Data Science (Oiled and Gassed)

Digitized Oil/Gas Field

This is a fictional story, based on true story. The true story is quite impressive by itself, but I decided to make it even more impressive by adding some color to it. In this story, everything highlighted in "bold" is a capability provided by TIBCO.

A oil company came to us with an interesting problem and some solution ideas. Their “business problem” was that their water pumps in Oil fields were dying leading to disruptions.

Why do you need water pumps in oil fields ? Because oil and water when they meet, they don't like each other and don't mix. Oil feels superior as it much slimmer/lighter than the fatty water. So you pump in water and you get Oil to rise. (Note - this is a highly simplified version of what actually happens in an Oil Field).

Thanks to TIBCO, they had “BigData” - detailed data sets of real time sensor data from the pumps in the oil field. These data sets had been connected via the TIBCO integration platform that connected pumps and other related devices into a "connected data/event fabric". This was used to build operational analytic dashboards so that they could optimize Oil field operations.

BigData if not managed properly can quickly become Big Data Swamp and can impede business operations, and in some cases lead to you to make sub-optimal business decisions. I coined a term called BIGO to describe this - BIG DATA IN, GARBAGE OUT.

Thanks to TIBCO they had accurate BigData data sets (pumps that had failed, pumps that were about to fail and healthy pumps). They enhanced these data sets with data sourced from other systems using data virtualization. Data virtualization allowed them to distill out relevant data from myriad of IT systems used to manage the oil field.

Using data sets, they used supervised machine learning to build an ensemble of data science models. These models were fine-tuned using an iterative and collaborative model optimization done by data scientists. These models focused on signals that the pump generates before it fails. Researchers were able to fine-tune these models to predict a probability of a pump failure.

These models were then plugged in connected data/event fabric for  pumps. Using  data visualization, they built real time dashboards that were being updated in real time with pump data. However these dashboards were not just passive displays of information. 

 Thanks to real time streaming analytics built into these dashboards, staff (technicians and other operational administrators) were now able to react in real time to an alert and take proactive action to stabilize the pump - send a technician in the field to fix/replace the pump. 

 Thanks to ingenuity of our customer's data scientists coupled with TIBCO’s products,  they were able to significantly reduce the number of disruptions due to pump failures.

 This connective fabric also has in-memory data and compute capabilities. This an essential component of real time architectures that need to respond in real time/right time. In this specific case, pump failure is a very disruptive event and has to avoided + proactively handled.

By using data integration we distilled and moved relevant data to in-memory data/compute grid that enabled real time analytics driven process orchestration. This allowed business processes to optimize and adjust based on real time events.

 The beauty of this architecture (Digitized Oil Field) is that it can be re-used for other use cases as well for preventative maintenance. For example, they used a similar approach to manage drill bit failures.

 To create a Digitized Oil Field we also helped them organize the data about their staff, IOT enabled devices, inventory and their supply chain partners to get a holistic view of Oil Field operations. We achieved this via master and reference data management. This provided operational staff with seamless access to data they needed with right quality.

 IOT (and upcoming 5G telco standard) is going to help make Digitized Oil Field even more efficient. IOT will allow more comprehensive monitoring and optimization of operations. The connective fabric can connect to IOT-enabled devices/assets in real time.

Oil Field staff want to share their research  and collaborate with other Oil fields and partners to create a bigger impact and learn from other Oil Field. They are planning to leverage cloud provider (AWS, Azure, GCP) to share information and collaborate. This will need cloud-native capabilities to move data to cloud , safe and secure API access and ability to build apps in the cloud that can also leverage on-prem data and applications.

 I mentioned lots of capabilities in this customer journey.

TIBCO P&T has organized these capabilities in various products organized in what we call “CONNECT, AUGMENT, UNIFY” theme. These products work seamlessly work with other via our integration DNA. These products are cloud-native, cloud-neutral (AWS, Azure, GCP) and support hybrid deployment options. These products can also be used via fully managed service (TIBCO Cloud).

#tibco