Upload
brent-biddulph
View
128
Download
0
Embed Size (px)
Citation preview
Big Data Analytics Utility Industry Business Drivers, Results December, 2016
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
The Pace of Change: Accelerate Innovation
Required power sector investments are massive • 2014-2035 cumulative investment of $16.4 trillion
needed globally IEA
Infrastructure improvements can’t wait • Power outages up 285% since 1984; U.S. ranks last among
the top nine Western industrialized nations in the average length of outages Bloomberg
Climate-related disasters increasingly costly • 7 out of 10 most costliest storms in US history occurred
between 2004 and 2012 President’s Council of Economic Advisers
Customers expect more…for less • Outage communications, service request updates and
high bill alerts…all at a low cost Accenture
2
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
New Sensors / Distributed
Computing on Transmission and
Distribution Lines alert operators, fix
problems, integrate large-scale
renewables generation
Smart Meters and Home Networks
help customers use energy wisely,
mitigate peak demand, integrate
local renewables
Strong Tailwind of Technology Trends Connected Customers, Connected Networks, Connected Assets
Generation
Transmission
Distribution
Consumers
BIG DATA
MOBILE CLOUD
3
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Common Business Drivers for Utility Companies
• Fraud Detection
– Predictive modeling of potential fraud in billing, meter tampering, power diversion
• Assets: Smart Meter and Grid Operations
– Preventive maintenance, determination of defective meters and equipment
• Energy Efficiency / Demand Response
– Targeted campaigns to reduce peak loads
– Grid utilization planning & energy trading
• Supply Chain Management Logistics
– Route optimization and optimized maintenance
– Optimally located supplies, distribution points
• Weather / Event Planning & Crew Deployment
– Optimal staging of manpower and equipment in response to changing weather patterns
– Safety analysis and optimization
• Customer Experience
– Call Center – anticipate reason for inbound calls, provide CS reps with heads-up display
– Customer Profiling – personalize campaigns, bill formats, preferred notification channels
– Intelligent Call Routing – matching customers with ‘best fit’ CS representatives
– Recommendations – target value added services, new programs, products, services
4
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Preventing Theft of Service (Fraud) Objective: Detect fraudulent power usage in current residential accounts, disintermediate theft of service activity, recover revenue and significantly reduce ‘false positives’
• Oracle Big Data Cloud Services, including: Cloudera Hadoop Dist., Connectors, Oracle R Advanced Analytics for Hadoop, Spatial & Graph, Storage and Data Integrator
• Field investigative teams dispatched based upon basic system alerts generated by 2 way power meters, which often (80% of the time) resulted in ‘false positives’
• Advanced Analytics teams struggle to access enterprise data, grain (detail) of data required is not available in DW and they spend 60-70% of time gathering, preparing data for analysis
• Executive leadership had investigated point solutions, but wanted to own the competency and leverage internal IT to establish a foundational capability (Data Lake / Data Lab) for broader enterprise analytic capabilities
Business Challenge / Opportunity
Proposed Solution
Customer Example
Regional Utility focused on Operational Efficiency and
Revenue Recovery from Theft of Services provided
Potential Benefits
• Improving fraudulent power usage detection in residential accounts by 80x viewed as very conservative Yr. 1 estimate
• Over a 3 year horizon, potential incremental revenue recovery from improved fraud detection and reduced ‘false positives’ ranged from $2.9m to $4.9m in benefits, and a 1,056% ROI
Analytic Sprint Results
• Developed behavior detector and false positive rules engine prototype for detecting potential instances of fraud. Achieved high level of confidence in back testing v. known fraud cases.
