Control Sheet Cosylab’s Newsletter Volume 29
ISSN: 1855-9255
December 2016 Table of Contents
2
Accelerator Systems for Science and Proton Therapy: Apples and Oranges?
4
Advances in Control System Design: Increased Reliability of Accelerator Systems
7
Fast Orbit Feedback for Pohang Light Source
10
Cosylab USA @ SLAC: Specialists Cover Peak Loads
12
Cosylab Supports Young Engineers
13
This issue features the next in our Cosylab Experts series of articles. In this issue, Gašper Pajor talks about accelerator systems for research and for proton therapy.
We present our approach towards the implementation of automated predictive diagnostics or prognostics and health management (PHM) as a component of the accelerator control system.
Cosylab implemented a fast orbit feedback system for the Pohang Light Source in Korea.
Laboratories, SLAC included, are increasingly opting to use contract specialists to help cover the peak workloads that arise when a major new construction project starts.
As part of Cosylab’s commitment to giving back to the scientific community, we sponsored the IEEE Student Branch (SB) of Ljubljana to attend the recent Student and Young Professional Congress.
New Year 2017
Cosylab d.d., Teslova ulica 30, SI-1000 Ljubljana, SLOVENIA Phone: +386 1 477 66 76 Email: controlsheet@cosylab.com URL: www.cosylab.com
page:
Control Sheet
Volume 29
2
ISSN: 1855-9255
Cosylab Experts
Accelerator Systems for Science and Proton Therapy: Apples and Oranges? We interview Gašper Pajor, Senior Project Manager at Cosylab. Gašper shares his insights on what control system integration projects in different accelerator applications have in common and what sets them apart. In particular we’ll be looking at the differences between machines used for particle therapy and those for pure research. About Gašper: Gašper has more than 15 years of experience with accelerator control system integration. His references range from the booster at the Australian synchrotron, to managing projects for medical applications, to overseeing the turnkey control system project for the gamma source for the Extreme Lightsource Infrastructure in Magurele, Romania. He currently heads the group that develops and maintains the medical systems portfolio of products of Cosylab. Control Sheet: Before we go into comparing various aspects of accelerator integration projects for different application fields, is this even a reasonable comparison? Or are we just comparing apples and oranges? Gašper: A proton therapy (PT) machine has quite a narrow application field, where the “scientific accelerator” is a very broad term. A fair comparison would be between machines of similar complexity. Projects with a time line of say 3 years cannot be compared to mega-projects of 10+ years. Let’s not consider those in this comparison. Control Sheet: OK, talking about time frames, is there a difference in the time frames when the different medical vs. scientific parties discover a need and start talking to Cosylab about them? Gašper: There is no hard difference
there is a more gradual achievement of maximum machine performance, often over many years. Of course you have significant milestones like first light and full energy. Compromises made along the way become upgrade or improvement possibilities for future updates. With medical accelerator projects, all tasks that are stalled or incomplete need to be sorted out before the first patient is treated. Control Sheet: Can we say then that scientific projects take longer? Gašper Pajor
between how the medical projects and the scientific accelerator projects approach us. It can happen at any stage of the development process, from way before a line of code is written, too late in the project when they figure out that there is too much work on the shelf for the in-house team. Control Sheet: Is there maybe a difference with strictness of deadlines? Gašper: It is more of a question what the deadlines mean. In PT, control system deadlines are more tightly coupled with the world outside the accelerator and there’s a sharp deadline, called first patient, where there’s obviously no room for negotiation. In comparison, in scientific machines
Gašper: You can’t say that. What we can say: because there is a very strong business aspect to the medical machines, the time line is usually more ambitious. How well this time line can be followed through depends on how well the engineering team is put together. Control Sheet: One could speculate that the less ambitious time line on scientific projects is based on prior experience? Gašper: It relates to the fact that on some accelerator machines, the need is to maximize physical performance, so the engineers push the boundaries of components including the Control System, and that process can be expected to take time. Control Sheet: How about various
page:
