA, LB, LC etc. are applied to a floor A/D converter 25, the
function of the latter being simply to identify the floor in which failure
event
occurs, but not the location of the breakage on the floor, which is
accomplished by triangulation by means of sensors Sl to S8 in the particular
floor. The multiplexed output signals from the A/Ds 24 and 25 are applied
to the processor 14, which keeps track of the events in real-time and outputs
the event record 26.
~1'~~3'~3
12
It was mentioned in conjunction with Figure 1, that instead of using a high-
pass
filter 18 a software filtering routine may be used. While there are several
methods of identifying valid failure signals, a preferred method is to use the
Fourier transform of the detector signals. The Fourier transform is the
conversion of signals which are time-varying voltage functions from detectors
to
frequency dependent voltage (or power) density functions. This is a well-known
operation in signal processing, but has been found to be particularly useful
in the
noisy operating environment of the present system. Figure 3 of the drawings
shows a graphic illustration of such Fourier transform of an actual detector
signal
at the time of an exemplary experiment (described under Example 1, infra). The
graph shows the spectral distribution of detector voltage levels versus time.
As
may be seen, there is clear spectral presence between approximately 6 kHz and
13 kHz in the period between approximately 60 ms and 200 ms after the record
commenced. Thus, the background noise monitor 21 continuously indicates
system sanity while the spectral components do not appreciably exceed the 6
kHz
threshold. The system detects a breakage event once the spectral profile
exhibits
the characteristics shown in Figure 3, even if only for milliseconds.
With reference to Figures 1 and 3, in an experimental embodiment of the
invention the three piezoelectric detectors Dl, DZ and D 3 (model 273-065A,
manufactured by the Archer Company) were placed on the surface of a 30 m x
~~ ~~3'~~
13
15 m slab of 20 cm thick concrete which formed one floor of a parking garage
(not in use and closed to the public) in Lethbridge, Alberta. The floor had
been
reinforced when constructed with post-tensioned steel cables. The detectors
D1,
DZ and D3 were placed in a triangular array. A microprocessor was connected
by conventional wiring to the three detectors. The detectors were arranged to
record seismic waves passing through the concrete floor. The garage, which was
slated for demolition, was known to have reinforcing cables in an advanced
state
of corrosion. One such cable was exposed, and breakage was induced. Figure
3 is the Fourier Transform graph showing the effect of this breakage. The
ordinate of the graph shows time after commencement of recording (in
milliseconds) and the abscissa shows recorded frequency in kHertz. The
vertical
direction represents the amplitude of detected signals in millivolts. The
graph
is a record of the output of one of the detectors. It will be evident that,
prior to
approximately 88 milliseconds, signals of relatively low amplitude and a
frequency under about 7 kHz were recorded. These represent background noise,
such as traffic passing, movement within the parking garage, etc. At 88
milliseconds, the induced cable breakage was detected. This caused signals to
be generated at frequencies as high as 13 kHz and above. Higher than
background noise frequencies occurred until approximately 160 milliseconds on
the measuring scale. Similar patterns were recorded by the two other
detectors.
~1'~~3 r
14
With reference to Figures 4A and 4B, in a second example, directed to
demonstrate the failure locating feature of the invention, a failure event was
simulated by dropping a 1 kg weight from a height of 30 cm onto a randomly
selected location 40 on a concrete slab floor of an office building in
Calgary,
Alberta (closed to the public). The portion of floor studied measured 8.0 m x
20.0 m. Three piezoelectric detectors, of the type used in Example 1
identified
as D, E and F, were located as shown in Figure 4A. The initial burst of high
amplitude and high frequency waves following the simulated failure was found
to arrive at detector E first. The waves arrived at detector D 0.7
milliseconds
later and then at detector F 1.2 milliseconds after the first waves had
reached
detector E. With reference to Figure 4B, the known locations of the detectors
and the relative differences in arnval times can be used to find the location
of the
break. To locate origin, the locations of sensors D, E and F are plotted on a
graph representative of the slab. Using the relative arrival times delays and
the
velocity of the seismic energy through the structure (found to be 2400 m/s),
the
comparative distances from origin when compared to detector E are determined
as follows:
Sensor D delay = 0.7 ms
Additional distance from origin as compared to sensor E
= 0.7 ms x 2400 m/s
= 1.61 m and,
Sensor F delay = 1.2 ms
~I'~~3~3
Additional distance from origin as compared to sensor E
= 1.2 ms x 2400 m/s
= 2.76 m.
These distance values are used as radii for the drafting of circles about
their
5 respective sensor as is shown in phantom at 36 and 38. A circle 41, shown in
phantom, is then drawn to intersect sensor E and to contact both circle 36
about
sensor D and circle 38 about sensor F. In this way the origin is determined to
be the center 42 of circle 41. The actual origin was, in this embodiment,
found
to be as indicated at 40. This was of course known, as the simulated breakage
10 had been induced. The actual origin compared very closely with the
calculated
origin, indicating an error of 0.23 m.
