Analytics Grow in Support of Strategic Asset Management

Knowledge of an asset and how it performs under a given circumstance is tribal, passed from engineer to engineer, as well as systematic, established through historical measurements that involve formal benchmarking.

Where applied human knowledge falls short or retires, computing-based systems are increasingly able to bring clarity and consistency to asset management objectives.

IDC’s Vertical Insights Survey 2015 found that almost two-thirds (62%) of utilities have implemented or are considering implementing big data and analytics capabilities. Because of the primary importance of operational assets, the focus is for utilities to use analytics and data tools to improve capital investment, operations and maintenance of assets.

Utilities now have access to more information than ever before and have come to recognize the need to make the best use of available data by applying analytics. In fact, investment in analytics is increasing. Recent research shows that the deployment of big data and analytics in production in business units and enterprise-wide has increased from 10.2% in 2013 to 30.3% in 2015.

Fig 2

Stages in Deployment of Big Data and Analytics in Utilities, 2013–2015

Q. At what stage is your organization today in the deployment of big data and analytics?
(% of respondents)

In production, enterprise-wide

14 %

8 %

2 %

In production, business units

17 %

3 %

9 %

Pilot/proof of concept

19 %

8 %

62 %

Researching

16 %

46 %

21 %

Considered, not yet pursuing

19 %

14 %

24 %

Not considering

16 %

22 %

38 %

2015 (n = 89)

2014 (n = 84)

2013 (n = 125)

Utilities have moved from an early focus on smart meter data to improved customer service and marketing as well as using multiple and varied data resources for planning, operations and maintenance. According to IDC’s 2013 Big Data and Analytics for Utilities Survey 72.9% of utilities with big data and analytics initiatives noted that the drivers for their investment in analytics in 2014 and 2015 were related to operations and maintenance.

Today, utilities are looking to analytics for answers to some very practical questions:

Objective	Questions
Reliability/Availability	How do we minimize unplanned outages on T&D networks, at generation plants? When should we repair an asset that is potentially failing?
Protecting health, safety, and the environment	Can we anticipate an incident about to happen and prevent it? How can we detect potential pipeline failures in time to prevent accidents? When should we repair an asset that is potentially failing?
Capital effectiveness	How can we safely defer capital investment in equipment (poles, pipelines, transformers, boilers, cables)?
Operational efficiency	How can we reduce the costs of maintenance and operations? How do we improve operational processes and thus the energy efficiency of our assets?
Market opportunity	How can we insure availability of generation when intermittent resources create more market volatility?

The types of analytics in play include:

Business intelligence: Used to display trends and drill down into data. Remote monitoring centers staffed by performance engineers are using drill-down techniques, visualization, and collaboration with onsite personnel.
Predictive analytics: Used to identify potential failure or indications of asset health. With predictive analytics alone, utilities have been able to identify and address single instances that have resulted in substantial avoided costs. Geo-spatial visualization that addresses operations and maintenance is especially helpful for distributed transmission and distribution assets. Predictive analytics can be:
- Mathematically based detection of anomalies in normal operating conditions through the analysis of condition data (e.g temperature, vibration, power usage).
- Engineering models based on degradation models or simulations under various conditions used to improve the degree of certainty.

Optimization analytics: Recommends what to do with limited resources. Advances in processing allow for utilities to prioritize based on their objectives and quickly access optimization results.
Cognitive analytics: Used to recommend actions based on learning best practice approaches from similar situations. The extra power of cognitive analytics helps in dynamically changing complex situations where computing can quickly identify repeatable and effective solutions.

A large North American power utility was able to save an estimated $4 million in avoided unplanned downtime and repair costs in one “catch.” Data science–based predictive analytics showed a pattern in vibration at a steam generation plant that would not have been identified with standard monitoring. A root cause analysis and inspection showed that the cause was loss of shroud material. The analytics applied did not include prescriptive (identifying the repair alternatives) or optimization (weighing alternative action paths), which likely would have yielded even more efficiencies.