Volume 29 project assumptions and dependencies, e.g. that the building construction time line will be met? Do the worlds differ in your experience? Gašper: The whole undertaking of putting a PT machine into operation includes everything from groundbreaking onwards. There is a strict and painful time line in place. Painful in the sense of reputations but even penalties, and to add to that the humanitarian aspect of patients waiting to be treated, that make the whole project a very tight collaboration. There is a lot at stake and that dictates where the management focus goes. Control Sheet: Are there best practices for meeting such tight deadlines? Gašper: For a medical project where you have 3 years and a green field, what happens is that you hire the most experienced people in every field. Best civil engineering, best equipment suppliers, best integration engineers. I have not heard of a medical project being delayed because of the building. A delay is so costly it is simply not an option. Control Sheet: Do the contractual setups reflect different sensitivities, or realities between partners? Gašper: Scientific projects are usually built with public funding. It means there are rigid rules how project managers are allowed to spend that money. It also means that when they order services or products, they have them already budgeted, so it means that, job-well-done, there is little business risk of a supplier not being paid. On the other hand, the stakeholders on the medical projects are often private companies, so there is always a certain business risk involved. Control Sheet: Changing the topic. Patient safety is essential in medical devices such as PT machines. There are strict processes, documentation
Control Sheet requirements and medical certification. Does that make the project harder? Gašper: For medical devices the certification is of course part of the game and processes are put in place to make sure certification is achieved. On PT projects, all the stakeholders acknowledge that the processes are in place for the right reasons, not as a means in itself, and that any shortcuts considered would come back with a vengeance before the end of the project. Requirements management, risk management, traceability, the testing strategy, V&V, etc. Although all this forms an additional upfront burden to engineering teams, it mitigates against the temptation of unmanaged deviations from the plan, the addition of nice-to-have features. The processes imply stricter planning to achieve them. In a way they pay themselves back. Control Sheet: Most Big physics machines are conceived as a one of a kind. PT machines are often conceived with multiple instances in mind, or as product instantiations. What implications does that have on the integration project? Gašper: You can indeed make a distinction between machines that are setup as one of a kind and machines that are conceived as a repeatable business. Staffing for the one of a kind type project is typically a mix of people who come from other (scientific, international) projects and local engineers. Both types stay for the long term. Companies that build medical accelerators have the goal and ambition to build a product with support so that it runs. In order to make that a sustainable business model, you have to do that many times over (pay back of R&D investment). It includes the use of experts, who only stay long enough to
3
ISSN: 1855-9255 do what is necessary and move onto the next project. In my experience here (with machines where the goal is to setup multiple instances) the setup never becomes rigid, just because of the business aspect. The business drives all the key decisions. As a result in such an environment one cannot expect that an organizational steady-state is ever reached. There is a market that trickles down to all aspects, including engineering. If business says “we’ll drop the PLC’s”, you drop the PLCs and find a new role for the PLC guy. In my experience, a lot of past decisions are questioned, challenged continuously. There is no illusion such a steady-state will ever be reached. Control Sheet: There’s an ROI, Return on investment, and it’s defined and measured differently between those projects? Gašper: On a PT machine the ROI is definitely clearer: the mission of the PT companies is to cure patients but to do so they have to make the bottom line, create continuity and shareholder value. It has been demonstrated that this can be done. As a result of the free market, there is more than one way to fund, and recover, the very large upfront investment. The picture (of ROI) is more complex on the side of scientific accelerators. Accelerators serve several tens of thousands of users, scientists, companies every year in the US, Europe and Asia alone. The impact of those research activities are only very indirectly measurable and very hard to express in monetary terms. Facilities also serve local users as well as visiting researchers in a spirit of open research. In addition, accelerators also have the benefit of helping to train the scientists and engineers of tomorrow and they unite the people from different countries and cultures today! I think
page:
Volume 29 we can all agree the progress through accelerator enabled research is there, but is ROI the right paradigm for assessing that? Control Sheet: Last question, do technologies of control systems differ?
Control Sheet Gašper: You find the same technological buzzwords pop up on either side. There are no technological “silver bullets” the other side would be unaware of :). As a result, project references in the one domain are equally relevant in the other. In the strict technological sense for sure, in the broader
4
ISSN: 1855-9255 project management sense also, if the integrator has the size, mileage and flexibility to cater to the differences. Control Sheet: Gašper, thank you for your time!
Advances in Control System Design: Increased Reliability of Accelerator Systems Matej Gašperin (Cosylab), Matjaž Omladič (Jožef Stefan Institute)
In commercial applications of accelerator systems, high availability of the accelerator is one of the key requirements. At Cosylab, we are actively developing strategies by which the control systems will be able to implement the required features and alleviate the issues related to reliability of accelerator operation in all the stages of the device life cycle.