While the above embodiment uses only seismic detection, both seismic and
acoustic detection may be used. Where both forms are used only two detectors
need be used to locate origin. Reflections of energy may also be used to pin-
15 point origin, however this requires precise locating of reflective
structures.
Further, while the arrival of the first wave pattern is clearly recognizable
reflective wave patterns are very complex.
Referring now to Figure 5 of the drawings, the operation of the system shown
CA 02178373 2006-08-22
16
in Figures 1 and 2 is described. The system continuously samples the sensors
(50) and once a signal from any sensor exceeds the background noise threshold
(51), it acquires the signal from all sensors (52) and performs a Fourier
transform operation thereon (53), whereupon the spectral density of the
frequencies to be obtained from such transformation is examined (54), and
the system saves the data (55), otherwise it returns to sampling routine (SO).
The data saved in step (SS) is then analyzed (56), for example, to find if it
matches the cable characteristics as shown in Figure 3. If it does not match,
then the event is located and saved (57) (once all sensor data is in) and a
record is compiled (58). If it does match then the cable breakage event is
located and Ragged (59), and, of course, a record is also compiled (60).
While the present invention is described using, as an exemplary embodiment,
the
monitoring of post-tensioning cables in a concrete structure, the invention
can
equally be applied to monitoring other tensioned structural reinforcements or
securing members where, for example, the reinforcements or securing members
are subject to breakage. Thus the present invention may be used to monitor
structures including suspension bridge wires; rivets or skin in airplanes;
bolted
structures; bonded cables in bridges; ship hulls and bulkhead; bolted
structures
such as cranes or towers; bonded cables cast adhesively in concrete; anchoring
cables and tie-backs. Preferably three or more detectors are positioned about
the
structure. Three detectors allow the origin of energy emissions to be located
quickly without employing an excessive number of detectors. The number of
~1'~~373
17
detectors employed is dependent on, for example, the expected amount of energy
released during the event and the sensitivity of the detectors.
The detectors are responsive to acoustic energy and/or seismic energy.
Suitable
detectors for use in the present invention include piezoelectric transducers,
capacitive transducers, accelerometers, microphones of all types, inductive
systems such as geophones, audio-acoustic transducers, acousto-optical
transducers, magnetic inductive devices or optical devices. Many of the
detectors respond only to one of acoustic energy or seismic energy. However,
a piezoelectric transducer responds to both. Most accelerometers do not
respond
well to acoustic emissions but do respond to seismic emissions. Accelerometers
may be linked to diaphragms to increase their acoustic sensitivity. Optical
detective devices such as acousto-optical transducers or optical
accelerometers
use a wide variety of methods to convert strain and stress into a change in
the
optical properties of a device, including the use of fibre optics or intensity
variations. Optical devices can be sensitive to both acoustic and seismic
emissions. An example of a suitable detector is the Lars 100 interferometer of
Gradient Lens Corp.
The detectors are positioned in close proximity to the structure and
preferably in
contact with the structure. The detectors can be embedded in the concrete of a
bridge or floor slab. However, to enhance the usefulness and simplicity of the
apparatus, it is preferred that the detectors remain on the surface of the
structure.
~~.783'~~
18
In this way the detectors may detect air-borne acoustic energy as well as
structure-borne acoustic energy arising from the event to be detected such as
a
cable failure. However, because of the slow velocity of air-borne acoustic
emissions when compared to those of structure-borne and seismic emissions, air-
borne acoustic energy is not of particular interest in the preferred
embodiment
of the present invention. As an example, the velocity of air-borne acoustic
energy is 300 m/s while acoustic and seismic concrete-borne energy pressure
wave and shear wave velocities are approximately 5000 m/s and 2300 m/s,
respectively, but, of course, knowing the exact velocity of propagation is not
necessary. The compression of concrete may cause variances in the velocities,
so tests may be carried out to determine the acoustic and/or seismic velocity
(depending on which is used in the particular installation) in the concrete of
the
structure if the velocity information is to be used to locate the site of the
failure.
In response to the detection of an energy wave each detector produces a
signal.
The detectors may collect the signal data independently, on some appropriate
collecting means such as magnetic tapes, until the information is required.
When
required, the collected data is analyzed to recognize a signal relating to
reinforcement failure. Preferably, however, the detectors are coupled to the
central processor 14 allowing continuous monitoring of the detectors.
When both acoustic and seismic responsive detectors are used, the time of
arrival
of energy of both forms may be recorded independently and compared to pin-
~1'~~3'~3
19
point the origin. Employing detectors which are responsive to more than one
form of energy is beneficial in finding origin as well as recognizing a cable
failure at times when there is excessive background noise in one of the energy
forms. If there should be excessive background noise in one energy form, the
other forms may be used to provide signals which are not distorted.
According to another aspect of the invention, a method is provided for
distinguishing between signals arising from events such as the breakage of a
tensioning cable and signals arising from other events, such as an object
falling on
a concrete surface.
