Engineers are developing a system that could save lives and billions of dollars in maintenance costs by using radar-like signals to detect when structures and mechanical parts, such as those in aircraft, are about to fail.
The ‘structural health monitoring’ system uses a network of devices that could be attached to, or embedded in, an aircraft or vehicle. Some of the devices, called actuators, send out high-frequency sound waves that create vibrations in surrounding materials. The other devices are arrays of sensors that detect acoustic signals bouncing off of the vibrating material.
By continually comparing data from the sensors, the system can determine whether a material is damaged, said Douglas E. Adams, an assistant professor of mechanical engineering at Purdue University.
Adams will present a research paper about the system December 13th during the International Conference on Smart Technology Demonstrators and Devices in Edinburgh, Scotland. The paper was written by Adams, undergraduate research assistant Rebecca L. Brown and engineer Mark Schiefer, a researcher from The Modal Shop, a company in Cincinnati specializing in test equipment related to vibration and acoustics.
Such systems may be used in future aircraft and commercial and defence vehicles to constantly monitor the condition of materials and mechanical parts. The monitoring will make it possible to determine whether a part is wearing out and how long it will last before failing, Adams said.
‘We want to move everything from time-based maintenance to what’s called condition-based maintenance,’ he said.
A condition-based strategy makes it possible to replace parts and materials only when they are getting close to failing, instead of simply replacing them after every so many miles or hours of use, which is the conventional approach.
‘You want to do this because it’s expensive to service systems,’ Adams said. ‘Let’s not service it just because the warranty says so. Let’s service it when it needs servicing.’
Adams created the software algorithms that enable the system to properly interpret structural vibration and acoustic signals. This interpretation is made more difficult by the fact that materials react differently to vibration as they heat up.
‘When you start up an engine on an aircraft it’s cool and a half-hour later it’s hot,’ Adams said. ‘Will your indicator show that there is damage just because the temperature has changed?
‘Materials also behave differently in the afternoon than at night because they’ve been in the sun all day. If we are monitoring materials in a bridge, we have to be able to tell the difference between damage and the sun rising or setting.’
A system being developed at Purdue was able to do just that: accurately determine when material is damaged, regardless of its temperature. The system also was able to determine the location and severity of the damage. It yielded accurate results even when researchers fluctuated how strong the sound waves and vibrations were; such fluctuations can confuse other monitoring systems.
The monitoring system uses a concept known as ‘structural radar.’
‘We take an array of sensors, and we use a technique similar to that used to track aircraft,’ Adams said. ‘We send out a burst of energy in 360 degrees, and that energy reflects off of damaged material and is picked up by sensors.’
The sensors could be used in aircraft or automotive components that are subjected to the most fatigue.
‘One example of a fatigue-critical area is aircraft landing gear,’ Adams said. ‘Every 50 landings, parts in the landing gear are routinely replaced.’
As commercial and military fleets get older, an increasing amount of money must be spent to maintain the ageing aircraft. A condition-based maintenance system could save billions of dollars a year for the commercial aircraft industry and the US Department of Defense, Adams said.
Such a system also could save lives by better monitoring the condition of aircraft, NASA launch vehicles and weapons systems like tanks and other vehicles.
Some components in weapons systems are made of recently developed composite materials. As they age, individual layers in these composites can separate, causing the so-called composite matrix to crack, making them weaker and posing a danger to military personnel, Adams said.
However, actuators and sensors could be embedded inside such materials during manufacturing so that the monitoring system is an integral part of components in everything from tanks to the family car, he said.
‘It costs a lot of money to schedule your car for service every so many miles, but does it need to be serviced every so many miles?’
The answer, usually, is no. The lifetimes of mechanical parts and fluids, such as lubricants, are largely determined by how hard they are driven, not just how long they have been in service, Adams said.
Future ‘self-healing’ technologies could use such a monitoring system to detect when a part or material is about to fail and then release a substance that provides temporary strength or support, preventing immediate failure. Such a technology might be used in aircraft to prevent catastrophic part failures.
The monitoring system relies on a theory known as structural diagnostics using nonlinear analysis, or sDNA.
By their very nature, non-linear systems are difficult to monitor.
‘Variability in all kinds of systems and uncertainties in their operating environments make it difficult to design general structural health monitoring systems,’ Adams said. ‘For instance, no two cars or aircraft are the same, even if they are the exact same year and model.
‘Problems that don’t appear in one can appear in the other.’
Engineers are far from being able to monitor all of these variables, especially in vehicles that are subjected to the most extreme conditions.
‘For instance, a reusable launch vehicle might see 1,500 degrees Fahrenheit during exit to, or re-entry from, space,’ Adams said. ‘We don’t have the capability yet to simulate these temperatures.’
Some types of monitoring systems currently are being used in manufacturing. One American automaker uses such a system to detect when machine-cutting tools, called spindles, are wearing out. And some commercial aircraft use the systems to monitor whether landing gear is working properly.
‘The primary difference between these other systems and our system is the algorithm and the way in which we implement it,’ Adams said. ‘Our system is a collection of structural health monitoring nodes, each of which implements the sDNA algorithm. The general idea is that we collect data all across the structure.’
In future work, Purdue research engineers plan to conduct long-term tests in NASA aircraft, as well as in vehicles related to military applications.