Using computers to monitor the health of structures
Prof. Tshilidzi Marwala
The continuous monitoring of the health of buildings, bridges and machines are
central to the safety of people who use, or come into close proximity of such
structures. A medical doctor has the advantage of a person being able to
communicate their ailment in a common language. How does a "building
doctor" understand?
The general philosophy of continuous monitoring of the health of structures
is guided by an old saying: "Everything has something to say and all we
need to do is to deconstruct its mode of communication and thereby understand
its language". Using this philosophy, it can be argued that all structures
have something to say about the state of their health. Therefore, all that needs
to be done is to understand their messages and then act on the information from
those messages.
In structural health monitoring, advanced dynamic methods are used to
deconstruct the information from the structures and artificial intelligence is
used as a brain that transforms the measured data into useful information.
Monitored Data
The data that is monitored indicates how the structure moves, the rationale
behind this being that the movement of a new structure and thus one of high
structural integrity, is rather different from that of an old and frail
structure. So comparing the old with the new provides a good indication of the
health of a structure. Scientists are able to quantify the difference between
the movements of young and old and this information is valuable in deducing how
healthy a structure is. There are many types of data that can be used to monitor
structures and one of the most widely used in structural health monitoring is
vibration data.
Vibration data
Vibration is an important measure of how a structure moves. Vibration data
may be measured from the structure using accelerometers. The picture below shows
an accelerometer being mounted on the surface of a bridge.
Intelligent Computers
At the heart of structural health monitoring is the extensive usage of
computers. After vibration data is measured using the accelerometer and
converted into digital format, they are then stored in a computer. The first
stage is to interrogate the data through filtering of noise and converting the
data into a more friendly format. Some of the ways of converting the data into a
user friendlier format is to apply signal processing tools such as Fourier
transform. Fourier transform takes the measured movement of the structure and
extract the harmonics of those movements. The rationale behind this technique is
based on the observation that harmonics of the movement of a structure reveal
information more clearly than the movement of the structure itself. The signal
processing tools are in the form of software which is located in the computer.
The data that has been processed needs to be transformed into usable
information - information as to whether the structure is healthy enough to
continue fulfilling its usual functions or not. Converting the data into this
kind of information requires some level of intelligence. Computational tools
have been developed that have some degree of cognitive intelligence and are
called artificial intelligence methods. Artificial intelligence is a field that
looks at natural systems and uses them for the benefit of mankind. There are
many types of artificial intelligence techniques. A well known type is called
neural network.
Neural network takes the concept in which our brain process information and
uses it to identify patterns. So in the context of monitoring a bridge for
example, neural networks take vibration data and convert them into information
which is used for decision-making. For example, if neural networks inform the
user that the bridge is going to collapse then proper intervention strategies
are put in place to not only save lives but infrastructure.
The image above illustrates the framework of the process.
Firstly, data is measured from a structure such as a bridge then processed
through the computer using signal processing tools and artificial intelligence.
The final output is information on whether the structure is healthy or not.
The entire process may be fully automated and, therefore, be operated
independent of human intervention. So next time you cross a bridge or enter your
office, spare a thought for the science behind keeping it stable.
More information
Prof. Tshilidzi Marwala
University of the Witwatersrand, School of Electrical and Information
Engineering
P/Bag 3, Wits, 2050
Website: http://www.dept.ee.wits.ac.za/~marwala
|
