An Interactive 5-Day Training Course

AI for Reservoir Performance & Production Forecasting

Evaluating Machine Learning and Physics-Informed Forecasts for Reservoir Decisions

AI for Reservoir Performance & Production Forecasting

Scheduled Dates

31 May - 04 Jun 2027
Dubai - UAE
$5,950
13 - 17 Sep 2027
London - UK
$5,950
04 - 08 Oct 2027
Online
$4,950
22 - 26 Nov 2027
Dubai - UAE
$5,950

Why Choose this Training Course?

This AI for Reservoir Performance & Production Forecasting training course from PetroKnowledge helps professionals evaluate how machine learning can support well and field production forecasts. Participants learn to assess surveillance data, compare forecasting methods and determine whether predictions are sufficiently reliable for their intended use. The course connects analytical performance with reservoir engineering judgement, helping asset teams assess forecast uncertainty and make informed decisions about model adoption.

Reliable production forecasting requires an understanding of reservoir behaviour and the operating conditions behind the data. Participants examine how engineering benchmarks, physical constraints and validation methods influence forecast credibility. They learn to recognise misleading accuracy measures, assess predictions under changing conditions and communicate uncertainty clearly. The emphasis is on establishing when an AI-supported forecast provides useful evidence for reservoir management, production planning and asset performance reviews. This training course is jointly offered with Excentra, combining reservoir engineering expertise with applied AI to support reliable production forecasting.

This AI for Reservoir Performance & Production Forecasting training course will highlight:

  • Production data quality and engineering benchmarks
  • Machine learning and physics-informed forecasting methods
  • Validation, back-testing and forecast error assessment
  • Probabilistic forecasts and uncertainty communication
  • Model approval, monitoring and retraining responsibilities

What are the Goals?

By the end of this AI for Reservoir Performance & Production Forecasting training course, participants will be able to:

  • Identify suitable applications for AI-supported forecasting.
  • Assess surveillance data quality and allocation uncertainty.
  • Establish appropriate decline curve benchmarks.
  • Compare machine learning and hybrid forecasting methods.
  • Evaluate forecasts using time-aware validation.
  • Interpret probabilistic outcomes and uncertainty ranges.
  • Define model approval, monitoring and retraining requirements.

Who is this Training Course for?

This AI for Reservoir Performance & Production Forecasting training course is designed for professionals who develop, review or use well and field production forecasts.

  • Reservoir Engineers
  • Petroleum Engineers
  • Production Engineers
  • Production Technologists
  • Reservoir Surveillance Engineers
  • Subsurface Data Scientists
  • Production Forecasting Specialists
  • Reserves Evaluation Engineers
  • Reservoir Engineering Managers
  • Subsurface Asset Managers

How will this Training Course be Presented?

The training course follows the decisions involved in assessing an AI-supported production forecast, from data preparation through technical approval and ongoing monitoring. Participants examine how production histories, engineering assumptions and validation results inform model selection. The course explains how forecast documentation, uncertainty assessments and performance records support review by reservoir engineers, asset managers and forecasting teams.

Organisational Impact

This AI for Reservoir Performance & Production Forecasting training course helps organisations establish consistent expectations for evaluating and governing forecasting models.

  • Standardise forecast validation and approval criteria.
  • Improve traceability from surveillance data to predictions.
  • Scrutinise vendor claims and model performance.
  • Detect deteriorating forecast performance earlier.
  • Communicate uncertainty more consistently in asset reviews.
  • Align model use with reservoir engineering evidence.

Personal Impact

This AI for Reservoir Performance & Production Forecasting training course strengthens participants’ ability to assess forecasting methods and explain their suitability for asset decisions.

  • Recognise unsuitable data and modelling assumptions.
  • Compare forecasts against engineering benchmarks.
  • Identify implausible predictions and validation weaknesses.
  • Interpret uncertainty ranges and forecast bias.
  • Communicate model limitations to decision-makers.
  • Define evidence needed for forecast approval.

Daily Agenda

Day 1: Reservoir Data, Benchmarks and Decision Framing
  • Forecasting for production plans and reservoir decisions
  • Rates, pressures, PVT data and well tests
  • Allocation uncertainty, shut-ins and operating changes
  • Missing data, outliers and data leakage
  • Arps decline models, assumptions and limitations
  • Prediction targets, forecast horizons and acceptance criteria
Day 2: Machine Learning for Production Forecasting

  • Time-series features and well-level attributes
  • Regression, tree ensembles and sequence models
  • Single-well and multi-well forecasting approaches
  • Time-aware training, validation and test datasets
  • Back-testing from successive historical forecast dates
  • Forecast bias, cumulative volume error and horizon accuracy
Day 3: Physics-Informed and Hybrid Reservoir Analytics

  • Material balance, flow regimes and physical constraints
  • Physical bounds and conservation terms in model training
  • Hybrid decline, simulation and machine learning approaches
  • Capacitance-resistance models for waterflood surveillance
  • Simulation proxies, history matching and scenario screening
  • Physical consistency checks and extrapolation limits
Day 4: Uncertainty, Scenarios and Model Assurance

  • Data, model and operating-condition uncertainty
  • Petroleum P10, P50 and P90 conventions
  • Calibration of forecast uncertainty ranges
  • Infill, workover, injection and facility scenarios
  • Sensitivity analysis and model explainability
  • Documentation, reproducibility, audit trails and oversight
Day 5: Deployment, Governance and Decision Support

  • Requirements for an operational forecasting pipeline
  • Integration with surveillance and planning workflows
  • Drift monitoring, retraining triggers and model ownership
  • Forecasting software and vendor evaluation criteria
  • Forecast evidence supporting reserves evaluation and reporting
  • Uncertainty communication and forecast approval responsibilities

Certificate

  • Upon successful completion of this training course, delegates will be awarded an official PetroKnowledge Certificate of Completion, signed by the course facilitator. The certificate confirms successful participation and records the total learning hours completed.
  • Continuing Professional Education credits (CPE): In accordance with the standards of the National Registry of CPE Sponsors, one CPE credit is granted per 50 minutes of attendance.

Would an alternative date be more suitable?

We offer a variety of tailored training options, customized to meet your organisation's needs. Delivered anytime, anywhere, we make it easy to bring expert training directly to your team.

Other Training Courses You Might Be Interested In

Follow Us:

Subscribe To Our Newsletter