07.06.2015 Views

Document - european association of national research facilities

Document - european association of national research facilities

Document - european association of national research facilities

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Maintenance and Reliability Workshop<br />

9-10 November 2011<br />

Synchrotron SOLEIL, Saint-Aubin, France<br />

Context<br />

One <strong>of</strong> the main tools used by large scale <strong>research</strong> <strong>facilities</strong> in maintaining the thousands <strong>of</strong><br />

items <strong>of</strong> equipment installed and deployed on site and to ensure that this equipment operates<br />

at a very high level <strong>of</strong> reliability is a Computer-based Maintenance Management System<br />

(CMMS):<br />

- by the Maintenance groups to manage all the equipment in their databases (from<br />

receipt <strong>of</strong> parts and stock control through to installation and renewal);<br />

- by the Buildings and Infrastructure groups to manage almost all preventive<br />

maintenance operations and regulatory controls;<br />

- and by several other groups (e.g. Vacuum, fluids, security, etc) primarily during<br />

machine downtime;<br />

- by a large number <strong>of</strong> support laboratories (e.g. Chemistry, Biology, Surfaces labs) are<br />

starting to enter their own equipment in the CMMS.<br />

All these issues surrounding equipment maintenance, i.e. all the aspects that contribute to<br />

maintaining equipment in optimal operating conditions, are in fact common to all major<br />

<strong>research</strong> <strong>facilities</strong> in Europe and a remarkable success-key <strong>of</strong> performance for the whole<br />

facility.<br />

For this reason, the CMMS working group <strong>of</strong> SOLEIL decided to organize a workshop at<br />

SOLEIL for benchmarking the experience <strong>of</strong> several <strong>facilities</strong> in Europe. The workshop has<br />

been focused on two topics, Maintenance and Reliability, with the aim <strong>of</strong> exchanging and<br />

sharing our experience with maintenance management tools, organizational requirements,<br />

working methods, difficulties encountered, and so on.<br />

1. Presentations<br />

SOLEIL: Presentation <strong>of</strong> the SOLEIL’s CMMS tool (Hélène Rozelot)<br />

Hélène Rozelot reviewed the tool’s main features (purchase requests, inventory, equipment<br />

location, history <strong>of</strong> movements, configuration management, traceability <strong>of</strong> work orders, stock<br />

management, and so on) and presented a few indicators extracted from the database using<br />

Business Objects (BO). She also looked back at the context in which the system was set up<br />

between 2002 and 2004. Lastly, she explained the organization set up to deploy the CMMS<br />

and highlighted the difficulties encountered, which stemmed primarily from the introduction <strong>of</strong><br />

new working methods and a new formalism to which SOLEIL’s teams were unaccustomed.<br />

MAX IV: Maintenance at the MAX IV laboratory (A. Månsson and J. Thånell)<br />

MAXLab may be described as a laboratory that has grown almost “organically”. This means<br />

that, right from the outset, maintenance has always been seen as one <strong>of</strong> many activities<br />

conducted by a small group <strong>of</strong> scientists and <strong>research</strong> engineers in charge <strong>of</strong> designing and<br />

building the laboratory. On the one hand, this reinforced the commitment <strong>of</strong> each person<br />

involved in maintenance, but on the other hand, as the laboratory has grown and the MAX IV<br />

project has got under way, the teams are now working with a maintenance system that is<br />

unsuitable for an organization <strong>of</strong> over 100 people.


ALBA: Maintenance at ALBA Synchrotron (L. Miralles)<br />

Lluis Miralles presented the organization <strong>of</strong> ALBA’s “Technical Support” division. The division<br />

is split into four “sections”: Infrastructure, R&D, Vacuum and Transversal. The group<br />

responsible for all maintenance at ALBA is in the Transversal section, and comprises only six<br />

technicians. Note that the Experiments division does not have any dedicated technicians.<br />

Maintenance operations are subcontracted, but maintenance management (scheduling <strong>of</strong><br />

operations and so on) is handled in-house. Lluis Miralles described the features <strong>of</strong> the<br />

PRISMA 3 s<strong>of</strong>tware used. This program features many <strong>of</strong> the typical functions <strong>of</strong> a CMMS,<br />

including topologies, stock, process workflow and statistics. He concluded with a<br />

presentation <strong>of</strong> conditional maintenance based on monitoring a number <strong>of</strong> parameters (pump<br />

pressures, motor vibrations, and so on) to trigger alerts that result in either preventive<br />

maintenance actions or warnings about potential breakages (predictive maintenance).<br />

SOLEIL: Maintenance & Reliability for the Process (P. Betinelli & L. Manciet)<br />