Introduction In commercial applications of accelerator systems, high availability of the accelerator is one of the key requirements. An example of such systems are medical accelerators for particle therapy. We present our approach towards the implementation of automated predictive diagnostics or prognostics and health management (PHM) as a component of the accelerator control system to increase reliability of accelerator operation. The main features of such a system are: ◊ a state-of-the-art reliability model of the accelerator, taking into account available data on the components such as estimated component lifetime, component recovery data and their mutual relation, ◊ an automated condition monitoring system, which enables timely
detection of underlying faults and predicts where a failure is likely to occur, so that a mitigating action can be taken in a more controlled manner. The anticipated effects of PHM systems are visible in a decreased number of failures, shorter service intervals and optimal utilization of maintenance resources.
PHM Systems for Particle Accelerators The accelerator contains several crucial components and subsystems that are subjected to wear, material stress and environmental influences. These eventually cause the equipment to fail and can result in beam trips or even emergency shutdowns of the accelerator. Furthermore, if such a fault is unexpected and unanticipated, main-
tenance procedures take longer than necessary. The implementation of efficient and effective maintenance strategies is therefore of paramount importance in guaranteeing a reliable, cost efficient and safe operation of the system. An important feature of condition monitoring systems is the prediction of the future evolution of the fault. This enables the plant personnel to accommodate maintenance actions well in advance. Even more, it can predict the remaining useful life of a component under changing operating conditions, thus providing information to operators on how the different operating regimes will affect the component’s useful life. Research shows that failures go through a distinct incipient phase [1]. This means there are some noticeable indicators, which provide advanced
page:
Control Sheet
Volume 29 warning about the onset of failure. The role of automated condition monitoring (CM) is to detect this onset, localize the root-cause and, possibly, trend its progression over time, in a timely manner. The remaining time until final breakdown can be long enough to allow for efficient maintenance service. The PHM system for a particle accelerator will consist of two main parts. The first part is the probabilistic model of the accelerator reliability. This model is used to represent the reliability of individual components and model their interactions in terms of error propagations. This model is then used to evaluate the design of the accelerator in terms of reliability and is used by the condition monitoring system during machine operation for fault localization. The second part of the PHM system is the automated condition monitoring system, which monitors the relevant signals on the accelerator during its operation and has the capability to detect the faults in their early stage of occurrence, isolate the component and the root cause of the failure. In specific cases (depending on the de-
vice type), the condition monitoring can also detect the fault in its incipient phase and can predict when the fault will occur in advance. The PHM system (Figure 1) is designed around the three main tasks, namely observation, analysis and action. Observation includes appropriate hardware and software components for signal acquisition processing in order to compute the required feature values. The features are then analyzed for presence of fault and possible trends. The anticipated effects of successful development of the Accelerator PHM system are up to a 50% reduction in mean time to recovery, up to a 30% decrease in unnecessary maintenance actions and increased availability of the accelerator.
Availability (out) matching Reliability and Predictability Mathematical modeling of complex systems in terms of reliability is a challenging task, mainly due to a large number of components and their interactions. Effective reliability model-
Figure 1: The PHM System
5
ISSN: 1855-9255 ing requires understanding of the failure mechanisms from a broad range of engineering skills. Additionally, advanced probabilistic modeling techniques have to be employed in order to model the relations of individual faults and the way they are propagated through the system. For this purpose, our aim is to build a copula-based shock model, taking into account available data of the estimated component lifetimes, their recovery data and their mutual relation. This model will be used in a simulation-based search for stress scenarios. After the start of operation of the plant, future maintenance and other operation data is used to update the model by using algorithms for autonomous machine learning. It has been pointed out in recent literature on system reliability that the joint law of component lifetimes (which is a crucial ingredient in evaluating availability and reliability) depends not only on the actual design of the system but also on stochastic relations between the components. For instance, a major power supply fail-
page:
Volume 29
Control Sheet
6
ISSN: 1855-9255
ure can seriously affect the MTBF approach for many components of the system. Other problems may arise in a more subtle way when the standard approach is in use. In general there are local shocks in the system that affect just one component and global shocks that affect many components and are sometimes even called Armageddon shocks. One of the main tools that has been recently recommended in this setting is copula theory which takes care of inter-component dependencies. A classical approach may lead to seriously biased estimates that can have unpredictable consequences on availability and reliability of the plant.