Example 4
As noted previously, signals arising from events such as cable breakage can be
distinguished from other signals by analyzing the frequency content and
treating as
probable breakages only those in which the frequency content exhibits
frequencies
above a predetermined frequency threshhold and which preferably have voltages
above a predetermined voltage threshhold. However, there is also a second
method
which has been developed, and which also is a good predictor of whether a
signal
is from a cable breakage or not. The second method can either be used to
confirm
the results of the first method, or can be used instead of the first method.
The second method is carried out by locating the apparent source of the event
as
discussed above (for example by triangulation) and plotting the distance of
each
sensor from the apparent source against the energy (expressed logarithmically)
20
which has been received at that sensor. Conveniently, the energy received is
plotted
in decibels when the sensors are acoustic sensors.
Relative energies of different sensors (which are all that is needed for the
practice
of this method) can also be calculated by multiplying the frequency by the
square
of the amplitude and integrating over the range of frequencies that include
substantially all of the energy recorded (for example, from 0.1 to 25000
Hertz.
The apparent source of the event need not be located exactly. Where there are
three
sensors, triangulation gives a single point. Where there are four or more
sensors,
it is not always possible to triangulate to get a single point, but
conventional
mathematical techniques can generate a region (known as the Vornoi region)
within
which the source of the event is located. This is typically an irregular
polygon. It
is sufficient for the purposes of the method to use, as the distance plotted
for each
sensor, the distance from the Vornoi region to the sensor.
If the sensors are all of approximately the same sensitivity, the energy
readings from
each sensor can be plotted directly, in logarithmic form, against distance.
For the
success of the method, minor errors in total energy recorded due to
differences in
sensitivity are not critical, and close calibration of sensors is not
required. Indeed,
variations of as much as an order of magnitude (i.e where one sensor registers
from
10% to 1000% of the signal of another supposedly identical sensor receiving
the
same signal) do not prevent the method from giving beneficial results,
although
errors of this sort are of course not desirable.
2~'~8373
21
If the sensors are of different rated sensitivities, their outputs must of
course be
expressed in the same units before plotting.
Once the distance and energy are plotted, the slope of the resulting graph is
examined. It is found that the graph produced by plotting results from a point
source non-breakage event (such as an impact on a concrete slab) is typically
a line
of fairly constant slope, in which the slope decreases with substantial
regularity with
distance from the point of impact. In contrast, the slope of the graph caused
by a
cable breakage event is found to be irregular, with the slope changing over
distance
and sometimes even increasing over short distances.
Results from as many sensors as are available should be plotted, to get as
accurate
a slope as possible. Results from as few as four sensors give meaningful
results,
however.
Although it is not desired to restrict the invention to any theory, it is
thought that
the variability of slope for the cable breakage event is caused because some
energy
is conducted through the cable itself and some through its sheath, and this
energy
passes into the concrete surrounding the cable sheath at varying distances
from the
point of breakage, This causes sensors located close to the sheath along its
length
to give higher readings than would otherwise happen. There is of course always
some non-linearity of response (for example, some energy may be reflected by a
discontinuity in the concrete or the like) but this is not usually enough to
affect the
overall slope of the curve in a non-breakage event.
~1'~~373
22
The effect is illustrated by Figures 6 and 7. Figure 6 shows a typical graph
made
by the impact made by operating a pneumatic concrete chipping hammer on a
concrete slab. The points representing energy (in decibels) against distance
from the
Vornoi region for eight sensors are plotted. If the closest and farthest away
of the
sensors are excluded, a graph of substantially constant slope "x" occurs with
the
remaining readings. As expected, this shows a declining signal with distance.
Even
including the closest and farthest away sensors, the "best fit" still shows
this slope.
Figure 7 shows a typical graph made by a cable breakage. In this case, results
were
recorded on 18 sensors. Applying "best fit" slopes to the results, there are
several
different slopes., There is initially an increasing slope "a", followed by a
decreasing
slope "b", followed by another increasing slope "c". The second closest
sensor,
B03, gives an anomalously high reading, but even this is smoothed out by the
"best
fit".
Thus, it is indicative of a cable breakage event to have a slope which varies
and
includes portions of increasing slope, where a logarithmic representation of
energy
(or relative energy) is plotted against the distance from the region of the
event.
Stated another way, where there is a substantially linear relation (which
corresponds to a straight slope of the graph) between decreasing energy
recorded
at the different sensors and increasing distance of such sensors from the
apparent
region of origin, then it is unlikely that the event was a cable breakage or
breakage
~'~'~83'~3
23
of tensioned reinforcement. However, when the relation is non-linear (i.e. the
graph has varying slopes) it is more likely that the event is the breakage of
a
tensioned reinforcement member. When some sensors farther from the event
record
more energy than those closer, it is still more probable that the event is a
tensioned
reinforcement breakage, This analysis can be used, either alone or as a
verification
of other methods, to determine when a recorded event is probably due to cable
breakage.
It will be understood that the forgoing description of the invention is by way
of
example only, and variations will be evident to those skilled in the art
without
departing from the scope of the invention, which is as set out in the appended
claims.