Pascale Betinelli started by explaining what SOLEIL wants from a CMMS and how it is used<br />

in practice. MAINTIMEDIA is an assembly <strong>of</strong> several modules (topology, stock, purchasing,<br />

process workflow and statistics). While each <strong>of</strong> the modules can be used independently, this<br />

approach is somewhat restrictive. The modules are in fact interconnected and the CMMS<br />

should be used as widely as possible to make the best use <strong>of</strong> the system. This implies that<br />

users must thoroughly understand and analyze their maintenance processes in order to<br />

make the best use <strong>of</strong> the CMMS. Pascale then displayed some data extracted with BO and<br />

showed that the use <strong>of</strong> the CMMS as a maintenance tool is very variable. She also<br />

mentioned other tracking tools used at SOLEIL, including MANTIS bug tracker, Excel, ELOG<br />

and MS Project. She emphasised the difficulty <strong>of</strong> imposing the use <strong>of</strong> a single tool or even<br />

defining rules for using each tool.<br />

Laurent Manciet then presented some more technical aspects <strong>of</strong> MAINTIMEDIA, including<br />

the Oracle database, the number <strong>of</strong> licences, the access rights defined, network architecture,<br />

and the technical problems encountered in particular last year (primarily slow database<br />

access). In their joint conclusion, Pascale and Laurent highlighted the general difficulties in<br />

using a CMMS, a program with complex functions that require users to define their<br />

requirements and processes properly beforehand. For this new modus operandi to be<br />

successful, teams must overcome their resistance to change.<br />

DIAMOND: Buildings & utilities maintenance – Contracted service or in-house staff (L.<br />

Walters)<br />

Lee Walters presented the organization <strong>of</strong> DIAMOND’s Support division, which includes the<br />

group responsible for infrastructure maintenance. He explained that maintenance operations,<br />

which had previously been contracted out, are now managed in-house, having fired the<br />

former contractors on site. He reviewed this situation, which has the advantage <strong>of</strong> enabling<br />

maintenance teams to be managed directly and makes the most <strong>of</strong> in-house resources.<br />

However, it requires considerable effort to set up, and also <strong>of</strong>fers less flexibility in terms <strong>of</strong><br />

available resources, in that external staff are required to handle peak loads.<br />

CERN: Maintenance for CERN’s cryogenic installations (S. Knoops)<br />

Sigrid Knoops presented the cryogenic equipment in use on the LHC, noting that the devices<br />

are prototypes designed by CERN. She reviewed how maintenance is organised, and who<br />

does what, depending on the complexity <strong>of</strong> operations:<br />

- Straightforward adjustments (e.g. setting a parameter) are handled by CERN in-house<br />

operators. Likewise, much more complex operations that require a redesign or that involve<br />

manufacturing new parts are handled in-house.<br />

- Conversely, standard operations (repairs, replacements, corrective and preventive<br />

maintenance) are contracted out.<br />

Her slides also included various operating statistics. INFORM EAM asset management<br />

s<strong>of</strong>tware is used (http://www.infor.com/solutions/eam/). Data are extracted with Business


Objects, and the database is linked to the technical documentation system (Engineering Data<br />

Management System or EDMS).<br />

LASER MEGAJOULE: Maintenance at the facility (H. Graillot & I. Granet)<br />

Alongside synchrotrons, LMJ is a representative <strong>of</strong> another class <strong>of</strong> major <strong>research</strong><br />

infrastructure (lasers). Although the infrastructure is still under construction, maintenancerelated<br />

aspects are already being taken into account. Almost 100,000 items <strong>of</strong> equipment<br />

have already been entered in the INFOR EAM database, ranging from ordinary devices to<br />

more complex systems such as the “target chamber“ with its 10-metre diameter. Hervé<br />

Graillot described the tool’s features (topology, stock, purchasing, process workflow and<br />

statistics) and explained that the CMMS is interconnected with the JIRA project tracking tool<br />

and the SIROCO technical database. Business Objects is also used to extract data which is<br />

analyzed to optimise maintenance procedures. In conclusion, Hervé noted that the CMMS as<br />

a whole (including purchasing, work order tracking and so on) is used throughout LMJ. He<br />

also noted that LIL, the Laser Integration Line used as a prototype for Laser MegaJoule, had<br />

used a different CMMS that its operators had found too complex to be workable.<br />

SOLEIL: Buildings & Utilities Maintenance (T. Didier)<br />

Thierry Didier, in charge <strong>of</strong> utilities at SOLEIL, presented the activities managed using the<br />

CMMS (general services, fluids and electricity), for monitoring and maintenance operations in<br />

particular. Certain preventive maintenance operations are mandated by regulations. This<br />

preventive maintenance entails drawing up work instructions, from which work orders are<br />

generated so as to schedule the group’s activities and required resources. Nonconformity<br />

sheets are drawn up where applicable. Thierry then presented the data extracted using BO<br />

and used for management purposes by team leaders. Lastly, he mentioned the motivational<br />

difficulties encountered and the steps taken to resolve them, including simplifying the<br />

technical data entered in the database, using work instructions suited to the group, and<br />

making extensive use <strong>of</strong> Business Objects.<br />

ESRF: Maintenance <strong>of</strong> Technical Infrastructures (T. Marchial)<br />