Model-based Condition Monitoring for Accelerator Components The role of an automated condition monitoring and prognostics system is to detect the presence of a fault (fault detection), localize the root-cause (fault isolation) and estimate the future progress of the fault and the time of failure (prognostics) in a timely manner. The basic stages of CM are feature extraction, feature evaluation and fault isolation. A feature is a function of the measured signal and should be sensitive only to the fault while being insensitive to the operating conditions. The output of the PHM system consists of a probabilistic assessment of the current risk of failure for individual accelerator components, obtained by fusing condition monitoring data and historical records of MTBF, identification and localization of the fault or failure in the case an anomaly is detected, estimated optimal failure prevention strategies and appropriate maintenance actions. Model-based fault detection relies
Figure 2: Model-based residual generation
on mathematical models to estimate the current condition of the system. Depending on the available prior information about the system physicsof-failure, the model can be either passed on physical equations (whiteor grey-box model) or data-driven (black-box model). The basis for fault detection and identification is the residual values, which are computed from comparing the model output and measured values, or comparing nominal and estimated system parameter values. The process workflow is outlined in Figure 2. Fault identification and localization is performed from the Fault Signature Matrix (FSM), which connects faults and specific residuals (symptoms). Given a set of symptoms and a set of considered faults, the theoretical fault signature matrix can be defined by binary coding the effect of a fault in every symptom. Then, fault isolation consists of looking for the closest
matching fault signature pattern in the FSM that matches the observed signature. The matching can be done by logical tests if the faults are uniquely identifiable or probabilistic inference in case of overlapping signatures and uncertainties. After the successful identification of the fault, the prognostics module tries to estimate the severity of the fault in terms of remaining lifetime of the component. One approach to prognostics is to analyze process variables that quantify inputs and outputs of components as archived by the accelerator’s distributed control system. These are not necessarily the same as symptoms but it may be the same variables as symptoms, since some features may have good diagnostic properties, but lack information relevant for prognostics. By quantifying and modeling the trend in the corresponding feature value, the presence of the impending fault can be estimated.
page:
Control Sheet
Volume 29
Conclusion We have discussed and presented a roadmap for the implementation of prognostics and health management for accelerator systems. PHM is a promising technology that can be used within the maintenance decision-making process to provide failure predictions, increase the operational availability of systems, lower sustainment costs by reducing the costs and duration of downtime, improve inspection and inventory management, and lengthen the intervals between maintenance actions.
The past decades have brought significant development in technologies for machine condition monitoring and fault diagnostics. Advances in vibration sensors, acoustic emission sensors, sensors for oil particle counts provide rich sources of diagnostic information by applying sensor fusion approaches. These opportunities have not been fully exploited yet. Our planned future activities in development of the Accelerator PHM System include: fault tree analysis of the Accelerator systems, Analysis of historical data on component MTBF and design and development of prognostic models for individual components.
7
ISSN: 1855-9255
REFERENCES [1] R. B. Randall, Vibration based condition monitoring, Wiley, Hoboken, NJ, 2011.
ABOUT THE AUTHORS Matej Gašperin, PhD, is a Software Developer at Cosylab. He has a strong background in control systems engineering and statistical signal processing. His work is mainly focused on development and implementation of advanced control and diagnostic strategies for medical accelerators. Prof. Matjaž Omladič, PhD, is a professor of Mathematics at Jožef Stefan Institute’s Institute of Mathematics, Physics and Mechanics and is collaborating with Cosylab as a scientific advisor. His research topics include probabilistic modeling, deep learning algorithms and big data analysis. His research results are published in over 70 scientific papers.
Fast Orbit Feedback for Pohang Light Source Blaž Kranjc (Cosylab) and Damjan Kumar (Cosylab)
To ensure that high quality synchrotron light is produced from a light source it is necessary that the beam orbit in the storage ring is stable. The noise introduced by the various synchrotron components can be partly compensated by an active orbit correction feedback system. Orbit correction feedback systems periodically adjust the current on the corrector magnet to compensate for the deviations to the beam orbit.
