Pursuing the digital railroad

© 2015 IBM Corporation

Pursuing the Digital Railroad

Ken Donnelly

IBMGlobal Systems Integrator Sales Leader

March 2015


Agenda


– Introduction and Overview

– Creating interconnected, instrumented and intelligent systems

– Improving performance by reducing unscheduled outages, predicting service demand and

increasing efficiency with predictive technology.

– Money talks – how much could your railway be saving by effectively leveraging sensor technology?


DRIVERS OF CHANGE CHALLENGES STRATEGIC IMPERATIVES

Economic Growth Global economy is expanding and citizens are getting wealthier. Rail providers will need to expand capacity to keep up with freight and passenger growth

Passenger GrowthAs the number and size of cities grows, pressure on rail system to move people between and within those cities grows.

GlobalizationThe growing interconnectedness of the world is driving growth in demand, with an expectation of improved service

Technology ImprovementsTechnology now enables the capture and analysis of real-time information about the status, location and condition of rail operations

Capacity and congestionMeet the growing, changing demand efficiently, consistently and profitably?

Empowered customersDeliver choices for ticket purchase, changes in travel plans, and presenting information in the way that passengers value

Efficient, green operationsReduce cost and dependency on scarce resources while reducing environmental impact.

Safety and securityIncrease the safety of operations, with less impact on customers and reduce exposure to security risks

Predict demand and optimize capacity and assets.

Dramatically improve the end-to-end traveler experience.

Improve rail operational efficiency while reducing environmental impact.

Assure safety and security of rail

Global Rail Business Drivers and Trends

© 2015 IBM Corporation4

Something to Think About: Is this how we approach asset management and optimization today?

Assets need to enable the services that maximize the profits of an enterprise.

The enterprise should manage asset maintenance and deployment to maximize profits, not to minimize asset

costs.

IBM Asset Optimization moves asset maintenance and deployment from a low cost paradigm to an

investment paradigm.

The investment paradigm means that maintenance and deployment costs are incurred at the

level necessary for the enterprise to operate at its profit-maximizing level.

4


ARC Report – The Internet of Things (IoT)


Agenda








Intelligent Sensors Growth Rate

7

1 - Sources: a. Berg, iSuppli, and ABI M2M forecast average for 2005-2011 extrapolated at 2010-2011 growth rate.b. FierceWireless “AT&T moves closer to embedded wireless vision” May 8, 2009


The project used data mining, machine learning and predictive modeling to predict impending failure/alarm of critical rail car components. The prediction drives proactive inspection and repairs, reducing operational equipment failure

Machine Vision Detector

Optical Geometry Detector

Truck Performance Detector

Wheel Impact Load Detector

Acoustic Bang Detector

Hot Box DetectorsWarm Bearing Detectors

Hot Wheel / Cold Wheel Detectors


Intelligent

A more intelligent transportation infrastructure will put all this new data to work. – Mobile monitoring systems will provide railroads with more intelligence through continuous real-time data

capture and analysis, such as the health of rolling stock and operational data from manifest verifications

to freight condition and intrusion detection.

– Sensors on cars will trigger messages based on decision modeling and analytics.

– Autonomic routines will then dispatch services, order parts, schedule maintenance and perform remote

diagnostics.

– Eventually, such technologies could reduce the need for fixed infrastructure along the wayside and give

railroads the flexibility and responsiveness they need to optimize crew schedules, and integrate

passenger and freight transport more seamlessly, with far fewer delays.

Locomotive

Health ScoreLocomotive

Life Span

Derived

Measurements

Primary

Predictors

Predicted

Monitored

Equipment

(Diesel Loco)- Alternator

- Engine

Alternator

Amperage

Alternator

Temp

Engine

TemperatureLUBE

Quality

Engine

RPM

Alternator

Amperage

Engine

Temperature

Predictive Model

Used

Linear Regression

Model

Cox Regression

Model

others..


