Isograph Packages A. Apollonio, 7/7/2016 andrea.apollonio@cern.ch Acknowledgements: M. Blumenschein, A. Fernandez Navarro, S. Hurst, O. Rey Orozco
ISOGRAPH Overview Reliability Workbench, modules: Prediction (standards, ) Fault tree (FT) Reliability Block Diagrams (RBD) Event tree Markov Weibull FMECA Availability Workbench, modules: AvSim Weibull RCM cost Andrea Apollonio page 2
Fault Trees: Example BIS FT Andrea Apollonio page 3
Fault Trees: More Complex Example BLMs FT Manual creation of fault trees and input data insertion Copy paste option available (2 different options) Option available for importing data from external files (not straightforward) Andrea Apollonio page 4
System Linac4 Reliability Block Diagrams Blocks connected in series/parallel (voting) depending on the functionality of the system RBD Linac4 MPS Li Linac4 ELECTRONICS.1.1.2.1 Converter BIS powering Linac4 MPS.1 BIS Linac4 MPS.2 WIC Linac4 ELECTRONICS.1.1.2.2 Converter A hierarchical structure of blocks can be created: Linac4 Source Linac4 RF Linac4.BI Linac4 MAGNETS Linac4.VACUUM Linac4 ABSORBERS Linac4 COOLING Linac4 MPS Same considerations done for FTs apply here Andrea Apollonio page 5
Definition of Failure Modes Each block/fault tree event can be assigned a failure mode manually Failure density function Failure recovery function Consequences FT/RBD and a maintenance strategy Corrective Inspection Planned Andrea Apollonio page 6
Failure modes, effects and maintenance RBD Andrea Apollonio page 7
Resources Definition Resources needed for maintenance are assigned to each task Task duration Personnel required Spare components available RBD Andrea Apollonio page 8
Maintenance: Scheduled Technical Stops Requires the definition of 2 phases: Operation, planned maintenance RBD Planned maintenance in Isograph: fix all components at a predefined time Andrea Apollonio page 9
Rules for Performing Maintenance RBD Andrea Apollonio page 10
Simulations Many details can be defined in Isograph, some examples: Storage cost for spares, capacity, etc. Logistic delays and costs RBD Once all parameters related to the blocks are defined, the simulation options are selected: Lifetime Number of simulations (Monte Carlo) Results are then presented in terms of: Mean (Un)Availability (system and sub-systems) Required resources Consequences of outages Andrea Apollonio page 11
Results RBD Options available to export results to text files Customization of reports possible but not straightforward Andrea Apollonio page 12
ISOGRAPH Simulation Watch Chopper powering failure Magnet powering failure RBD RF Tuner failure Andrea Apollonio page 13
Failure Rate Predictions - Standards Prediction Calculation of failure rate of standard components from commonly used reliability handbooks Andrea Apollonio page 14
Example of CLIQ Trigger Card Prediction Andrea Apollonio page 15
Failure Rate Predictions Prediction Manual procedure, but definition of libraries possible Andrea Apollonio page 16
Weibull Analysis of Historical Data 99.9 99 90 70 Eta estimator 50 FANS Cumulative Probability Weibull 2-parameter Weibull Median rank h: 4154 b: 1.821 g: 0 30 r: 0.9762 Unreliability (%) 20 10 5 3 2 e: 0.0464 B10: 1207 P0: 0% 1 0.5 0.3 0.2 0.1 259.8 873.5 2937 9874 Time Importing data from e.g. excel files possible Andrea Apollonio page 17
FMEA Template FMEA Andrea Apollonio page 18
Costs Andrea Apollonio page 19
Summary Rather good experience with Isograph so far At CERN, especially used for: Fault tree modelling Failure rate predictions Availability modelling with RBD Some manual work required when: Creating models with a high number of components Defining maintenance strategies common for accelerator applications Retrieving useful information from the available results (strange terminology) Importing of input data (not intuitive) Good interaction with the support Andrea Apollonio page 20
THANKS A LOT FOR YOUR ATTENTION! Andrea Apollonio page 21