Oil & Gas Analytics Revolution with Dataraft

Oil & Gas generates petabytes from seismic, logs, SCADA, OBD, and production reports - but translating into decisions requires PhDs or months of work. Dataraft delivers no-code analytics tailored for upstream, midstream, and downstream, turning raw data into production gains, maintenance savings, and optimized CAPEX.

1. Predictive Maintenance: Equipment + Pipelines

Upstream: Mud pumps, compressors, turbines on drilling rigs/production platforms.

Midstream: Storage terminals, custody transfer points.

Downstream: Fractionating towers, hydrotreaters.

Dataraft Approach:

SCADA/PLC Data (vibration, RPM, thermodynamics)

+ Telemetry (pipeline pressure/flow)

↓

Anomaly Detection (Isolation Forest, Z-score)

↓

Pattern Matching vs. Healthy Baselines

↓

Predictive Alerts + Pigging Targets

Real Results:

• Compressor Loading: Vibration + RPM → Failure prediction 7–14 days early.

• Pipeline Pigging: Pressure/flow deviations pattern-matched to locate blockages.

• ROI: 25–35% maintenance cost reduction by replacing calendar-based schedules.

2. Porosity Estimation from Neutron/Thermal Neutron Logs (85% Accuracy)

The Problem: Coring for porosity validation requires tripping pipe and lab analysis under in-situ conditions - expensive and slow.

Dataraft ML Classification:

Predictors: Neutron counts (NPHI, CNTC), Thermal Neutron (CFTC, TNRA)

Target: NPOR (Measured Porosity)

Models: XGBoost, Random Forest → 85% accuracy on unseen data

Reservoir Insights:

Porosity Distribution:

• 50–75% section: 9–14% porosity

• High-pay streaks: Up to 27%

Decision Rules Generated:

if CFTC < 3996 AND CNTC < 5041 AND TNRA > 1.29 → NPOR = 1% (Poor)

if CFTC < 3996 AND CNTC < 5041 AND TNRA < 1.29 → NPOR = 5%+ (Desired)

Scatter Insights: NPOR inversely proportional to far-neutron counts; thermal neutron porosity directly proportional—physics validated.​

3. Well Performance Analysis & Artificial Lift Optimization

The Challenge: Wells evolve—natural flow → artificial lift (ESP, gas lift, rod pumps). Wrong ALS choice kills production; physical models don't capture dynamic conditions.​

Dataraft Well Analytics:

5 Wells × Multi-Year Data:

DaysOld, AvgWHT (Wellhead Temp), AvgChokeSize, AvgWHP (Wellhead Pressure)

↓

Predict BOPD (Barrels Oil Per Day)

↓

What-If: "Optimize choke for Well 14H?"

Key Findings:

Production Distribution (Boxplot):

• Mean 2x Median → Few super-producers, many average/poor wells

Top Predictors (50% Power):

1. DaysOld (35%)

2. AvgWHT (8%)

3. AvgChokeSize (4%)

4. AvgWHP (3%)

Actionable Rules:

if ChokeSize 60–98 AND WHP < 38 AND DaysOld 429–921 → High Production

if ChokeSize < 9 AND WHP < 38 → Poor Production

Well 14H Intervention: Early production hiccups traced to choke restriction—optimized post-analysis.

4. Integrated Asset Management: Reservoir → Production → Refining

Vision: Replace siloed analysis with end-to-end optimization across the value chain.

Dataraft Architecture (with GE iFIX, Hortonworks, Azure):

Seismic + Logs + PTA + Production Data

↓

Big Data Ingestion (Hadoop/Spark)

↓

ML Surrogates Replace Mechanistic Simulators

↓

Integrated Optimization (Drilling → Field Dev → Refinery)

Upstream Examples:

• Sweet Spot ID: Seismic + logs → Optimum drilling locations.

• Field Development: Reservoir + drilling + production → Phased plans.

Downstream:

• Refinery Yield: FCC, distillation economics + demand trends → Optimal crude slate.

5. Midstream Preventive Maintenance (Real Deployment)

Client: Midstream O&G operator treating natural gas/liquids.

Dataraft Solution:

Minute-Level Data (45 Wells) → OSI PI Server → MSSQL

↓

Outlier Handling + Dynamic Thresholds

↓

PowerApps/PowerBI Dashboards (Role-Based)

↓

Automated Alerts + Water Chemistry Correlation

Business Wins:

Processing: 45 wells' minute data in <3 minutes

Visualization: Pressure, flow, filter performance

Savings: Proactive failure alerts → Optimized maintenance

Dataraft Oil & Gas: Technical Differentiators

Book Your Dataraft Oil & Gas Analytics Demo Today.