CERN openlab technical workshop Siemens collaboration 5th Nov 2015 Filippo Tilaro, EN/ICE CERN
Analytical framework for the CERN control system Scalable and fault-tolerant !!! Data Analysis Framework Expert Data Processing Modules Fieldbus TN PLCs Sensors & Actuators MOON (Monitoring) High Voltage DIM/CMW OPC Field layer Process layer Supervision layer FFT Neural Network (R) +600 Application Services Analysis memory and configuration Machine Learning CEP (Java) Visualisation Patterns (LabView) (WatchCAT) + 2000 PLCs + 650 FECs Data collection & feedback Historical Data +50M channels 5th Nov 2015 Siemens CERN openlab
Analytical areas of interest Online monitoring Continuous service to analyse the system status and inform operators in case of fault detection Fault diagnosis “Forensics” analysis of system faults that have already happened in the past. In some cases root-cause analysis Engineering design Analysis of historical data to draw conclusions about system behaviours which could be helpful to improve / optimize the system under analysis Oscillation analysis for cryogenics valves Anomaly detection by sensors data mining PID supervision 5th Nov 2015 Siemens CERN openlab
LHC Cryogenics system and valve oscillation Keep magnets under Superconductivity Liquid helium bathing the LHC’s magnets cooled down to 1.9K 34000 physical instrumentations and channels 12136 AI, 4856 AO,4536 DI,1568 DO 4000 analogical control loops More than 120 PLCs Siemens S7-416-2DP 30000 conceptual objects/parameters Valve oscillation can affect: Control system stability Maintenance (stress on the equipment) Performances & Safety Increase in the data load Wrong valve oscillation 5th Nov 2015 Siemens CERN openlab
Oscillation analysis for cryogenics valve High computational cost Time window length Number of channels Filters degree Running system: ~3000 channels under analysis daily Algorithm outcome validated by system experts (in terms of false positives) 5th Nov 2015 Siemens CERN openlab
PID supervision In collaboration with the University of Valladolid (Ph.D. R. Martinez) Based on: “Performance monitoring of industrial controllers based on the predictability of controller behaviour”, R. Ghraizi, E. Martinez, C. de Prada CERN control systems contains hundreds of thousands of control loops in operation PID performance has an impact on: Process security Quality of physics Maintenance (stress of the equipment) Issues: Many sources of faults / malfunctions External disturbances / factors Bad tuning Wrong controller type / structure Slow degradation System status dependency Process Controller u w y SP CV v MV 5th Nov 2015 Siemens CERN openlab
Evaluation of PID supervision Analysis algorithm: Based on performance Harris index and error prediction No a priori knowledge about the system under analysis Running system: More than 3000 control loops under analysis daily Algorithm outcome validated by system experts (in terms of false positives) High computation cost: Order of the regression model Prediction level Size of the process data history Time window size Bad Good 5th Nov 2015 Siemens CERN openlab
Integration of data analytical solutions into Siemens’ Smart Data Technologies New version of WatchCAT Extended support for R and Octave Still a prototype version Plug and play architecture R implementation of the oscillation detection algorithm Octave implementation of the algorithm for evaluation of PID supervision Evaluation of the new version: Improved performances and memory allocation Feedback to Siemens Code name “WatchCAT” Data Fusion of events & sensors Complex Event Processing Automated Learning of fault patterns Logical Reasoning for Fault Detection & Isolation Fault prediction based on recognizable patterns 5th Nov 2015 Siemens CERN openlab
Cloud-based analytical solutions & Siemens’ Smart Data Technologies Code name “ELVis” Cloud-based BIG Data Analytics for Time Series Sensor Data Real-Time Stream Processing at customizable KHz-Rates High Performance Online Visualization in Rich Web-based UI Intelligence for Sensor Data Validation Job-based Online Data Analysis Jupyter & Dockers Slave Slave Master Slave Slave NFS Scripts Data Sharing code and results Distribute computational load Interactive analysis Multi-language development environment 5th Nov 2015 Siemens CERN openlab
Status & Next Steps CERN & Siemens collaboration : New analytical algorithms designed, implemented and integrated into Siemens analytical framework Signal oscillation detection Evaluation of PID supervision Anomaly detection by sensors data mining [on going] Evaluation of Siemens diagnostic tools Cloud computing ELVis as a Storm based solution Jupyter and Docker as a cloud development environment for code sharing “Formalizing expert knowledge in order to analyse CERN control system”, ICALECPS 2015 Continue the integration of CERN specific extensions & data analysis algorithms / solutions Extensive deployment of Siemens cloud-based solutions for Big Data analytics as a Service 5th Nov 2015 Siemens CERN openlab
Thank you for your attention! Any Questions Thank you for your attention! 5th Nov 2015 Siemens CERN openlab