Introduction Orbit feedback systems are usually composed of two subsystems, a slow orbit system (SOFB) running at frequencies around 1 Hz and fast orbit feedback which corrects for high frequency errors. The typical FOFB systems run at 1 kHz or faster. Cosylab implemented a FOFB for the Pohang Light Source (PLS) in Korea. The FOFB was designed to fit the hardware constraints of the PLS and it is able to run at frequencies up to 1
kHz. Due to the requirements of the FOFB a SOFB application was also rewritten with the new functionality as a softIOC. A diagram of the orbit correction is shown in Figure 1. The source code for the whole system was provided to the customer which enabled the customization of the system after the delivery.
Basics of Orbit Correction With orbit correction we are correct-
ing small deviations from the reference orbit ∆x with a current change on the corrector magnets (∆I). In the regime of small changes we can linearize the relation between the two and introduce the response matrix R in a way that: ∆x=R∙∆I The response matrix is a property of the machine and can be either obtained from simulations or can be measured. As the response matrix is usually non-square, a direct inverse
page:
Control Sheet
Volume 29
8
ISSN: 1855-9255 requires an additional communication between the SOFB and FOFB system.
FOFB Architecture The architecture design was constrained by the PLS selection of the VxWorks based VME single board computers for the nodes and no access to the external timing system. The FOFB system consists of 12 FOFB nodes equally distributed around the storage ring near the positions of the beam position monitor (BPM) readout and 1 FOFB master. The FOFB master is required due to lack of access to the external timing system. The main role of the master node is to generate events at precise intervals. These events are used for the synchronization of the FOFB nodes. In addition to generating events, the master is also responsible for system health monitoring, configuration distribution and interfacing to the existing control system with EPICS interface. An outline of FOFB architecture is depicted in Figure 2. Figure 1: PLS orbit correction systems.
of the matrix cannot be calculated and a singular value decomposition (SVD) based pseudoinverse R-1 is used. As the matrices are usually ill-conditioned some of the smallest singular values are discarded during the pseudoinverse calculation. The current changes on the correctors that correct for the orbit errors are then
in a way that the SOFB also corrects for all of the corrections of the FOFB. This is called current unloading and it
∆I=R-1∙∆x. This kind of algorithm only accounts for the storage ring dynamics, so usually an additional PID controller is added to the calculation. The currents on FOFB magnets would saturate quickly if the FOFB would run on its own which would prevent any further corrections. To solve this problem we integrate the FOFB and SOFB
Figure 2: FOFB architecture overview.
A node is designed not to include any form of state machine and only operate as a listener to the requests issued by the master in the form of notifications of start and stop of the cycle. This enables the nodes to be realized
page:
Volume 29
Control Sheet
9
ISSN: 1855-9255
Figure 3: Timing diagram of FOFB.
as simple and fast actuators where it is crucial to save on processing time since it affects the repetition rate of the whole system. The master and all the nodes are connected in a ring with reflective memory (RFM) cards to share the data and events between them. Each node is connected to the BPM group to read the positions of the beam and to the magnet power supplies (MPS) to which the current is applied after calculation. All possible calculations such as SVD, disabling of the BPMs and correctors is done before the start of the correction loop to save time during the FOFB cycle. The results of these calculations are provided to the nodes as configuration via RFM.
FOFB clock drifts relative to the fast data acquisition BPM output, and the FOFB nodes might start reading data that is one cycle older. This in turn adds 100 µs of jitter into the whole system. The time line of the FOFB cycle is depicted in Figure 3. The FOFB master application responsible for dispatching events and running EPICS applications on the FOFB master are decoupled. The responsibility of the EPICS interface is to provide an interface to the rest of the control system and to the operators to
control and monitor the system. The EPICS interface is designed only to be available on the FOFB Master; it handles all relevant operations to provide the interface to the whole FOFB system as one – effectively joining master and nodes into an entity. This way, the operator can configure and monitor the system as one without the need to interface multiple components of FOFB. The FOFB EPICS interfaces exposes the following components: ◊ I nitialization MPS set-points before the orbit correction algorithm
The cycle starts with the FOFB master sending a FOFB_CYC_START event to the nodes via RFM. The nodes then read the BPM values associated with their group via UDP and dispatch them to all other nodes via RFM. Then, the remaining calculation is performed. When the node receives a FOFB_CYC_ STOP, the calculated currents are applied to the magnet power supplies. The data on the applied current values is also distributed via RFM to the FOFB master as it is required for current unloading to the SOFB. In order for the FOFB to remain in sync with the rest of the accelerator, especially the BPM readout, the master node is connected directly to the 10 Hz BPM trigger line via a DIO card. Without this synchronization, the Figure 4: Orbit stability comparison with and without FOFB [1].
page:
Control Sheet
Volume 29 ◊ ◊ ◊ ◊
Node status Desired orbit Response matrix Node and master configuration, such as disabled BPM/MPS ◊ Average current on correctors is exposed with 10 Hz to SOFB for current unloading.
