Predictive Maintenance Using Hadoop for the Oil and Gas Industry

Reducing operating costs and increasing efficiency are, and will always be, priorities for any business, but they become imperative when an industry is facing cyclical challenges. Given the current volatility in the oil market, the oil and gas industry is looking for solutions that can proactively address inefficiencies through better asset tracking and predictive maintenance.

In the Digital Oilfield

Sensors in the digital oilfield produce a vast amount of data that is often underutilized. McKinsey estimates that a typical offshore production platform can have more than 40,000 data tags, though many may not be connected or used. Companies typically use their oilfield sensors to monitor real-time operations status, but the data is not often stored and analyzed to help predict potential equipment problems.

As the Society of Petroleum engineers notes on their wiki, digital oilfields are just beginning to move toward a focus on “operational efficiency and the optimization of processes that are not directly related to the core petroleum engineering activities for which they are currently used.”

The most logical place to start extending the digital oilfield to improve operations efficiencies is through proactive maintenance—using big data to confidently predict equipment failure and address issues before they become costly problems.

As oil and gas operations become more complex, it’s often difficult to have visibility into the condition of equipment, especially in remote offshore or deep-water locations. McKinsey found that rigs in the North Sea were performing as planned only 82 percent of the time, far less than the typical up-and-running target of 95 percent.

Physical inspection of equipment in remote locations is typically an expensive process. This lack of visibility can lead to equipment failure and costly unscheduled maintenance and non-productive time (NPT), as well as oil spills or accidents resulting from failing equipment.

Obviously, warding off a catastrophic failure results in huge across-the-board benefits. But even small improvements in efficiency can yield significant savings. McKinsey estimates that improving production efficiency by ten percentage points can yield up to $220 million to $260 million bottom-line impact on a single brownfield asset.

Technical Challenges and Solutions

The problem with big data is that it is… big. Loading large data sets can consume significant time and resources, forcing unacceptable delays before analysis can start.

But next-generation solutions, designed specifically for use with today’s massive data sets, make it possible for analysts to perform remote monitoring and analysis on data stored in a central repository. An analyst can therefore work with volumes of data retrieved from many assets at many locations. This robust level of availability enables entirely new levels of predictive maintenance, allowing companies to mitigate problems early, prevent equipment failure, and increase net product output.

To predict potential equipment failures before they can occur, the solution must provide the following capabilities:

  • Store and processes real-time and historical sensor data
  • Ingest and analyzes real-time sensor and historical data alongside maintenance data generated from industrial equipment for oil rigs, chemical plants, or mining operations
  • Proactively learns patterns of normal and errant behavior across various types of equipment to provide warnings of minor degradation.

A next-generation solution will also provide data analysts with actionable insight into machine and process operations efficiency, quality, and utilization. Understanding these factors will point to additional ways to ramp up productivity, increase asset ROI, and enhance safety.

Industry Case Study

An oil and gas customer chose a new big data platform to help them be more proactive in their approach. The customer’s goal was to improve predictive maintenance along with reducing operational costs and Non-Productive Time (NPT).

The cost of Non-Productive Time per asset during drill to completion is $500K to $1M per day; post-completion is $40K to $300K per day on average. The number of days per year this customer experiences downtime varies, but the industry average is anywhere from one to three days per year per asset pre-completion, and two to five days per asset post-completion. This customer has several hundred assets in production.

The customer’s new enterprise-ready solution is able to capture massive amounts of data in a cost-effective and scalable way, easily ingesting both structured and unstructured data. The centralized data center supports applications with a predictable scaling model. The customer now has a powerful mechanism to rapidly analyze all variable inputs against existing failure rate models and alert them before equipment is likely to fail. Their previous solution could not scale to accomplish this in a cost-effective way.

This customer is experiencing a significant productivity boost and benefiting from reduced operational costs thanks to their deployment of a predictive maintenance solution built on top of an enterprise-ready Hadoop deployment.



Predictive Maintenance using Hadoop for the Oil and Gas Industry

Streaming Data Architecture:

New Designs Using Apache Kafka and MapR Streams




Download for free