When it becomes operational in 2007, the Diamond synchrotron particle accelerator will provide the UK's first third-generation light source, superseding the ageing facility in Daresbury. A doughnut shaped super-microscope the size of several football fields, it will produce incredibly intense light - mainly as X-Rays - that will help researchers in the development of new medicines and high tech materials, as well as in the investigation of environmental issues such as climate change.
As a third-generation design, the Diamond synchrotron will provide one of the brightest and most powerful sources of synchrotron radiation available, attracting research projects from across the globe. By 2007, phase one of the project - the UK's biggest research project for three decades - will have been completed, and up to seven synchrotron beamlines will be available to researchers. By the time of its completion in 2012, 22 beamlines will be operational, with the scope to add 18 more according to demand.
At the heart of the doughnut shaped structure is a linear particle accelerator which fires electrons into an inner ring. With powerful magnets steering the electrons around the ring, RF fields accelerate them close to the speed of light. From the inner ring, the electrons pass into a larger storage ring, again guided and excited by oscillating magnetic and RF fields. As they travel around, they emit synchrotron radiation which is channelled into beamlines where researchers can draw off the light frequencies and energies they need to perform their experiments.
Built up from 24 individual cells, the storage ring has a circumference of 561.1m, storing 3.0GeV of energy in a particle stream that has a life of around 10 hours. To ensure maximum efficiency, the machine operates at an extremely high vacuum of 0.000000001mbar. With such a high vacuum, and with so much energy from the oscillating magnetic and RF fields, the provision of effective machine protection was a key design consideration.
Overall control of the facility is provided by an EPICS distributed control system - an established choice within the accelerator community, offering all the necessary functionality for accelerator-related systems plus full support for the required hardware interfaces. However, it was quickly realised that such a system was not the best option to provide the critical interlocking systems needed to protect the machine. This was far better left to dedicated hardware. A PLC subsystem was considered a more reliable means of managing the protection functionality.
Whilst the use of the EPICS distributed control system was taken as read almost from day one, the same could not be said for the particulars of the PLC subsystems, as electrical project engineer Simon Lay explains: "It was clear that we needed a control system that was modular in design, and distributed. That was the best way to assure the overall reliability of the control system - our target is to achieve 99% availability for the machine for operation in excess of 6000 hours per year. Single point failure systems take too long to diagnose and too long to repair.
"At the same time we also recognised a need for the control system to be scalable. The facility is expected to have a 30 year life, and any necessary expansion of the control system in tandem with the evolution of facility must not impact on the performance. Finally, again because of the long lifespan, it was important that we built the control system on open standards and an open architecture. The control system had to have the ability to evolve to incorporate upgrades and new technology, and had to ensure seamless integration at all levels.
"We investigated a number of PLC options, but it was Omron's products that not only met our immediate requirements, but also offered the greatest potential for expansion in line with the evolution of the facility."
With input from Omron's engineers to evaluate different PLC options, it was decided to base the protection subsystems around the company's CJ1 series PLCs. "Reliability is obviously paramount in an application like this," says Lay. "And in 15 years I've never seen an Omron PLC fail, so I had a high level of confidence in their reliability. Also, in setting up the subsystems, we wanted to be able to keep things as simple and tidy as possible. The extensive on-board I/O capabilities offered by the CJ1 PLCs meant we could eliminate the need for secondary cabinets."
Equipment protection
The machine protection concept is built around a series of interlocks on each of the 24 cells of the machine, with the idea being to protect the machine by isolating any individual cell as quickly as possible. Parameters being monitored include the flow of water that cools the magnets, plus the temperature trend data and any thermal shocks, the presence of obstructions in the pipework, system pressures, and vacuum leaks within the cells.
As well as monitoring critical parameters, PLCs also control all of the vacuum valves, preventing a vacuum valve being opened without there being good vacuum on both sides. In all there are 24 machine protection PLCs, 24 4-valve control PLCs and 24 6-valve control PLCs. All the controllers are networked over a fibre optic star.
"We wanted to separate the machine protection functions from the valve control for a number of reasons," explains Lay. "Building up such a high vacuum takes a long time, and when the machine protection PLC trips you don't necessarily want to lose the vacuum as well. So using separate PLC systems helps to maintain machine availability. Also, breaking down the functionality onto a number of discrete PLCs means we can build and test the machine in blocks."
The physical realisation of EPICS at Diamond is as multiple embedded VME systems, called IOCs, and these communicate with the CJ1 PLCs over Ethernet. Lay comments: "The seamless networking means that, if there is a trip condition, we can interrogate the PLCs from within EPICS to see which interlock has tripped and why."
With the PLCs quickly proving their worth, Lay reports that the CJ1 is now the site standard for control at subsystem level. Realising the interlocking in PLCs as opposed to the IOCs removes the need to protect the IOC outputs in the event of an IOC crash, so improving reliability. And having to restart an IOC does not disturb the operation of the equipment. Further, the strategy provides a level of isolation from the operational parameters of the control system, ensuring that changes are not overly easy to make by unwitting operators whilst maintaining flexibility to enable the technical group to make changes when required.
Lay is now investigating the use of the CJ1 PLCs to provide dynamic control of the positioning of the magnets. "In guiding the particle beam, there can come a point when you can run out of adjustment from just varying the magnetic field alone. But we have an idea that by mounting the magnets on girders actuated by a cam and jack system, we could effect some subtle positional changes which would increase the level of adjustment that could be made.
"This is a very dynamic project," Lay concludes. "In some areas we don't even know yet all the final requirements. Omron's PLCs not only give us the reliability and performance for the systems that we know we need, they also give us the maximum possible flexibility to accommodate the ones we don't."
