The Cambridge Monitoring System (CMS) is a new, ultrasonic-based monitoring system primarily for concrete structures but with possibile application to structures made of other materials.
Cambridge Ultrasonics has successfully developed a R&D prototype system that works well.
We are looking for investors and partners for pilot scheme evaluation tests of the curent CMS.
What CMS provides
Benefits of using a CMS system
The need for CMS monitoring
There is a growing problem within Europe and other countries in the world concerning the maintenance of large concrete structures. The most notable problem is the unexpected deterioration, due to corrosion, of steel stressing-tendons, causing the tendons to reduce their load-carrying capacity. This is serious because there is no reliable, non-destructive measurement for testing the degree of corrosion. CMS provides for the first time an ultrasonic monitoring system for large concrete structures that is sensitive to corrosion of tendons at depths up to 2m in concrete.
CMS has been developed as part of a collaborative European research project supported by the European Commission with the following partners: SP (Sweden), Necso (Spain), IETcc (Spain), Queen's University Belfast (UK), Sonatest Plc (UK). Cambridge Ultrasonics is the co-ordinator of the project and it is our know-how that is being developed and evaluated in the project. CMS is one of fruits of the project; CMS has been developed by Cambridge Ultrasonics and is being evaluated by four of our partners.
How CMS works
A network of intelligent, CMS sensors is attached to the structure to be monitored and connected by cable or radio back to an Archive PC. Each sensor injects a known pattern of waves into the part of the structure to which it is attached and each sensor collects echo-waves back from the interior of the structure. CMS is not an acoustic monitoring system, which is a passive monitoring system and which is potentially very sensitive to operational noise. In CMS echo-waves travel from a range of about 2 m in all directions back to the sensor; the precise range depends upon the local geometry of the structure and the concrete mix and condition. Echo-signals are processed to remove noise from operational sources, such as: road traffic noise, turbine noise, boiler noise; this process is aided by the use of specially selected signals to be transmitted and preferentially received using advanced pattern recognition algorithms. Signals are sent from each sensor to the Archive PC according to a schedule of tests controlled from the Archive.
In the Archive each sensor has a specially trained artificial neural network to check each new signal from a sensor, one neural network for each sensor. Fundamentally, the signal from each sensor should remain the same so that any changes can be detected - it's not quite as simple as that, unfortunately. Firstly, temperature changes in the structure cause the signal to change, secondly the piezoelectric elements in the sensor tend to deteriorate with time but the CMS has special methods to compensate for those changes.
Each neural network gives an output which is a probability of significant structural change. These are combined in the Archive in a special decision tree, in which results for sensors can be grouped, corresponding to the structural elements to which the sensors are attached. Consequently, the decision tree creates a probability of significant structural change for each structural element; these structural elements are combined in the same way to produce a result for the whole structure. There are many ways of combining the different probabilities and CMS provides a very flexible set of methods; precisely how to combine the probabilities is something for a structural engineer to perform during commissioning of the monitoring system.
Once the CMS is installed and commissioned it runs more or less automatically. The Archive PC is self booting so even if there is power interruption the PC will re-start itself, re-start the Archive program, re-start the testing schedule and re-start the decision-making procedure. Clearly, during any periods of power-loss CMS does not function and monitoring is suspended. Otherwise CMS collects and monitors the structure and processes the information, up-dating its risk analysis for the structure after every sensor collects new data. It takes about 15 seconds for each sensor to collect data and send it to the Archive so a CMS of 100 sensors can test at all locations once in about 30 minutes - meaning a structure can be tested 50 times a day if necessary but this is probably much more frequent than necessary.
CMS can provide alerts or warnings
in a variety of ways. It can communicate with a CAD program, upon
which a model of the structure can be created and then sensor
objects can be added to the model to model the positions of real
sensors on hte real structure. The Archive will send results of
decisions from the artificial neural networks to the CAD program
and the appropriate sensor objects are modified (for example by
a change in colour) to alert an operator about any significant
changes. Archive can alert in a variety of ways including sending
e-mails, so it is not necessary for an operator to be physically
next to the Archive. One operations centre could monitor many
structures each running a CMS.
One form of output from CMS - a CAD program shows a model of a structure with its network of sensors, these change colour when significant change is detected.