Rob Miller General Manager, Global Solutions

Transcription

1 Machine Learning and Cognitive Analytics to improve pump performance Rob Miller General Manager, Global Solutions

2 Flowserve Background Leading manufacturer and aftermarket service provider of comprehensive flow control systems History dates back to 1790 with more than 50 well-respected brands such as Worthington, IDP, Valtek, Limitorque, Durco and Edward Design, develop, manufacture and repair precision-engineered flow control equipment for customers critical processes Portfolio includes pumps, valves, seals and support systems, automation and aftermarket services supporting global infrastructure Focused on oil & gas, power, chemical, water and general industries Worldwide presence with approximately 17,000 employees 71 manufacturing facilities and ~200 aftermarket Quick Response Centers (QRCs) with Flowserve employees in more than 50 countries SIHI Group acquisition in January 2015 includes 28 facilities (15 service center) Long-term relationships with leading energy customers National and international oil & gas, chemical and power companies, engineering & construction firms, and global distributors Established commitment to safety, customer service, and quality with a strong ethical, compliance and performance culture VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 2

3 Business Impact Limitations The Business Problem and Current Limitations Threshold-based monitoring systems could only identify failures a few hours prior to them occurring Custom engineered algorithms for predictive capability for each pump type and application is a lengthy process. Prevention of unplanned failures and related downtime Insufficient time to respond effectively Lengthy process to build and then maintain custom engineered algorithms Inadequate for detecting unknown states with various process conditions VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 3

4 Valuable Insights - Automated Pump Failure Prediction Identification of pump state Is there a problem? If so, what kind of problem? Is there a problem we ve never seen before? (signature DB approaches don t work well here) Is the pump operating in the right state? Fleet State Is the entire system operating well? Is the entire fleet optimized? Prediction When will the problem or catastrophic failure occur? When will it require maintenance? Forensics What factors were most responsible for a failure or sub-optimal operation? Is this due to natural or malicious reason (cyber attack)? VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 4

5 Derived Feature Multi-dimensional analysis Multi-dimensional analysis feature provides early and accurate warning with minimal false positives Optimized threshold to reduce false positive alerts Early Warning ~5 days before failure Dynamic and adaptive threshold that will continue to learn and adjust with more data Point of Failure Date VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 5

6 Proven Use Case in Pump Monitoring Objectives 1. Recognize Operating States 2. Detect Anomalies 3. Predict Pump Failure Data Analyzed Pre-filtered vibration data (feature data) 3 Years worth of run data on production asset Results Identified operating modes with >99% accuracy Accounted for four criteria defined by client to handle imperfect data and operating conditions Predicted failures 5 to 6 days in advance (20x improvement) Previous method predicted only 12 hours in advance Completed with less than 2% false positive rates VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 6

7 With New Machine Learning Technologies The Future Is Now Traditional maintenance & reliability programs have done a great job advancing reliability, but at many sites our long time customers are asking what s next? Leveraging new machine learning technologies provides opportunity for step-change in approach for equipment performance and reliability MTBF MTBF Reactive Minimal and disparate system Victim mentality MTBF Some planning and scheduling Formal and informal systems Dabbles in Predictive Maintenance (PdM) Lack of documented asset management strategies MTBF Documented Strategic Body of Knowledge (BoK) based approach Leverages Predictive Maintenance Disciplined behavior 80 + MTBF Enterprise approach Exploits technology Flexible market capture Full and robust life cycle management Culture driven VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 7

8 Why Machine Learning? VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 8

9 Cognitive Analytics Cognitive Analytics refers to a set of innovations that are inspired by the way the human brain thinks: Processes information Draws Conclusions Benefits Codifies instincts & Experience into learning Enables machines to penetrate the complexity of data to identify associations Presents powerful techniques to handle unstructured data Continuously learns from new data entering the system and from previous insights Provides NLP support to enable human to machine and machine to machine communication Does not require rules, instead relies on hypothesis generation using multiple data set VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 9

10 What Machine Learning & Cognitive Analytics can deliver several benefits The integration of machine learning provides many different benefits Scalability Automated model building capability does not require manual model building of every asset/component Adaptability Adapts to new and changing conditions automatically Higher Accuracy Automated feature enrichment and extraction that can deliver better insights and higher accuracy External Factors Can incorporate external factors (e.g. environmental issues) In-context Remediation Advisor that understands natural language to help technical teams Security Out-of-band, symptomsensitive approach beyond IT security VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 10

11 SparkCognition reduces the cognitive burden for security and prognostics SMEs Sensors DAQ System Asset Monitoring Actionable Insights Fleet Wide Optimization Cyber Security VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 11

12 Additional Questions or Information: Rob Miller contact info: Office# Thank You! VISION, EXPERIENCE, ANSWERS FOR INDUSTRY ARC Advisory Group 12