Thierry Marchial briefly presented the activities <strong>of</strong> the 30-strong Buildings and Infrastructure<br />

group, part <strong>of</strong> ESRF’s Technical Infrastructure Division. He described the equipment<br />

managed in the CMMS (using CARL Master s<strong>of</strong>tware). This equipment is classified in 3<br />

levels:<br />

- Equipment that could cause beam losses (accelerator and source)<br />

- Equipment that could affect users (beamlines and mains power)<br />

- Other equipment.<br />

He then reviewed the outsourced maintenance contracts managed in the CMMS (which do<br />

not include “primary maintenance” or regulatory inspections). These account for:<br />

- 33 documents<br />

- 4-year contracts, revisable annually<br />

- 14,000 hours <strong>of</strong> work on site in 2011.<br />

Contracts are managed using the Plan-Do-Check-Act (PDCA) quality system. Each incident<br />

is analyzed by the subcontractor and the Buildings and Infrastructure group, and corrections<br />

are made:<br />

- Identification <strong>of</strong> expected performance,<br />

- Quality control and tracking (in particular through indicators).<br />

A bonus/penalty system has also been set up for subcontractors.<br />

Orsay Proton Therapy Centre: Maintenance and reliability (S. Meyroneinc)<br />

Samuel Meyroneinc briefly presented the principles <strong>of</strong> proton therapy and the <strong>facilities</strong> at the<br />

Orsay Proton Therapy Centre (CPO). He emphasized the operational constraints (medical<br />

environment) which restrict maintenance to scheduled slots, and the need for a decisionmaking<br />

structure to analyze the potential impacts <strong>of</strong> maintenance work and how treatment<br />

results may be affected. He explained how reliability and maintenance can result in


incompatibilities with R&D: when maintenance and upgrades affect several fields <strong>of</strong><br />

expertise, it is difficult to guarantee that the results will comply with specifications (and be on<br />

schedule and within budget). With more recent technologies, preventive maintenance is<br />

necessarily limited and dependent on experts.<br />

Maintenance at CERN: On the brink <strong>of</strong> a new maintenance project (G. Perinić)<br />

Goran Perinić briefly presented the LHC and detectors. He emphasized the fact that even a<br />

minor incident on a device during operation can require the entire accelerator to be shut<br />

down, and cited examples where this has occurred. Accordingly, preventive maintenance<br />

activities are given increasing importance. Emergency repairs must therefore be<br />

distinguished from maintenance operations, and the latter must be taken into account<br />

properly to improve operation. He also explained the link between the maintenance tool,<br />

INFORM EAM, and the EDMS technical documentation system. CERN is currently setting up<br />

a new cross-department maintenance project, with the aim <strong>of</strong> reducing the frequency <strong>of</strong><br />

scheduled technical downtime from once every six weeks to once every twelve weeks. With<br />

this in mind, an audit was conducted (based on a questionnaire sent to support groups in<br />

order to assess their approach to maintenance), which showed the diversity <strong>of</strong> maintenance<br />

methods and tools used by support groups. The aim is now to define a modus operandi<br />

(methods, tools, indicators) common to all these groups. A dedicated organization and staff<br />

have been set up to complete this project within two years. A CMMS working group has been<br />

formed to define the common modus operandi, which an expert group will help to transpose<br />

into operational procedures.<br />

2. Round tables<br />

FIRST ROUND TABLE: What are the necessary resources and organization for an effective<br />

maintenance management system?<br />

First discussion on this topic: “centralised” or “distributed” maintenance?<br />

If maintenance is distributed between several groups or departments, there is a risk <strong>of</strong><br />

diluting responsibilities, duplicate work, different methodologies and tools, resulting in lower<br />

efficiency. Centralised maintenance appears preferable, but should be considered as a<br />

process in its own right<br />

Second discussion on this topic: who is responsible for maintenance?<br />

For the majority <strong>of</strong> participants, the owner <strong>of</strong> the equipment is responsible for the function<br />

that it provides and should therefore bear responsibility for maintaining it. The maintenance<br />

team, in turn, is responsible for performing the maintenance requested by the person<br />

responsible for the equipment. It is also the maintenance team’s responsibility to propose<br />

equipment improvements to the development department.<br />

Third discussion on the topic <strong>of</strong> common tools.<br />

Some participants believe there is no need to use the same tool in all areas <strong>of</strong> activity.<br />

CERN, on the other hand, is working to enhance the reliability <strong>of</strong> its <strong>facilities</strong>, and is moving<br />

to a single common tool for all tracking.<br />

Fourth discussion on the topic <strong>of</strong> rights management<br />

Most participants feel that topology management rights (creation / deletion) should be<br />

restricted in order to minimise accidental errors and ensure consistency. CERN does not<br />

restrict user rights to any great extent and counts on training and understanding <strong>of</strong> the<br />

system.