Conclusion Since the deployment of the FOFB, PLS has changed the implementation of the algorithm to use TSVD based algorithm instead of SVD to improve
its performance [1]. The measured noise with and without the FOFB is displayed in figure 4. Noise at frequencies under 60 Hz is successfully suppressed by the FOFB. The FOFB is now used during the user operation [1].
REFERENCES [1] S. C. Kim, J. Y. Lee, J. W. Lee, et al., Fast Orbit Feedback at PLS-II Storage Ring, Proceedings of IPAC2016 held in Busan, Korea, 8-13 May 2016 (http:// accelconf.web.cern.ch/AccelConf/ ipac2016/papers/wepor004.pdf )
10
ISSN: 1855-9255
ABOUT THE AUTHORS Blaž Kranjc studied at the Faculty of Mathematics and Physics and joined Cosylab in 2013 as a developer. Currently, Blaž is supporting beamlines at Diamond Light Source. He spends his days with EPICS and Java. Blaž relaxes by playing board games, exploring the depths of the internet and cooking. While he is not too keen on sports, in general, he does enjoy inline skating and hiking. Damjan Kumar is currently a senior software developer at Cosylab. He has more than 20 years of experience in C, C++ and Java development, Linux system development and administration, and implementing systems ranging from CORBA-based, database-backed distributed business applications to real-time distributed controllers integrated with EPICS. Damjan understands the importance of software quality, and is a passionate advocate of test driven development, automated unit testing, continuous integration and as-high-as-feasible code coverage.
Cosylab USA @ SLAC: Specialists Cover Peak Loads Gašper Janša (Cosylab) and Rok Šabjan (Cosylab)
Large and established national laboratories, SLAC included, are usually well-staffed with controls engineers. However, when a major new construction project starts, the in-house Controls team can find themselves in a situation where they need to cover the workload of the new project and still operate and maintain the current machines. New hires are risky and it takes time to get a new person up to full speed, so labs are increasingly opting to use contract labor from specialist companies like Cosylab to help cover the peak workloads during new projects. SLAC National Accelerator Laboratory, a US Department of Energy National Laboratory, is operated by Stanford University and is located in Menlo Park, California. SLAC’s history started in 1962 when construction began on the world’s longest (3 km) linear accelerator. Today SLAC operates several machines such as the Stanford Synchrotron Radiation Lightsource (SSRL) and its flagship free electron laser (FEL), the Linac Coherent Light Source (LCLS). LCSL is the world’s brightest FEL capable of producing X-rays a billion times brighter than those available before its construction at 120 times
per second. LCLS occupies the last 1/3 of the original linear accelerator. Currently the laboratory is undertaking a major upgrade of the LCLS (known as LCLS-II). LCLS-II will operate at a rate of 1 million pulses per second, a staggering 8 000 times better than LCLS as well as a factor 10 000 brighter! At its heart will be two new lines of undulators which will increase the available energy range of LCLS. LCLS-II first light is expected in the third quarter of 2020. Continuous operation of LCLS during the construction of LCLS-II and later coexistence of both accelerators presents a challenge both in an orga-
nizational as well as a technological sense. Cosylab started working with SLAC on a few development projects in 2007. A larger collaboration with SLAC started in 2012 when Cosylab USA, a daughter company of Cosylab, was established. A Cosylab expert joined the SLAC controls team on site and started working on operational maintenance and support as well as new developments for both LCLS and LCLS-II. Since then the Cosylab USA team has grown to 7 people permanently on site at SLAC. The Cosylab USA team is a balanced
page:
Control Sheet
Volume 29
11
ISSN: 1855-9255 driver development), which can only be obtained from a larger pool of engineers, the Cosylab USA team draws on the experience at Cosylab headquarters in Slovenia. In such cases, one of the local engineers acts as a liaison between developers in Slovenia and SLAC customers, thus ensuring a smooth execution of the project. Several such projects have been successfully completed, e.g. development of the Linux kernel driver automatic test environment for LCLS magnet controller boards, merging and cleanup of event and event2 EPICS modules by rewriting the kernel module, user space library, EPICS device and record support, and EPICS integration of a digitizer using AreaDetector module.
