Contact Us
We are taking you to another website now.
April 30, 2018

The future of oil and gas (O&G) in the energy economy is uncertain. Renewable sources, according to a recent World Economic Forum report, is set to disrupt the market and change the conventional energy demand and supply pattern. In response, some companies like General Electric (GE) is heavily investing in expanding their O&G footprint through acquisitions. But, as wind and solar sources threaten to dominate the industry over the next decade, can predictive and prescriptive asset management breathe new life into an industry which is bogged down by soaring upkeep costs?

O&G production and transmission – from wells deep inside the Earth’s crust to the end-consumer – is governed by a long and complex supply chain comprising many moving components. This supply chain is incomplete without industrial assets such as rigs for exploration, pipelines for midstream, and refineries for downstream processes. Together, they form a majority of the capital and operating expenses borne by an enterprise.

Owing to the dearth of skilled engineers, these assets are not always maintained as well as they need to be. Drilling equipment and subsea pipes are often subjected to extreme environments, which also make regular manual maintenance both difficult and dangerous.

Drilling through maintenance data

Herein, predictive analytics based on statistical techniques can play a vital role in forecasting equipment lifecycles and estimating time to failure. The underlying principal for such an analytics-led predictive maintenance framework is aimed at calculating future failure timelines with respect to an asset’s present state and historic data.

With strong data management and quality assessment processes in place, information collected from asset intelligence platforms such as OSI PI and Maximo and unstructured maintenance logs can be used for analysis. Failure rates and root causes can then be discovered using a condition-based maintenance model, while a risk-based model predicts time to failure and computes the subsequent impact on OPEX. This in turn not just simplifies inventory management of spare parts, but also helps operations and maintenance (O&M) teams prepare for and accurately plan repairs, welds, and other such activities. As the framework’s components mature over time, they will come together to deliver an end-to-end view of field equipment and lay the foundation for prescriptive asset maintenance.

Selecting a predictive analytics model

Various advanced analytics algorithms such as survival models and reliability models can be used to drive predictive asset management. A survival model, for example, is capable of determining an equipment’s next failure date using known parameters, which contribute to its breakdown. For pipelines used in refineries and for midstream oil transportation, corrosion is an important factor for determining rate of depreciation. Since it is not always possible to schedule periodic maintenance based on the engineering team’s projections, Cox’s proportional hazards model for calculating survival-time can reveal when a pipeline might give away based on existing and historical corrosion rates.

Alternatively, a reliability model based on Weibull analysis can be used to estimate important life characteristics of an asset, such as reliability or probability of failure at a specific time, the mean life, and the failure rate.

Once a model is selected for further usage by the maintenance teams based on the model accuracy and other engineering parameters, the following activities are usually conducted:

  • Define if the values of the KPIs are in line with the industry standards
  • Rectify gaps in the data with respect to data collection, integration, and error logging methods as understood during advanced analytics engagement. Data profiling and Master Data Management are recommended to ensure getting trust worthy data for advanced analytics.
  • Define anomaly detection or identification of factors that can lead to failure
  • Create failure dictionary for a consolidated view of failures across all assets
  • Define asset health indices for every asset and equipment
  • Execute survival and Weibull models to define the overall failure timelines and useful life of the asset
  • Use the failure information to optimize maintenance schedules
  • Analyze financial impact due to asset failures – measure the losses incurred due to historical failures and derive the estimated financial losses if the critical equipment fails
  • Determine potential savings based on a combined computation of
    • existing maintenance costs
    • capital cost of assets
    • valuation of inventory stock

As per industry standards a proper asset management strategy can help an enterprise reduce around 10-30% of its operating expenses. Needless to say, predictive data empowers you to course correct from time to time and steer your enterprise in alignment with its overall vision, strategy, and goals.


As a part of the Analytics and Insights (A&I) team at Tata Consultancy Services (TCS), Sugato specializes in developing solutions for energy, resources, and utilities space. For nearly 14 years, he has worked across several other industry verticals including banking, retail, life sciences, and telecom. He holds a degree in computer engineering from Kalyani University, West Bengal and has experience working with data warehousing and business intelligence, social network analysis, and statistical analytics and big data analytics practices.



Thank you for downloading

Your opinion counts! Let us know what you think by choosing one option below.