Tradition meets modernity Predictive Analytics with Artificial Intelligence Asia Pacific Rail 2017 Sven Nowak Slide 1
Overview 1 TÜV SÜD Group 2 TÜV SÜD Rail 3 Predictive Analytics 4 Contact Details Slide 2
Choose certainty. Add value. The world s number one brand of choice for premium quality, safety and sustainability solutions that add tangible value to our clients. Slide 3
TÜV SÜD in numbers: Growing from strength to strength 1 One-stop technical solution provider 150 years of experience 1000 locations worldwide 2,220 24,000 million Euro in sales revenue for 2015 employees worldwide as of February 2016* Note: Figures have been rounded off. *As of 29.02.2016: Inclusive of acquisition in January 2016. Slide 4
Global expertise. Local experience. Global Headquarters Munich, Germany Legend: Countries with TÜV SÜD offices Regional headquarters Note: Figures have been rounded off, as of 31.12.2015. GERMANY Euro 1,283 mio 11,600 staff INTERNATIONAL Euro 939 mio 10,800 staff Slide 5
Technical expertise & broad industry knowledge Testing & product certification Chemical, physical, mechanical, electrical and environmental testing and product certification. Inspection Product, system, building, plant and infrastructure inspection. Auditing & system certification Audits system certification in a variety of fields including quality, safety, energy, IT security, social compliance and environment. Knowledge services Safety, quality, risk, environmental protection and regulatory advisory. Training Training in work safety, technical skills, management systems and executive programs. Slide 6
Increase Decrease Tangible economic benefits for your business Market access Costs & inefficiencies Productivity & profits Time to market Brand reputation Business risk Slide 7
Adding value through quality, safety and sustainability solutions across the 3 Strategic Business Segments MOBILITY INDUSTRY CERTIFICATION Ensure 100 years in safe mobility: Periodical Technical Inspections Car business services Fleet management Automotive Maximise reliability, safety & efficiency for: Chemical, oil & gas Power & energy Manufacturing & industrial machinery Rail Real estate and infrastructure Achieve market access for: Manufacturing & industrial machinery Consumer products & retails Healthcare & medical devices Telecommunications & IT Transportation (Automotive, Aerospace and Marine Component) Slide 8
Overview 1 TÜV SÜD Group 2 TÜV SÜD Rail 3 Predictive Analytics 4 Contact Details Slide 9
TÜV SÜD Rail: Holistic solutions across the value chain From planning through to engineering, installation, testing, commissioning, operation and decommissioning. TÜV SÜD Rail ensures increased safety, efficiency and added economic value for manufacturers, operators and authorities. TÜV TÜV SÜD SÜD 04/04/2017 Rail Presentation Tradition meets modernity Slide 10 Slide 10
Rail Services Rail services Engineering Services (SIGNON) Consulting Services (SIGNON & TÜV SÜD) Testing Services ISA & Certification Services Training Services Rolling stock High speed conventional rail light rail & metro Infrastructure Earthworks track works civil works & tunnel Signalling Train protection systems Signalling Telecom & SCADA Energy Power supply & Catenary Embedded systems Programmable electronics safety components rail automation (HW, SW) industrial IT security Slide 11
TÜV SÜD Rail: Holistic solutions across the value chain Overall Services Rolling Stock Metro/Urban Rail Systems Infrastructure Complete Systems Signalling Systems Rail Energy Embedded Systems Slide 12
Overview 1 TÜV SÜD Group 2 TÜV SÜD Rail 3 Predictive Analytics 4 Contact Details Slide 13
TÜV SÜD Digital Service Centre of Excellence in Singapore The proliferation of new technology can only happen if it can be used safely & securely Provide innovative solutions around Safety, Security, Reliability, Efficiency and Interoperability for complex systems Smooth integration of new IoT technology Enabler for Digital Transformation Slide 14
Strategic Areas Autonomous Vehicles: Safety and Cyber Security for AVs. Testing and Certification framework. Smart Sensor Lab Self calibrating / healing and realtime validation of sensor networks IoT Platform Safe and secure data handler, data analytics Smart Lifts Make smart lifts safe, secure, reliable and efficient. Smart Buildings Requirements definitions, advisory for technology integration, safe and secure operation. CoE Digital Service Industrie 4.0 Make I4.0 and robotics applications robust, interoperable and available. Assess Industrie 4.0 Maturity Smart Healthcare Make smart healthcare installations reliable, safe and secure. Digital Rail Predictive Analytics, automated train operation, cyber security Slide 15
Digital expertise PREDICTIVE ANALYTICS ARTIFICIAL INTELLIGENCE IN REAL-TIME Advanced Digital Sciences Center Slide 16
Capacity and efficiency Greatest challenges are in capacity and efficiency Urbanisation Utilisation Construction of of traditional infrastructure alone alone will not be will able not to be keep able up to with keep the up rapid with rate the of rapid urbanisation rate of urbanisation Tipping point at which additional supply will no longer provide an efficient means to service demand Tipping point at which additional supply will no longer provide an efficient means to service demand Intelligent transportation infrastructure Technology will play a major role by adding intelligence to the infrastructure and enable it to increase its capacity at a faster rate Provision of physical infrastructure versus underutilised capacity of existing infrastructure Slide 17
SMART Mobility Solutions Integration Automation Sophistication Multi-modal transport Connectivity Flexibility to use best-fitting transport mode Safer transport systems Operational performance Environmental performance Better utilisation of transport assets Monitoring health of physical assets Response in real-time to manage capacity and avoid/predict disruption Slide 18