Agenda








Total Cost of Ownership

End of Life Replacement CostsWarranty Tracking InabilityManufacturer TCO DifferencesUnknown factors that reduce TCOUncertain Inventory RequirementsNo Closed Loop Measures/Metrics

Maintenance Inefficiencies

High Backlog ConfusionCrew, Tool, and Resource UtilizationInadequate Response TimeInadequate level of MaintenanceEmergency vs. Routine RedundancyComplex Work Planning

Customer Behavior Uncertainty

Asset Utilization FluctuationsCustomer Satisfaction ImpactsLoss of RevenueService Level Agreements not met

Reliability Consequences

Inability to Predict Failure Effect of Usage and Environment Restoration times are uncertainRoot Cause Failures not understoodRisk Factors not Clear

Enterprise Assets

Typical Pain points


What if you could accurately predict which characteristics tend to lead to an increased frequency of failures?

What if, when an asset is scheduled for maintenance, you could predict what parts are likely to fail in the near future?

What if you could identify the characteristics that tend to increase ownership cost and downtime over the life of a system?

What if you could replace those parts that have not yet failed and avoid further unscheduled downtime?

What if you could quickly mine the thousands of logs that describe the maintenance performed on systems and determine what important observations are being logged by the maintenance team?

What if you could unearth patterns in maintenance operations over time that could point to opportunities for improvements?

Are you facing such challenges? And/or other ones?


Predictive Analytics is essential to answering these question ….

Captured

Detected

Inferred

Use Structured Data & Unstructured Data

Descriptive Analytics

Prescriptive Analytics

Predictive Analytics

Simplified to be consumable and accessible to everyone, optimized for their specific purpose, at the point of impact, to deliver better decisions and actions through:

What trends will continue? Forecasting

How can we achieve the best outcome and address variability? Stochastic Optimization

What happened?

What exactly is the problem?

How many, how often, where?

What actions are needed?

What could happen if? Simulation

How can we achieve the best outcome? Optimization

What will happen next if? Predictive Modelling

Analytics Sophistication

• From multiple (valuable) sources

Sensors (temp, PSI,..)

Maintenance Records

R&M, OMS, Warranty,..

Operating Conditions

MultipleData Sources










Agenda








“Gentlemen, we have run out of money. Now we must think!”

- Sir Winston Churchill

A First Thought ……


An Example

Increase reliability 30% improvement by 2015

Cost savings of £2.2B

Annual passenger growth since start of the 2008 recession 3.8%

700% safety improvement

40% asset reliability performance improvements

£400 million saving due to asset life extension

£4.6 million saving due to mobilization of inspection processes

Improvement in reliability (MDBF) 478% across specific asset classes

This is more than just technology and process, it is aboutpeople and organizational culture. For the business transformation to deliver the results of world class this has to be

led from the very top of an enterprise.


Organizational Challenges

Probably the most important and challenging aspect of an a program is addressing the needs of people within

the organization.

Because asset management is holistic, it depends upon comprehensive coordination and communication. Most

organizations are functionally segregated.

The challenge is to help the people in the organization understand and appreciate the benefits of the process

from the perspective of the entire organization rather than the viewpoint of their individual units.

Another challenge is building organization-wide commitment to change. Creating buy-in at both the executive

and operations levels of the organization is critical to success.

Ad hoc

Foundational

Competitive

Differentiating

Break away

•Spreadsheets and extracts

•Data warehouses and reporting

•Contextual business rules and pattern recognition•Content analytics of unstructured data

•Master data•Managing structured Data•Metrics, dashboards, scorecards

•Predictive, real time analytics

Big Data Value Proposition


The Digital Railroad


Summary

The World is changing rapidly ….. Intelligent Devices and the ability to effectively harvest the right data

and convert it to meaningful Information is now more crucial than ever.

Technology is easy …. Organizational changes are CHALLENGING– Information Silo’s must be eliminated.

Paper Records are archaic … Do you have a strategy to capture data real time?

Seconds count to delivering a Safe, Reliable, and Profitable Service.

Pursuing the digital railroad

Engineering

ibm corporationagendapursuing

ibm corporationpursuing

ibm corporationarc

ibm corporation4something

asset maintenance

rail system

rail providers

service demand