SECOND ROUND TABLE: What common indicators could <strong>research</strong> infrastructures set up<br />

and share?<br />

First discussion: motivation <strong>of</strong> teams involved in maintenance (who are also involved in<br />

R&D work which is generally felt to be more rewarding).<br />

• Remark on financial aspects: it is possible to have two separate teams (R&D and<br />

maintenance) provided sufficient resources are available.<br />

• For complex equipment, the teams must have well-honed skills Accordingly it is important<br />

to involve them in development as well as maintenance.<br />

• Training and educational aspects are important: the value <strong>of</strong> maintenance work must be<br />

highlighted to show that it contributes to enhancing reliability.<br />

• Senior management support is essential. The participants noted that defining indicators is a<br />

way <strong>of</strong> providing recognition that maintenance has an important role to play in a facility’s<br />

performance. Indicators such as those set up by CERN are a motivating factor for staff.<br />

The bonus/penalty system set up by ESRF was mentioned again, though it was pointed out<br />

that the system is difficult to implement on an in-house basis.<br />

Second discussion: what should be the ratio <strong>of</strong> preventive maintenance to corrective<br />

maintenance?<br />

This ratio has consequences on the financial cost and the cost in terms <strong>of</strong> resources. How<br />

can this be quantified? The question remains open.<br />

Third discussion on this topic: Indicators<br />

The participants noted that a number <strong>of</strong> indicators are defined by standards (MTBF,<br />

equipment availability, EN 15341, EN 1306, etc). While it is easy to implement such<br />

Indicators in some fields (infrastructure, accelerators, etc.), they are not easily accepted by<br />

scientists for whom publications are the key indicator.<br />

3. Miscellaneous<br />

ESRF proposes to make its technical documentation relating to maintenance contracts<br />

available to the participants. DESY would like feedback about the CMMS s<strong>of</strong>tware used by<br />

the various institutes. CERN proposes to share its indicators.<br />

4. Conclusions<br />

The workshop has highlighted the importance <strong>of</strong> maintenance, although there are still many<br />

open issues. Various points have been revealed:<br />

• CMMS s<strong>of</strong>tware is relatively widely used in departments which outsource part <strong>of</strong> their<br />

activities and which have contractual relations with subcontractors; however, it is<br />

much harder to impose the use <strong>of</strong> a CMMS “internally” (within a department or a<br />

<strong>research</strong> body, for example).<br />

• Senior management commitment is essential to ensuring CMMS use.<br />

• Most <strong>of</strong> the organizations in attendance have IT support personnel dedicated to<br />

setting up the CMMS (and responsible in particular for defining and administering<br />

user access rights to topologies)<br />

• The participants proposed to review in a year’s time all the questions that arose;<br />

• CERN is volunteer for hosting the next workshop in 2012; The intention is to hold<br />

more round table sessions in parallel next year, so that all issues can be discussed


Participants<br />

Institute Surname Name<br />

ALBA CAMPS Antoni<br />

ALBA CARLES David<br />

ALBA MANOTAS Juan Jose<br />

ALBA MATILLA Oscar<br />

ALBA MIRALLES Lluis<br />

ALBA NARDELLA Alberto<br />

ALBA SALVAR Daniel<br />

CEA/LMJ GRAILLOT Herve<br />

CEA/LMJ GRANET Isabelle<br />

CERN KNOOPS Sigrid<br />

CERN MUCNIER Christophe<br />

CERN PERINIĆ Goran<br />

CERN PILON Regis<br />

CPO MEYRONEINC Samuel<br />

DESY HATJE Jan<br />

DESY KUDLA Ignacy<br />

DESY PENNING Joerg<br />

DESY WEDDIG Henning<br />

DIAMOND WALTERS Lee<br />

ESRF FAVIER Frederic<br />

ESRF MARCHIAL Thierry<br />

ESRF ROUX-BUISSON Pascal<br />

IPNO BZYL Harold<br />

MAXLAB MANSSON Anders<br />

MAXLAB THANEL Johan<br />

SOLEIL BETINELLI Pascale<br />

SOLEIL CORTES Bruno<br />

SOLEIL DELETOILLE Xavier<br />

SOLEIL DIDIER Thierry<br />

SOLEIL GAGEY Brigitte<br />

SOLEIL HERBEAUX Christian<br />

SOLEIL MANCIET Laurent<br />

SOLEIL ROZELOT Helene

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!