Major challenges of future mobility solutions Security Increasing number of access points for attacks Attacks may propagate across connected systems Complexity Data management Cognitive technology with dynamic system properties Testability becomes more limited & system behaviour in event of failures less controllable Interoperability Reliability System complexity makes assessment of reliability more complicated Automated systems with limits handling abnormal operation Slide 19
Current Best Practices for Advanced Analytics with limits Installation of external sensors using predictive analytical methods to monitor the health of physical assets Interpretation of data Standalone solutions impede holistic realtime analysis Interpretation of data is often deterministic based on fixed rule sets and thresholds Installation of components into an operating safety-critical infrastructure may raise massive safety concerns of operators and authorities System reliability may be significantly reduced by false positive results Mismatch between right processing (data analytics) and interpretation (subject matter expertise) to generate useful information for decision making process Slide 20
Tapping available data sources Existing technologies already generate and collect data that simply needs to be reinterpreted Accessing this data does often not even require modifications of the certified system Application of cognitive technologies will avoid false positives Algorithms with the ability to learn will make systems safer and more reliable over time Identification of critical anomalies in parts of the system, which may not be even specifically monitored J.W. von Goethe: You only see what you know Slide 21
Advanced Analytics Data Decision Action Business Intelligence Data analytics Data analytics is the process of examining data sets to draw conclusions about the information they contain Predictive analytics uses both new and historical data to forecast system behavior Machine learning provides computers with the ability to learn without being explicitly programmed (detect patterns in data to adjust program action apply to new data or draw inferences) Artificial intelligence is the simulation of human intelligence processes by machines including Learning (acquisition of information and rules for using the information), Reasoning (using the rules to reach conclusions), and Self-correction Analysis Descriptive What happened? Diagnostic Why it happened? Predictive What will happen? Suggestive What you should do! Prescriptive What I will do! Human Input Artificial Intelligence Human Intelligence Slide 22
Advanced Analytics Platform continuously and in real-time Data acquisition Analysis Representation Optimisation Heterogeneous data sources Recognise correlations Predict events Anticipates even previously unknown anomalies Requests human feedback regarding anomalies in order to gain more knowledge Adaptive datadriven optimisation algorithms Slide 23
Machine Learning Slide 24
Artificial Intelligence Slide 25
Predictive Analytics Platform - User Interface Slide 26
Use Case Failure Indication of Train Doors Slide 27
Use Case Failure Indication of Switches Slide 28
Use Case Optimisation of Passenger Flows Slide 29
Use Case Prevention of human errors Significant number of accidents can be attributed to human errors Accidents show the magnitude of misconducts (both intentional and unintentional) Human interaction with technology is usually clearly specified and deterministic Advanced analytics algorithms recognise patterns and evaluates the human behaviour Human errors are mostly more subtle than operating errors Latent failures may lie dormant for years Physical human-machine interfaces pose security threats which may remain undetected for a long time Source: https://en.wikipedia.org/ Slide 30
Outlook Construction of traditional infrastructure alone will not keep up with the rapid rate of urbanization Intelligence technology will enable infrastructure to increase its capacity at a faster rate Increasing complexity, integration and growing dependency complicates control of operations Common approaches to monitor the health of physical assets has limits Existing technologies already generate and collect data that simply needs to be reinterpreted Predictive Analytics in real time using intelligent algorithms from the field of Machine Learning and Artificial Intelligence enables the extraction of hidden knowledge in vast data sets Learn any complexity of a system's normal behaviour independently and continuously Anticipates even previously unknown anomalies Adaptive data-driven optimization algorithms. TÜV SÜD with its partners combine the traditional knowledge of railway systems with elements of advanced data analytics using artificial intelligence to meet challenges of the digitalisation TÜV SÜD is looking for partners to co-innovate methods to enhance safety and reliability using new digital solutions Slide 31
Overview 1 TÜV SÜD Group 2 TÜV SÜD Rail 3 Predictive Analytics 4 Contact Details Slide 32
TÜV SÜD Why choose us? Gain world-class expertise Get independent and impartial services Receive value added services Save money and time Minimise risk The TÜV SÜD experts are recognised by and work in close collaboration with numerous national and international authorities including EBA. TÜV SÜD Rail is recognised worldwide as an independent and neutral technical service provider. We offer a complete range of customer-focused services on safety & quality throughout the entire lifecycle of projects. Our holistic approach covers the whole railway system, including gap analysis and maintenance procedures. Our experts support to increase safety, save time and add tangible economic value to your railway operations. As one-stop solution provider we ensure safety of the entire rail system. We are working together with a global network of international and local experts in all key markets worldwide who provide global knowledge of processes and procedures in the rail sector. TÜV TÜV SÜD SÜD 04/04/2017 Rail Presentation Tradition meets modernity Slide 33 Slide 33
How can we help you? Predictive Analytics with Artificial Intelligence Asia Pacific Rail 2017 Sven Nowak Vice President Digital Service Principal Engineer Railway Signalling FIRSE, MIET sven.nowak@tuv-sud.sg www.tuv-sud.com/rail Slide 34