Cosylab’s Žiga Oven infront of the LCLS-II prototype Undulator
ABOUT THE AUTHORS
Cosylab’s Robert Hoffman and Thomas Kurty
combination of expats from Cosylab as well as local engineers employed directly by Cosylab USA. Our involvement through years has ranged from operational maintenance and support for a wide range of motion control for LCLS and SSRL, high level Python development for the LCLS machine protection system configuration editor, web development
The Cosylab USA team is quite well integrated with the SLAC controls team. An established task order process helps both teams to carefully consider the scope and requirements of specific projects. We work on projects together and jointly address what solutions are best for particular cases. The added bonus of working so closely together with the SLAC controls team is that we are helping to contribute to SLAC’s rich scientific history.
for LCLS users, EPICS IOC engineering work for the LCLS-II laser and laser heater controls, LCLS-II undulator motion control, and FPGA development for the LCLS-II common platform. Our team has also presented several EPICS courses for SLAC operators and new employees. Sometimes, when a project needs a specific expertise (e.g. Linux kernel
Gašper Janša has been with Cosylab since 2003. Before managing Cosylab USA, he spent several years at GSI working as the lead for the controls part of the Slovenian in-kind contribution to FAIR. Rok Šabjan is a co-founder of Cosylab and has been working on big physics control systems since 1996. After specializing in EPICS development, Rok led the initial Cosylab EPICS projects and has established and led a team for the first big account - ITER. In the past few years, he has been focussing mostly on business development and sales. Rok has a background in physics and management.
page:
Volume 29
Control Sheet
12
ISSN: 1855-9255
Cosylab Supports Young Engineers By : Katja Perme (IEEE SB Ljubljana)
As part of Cosylab’s commitment to giving back to the scientific community, we have sponsored the IEEE Student Branch (SB) of Ljubljana to attend the Region 8 (Europe, Africa, Middle East) Student and Young Professional Congress, held in Regensburg, Germany in August 2016. The following is their report. The Student and Young Professional (SYP) Congress is an international event organized by the Institute of Electrical and Electronics Engineers (IEEE). This year’s congress was held in Regensburg, Germany from 17 to 21 August where more than 400 students and young professionals gathered from more than 350 universities across Europe, the Middle East and Africa. The goal was to experience different cultures and to make new friendships which could lead to future professional cooperation. A huge boost was given to Career, Entrepreneurship and Social skills development and the role of IEEE in our local community.
After our arrival in Regensburg, an opening ceremony was held at the Thon-Dittmer Palais theater. Afterwards we went on a 4-hour-long cruise on the Danube River. Each day we had different workshops during the afternoon and interesting social events in the evening. The Gala Dinner with the award ceremony on the region 8 level (Europe, Africa and Middle-East) was the main social event. We should also mention that our work and passion, with more than 15 events during 2015, finally paid off with winning the
exemplary student branch award. On our last evening we had a chance to show the other participants a taste of our cultures. During the multicultural evening – as it was called - we had a chance to taste food and beverages from a variety of countries and to share our country’s traditions with other participants. Experiencing such a congress is simply unforgettable. With so much to gain there is really no reason not to go. Thank you again, Cosylab, we had a great time.
Three electrical engineering students from the IEEE Student Branch Ljubljana and Maribor participated in the congress. It was an amazing experience for all of us. Apart from all the new friendships and knowledge, we have also gained plenty of new ideas for activities which we will now try to incorporate in our local student branch. All of this would not be possible without the help from Cosylab, who funded a substantial part of our trip. We would like to thank them for giving us an opportunity to meet engineers from all over the world and to gain valuable skills which are a must-have in modern day business. The congress itself took part in the UNESCO city of Regensburg, at the biggest technical university in Bavaria - OTH Regensburg. One could describe it as a place filled with enthusiasm and art.
The Slovenian Team at the at the 2016 SYP Congress: Daniel Pavlovski, Katja Perme (IEEE SB Ljubljana) and Aleš Breznik (IEEE SB Maribor). Cosylab sponsored the IEEE SB Ljubljana members.
Cosylab d.d., Teslova ulica 30, SI-1000 Ljubljana, SLOVENIA Phone: +386 1 477 66 76
Email: controlsheet@cosylab.com
URL: www.cosylab.com