April 28, 2015

How we survived a near death experience in a Big Data Project ? A Story from the trenches

One month back Flutura was in  a meeting with a customer in Houston where we were involved in an IOT big data project . It was a very tense meeting ... Budgets were being slashed, heads were rolling because of the steep drop in Oil prices. The CEOs office was forensically examining all projects thru a massive magnifying glass. The tension in the room was palpable. The cross functional team had spent a lot of time building a platform and the CEO wanted to see tangible business outcomes and how it made money for them. Given the intense market pressure, if the team did not articulate an ultra specific path to monetization, the project was definitely going to be axed. After about 3 hours of intense conversation the cross functional team of Engineering and IT folks were able to hammer out a 1 slider to make the case to the CEOs office .The meeting happened a week later.The project survived. Everyone breathed a sigh of relief ....

So what did we learn from the near death experience ?

The beauty about this near death experience in executing the IOT big data project ( We must admit ...It was a quite nerve wracking experience for us at Flutura as we went thru the turbulence ) was that it forced us to focus on the core . We have synthesized our learnings and we would be the first ones to admit that many of them are obvious but we felt the strong need to reinforce and internalise them.
Lesson-1 :  Think Outcomes ( for example : The project would reduce fuel consumed by 3 % resulting in $ 34,000,000 savings per year in first 2 years )
Lesson-2 : Are we  solving a top 5 problem which keeps business awake at night ( for example one should not be solving problem no 18  in the problem list from a business perspective ... forcibly ranking problems helps )
Lesson-3 : Is Status Quo an option ? ( One would be surprised as to how status quo is an option for many industries, businesses and processes. You may be excited about the possibilities but )
Lesson-4 : Don't use the following words while speaking to CXO team
  • "Big data"
  • "Analytics", "Data Science", "Algorithms", "Statistics" etc etc
  • "Hadoop", "Spark" , "Data Lake", "Lamba Architecture" etc etc
These are sure ways to short circuit a project as it activates the wrong  neural pathways for attention starved executives
Lesson-5 : Spend 70 % of conversation in first 4-6 weeks of a big data project ontangible Business case and not solution or architecture.

To conclude

These are tough times ...Markets are undergoing tectonic transition ... Industries which were immune to recession are tightening the belt.Big data and analytics has a huge role to play in driving efficiencies in many industries. The simple but obvious things evade us. We are reminded of a quote by T S Eliot
What we call the beginning is often the end. And to make an end is to make a beginning. The end is where we start from.
We hope the above learnings reinforce your experience and happy to hear your stories :)

April 21, 2015

Oil & Gas Intelligence 2.0: The Right time for a Big Data Revolution.

A silent but definite shift is taking place in the Oil & Gas industry. This shift has been triggered because of the convergence of 3 important forces that have a cascading multiplier effect on the health of all other industries. So, what are the hidden forces powering the transformation of the Oil & Gas industry? What real life opportunities present themselves from a big data analytics perspective on the upstream and downstream side? What impact do these use cases have on the cost and bottom line economics? In this blog, Flutura would share a point of view based on extensive interactions with leaders at the global hub of the Oil & Gas industry – Houston.

What is the stimulus for Oil & Gas transformation?

If we look deep down at some of the fundamental irreversible market forces at work within the industry, we see 3 important themes causing a tectonic shift to impact the industry’s economics.

Oil & Gas Trend-1: Ageing work force

Of the most important problems looming large at the industry is the fact that almost half of the Oil & gas workforce will retire in the next 2-3 years according to reports. As this valuable generation exits the work force, a lot of experiential knowledge regarding various aspects of oil operations goes out with them. This is driving Oil & Gas companies to look at “codifying” the experiential knowledge of these veterans who will soon exit the system.

Oil & Gas Trend-2: Extreme Pricing Pressures

The steep plunge in Oil prices to almost 50 percent as OPEC refused to cut production in response to the highest U.S. output in three decades. With Oil prices at a historic low, cost optimization to minimize margin erosion is of paramount importance – the need of the hour is to increase overall margins/profitability and generate cash flow.

Oil & Gas trend-3: Accelerated asset digitization

The last mile in the oil industry is a sensor jungle which consists of a variety of MWD (Measurement while drilling) and LWD (Logging while drilling) sensors and SCADA devices. Thousands of these sensors are measuring multiple things from drill rpm, mud flow rates, CO2 gas emissions, valve positions, pump states, and in the process emit billions of events. These dramatic times present perfect opportunities to drive efficiency programs, and data is a competitive weapon that can be used to achieve this business outcome.

What are 3 Game changing big data use case families which impact oil & gas cost outcomes?

1.     Real time drilling operations intelligence
2.    Rig Asset calibration
3.    Single 360 degree view of well intelligence

Use case family-1: Real time drilling operations intelligence

As upstream operations are being performed, a variety of operating parameters are captured in real time like torque, rpm, pump pressures, flow out etc. These data points are captured using historians, condition monitoring systems and SCADA devices.

·         Some of the business questions which can provide deeper insights and situational awareness into operational performance are:
  • What is the frequency and recency of out of bound conditions for the various parameters outlined above?
  • What is the spread of Anomaly alerts across various alert types and what is their relative direction in terms of velocity? Is alarm velocity increasing / decreasing?
  •  Is there a correlation between the out of bound of drill rpm rhythms from MWD Sensor data and operators experience?
  •  What are the frequent sequence of events (mud flow, Drill pressure, temperature, CO2) exhibited in rhythm disturbances prior to a near miss event which can trigger shutdown actions before an adverse event occurs?

Use case family-2: Calibrating Rig Assets

One of the most frequent problems encountered in the rig is the occurrence of assets that are poorly calibrated. For instance, hydro-mechanical gauges tend to over-report hookload. The Impact? à  False perception of getting more WOB than in reality. So, the basic question one needs to answer using big data analytics is what calibration parameters discriminate the healthy devices from the unhealthy devices. Once we analyze the degree of variances, one can trigger a re-calibration intervention in the front-lines. Advanced visualization techniques like Box plots can amplify variation pattern differences to the engineer which can trigger a front-line asset re-calibration event through maintenance workflow engines like Maximo.

Use case family-3: Rig 360 intelligence

Currently the gap in intelligence is due to a holistic 360 degree view of rig. Data is fragmented across a variety of systems like Condition monitoring systems,  Historians, Maximo, MWD and LWD logging systems, Seismic data, Weather Data and Sensor data ( flow meters, pressure, ROP ) in SCADA. Having a holistic view of the rig in an Oil & gas data lake can reveal considerable opportunities for cost optimization. For instance, unplanned downtimes bleed the industry of nearly a $100 Billion every year – the failure of a critical equipment such as an ESP can cost an operator millions of dollars per unit and add to safety risk as well. Cutting across multiple data sets can help identify key early warning signals and failure signatures that are vital for failure prevention.

Concluding thoughts

Tough times don’t last but tough organizations do. The present situation offers an opportunity for Oil & Gas companies to reimagine how they approach data intelligence. A good first step would be to create an Operational data science group to institutionalize analytics and a data platform to industrialize analytics.