• Demonstrated the ability to easily add new data, business rules and scale analytic model beyond current sampling process
5
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Preventing Infrastructure Damage from Pipe Leakages Objective: Develop a flexible AI system to guide infrastructure operations and strategic planning that provides mathematically concrete inferences and predictions
• Oracle Big Data Cloud Services, including: Cloudera Hadoop Dist., Connectors, Oracle R Advanced Analytics for Hadoop, Spatial & Graph, Storage and Data Integrator
• Every year billions dollars of economic damage, environmental damage, and opportunity costs are caused by preventable pipe leaks
• Pipe Leaks cause pipes to burst, water main breaks, and destabilize surrounding pipes, in addition to being a threat to public health
• Firm's existing system was too theoretical, non-data driven, relied upon human expert judgment, and most importantly - couldn't accurately predict whether a pipe would leak
Business Challenge / Opportunity
Proposed Solution Customer Example
One of the largest public water utilities in the US focused on reducing the
economic and environmental impact of pipe leakages
Potential Benefits • Billions of dollars in savings in infrastructure costs, reduction in
false positives/false negatives, technological enablement, and the development of new manufacturer purchasing standards
Analytic Sprint Results
• Developed 100% data-driven AI system that can predict whether a pipe would leak with 99.9% accuracy, significantly reducing the cost of both false positives and false negatives
• Demonstrated ability of AI system to provide mathematically concrete casual inferences and predictions, confirmed expert knowledge, and taught experts things that they did not know
• Enabled the firm to perform infrastructure stress testing and scenario modeling using the AI system's probabilistic outputs as inputs and move from a reactive stance to a proactive one
6
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Optimizing Data Management, New Analytic Capabilities Objective: Reduce data management costs by optimizing workloads across analytic platforms, while simultaneously ‘democratizing’ new data sources and introducing new analytic capabilities
• Oracle Big Data Appliance
• ETL & Data Migration Services (partner)
This utility client is working to reduce their cost of data management and introduce new analytical capabilities to their enterprise.
• Reduce significant $ spent on ETL processing on the current legacy data warehouse (DW)
• Integrate new data sources such as Smart Meters into new product offerings and pricing on the grid
Business Challenge / Opportunity
Proposed Solution Customer Example
Potential Benefits
• By expanding their existing Oracle BDA footprint to support the legacy data warehouse ETL workloads and introducing capacity for new data, they will reduce the long term capital and opex costs of their existing Data Warehouse (DW)
Big Data Optimization Results
Client was looking for guidance on sizing and capacity questions related to legacy DW and emerging multi-structured data needs. Within 3 weeks, Oracle’s Big Data Optimization team provided: • Planning and Discovery • DBQL & PDCR Interrogation • Documented Data Flow • Key findings and Recommendations
The largest provider of electricity in deregulated
Western US market, serving 14 million customers. Also operates a regulated gas and
water utility.
7
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Enterprise Data Lab, Enabling New Analytic Capabilities Objective: Enable new analytic capabilities and business insights to improve operating efficiencies, asset management, grid reliability and overall customer experience
• Enterprise Data Lab, consisting of: Oracle Big Data Discovery and Oracle Big Data Appliance
Client enterprise steering committee established and drove development of a 3 year strategic roadmap to support Utility of the Future growth initiative. • Define enterprise analytic strategy and required technology
ecosystem to support the long-term strategic objectives • Design reference architecture and phased implementation
approach to integrate all domain subject areas
Business Challenge / Opportunity
Proposed Solution Customer Example
Potential Benefits
• By establishing a Data Lake and Data Lab, client will be able to finally collect vast amounts of meter data and customer data, enabling new analytic capabilities and insights to improve operating efficiencies, reliability and revenue recovery
Big Data Business Use Cases
• Predictive maintenance and remote monitoring health of meters, transformers, comms.
• CVR - Voltage optimization, leveling the voltage profile
• Real time asset health index scores (system and SCADA data)
• Water leakage detection and usage monitoring
• Customer Trading – energy usage forecasting and arbitrage
The largest electric holding company in the US by revenue, the largest
regulated utility in the US with approximately 10
million Eastern customers.
8
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | 9
Questions, Comments?
Reach Out, Let’s Talk!
https://www.linkedin.com/in/brentbiddulph
@BrentBiddulph
Reach Out, Let’s Talk!
https://www.linkedin.com/in/brentbiddulph
@BrentBiddulph