Big data and Process Safety
- 1. Proactive Risk Leadership – using data analytics to prevent serious process safety incidents Coen van Driel Robert Kauer TÜV SÜD and Kienbaum Management Consulting October 6th, 2016 Process Safety Transformation Management Technical Process Safety Expertise& 2016 European Conference on Process Safety and Big Data
- 2. 2 Big Data is “in” and everyone wants to get into it but most don’t understand it ... Big Data, Big Expectations, and a lot of opinions
- 3. Big data the holy grail for the industry…. …or is it of great value for risk management/ proactive risk leadership?
- 4. 4 Our reference when we are talking about big data in the context of risk management Understanding the business Data Analytics The use of algorithms and statistical modeling to draw meaning and insights from data Supporting the decision making circle Big data Data generated from various sources in different formats in an ever increasing speed
- 5. 5 Big Data defined by the 5 V’s Big Data Definition » Volume » The quantity of generated and stored data. The size of the data determines the value and potential insight- and whether it can actually be considered big data or not. » Variety » The type and nature of the data. This helps people who analyze it to effectively use the resulting insight. » Velocity » In this context, the speed at which the data is generated and processed to meet the demands and challenges that lie in the path of growth and development. » Variability » Inconsistency of the data set can hamper processes to handle and manage it. » Veracity » The quality of captured data can vary greatly, affecting accurate analysis.
- 6. 6 The transformation process to Data analytics Excellence takes an multidiciplinary approach Five levels of data analytics Level 1: Basic analytics » Standard RCA and application of PDCA » Firefighting to deal with problems Level 2: Descriptive analytics » Applying structural lean six sigma on single source data » Descriptive statistics and regression analyses to understand past trends Level 3: Predictive analytics Level 4: Prescriptive analytics » Start modeling future outcomes with past data » R-analyses on multiple sources » Converting trends into future scenarios to make decisions » Complex algorithms using multiple source of data to convert into action Level 5: Data Analytics excellence Action to take Looking Ahead Looking Back Average position of most companies in chemical sector. Compliant TransformationJourney
- 7. 7 Using the data analytics it will be possible to grow from static audit type evaluation to continuous information and real time decision making Technological developments Audit type results Real time Process Safety decision making Process Safety Performance Measurement
- 8. 8 To transform to real time decision making using data analytics various sources need to be combined based on a risk management approach Integrated risk process safety management system
- 9. Linking risk based thinking with process safety performance based on big data preventing serious process safety incidents from happening Proactive risk management Hazard analysis Risk based thinking Process Safety performance Select High Impact scenarios Review with the work teams Integrate in work process Visual board shop floor Leadership rounds Performance monitoring Performance Site/Europe • Risk identification through efficient and tailored hazard analysis (PHA/PHR process) • Aggregation of the performance to develop policy and strategy • Highest Consequence Scenarios • Extract key scenarios for each section • Use the KPI’s to set priorities and show the developments • Prevention activities & reduce consequences • Review in control room and other operational area’s • Develop risk awareness with personal • Verification of risk awareness of personal in the organization • Verify with management system elements and element owner • Integrate in Shifttour maintenance / production, Inspections, Alarms • Visualize near misses based on the shift tours, process deviations, etc. • Discuss findings in shifts and meeting structure Risks need to be clear to define leading KPIs on near miss level of the API pyramid (level 3) and controlling the barriers (work processes) 9 Big data DADA DADAData – Analyze – Decision – Action
- 10. The decision making process based on the risk based management thinking enable proactive risk leadership 10 Big data based risk based performance management Data Analysis Decision Action Business Understanding Deployment Evaluation Modeling Data Understanding Data Preparation Data
- 11. 11 Example From standard inspection to multi source risk based decision making Level 1: Basic analytics Level 2: Descriptive analytics Level 3: Predictive analytics Level 4: Prescriptive analytics Level 5: Big data excellence OK 5y OK 5y » isolated approach » subjective, relying on expertice » not transparent CML Limit State » single source » frozen until next inspection results » isolated approach CML on-stream monitoring » multi source » multi-disciplinary » more flexible decision making screening in combination with hot spots » risk-oriented and focused » interlinked multi-sources » „real- time decision making and links to other initiatives » learning about relations
- 12. Big data is not the holy grail for the industry…. But the opportunity to enable proactive risk leadership at all level using prescriptive analytics