Valued at $103 billion in the U.S., the oil and gas industry is also one of the early adopters of the Internet of Things (IoT). Oil and gas companies combine the IoT, machine learning and the cloud for greater management of remote facilities and tank collection sites so they can act in real-time as safety and regulatory issues arise.
Predictive maintenance and reactive remediation at remote sites
With far-flung extraction facilities and tank collection sites, oil and gas companies must manage spills, emergency shutdowns and regulatory issues throughout their remote field operations. Today, most facilities leverage predictive maintenance to react to problems in tank levels, collecting pressure and flow rates from well sensors on an hourly basis to respond to potential issues. Once an issue is identified, intermediation and corrective action must be taken immediately. However, that can be impacted by the speed at which it can be applied in the field. This is where tank level forecasting helps manage and abate these and other issues, identifying problems in enough time to perform a quick remediation.
Safety issues aside, the most prominent problem affecting production in gas wells is liquid loading, which is the inability of gas to remove liquids being produced in the wellbore and a hole drilled for the purpose of exploration or extraction of natural resources. This occurs when the velocity (speed and direction) of the gas being produced drops below “critical velocity.” The produced liquid accumulates in the well, creating a static column of liquid (see diagram below). The liquid creates a back pressure against formation pressure that forms within the pores of a formation rock, increasing until the well ceases production. Early prediction is vital for preventing economic losses. In this case, time is money and a call to action needs to be prompt. This is where machine learning comes into play.
Oil and gas IoT and machine learning use case
Machine learning refers to generalizable algorithms that enable a computer to carry out a task by examining data rather than through hard programming. Algorithms are used to make predictions based on historical data, focusing on the outcome indicators for future events. Machine learning enables analytics required for predictive modeling using statistical methods (see diagram below).
In recent years, the IoT has evolved into a critical tool for ambitious enterprises and critical industries such as energy, oil and gas production. It has helped companies create revolutionary technologies and improve end-user experiences. In the process, it has also helped produce valuable data that, when analyzed, can be leveraged to improve business operations and offerings.
Applying predictive analytics, oil and gas businesses can deploy an IoT solution using sensors in different locations to generate on-premises simulated data. This data is used to perform an extraction, transform and load (ETL) process that moves the data to a cloud-based analytics platform, accurately identifying locations among multiple sites that require timely maintenance.
Companies can decrease their application response time with an interactive notification system utilizing IoT technology. An interactive mobile application using an analytics platform with machine learning as a central output can send notification messages and identify priority sites in real time, as illustrated in the diagram below.
How interconnection and the cloud power the IoT infrastructures
We offer direct and secure access to multiple cloud and network service providers in every one of our Equinix International Business Exchange™ (IBX®) data centers via interconnection solutions and services that enable the private data exchange between businesses. Our energy, oil and gas customers can also leverage an interconnection Oriented Architecture (IOA™) strategy, deployed on Platform Equinix™, to enable a robust IoT network, extending the digital edge through connected devices, and then establishing an IoT interconnection hub to process the collection of local data and send summarized results to those who need to quickly act upon them.
The dense ecosystem of energy, oil and gas companies that do business on Platform Equinix enable companies in this industry sector to exchange information and interact in real time over direct and secure interconnection. Greater interconnection within the energy and utility industry is increasingly becoming a key requirement for businesses working in this sector. In fact, “The Global Interconnection Index,” a market study published by Equinix, estimates the installed, global Interconnection Bandwidth capacity within the energy and utility industry could grow annually by 73% between 2016 and 2020.
Leveraging interconnection, cloud and machine learning professional services
To help our customers in this critical industry sector quickly and securely deliver meaningful business results from IoT infrastructures, Equinix Professional Services (EPS) guides them through complex interconnection requirements, IT infrastructure changes, network transformations, and hybrid and multicloud deployments. EPS provides expert consulting that enables our customers to optimize cloud migration by matching service providers and architectures to individual business needs. Our cloud and network specialists cooperatively work with enterprise IT teams to develop and apply new capabilities, such as the IoT, machine learning and predictive analytics, on Platform Equinix™.
Machine learning and cloud consulting engagements provide our energy, oil and gas, and other enterprise customers with:
- Examines current state architecture, identifies machine learning use cases, discusses assumptions and constraints of data types
- Identifies cloud platform(s) based on the best match for specific workload and data types
Proof of Concept:
- Analyzes, prepares and evaluates data, resulting in a machine learning application of algorithms to improve results of analysis
Summary and Roadmap:
- Provides a top-down review of a defined applied machine learning process
- Presents results, including specific illustration of context, problem, solution, findings, limitations and conclusions
- Creates a roadmap on how to operationalize and integrate the machine learning and cloud migration processes into existing capabilities