In future, engineered systems will need to be designed so they can be repaired, not replaced. This requirement is being driven by resource efficiency, material costs, warranties and the rising popularity of service-led business models. Such is the growth of product-service systems that Andy Harrison, chief life cycle engineer at Rolls-Royce, told industry colleagues at a strategy day run by the EPSRC Centre in Through-Life Engineering Services (TES) in May that there is “a vision of the not too distant future that nobody will ever buy an asset again ... they will simply pay for some kind of service or functionality”.
The ‘power by the hour’ model made famous by Rolls-Royce is now becoming de rigueur across non-aerospace and defence sectors, including transport – think Bombardier Transportation’s Voyager project – shipbuilding and operation, energy and power generation, and automotive.
Researchers at the TES Centre, led by Professor Raj Roy at Cranfield University and colleagues at Durham University, are working on a national strategy for TES. They contest that Britain has a lot to win or lose in the exploitation, or not, of product-service systems as a national speciality.
Many through-life strategies require some clever technology. One of the most eye-catching areas is autonomous, or self-maintaining, repairable systems. Modern engineering must perform reliably in the event of random ‘upset’ events that threaten to induce malfunction or unpredictable behaviour. These needs are fuelling the integration of fault-tolerant and self-repairing techniques into electronic systems at the point of design.
Within electronics, the two main methods to maintain in-service operation in the presence of faults are fault masking – or fault tolerance – and self-reconfiguration to remove the faulty logic. Each has its own method for fixing the problem, and both lead to increased availability.
Fault masking – where the fault is still there, but a process circumvents the problem caused by it – has been around since the 1950s, when research by John von Neumann postulated that engineered systems could have ‘massive redundancy’ built-in to bypass faults that arise. “It didn’t take off because it’s expensive,” says Dr Richard McWilliam, research fellow at the Centre for Electronic Systems at Durham University – part of the EPSRC’s TES Centre. “In fact, Nasa went a different route, pursuing ultra-reliability, using a technique called component screening, where every part was carefully selected.”
Fault masking is experiencing a renaissance, but in a type of nanoscale-device manufacture that will differ radically from traditional methods for making computer processors, for example. The low manufacturing yield of this new paradigm makes fault masking necessary. Professor Alan Purvis, head of the Durham centre, and McWilliam are interested in how adoption of fault masking in manufacture could improve through-life strategies for keeping big assets running.
The other method, self-reconfiguration, re-establishes a fault-free system by deactivating logic that has suffered permanent faults. This area is much newer, and requires a physical action to eliminate what is faulty. How can this be performed in-service?
Self-configurable repair in electronics has become possible with the emergence of FPGA chips – field-programmable gate arrays that were originally conceived to make the design of these electronics much cheaper. “You originally had a chip that you could reconfigure, done at the point of manufacture, but fixed thereafter,” says Purvis. “Now, FPGAs can be reconfigurable in the field.”
Nasa, for example, reconfigures electronics remotely in satellites and other space systems. This method could be rewarding if applied to aviation, defence, renewable energy and other industries. However, it can be risky, and avionics certification is rigorous. “You could be compromising the integrity of the system,” says Purvis.
Repair In The Air: Durham University's system to stimulate the effect of self-reconfiguring electronics in avionics While industry is cautious about adopting this technology wholesale because it fears a breach of system integrity, the potential prize is huge. Accurate self-configuring electronics could eliminate the problems associated with ‘no fault found’, where aircraft and assets are hauled out of service to repair a fault that does not exist but was detected.
Another reason that self-repair could become a boom sub-industry is the growing trend for relying on commercial off-the-shelf components in mission-critical systems. These components are becoming more vulnerable to faults, or ‘upsets’, because of diminishing feature size and lower operating voltage. “The density of the electronics means adding fine-grained ‘spares’ within the fabric is much more plausible than it would have been a few years ago,” says McWilliam.
As a result of this trend, the hardware demonstration work developed at the TES Centre has concentrated on both new detection capabilities embedded in the design and fine-grained fault tolerance.
So, will big aerospace companies be investing in self-repairing electronics soon? Self-repair needs multi-tier acceptance. The TES Centre advocates a step-by-step approach in this safety-obsessed market, where few companies make step-change technology investments. First, persuade industry to invest in far better fault detection – not a trivial thing, says McWilliam. “A lot of systems have insufficient fault detection, and self-repair depends on more accurate, localised fault detection. That’s the bedrock that could see a trickle-down.”
The TES Centre is talking to semiconductor companies that install processors in cars, where the aim is to reduce hundreds of chips per car to just a few advanced chips that could perform multiple applications.
Beyond self-healing systems, engineers are also working on autonomous robots to maintain complex assets.
The University of Nottingham co-ordinates a European research team for the FP7 MiRoR project – Miniaturised Robotic System for In-Situ Repair – which has designed and realised a hybrid walking hexapod with a small robot that can perform machining operations on large assemblies such as bridges and nuclear power stations.
First reported in PE in 2010, the project has really moved forward – literally. The hexapod can now walk at about one metre per minute and stop at a required location, reference itself against a ground feature, and start machining. The final working prototype has been built by Spanish company Tekniker, but the Nottingham team has designed and built a snake-like ‘continuum robot’.
The 1.2m arm operates through 24 degrees of freedom and can support an end-effector for machining tasks in hard-
to-reach locations. The two systems – the hexapod and the continuum robot arm – can work independently or coupled in a system called Mini-RoboMach, with a unitary control unit developed by the IPA Fraunhofer Institute in Germany.
The prototype offers promise for autonomous repairs of engineering structures, and this potential is reflected in the project’s industry partners, which include Rolls-Royce, Spanish construction firm Acciona, oil and gas group Petrom and IntelligeNDT Systems in Germany. “No one else has developed a walking hexapod machine tool with a continuum robot on top, where both can perform machining operations in combination or separately,” says MiRoR project co-ordinator, Professor Dragos Axinte of Nottingham University.
The hexapod is equipped with a wide-view laser sensor for navigation and a series of smaller cameras for referencing itself against the machining job, for which a high-speed spindle is mounted on its upper platform. The continuum robot has several innovative features, including a hollow core that can house the mechatronics for different appliances.
Key features of the latest prototype are fewer actuations, so the continuum can be compact and fitted to the hexapod. It also has a rigidising system “that can be selective – it can stiffen parts of the continuum robot while leaving others free to perform machining operations”, says Axinte. The rigidising feature enables the robot to take relatively high payloads.
Ultimately, MiRoR aims to conduct repairs on structures where the robot can position itself, work autonomously and ‘reason’ on how to access the working area, while learning from its experiences.
MiRoR was funded by the EU’s FP7 fund for the Factory of the Future, and the team’s next big deadline is January 2016, when it will exhibit its capabilities to industry partners. Nottingham University has built a demo ‘playground’, replicating an offshore platform, parts of a nuclear power station and large engines.
It’s clear that all these through-life strategies still present risks and challenges. But if these can be tackled, and the potential of the technologies realised, then the vision of Rolls-Royce’s Harrison looks a step nearer arriving.