GANESH10203@GMAIL.COM

Deep Sea
Digital Oilfield System Leak Detection Production Forecast Visualization

One-stop tool for several deep sea oil extraction businesses including Leak Detection, Field Automation, Telemetry, Data Management, Data Analytics, Visualization, Business and Technical Workflows, Modeling, Simulation, and Optimization.

Leak Detection
Leak Detection

Challenge

Build an integrated digital system for deepwater oil production, surveillance, and reservoir management that caters to more than 10 different workflows and different types of users.

Outcome

Integrated system with multiple workflows including
1. Production Optimization
2. Realtime Surveillance
3. Deep Sea Leak Detection
4. Production Forecasting
5. Field Automation
6. Data Management
7. Performance Analysis
8. Well Routing Optimization


In this case study I will focus on one of the workflows - Leak Detection.

Leak Detection

Petroleum engineers needed a system for remote oil and gas leak detection in the deep sea oil extraction process.

Before: When the pipe pressure exceeded a certain threshold, the well would be shut down. Then there was a wait for the pipe or well to come to hydrostatic pressure (seawater pressure outside the pipes). If it did, then it was considered a true leak.

After: When the pressure exceeded a certain threshold, a Production Engineer would compare "Rate of Change" chart with multiple sensor data using the new tool and decide if there was an actual leak.

Millions of dollars were saved daily by not shutting down the production well for every false alarm.

My Role and Process

As a Product Manager, I was tasked with collaborating with executives, business analysts, petroleum engineers, development team, and UX designers to ideate, prioritize, design, and develop the Leak Detection system.

Conduct user research and data analysis to identify the needs


Ideate, define the features and roadmap


Manage and prioritize information architecture, interaction design, and content strategy efforts to develop and iterate.

Research

User interviews and surveys, contextual enquiry, data from existing systems

Synthesis

Brainstorming sessions, severity ranking, affinity diagram, prioritization

Design

Wireframes, visualization types, layouts

High Fidelity Prototype

High fidelity mocks, rapid prototypes of visualizations

Evaluation

Heuristic evaluation, severity ranking, task analysis

Development and Improvements

Iterate, feedback, additional features

Success Metrics

Adoption, task success, cost reduction, number of false shut downs prevented, increase in production

Leak Detection
sketch

Research

Conducted user interviews with oil production engineers, product owner, business analysts, and other stakeholders to collect product visions, goals, and business and regulatory requirements.

Conducted contextual enquiry to gather in depth understanding.

Organized focus group sessions consisting of 4-8 participants throughout different stages of UX design. Led the discussion and exercises on datapoints collected in previous stages. Received verbal and written feedback from participants.

User Interviews: wants ≠ needs

Synthesis

Synthesis

After collecting recordings from the user interviews, we conducted affinity mapping with the team to synthesize structure of the tool. We grouped these problems under common themes and features in the platform. Then severity analysis was performed on each of these tasks and features.

Task Criticality x Impact x Frequency = Severity


Task criticality - how important is the task to the user? (1 = low, 5 = critical) Impact - how much of an impact does this issue have on the user's task? (1 = suggestion, 5 = blocker) Frequency (%) - how many times does this come up out of total participants?

Roadmapping and Prioritization

Using BUC (Business benefits, User benefits, Cost) analysis coupled with RICE (Reach, Impact, Confidence, Effort) analysis, categorized solutions into broader Epics to make product roadmap. Product, design, and engineering teams came to the conclusion that the first Epic we would prioritize would be the interactive chart.

Leak Detection
Leak Detection
Leak Detection
Leak Detection

Design

Design Requirements: Keep it simple, focus on functional requirement, and design for performance


The design phase of the project was collaborative (involving input and ideas from UX team, petroleum engineers, development team, etc.) and iterative (cycling back upon itself to validate ideas and assumptions).

Building on the user feedback loop established in previous phases, the premise of the design phase was to put ideas in front of users, get their feedback, refine them, and repeat.

Put design in front of users, get their feedback, refine them, and repeat


We had multiple versions of "Pressure" and "Rate Of Change" visualization wireframes to gather feedback from Production and Engineering users. This involved establishing a standardized visual hierarchy and different types of visualization.

leak detection
leak detection

High Fidelity Prototype

A Hi-Fi prototype was developed before user testing because interactive wireframes for user testing had not been successful in the past.

Users found it difficult to visualize how these grey boxes would magically transform into their beautifully crafted interactive visualization tool


Users often complained that our low-fidelity prototypes did not match up to their expectations of an interactive tool.

Through iterative testing and development, it was identified that, most of the time, one particular chart needed to be compared to 10 other visualizations. Locking that chart to the top solved the problem as it was the driving factor for leak detection.

Evaluation

Weekly user testing sessions on task analysis for some of the complex tasks resulted in a 35% improvement in task success.

Discovery: Confusing to match dots as events on the chart.
Solution: Scrollable lines as events on stackable charts.

Discovery: Users had a hard time identifying what stage of the process they were currently in.
Solution: Clear visual hierarchy that aligned with the mental map of the users.

Discovery: Users wanted to see before and after an event.
Solution: Added buffer and a button to shift the timeframe.

Discovery: Wanted to compare "Rate of Change" with a number of other parameters.
Solution: Make the top chart sticky.

Development and Additional Features

Worked very closely with the design team, engineering team, and stakeholders to develop prioritized backlog items. It was a continuous discovery and delivery process as we iterated through each sprint.

Result and Reflection

Using this tool, the production engineers could detect a leak sitting at their desks and act on it without shutting down the well, thus saving millions of dollars to the company everytime there was a false leak alarm.

The lesson learned is that it is important to involve the development team early in the process to understand technical challenges.